Risk and Exposure Assessment for Review
of the Secondary National Ambient Air Quality
Standards for Oxides of Nitrogen and
Oxides of Sulfur
Final
Appendices
Photo courtesy of the National Park Service
-------
-------
EPA-452/R-09-008b
September 2009
Risk and Exposure Assessment for Review of the Secondary National Ambient Air Quality
Standards for Oxides of Nitrogen and Oxides of Sulfur
Final
Appendices
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC
-------
DISCLAIMER
This document has been prepared by staff from the Health and Environmental Impacts
and Air Quality Analysis Divisions of the Office of Air Quality Planning and Standards, the
Clean Air Markets Division, Office of Air Programs, the National Center for Environmental
Assessment, Office of Research and Development, and the National Health and Environmental
Effects Research Laboratory, Office of Research and Development, U.S. Environmental
Protection Agency. Any opinions, findings, conclusions, or recommendations are those of the
authors and do not necessarily reflect the views of EPA. This document is being circulated to
obtain review and comment from the Clean Air Scientific Advisory Committee (CASAC) and
the general public. Comments on this document should be addressed to Dr. Anne Rea, U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, C539-02,
Research Triangle Park, North Carolina 27711 (email: rea.anne@epa.gov).
-------
September 2009
Appendix 1
Aquatic Nutrient Enrichment Case Study
Community Multiscale Air Quality (CMAQ) Model
Final
Prepared by
U.S. Environmental Protection Agency
Office of Atmospheric Programs
Washington, DC
-------
-------
Community Multiscale Air Quality (CMAQ) Model
TABLE OF CONTENTS
Acronyms and Abbreviations 1-iv
1.0 Introduction 1-1
1.1 Overview of CMAQ Model Application 1-1
1.2 CMAQ Model Performance Evaluation 1-3
1.2.1 CMAQv4.6 2002 Annual Average Predictions versus Observations for
the Eastern United States 1-5
1.2.2 CMAQv4.6 2002 Annual Average Predictions versus Observations for
the Western United States 1-9
1.2.3 CMAQv4.7 2002 through 2005 Annual and Monthly Average
Predictions versus Observations 1-13
1.2.4 Extended Evaluation of CMAQv4.6 2002 Predictions versus
Observations 1-21
2.0 References 1-29
LIST OF FIGURES
Figure 1.1-1. CMAQ continental United States and eastern and western modeling
domains 1-2
Figure 1.2-1. 2002 CMAQv4.6 annual average SO2 predicted concentrations versus
observations at CASTNet sites in the eastern domain 1-6
Figure 1.2-2. 2002 CMAQv4.6 annual average SO2"4 predicted concentrations versus
observations at CASTNet sites in the eastern domain 1-6
Figure 1.2-3. 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the eastern domain 1-7
Figure 1.2-4. 2002 CMAQv4.6 annual average NH4+ predicted concentrations versus
observations at CASTNet sites in the eastern domain 1-7
Figure 1.2-5. 2002 CMAQv4.6 annual average SO42" predicted wet deposition versus
observations atNADP sites in the eastern domain 1-8
Figure 1.2-6. 2002 CMAQv4.6 annual average MV predicted wet deposition versus
observations atNADP sites in the eastern domain 1-8
Figure 1.2-7. 2002 CMAQv4.6 annual average NH4+ predicted wet deposition versus
observations atNADP sites in the eastern domain 1-9
Figure 1.2-8. 2002 CMAQv4.6 annual average SO42" predicted concentrations versus
observations at CASTNet sites in the western domain 1-10
Figure 1.2-9. 2002 CMAQv4.6 annual average SO2 predicted concentrations versus
observations at CASTNet sites in the western domain 1-10
Figure 1.2-10. 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the western domain 1-11
Final Risk and Exposure Assessment September 2009
Appendix 1 - i
-------
Community Multiscale Air Quality (CMAQ) Model
Figure 1.2-11. 2002 CMAQv4.6 annual average NH4+ predicted concentrations versus
observations at CASTNet sites in the western domain 1-11
Figure 1.2-12. 2002 CMAQv4.6 annual average SC>42" predicted wet deposition versus
observations atNADP sites in the western domain 1-12
Figure 1.2-13. 2002 CMAQv4.6 annual average N(V predicted wet deposition versus
observations atNADP sites in the western domain 1-12
Figure 1.2-14. 2002 CMAQv4.6 annual average NH4+ predicted wet deposition versus
observations atNADP sites in the western domain 1-13
Figure 1.2-15. 2002-2005 Domain-wide average SC>42" predicted concentrations and
observations by month at CASTNet Sites in the eastern domain 1-15
Figure 1.2-16. 2002-2005 Domain-wide monthly aggregate model performance statistics
for SC>42" concentrations based on CASTNet sites in the eastern domain 1-15
Figure 1.2-17. 2002-2005 Domain-wide average TNOs predicted concentrations and
observations by month at CASTNet sites in the eastern domain 1-16
Figure 1.2-18. 2002-2005 Domain-wide monthly aggregate model performance statistics
for TNOs concentrations based on CASTNet Sites in the eastern domain 1-16
Figure 1.2-19. 2002-2005 Domain-wide average NH4+ predicted concentrations and
observations by month at CASTNet sites in the eastern domain 1-17
Figure 1.2-20. 2002-2005 domain-wide monthly aggregate model performance statistics
for NH4+ concentrations based on CASTNet sites in the eastern domain 1-17
Figure 1.2-21. 2002-2005 domain-wide average SC>42" predicted deposition and
observations by month atNADP sites in the eastern domain 1-18
Figure 1.2-22. 2002-2005 domain-wide monthly aggregate model performance statistics
for SC>42" deposition based on NADP sites in the eastern domain 1-18
Figure 1.2-23. 2002-2005 domain-wide average N(V predicted deposition and
observations by month atNADP sites in the eastern domain 1-19
Figure 1.2-24. 2002-2005 domain-wide monthly aggregate model performance statistics
forNCV deposition based on NADP sites in the eastern domain 1-19
Figure 1.2-25. 2002-2005 domain-wide average NH4+ predicted deposition and
observations by month at NADP sites in the eastern domain 1 -20
Figure 1.2-26. 2002-2005 domain-wide monthly aggregate model performance statistics
for NH4+ deposition based on NADP sites in the eastern domain 1 -20
Final Risk and Exposure Assessment September 2009
Appendix 1 - ii
-------
Community Multiscale Air Quality (CMAQ) Model
LIST OF TABLES
Table 1.1-1. CMAQ Nitrogen and Sulfur Deposition Species 1-3
Table 1.1-2. Formulas for Calculating Nitrogen and Sulfur Deposition 1-3
Table 1.2-1. Normalized Mean Bias Statistics for Predicted and Observed Pollutant
Concentration 1-13
Table 1.2-2. Normalized Mean Bias Statistics for Predicted and Observed Pollutant Wet
Deposition 1-14
Table 1.2-3. CMAQv4.6 Model Performance Statistics for PM2.5 and Selected
Component Species Concentrations 1-22
Table 1.2-4. CMAQ v4.6 2002 Model Performance Statistics for Nitric Acid and Sulfur
Dioxide 1-25
Table 1.2-5. CMAQv4.6 2002 Model Performance Statistics for Annual Sulfate and
Nitrate Wet Deposition Model Performance Statistics 1-26
Table 1.2-6. CMAQv4.6 Model Performance Statistics for Hourly Ozone Concentrations 1-27
Table 1.2-7. CMAQv4.6 Model Performance Statistics for 8-hour Daily Maximum
Ozone Concentrations 1-28
Final Risk and Exposure Assessment September 2009
Appendix 1 - iii
-------
Community Multiscale Air Quality (CMAQ) Model
ACRONYMS AND ABBREVIATIONS
particle nitrate
ASC>4 particle sulfate
CMAQ Community Multiscale Air Quality
EPA U.S. Environmental Protection Agency
ha hectare
HONO nitrous acid
HNOs nitric acid
kg kilogram
NADP National Atmospheric Deposition Program
N2Os nitrogen pentoxide
NH4+ ammonium
NHX total reduced nitrogen
NO nitric oxide
NC>2 nitrogen dioxide
MV nitrate
NOX nitrogen oxides
NOy oxidized nitrogen
NMB normalized mean bias
NTR organic nitrate
PAN peroxyacl nitrate
PM particulate matter
ORD Office of Research and Development
SO2 sulfur dioxide
SO42" sulfate
SOX sulfur oxides
Final Risk and Exposure Assessment September 2009
Appendix 1 - iv
-------
Community Multiscale Air Quality (CMAQ) Model
1.0 INTRODUCTION
This appendix provides an overview of the Community Multiscale Air Quality (CMAQ)
model and the modeling system used for simulating pollutant concentrations and deposition for
the years 2002 through 2005. Included in this appendix are the results of a model performance
evaluation in which model predictions of sulfur dioxide (802), nitrogen dioxide (NC^), SC>42",
total NCV*, and ammonium (NH4+) concentrations and SC>42", N(V, and NH4+ wet deposition are
compared to observations.
1.1 OVERVIEW OF CMAQ MODEL APPLICATION
The CMAQ model is a comprehensive, peer-reviewed (Aiyyer et al., 2007), three-
dimensional grid-based Eulerian air quality model designed to simulate the formation and fate of
gaseous and particle (i.e., particulate matter or PM) species, including ozone, oxidant precursors,
and primary and secondary PM concentrations and deposition over urban, regional, and larger
spatial scales (Dennis et al., 1996; U.S. EPA, 1999; Bryun and Schere, 2006). CMAQ is run for
user-defined input sets of meteorological conditions and emissions. For this analysis, we are
using predictions from several existing CMAQ runs. These runs include annual simulations for
2002 using CMAQv4.6 and annual simulations for each of the years 2002 through 2005 using
CMAQv4.7. CMAQv4.6 was released by the U.S. Environmental Protection Agency's (EPA's)
Office of Research and Development (ORD) in October 2007. CMAQv4.7 along with an
updated version of CMAQ's meteorological preprocessor (MCIPv3.4)t were released in October
20081. The CMAQ modeling regions (i.e., modeling domains), are shown in Figure 1.1-1. The
2002 simulation with CMAQv4.6 was performed for both the Eastern and Western domains. The
horizontal spatial resolution of the CMAQ grid cells in these domains is approximately 12 x 12
km. The 2002 through 2005 simulations with CMAQv4.7 were preformed for the eastern 12-km
domain and for the continental United States domain, which has a grid resolution of 36 x 36 km.
The CMAQv4.6 and v4.7 annual simulations feature year-specific meteorology, as well as year-
specific emissions inventories for key source sectors, such as utilities, on-road vehicles, nonroad
* Total NO3" includes the mass of nitric acid gas and particulate nitrate.
t The scientific updates in CMAQ v4.7 and MCIP v3.4 can be found at the following web links:
http://www.cmascenter.org/help/model_docs/cmaq/4.77RELEASE_NOTES.txt
http://www.cmascenter.org/help/model_docs/mcip/3.4/ReleaseNotes
t The differences in nitrogen and sulfur deposition in the case study areas between CMAQ v4.6 and v4.7 for 2002
are small, as described in Chapter 3.
Final Risk and Exposure Assessment September 2009
Appendix 1-1
-------
Community Multiscale Air Quality (CMAQ) Model
vehicles, wild fires, and natural biogenic sources. Emissions for other sectors of the inventory for
each of the years modeled rely on inventories for 2002. Details on the development of emissions,
meteorology, and other inputs to the 2002 CMAQv4.6 runs can be found in a separate report
(U.S. EPA, 2008). Inputs for the CMAQv4.7 runs for 2002 through 2005 were derived using
procedures similar to those for the CMAQv4.6 2002 runs.
CONUS Domain
Western Domain
Eastern Domain
Figure 1.1-1. CMAQ continental United States and eastern and western modeling
domains.
Each CMAQ model run produces hourly concentrations and wet and dry deposition of
individual pollutant species in each grid cell within the domain. Concentration predictions for
NOy* and SO2, both in units of parts per billion (ppb), are produced as part of our standard model
output. The CMAQ deposition data for nitrogen and sulfur species are used to calculate oxidized
and reduced wet and dry nitrogen deposition, wet and dry sulfur deposition, and total reactive
* NOy is defined as the sum of CMAQ predictions for NO, NO2, HNO3, and PAN.
Final Risk and Exposure Assessment
Appendix 1-2
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
nitrogen and total sulfur deposition. These composite deposition variables are derived from the
species identified in Table 1.1-1 as applied in the formulas shown in Table 1.1-2. The CMAQ
deposition data are in units of kilograms per hectare (kg/ha). We are also including in the
analysis gridded precipitation data that were input to the CMAQ runs to help understand the
temporal and spatial behavior of wet deposition.
Table 1.1-1. CMAQ Nitrogen and Sulfur Deposition Species
CMAQ Species
ANO3
HNO3
N205
HONO
NO
NO2
PAN
NTR
ASO4
SO2
Chemical Name
Particle Nitrate
Nitric Acid
Nitrogen Pentoxide
Nitrous Acid
Nitric Oxide
Nitrogen Dioxide
Peroxyacyl Nitrate
Organic Nitrate
Particle Sulfate
Sulfur Dioxide
Table 1.1-2. Formulas for Calculating Nitrogen and Sulfur Deposition
Deposition Type
Oxidized Nitrogen
Reduced Nitrogen
Sulfur
Formula
0.2258*ANO3 + 0.2222*HNO3 + 0.4667*NO + 0.3043*NO2 + 0
0.1157*PAN + 0.2978*HONO + 0.1052*NTR
2592*N2O5 +
0.7777*NH4 + 0.8235*NH3
0.3333*ASO4 + 0.5000*SO2
1.2 CMAQ MODEL PERFORMANCE EVALUATION
This evaluation of CMAQ focuses on model performance for concentrations and
deposition of nitrogen and sulfur species. For the most part, these comparisons are based on
predictions and observations using annual average concentrations and deposition, to be
consistent with the use of annual predictions in this assessment. A more comprehensive set of
Final Risk and Exposure Assessment
Appendix 1-3
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
model performance information for the 2002 CMAQ simulation is provided in Section 1-2.5,
below.
The purpose of this evaluation is to provide information on how well model predictions
of performance for concentrations and deposition of nitrogen and sulfur species match the
observed data on a regional basis, and not to evaluate performance for an individual location or
area. The CMAQ predictions of SC>2, SC>42", total N(V* and NH4+ concentrations, as well as
SC>42", NOs", and NH4+ wet deposition, were compared to the corresponding measured data for
the years 2002 through 2005t. This analysis compared the annual average predictions of 862,
SC>42", total N(V, and NH4+ concentrations to measurements from CASTNet sites. It also
compared the CMAQ annual total SC>42", N(V, and NH4+ wet deposition to measurements of
these species at National Atmospheric Deposition Program (NADP) sites. In all cases, the model
predictions and observations were paired in space and time to align with the corresponding
observations.
For the 2002 CMAQv4.6 runs, model performance results are provided for both the
eastern and western modeling domains. For the 2002 through 2005 CMAQv4.7, runs have
performance results for concentrations and deposition for the eastern modeling domain only.t
The CMAQ v4.7 performance results for 2002 through 2005 are courtesy of EPA's ORD. The
equations used to calculate model performance statistics for the CMAQv4.6 and v4.7 simulations
are described elsewhere (U.S. EPA, 2008).
The "acceptability" of model performance is judged by comparing the CMAQ
performance results to the range of performance found in other recent regional photochemical
model applications. (U.S. EPA, 2006a; U.S. EPA 2006b; Dentener and Crutzen, 1993) These
other modeling studies represent a wide range of modeling analyses, which cover various
models, model configurations, domains, years and/or episodes, chemical mechanisms, and
aerosol modules. The CMAQv4.6 and v4.7 performance results are within the range found by
these other studies. A key limitation of this evaluation, as noted in Section 3.5, is the lack of
"true" dry deposition measurements for nitrogen and sulfur that are appropriate for comparison
to CMAQ deposition predictions. Thus, this evaluation relies upon the model performance
* Total nitrate includes nitric acid gas and paniculate nitrate.
t There are insufficient non-urban measurements of NO2 to provide for a meaningful regional evaluation of this
pollutant for the purposes of this assessment.
t CMAQv4.7 was not run for the Western 12-km domain for 2002 through 2005.
Final Risk and Exposure Assessment September 2009
Appendix 1-4
-------
Community Multiscale Air Quality (CMAQ) Model
results for various gas and particle species concentrations as well as wet deposition, collectively,
to indicate the extent that the modeling system provides a scientifically acceptable approach for
use in this assessment.
1.2.1 CMAQv4.6 2002 Annual Average Predictions versus Observations for the
Eastern United States
Figures 1.2-1 through 1.2-7 display the comparison of CMAQv4.6 predictions of annual
average concentrations and annual average wet deposition for monitoring CASTNet
(concentrations) and NADP (wet deposition) sites the eastern domain. Each data point in the
figures represents an annual average paired observation and CMAQ prediction at a particular
CASTNet or NADP site. Solid lines indicate the factor of 2 around the 1:1 line shown between
them.
Using the normalized mean bias (NMB) statistic, we see the 2002 CMAQ run tends to
underpredict concentrations of SO42" (NMB = -12.0%) and NH4+ (NMB = -3.4%) and
overpredict SO2 (NMB = 33.8%) and TNO3 (NMB = 11.2%) within the eastern domain for 2002.
For wet deposition, the 2002 CMAQ run tends to underpredict NO3" (NMB = -11.4%) and NH4+
(-6.1%) and overpredict SO42" wet deposition (9.4%).
Final Risk and Exposure Assessment September 2009
Appendix 1-5
-------
Community Multiscale Air Quality (CMAQ) Model
20Q2at met2v33 12km ESO2 for 20020101 to 20021231
1-
O
o CASTNet (2Q02ac. met2v33 12kmE)
Period Average
SO2 { ug/m3 )
6 8
Observation
Figure 1.2-1. 2002 CMAQv4.6 annual average SC>2 predicted concentrations versus
observations at CASTNet sites in the eastern domain.
2002ac met2v33 12kmESO4 for 20020101 to 20021231
CASTNet 2002ac met2v33 12kmE)
0.0 0.5 1.0
2.0 2.5 3.0 3.5
Observation
4.0 4.5 5.0
\2-
Figure 1.2-2. 2002 CMAQv4.6 annual average SO "4 predicted concentrations versus
observations at CASTNet sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-6
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
2002ac met2v33 IZkroETNO3 for 20020101 to 20021231
1
O
a CASTNef (2Q02ae. met2v33. 12krnE)
Period Average
TNO3 ( ug/m3 j
0.0 0.5 1.0
1.5 2.0 2.5 3.0 3.5
Observation
4.0 4.5 5.0 5.5 6.0
Figure 1.2-3. 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the eastern domain.
2002ac met2v33 12kmE NH4 lor 20020101 to 20021231
CASTNet (2002ac met2v33 12kmE
1.0 1.5
Observation
Figure 1.2-4. 2002 CMAQv4.6 annual average NILt+ predicted concentrations versus
observations at CASTNet sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-7
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
20Q2ac met2v33 l2kmESO4f0r 20020101 to 20021231
n - o NADP dep (2002ae.. met2v33 12tonE)
(*fl*aj <*-)
IA * 0 g$ NMB « S 4
RMSE * 4,33 NME s 23 1
RMSEs = t S WAdnB = 5.8 S.
RMSEy = 3.SS Ng^cSnE = 17.3
MS = * Z7 F@ = 33 °'
ME « 311 FE * 2t 0
M4nB * 0 73
MdriE s 2,22
1
u
15 20
Observation
Period Accumulated
SOI ( kg/ha)
2-
Figure 1.2-5. 2002 CMAQv4.6 annual average SC>4 " predicted wet deposition versus
observations at NADP sites in the eastern domain.
SOOZac mel2v33 12km£NO3for 20020101 to 20021231
a NADP dep(2002ac met2v33 12kmE)
RMSE a 2.99 F«E * 204
RMSEs * 132 W!*(S a -IS 2
RMSEy = 2.68 NMdnE = 56.1
MB = -t 31 FB = -16.7
ME 2 34 FE = 25 ©
MdnB • -t 74
MdnE a 2,0?
Period Accumulated
NO3 { kg/ha j
10 15 20 25
Observation
Figure 1.2-6. 2002 CMAQv4.6 annual average NCV predicted wet deposition versus
observations at NADP sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-8
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
2002ac
12kmE NH4 for 20020101 to 20021231
NADP dep(2QG2ae.met2v33 12hmE)
Observation
Figure 1.2-7. 2002 CMAQv4.6 annual average NH4+ predicted wet deposition versus
observations at NADP sites in the eastern domain.
1.2.2 CMAQv4.6 2002 Annual Average Predictions versus Observations for the
Western United States
Figures 1.2-8 through 1.2-14 display the comparison of CMAQv4.6 predictions of
annual average concentrations and annual average wet deposition for monitoring CASTNet
(concentrations) and NADP (wet deposition) sites the western domain. Each data point in the
figures represents an annual average paired observation and CMAQ prediction at a particular
CASTNet or NADP site. Solid lines indicate the factor of 2 around the 1:1 line shown between
them.
Using the NMB statistic, the 2002 CMAQ run tends to underpredict concentrations of
SO42" (NMB = -20.8%), NH4+ (NMB = -16.2%), and TNO3 (NMB = -19.6%) and to overpredict
SO2 (NMB = 31.2%) within the western domain for 2002. For wet deposition, the 2002 CMAQ
run tends to underpredict NO3" (NMB = -43.6%), NH4+ (NMB = -43.4%), and SO42" wet
deposition (NMB = -20.5%). In general, model performance for the western domain is degraded
somewhat compared to the results found for the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-9
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
2002ae 12km WUS SO4 for 20020101 to 20021231
"" a CASTNet (2Q02ac 12km WUSi
1.5 2.0
Observation
2-
Figure 1.2-8. 2002 CMAQv4.6 annual average SO4 " predicted concentrations versus
observations at CASTNet sites in the western domain.
2002ac 12km WUS SO2 for 20020101 to 20021231
a CASTNet (2002ac 12km WUS5
Penod Average
SO2 | yg/m3 }
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Observation
Figure 1.2-9. 2002 CMAQv4.6 annual average SC>2 predicted concentrations versus
observations at CASTNet sites in the western domain.
Final Risk and Exposure Assessment
Appendix 1-10
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
Z002ac 12km WUS TNO3 for 20020101 to 20021231
s
5
O
o CASTNet (2Q02ac 12km WUS)
Period Average
TNQ3 ( ug'mS }
0.0 0.5
3.5 4.0
Observation
Figure 1.2-10. 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the western domain.
2002ac 12km WUS NH4 for 20020101 to 20021231
- o CASTNet (2002ac 12km WUS)
-o 04
MdnE SB 0 06
Period Average
NH4 | ug'm3 )
0.0
0.6 0.8
Observation
1.2
Figure 1.2-11. 2002 CMAQv4.6 annual average NH4+ predicted concentrations versus
observations at CASTNet sites in the western domain.
Final Risk and Exposure Assessment
Appendix 1-11
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
2002ae 12km WUSSO4lor 20020101 to 20021231
o
o NADP dep(2GG2ac 12km WUS)
IA * 085 r
-------
Community Multiscale Air Quality (CMAQ) Model
2002ae 12km WUS NH4 for 20020101 to 20021231
s
o NADP dep!2QG2ae 12km WUS
Period Accumulated
NH4 f kgtia )
1.5 Z.O Z.5
Observation
Figure 1.2-14. 2002 CMAQv4.6 annual average NH4+ predicted wet deposition versus
observations at NADP sites in the western domain.
1.2.3 CMAQv4.7 2002 through 2005 Annual and Monthly Average Predictions
versus Observations
The annual normalized mean bias statistics for the CMAQv4.7 2002 through 2005
simulations are presented in Table 1.2-1 and Table 1.2-2 for annual average concentrations and
annual total wet deposition, respectively. In general, model performance for each species is
similar in each of the 4 years. The CMAQv4.7 performance for 2002 is similar to that of
CMAQv4.6 for most species. Notable differences are seen for concentrations of TNOs and NH4+
For TNOs, the overprediction in CMAQv4.7 is about twice that of CMAQv4.6. For NH4+,
CMAQv4.7 slightly overpredicts observations (4%), while CMAQv4.7 slightly underpredicts
observations (-3%).
Table 1.2-1. Normalized Mean Bias Statistics for Predicted and Observed Pollutant
Concentration
Pollutant
Concentrations
SO2
SO42
2002
45%
-13%
2003
39%
-9%
2004
47%
-13%
2005
41%
-17%
Final Risk and Exposure Assessment
Appendix 1-13
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
Pollutant
Concentrations
TNO3
NH4+
2002
22%
4%
2003
26%
11%
2004
22%
7%
2005
24%
2%
Table 1.2-2. Normalized Mean Bias Statistics for Predicted and Observed Pollutant Wet
Deposition
Pollutant
Deposition
SO42
NO3
NH4+
2002
11%
-13%
-11%
2003
6%
-17%
-16%
2004
11%
-14%
-9%
2005
6%
-17%
-13%
Figures 1.2-15 through 1.2-26 provide a comparison of observed and predicted monthly
concentrations and monthly total wet deposition for an aggregate of monitoring sites the eastern
domain. This time series format is intended to reveal differences and similarities in performance
within and across the 4-year period. Figures 1.2-15 and 1.2-21 indicate that the predictions of
SO42" concentration and SO42" wet deposition closely track the temporal patterns exhibited by the
observations. However, the correlation is higher and the error is lower for SO42" concentrations
than the corresponding statistics for SO42" wet deposition (see Figures 1.2-16 and 1.2-22).
Predictions of NCV concentrations, although highly correlated with the observations in most
months, show relatively large error and positive bias in the fall with a peak in October for each
year. Model performance for wet deposition of MV also has a seasonal pattern, with
underprediction of approximately 40% in the late spring and summer and overprediction from
October through December. Observed concentrations of NH4+ are overpredicted by CMAQ in
the spring and fall and underpredicted in the summer. The overprediction in the spring peaks in
March/April, while the peak overprediction in the fall occurs in October/November of each year.
Model predictions of NH4+ wet deposition more closely track the temporal patterns of
observations than do the predictions of NH4+ concentrations. There does not appear to be strong
seasonal differences in performance across the 4 years as seen in NH4+ concentrations; however,
the greatest underprediction appears to occur in May in each year. The differences and
similarities in the seasonal patterns in model performance for various species are being analyzed
by EPA to understand and explain these relationships, with the goal of improving model
Final Risk and Exposure Assessment
Appendix 1-14
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
performance through improvements to emissions and meteorological inputs and scientific
formulation.
CDC^PHASE^RUNS CASTNET SO4 tor 20020101 to 20051231 ; State: All; Site: All
7.0 -
6,5 -
6.0 -
5.5 -
5.0 -
— 4.5 -
n
o> 4-° ~
a
3 3.5 -
i
3.0 -
2.5 -
2.0 -
1.5 -
1.0 -
0.5 -
-A- CASTNET
-* CMAQ
& ' "? ' &'••
A H A'
A' A\ l\ -A ?
1 xA • 1 Ax\ ' 1 x
1 x Nl - 1 ' / './> ! ' ''-
|x' i x k ;x V', i' x 1
I/ * )' 4 /-> .', • !/ 1
l| | 1 I Ax \\ ';'!
i - it I i1 / Vi h
A ' 1 • ' 'A (i
fA i \ ' r f
<' X ) * , -i ^
<. ' -1 - • , A, A A .'/ i 4 'J,
A« 1 i , ".:-/ .
''/ \ x ' '>
i' / \ iX X ''A.
AX'x fc. i1) 1 ' k'
'J/ - i xAx'l ', f"
w \/
I.5 X
2002 2CO3 H)04 2005
1 3 5 7 9 11 1 3 S 79111 3579111 3 57911
Months
2-
Figure 1.2-15. 2002-2005 Domain-wide average SC>4 " predicted concentrations and
observations by month at CASTNet Sites in the eastern domain.
COC_PHASE_RUNS CASTNET SO4 tor 20020101 to 20051231; State: AH; Site: All
40 -
30 ~
20 -
?
— 10 -
IE
Z
S
z.
-10 -
-20 -
-30 -
40 ~
Correlation
-A- NMB
-»- NME ;
i *.
* ', ' ' ' •
, . ' XX
A
/A
AA .", A \ - A
A '"' f ^ A'A ^ "'''t -, ^ ' """ /•
'' ^ >! \ / \ " (' 1 f"' / ^ !. ^'^.i'i
.'' \ 'J' ^ \ / '' l \ 1 \ 'A'^ \i \ A ^
^ A' A \ • a I / a , ''. / V
i / I / I "A "
'A A
2002 2003 2004 2005
- 1.0
- 0.9
- 0.8
- 0.7
- 0.6 i
c
o
- 0.5 OJ
-0.4 0
- 0.3
- 0.2
- O.I
- o.o
1 3579551 3579111 3579111 3 6 7 9 11
Months
Figure 1.2-16. 2002-2005 Domain-wide monthly aggregate model performance statistics
for SC>42~ concentrations based on CASTNet sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-15
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CDC_PHASE_RUNS CASTOET TNO3 for 20020101 to 20051231 ; State: All; Site: All
5.5 -
5.0 -
4.5 -
4.0 -
n
o> 3.5 -
3_
Z 3.0 -
2.5 -
2.0 -
1.5 -
-A- CASTNET
•-*- CMAQ
X.
11
'
( x *
1 i /I
« ' ft 'I -1
x « A XA ' x ll> x «•'
V / III Ml A k /
A | ' / if i / \ 1' "! x x
\ 1 /" r" -1 Mil ] *,| 11 | f
AA\X\ x ' / 4 x / */ 4 / f f\
A\/VX A 1 /i x / H / / i x 1
W"\A/ V\/ AA 1 / A. ^ |
\' i!A\/ A i v / '-\ x J
A &£^x l ft,. * A ^ / A
AAA xV ^fi* : 4* .A'
"A'
20C2 2033 2004 2005
' r i r [ '! r i" i r f r i rri '[ '[ n ri FT r j r~i y r r T'T rr rr'i [ 5 r'rr'i'rr? f r
1 3579111 3579111 3579111 357911
Months
Figure 1.2-17. 2002-2005 Domain-wide average TNOs predicted concentrations and
observations by month at CASTNet sites in the eastern domain.
CDC_PHASE_RUNS CASTNET TNO3 for 20020101 to 20051231; State: All; Site: All
110 ~p
100 -
90 -
80 -
70 -
— 60 ~
8s
i 5°~
Z 40 -
•^
•fi
5 30 -
2
20 -
10 -
0
-10 -
-20 -
Correlation
-A- NMB
-K- NME :
x ! x
4 • 4 A
1 fl f!
A 1 •
H 1 'i'
A J ^ ; ••' i | s
1 1 /A 1 x7f •'- " ^
xf T v j 1 - x f t '! i
/ 1 // ! /'' x^
f i / / I v ' i i!
x XA ^ /' 1 / i I- x /
4V\ // 11 xx, /xf t / ,A H /"x x A |1
,x/\ 'x ' / i / Vx / '\' xxxx' /I A\ >XA x ^ r
/ \ i X ^ -'^ J 1 ^ ^ , 'i "J / v/^ ^
x^ \ A \//\ A/ A:*: / \ / \ / '
A' A / "f A / A \ /\ A 1 / A, A
\ / ^ \t / 'i ^ A jf \ « \ /
Aj A *'" L/ t A
A Tl
\/
A
2002 2003 2004 2005
- 1.0
- 0.9
- 0.8
- 0.7
- 0.6 i
.1
- 0.5 to
CD
~ Q Q O
' O
- 0.3
- 0.2
- 0.1
- 0.0
1 3579 11 1 3579 11 1 3579 11 1 3579 11
Months
Figure 1.2-18. 2002-2005 Domain-wide monthly aggregate model performance statistics
for TNOs concentrations based on CASTNet Sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-16
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CDC_PHASE
2.0 -
1.8 -
1.6 -
ff
E
o> 1.4 -
13
^
X
2 1.2 -
1.0 -
0.8 -
0.6 -
-A- CASTNET
-*- CMAQ
A
A
f k
K f- ^
/\ /v .\
x, 1 x \
I
1
f J
«• J '
* x A A
\f
/
A
2002
13579
RUNS CASTNET NH4 for 20020101 to 20051231; State: All; Site:
,x ^
x7'* I
I F
k I
F i
^ M I
A Vj
A
11 1
X
11
1
A
/!
l\
1
T"
3
£
f
i
1
j
[
, X !
\ A'x
I K
»/
1
A|
2003
^T™r">
5 7
1
|
A 1
I
1
'1
1
1
i
1
1
A
'A
9 11
X
A
'
A /
l\ A^A / S/
M / / \ ' xxi
MX /x 4 *A I
I V ^ \ X I A
\ »A V \ 1 I
V M i /
i./ \ / 1
i \/ i » ')
| * S xf ^
$ 1 /
A A A
2004
1 3 5 7 9 11 1 3 5
A
/ \
X A
^ \
|
\ / x
X
2005
7 9
All
k
y.
x
A A
1 I
]
1 f
11
1
ll
A
1 1
Months
Figure 1.2-19. 2002-2005 Domain-wide average NH4+ predicted concentrations and
observations by month at CASTNet sites in the eastern domain.
90 -
BO -
70 -
60 -
50 -
g 40-
S 30-
"~ 20 -
Z 10 -
0 -
-10 -
-20 -
-30 -
-40 -
CDC_PHASE_HUNS CASTNET NH4 lor 20020101 to 20051231; State: All; Site: All
Correlation
-&-. NMB
-*- NME
: : 1
f, : : i
k - I
l\ ' A ' A 1
/\x xa ,A t x-"\ /|' /^ J|
A/\x l-^x^/ix //fx"x-x-x ]\^xxx/ f x^\ Vx/] j.
x / V M x i j1 i / \
A A / / / ' 1 ' 1 ^
/A / ' / 1 ''1 ^\ / 't A
1 / t' ' ^A ,1 \ /":' , L t \
A />i ' A i "" \ AA' " i 1
\X A' \l ' A \ 1
A A \ 1
AA"
2002 2003 2004 2005
13579 11 1 3579 11 1 3579 11 1 3579 11
- 1.0
- 0.9
- 0.8
- 0.7
- 0.6 £
.1
- 0.5 to
CD
^0.4 0
- 0,3
- 0.2
- 0.1
- 0.0
Months
Figure 1.2-20. 2002-2005 domain-wide monthly aggregate model performance statistics
for NH4+ concentrations based on CASTNet sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-17
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
OS
CDC_PHASE_RUNS NADP SO4_dep for 20020101 to 20051231; State: All; Site: All
450
400
250 -
3 150
100
50
0
NADP
CMAQ
2004
TTTT
~r~n~
1 3 5 7 9 11 1 3 5
9111357911135
Months
Figure 1.2-21. 2002-2005 domain-wide average SC>42~ predicted deposition
observations by month at NADP sites in the eastern domain.
and
CDC_PHASE_RUNS NADP SO4_dep for 20020101 to 20051231 ; State: All; Site: All
100 -
90 -
80 -
70 -
f 60 -
£.
11 50 -
Z
- 40 -
^
2 30 -
20 -
10 -
-10 -
-20 -
Correlation
_^_ NMB
-x- NME
x
l\ x :
/ V A X x X
' " \ * X |.\ / \. X . X X ' "x
f x x A x \ M / x / \ x- \ / x x
/ V\ « X' X x x*, / x' '^ x; * 1
" X x 1 K \ 1 'x * X ' x X \ /
x' x ' x ' x
A
A ,'i
l'NA '' \
/ \ I \ A
A \ A 1 f
A A • / • A" \ 1 • i 1
A A'"\ A'\ A \ . 1 A ,'
^V V XAA/^ "\_AL U\ ./
A A A
2002 2003 2004 2005
- 1,0
- 0.9
- 0.8
- 0.7
^-~
- 0.6 ^
.1
- 0.5 to
0)
k—
o
• o
- 0,3
- 0.2
- 0.1
- 0.0
1 3579 11 1 3579 11 1 3579 11 1 3579 11
Months
Figure 1.2-22. 2002-2005 domain-wide monthly aggregate model performance statistics
2-
for SC>4 " deposition based on NADP sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-18
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CDC_PHASE_RUNS NADP NO3_dep for 20020101 to 200S1 231 ; State: All; Site: All
280 -
260 -
240 -
220 -
200 -
S-180-
j;
^ 160
8-140 -
D
n 120 -
3
Z 100 -
80 -
60 -
40 -
20 -
0 -
-e*~ NADP
-x- CMAQ
A
[1 A
H f . A n
A / i\ /
A / i /I
\A MX ' M f,
\ /' A A A ''
x / V ! '-' ! A J 1
a ' i X IX n 1 A I
f\A A A \ A A A
/ x [ *x l\l/\ x' Vx* i '" jx1 A 1
1 >h|/'\ 1 r\ J >< x ?? f/\ 1 \/\ 1 r ' j s x i
1 x \ M AJ £ \ / F ii f " \ II ^1 ' S ;\ !
1 V v A j 4 1 ^ I f d / ^ \ 1 / ^ ( I / \
f X X - -1 |I v V\ j ^\ '/ ^ 1 / \ A ^
/ v T t- \t x\v t N/W^
* ill I/ H 9 \ '
II « U A \ /
II x U \/
1 X ' x
I
*
2002 2003 2004 2005
r i ' r [ "j r'Tji'T'i'i rr i i i"i"i"i FTTTTI i r r'i r T' r"! T'i'T'i i r'i r'! FT ri r
1 3579111 3579111 3579111 357911
Months
Figure 1.2-23. 2002-2005 domain-wide average N(V predicted deposition and
observations by month at NADP sites in the eastern domain.
CDC_PHASE_RUNS NADP NO3_dep for 20020101 to 20051 231; State: All; Site: All
100 -
80 -
60 -
g 40-
u
5
2 20 -
M
5
o
-20 -
-40 -
-60 -
Correlation
&-- NMB
-*- NME
x
A x *
\ / ' XX ' \/ /^ / "x ^ ,xxx '••- /X*' XXK/
xx_ / x >v v-x xx \ xx • yx
'A. x' ' " X
..
AA AA . '
/ 1 '' '
A , f ' . r A A AA
\ j / '' 1 I/ \ l T
\ 1 ^ A ^ A A s
1 i V f' • \ j \
A\7 • iA "AAA' . ^/\AA
2002 2003 2004 2005
- 1.0
- 0.9
- 0.8
- 0.7
- 0.6 S
.1
- 0.5 to
V
-0.40
- 0,3
- 0.2
- 0.1
- 0.0
13579111 3579111 3579111 357911
Months
Figure 1.2-24. 2002-2005 domain-wide monthly aggregate model performance statistics
for N(V deposition based on NADP sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-19
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
110 -
100 -
90 -
80 -
70 -
3°
§>
g- 50 -
30 -
20 -
10 -
o -
-10 -
CDC_PHASE_RUNS NADP NH4_dep for 20020101 to 200S1231 ; State: All; Site: All
-&- NADP
•-*- CMAQ
A
)i
I
I -A
,'l
• u
I ft X
A I A A ' ' '/I ' A- ft
^ ' ' \i ' ^ X>l A' '* I I ' ^ r
X x t\ I** I '
1 1 ii '«
i ' ' \ / \
A A / 1 A ^ | i | \ *' /.,
/\ /- f f\/UA * \i\r\ i\ j
A A \ / \ / AA/iA-A | & \ I, A' A'. A / A / "
\/ " ~\ '< A 'X '' ^ ''
A
2002 2003 2004 2005
- 1.0
- 0.9
- 0.8
- 0.7
- 0.6 £
.1
- 0.5 to
£
„ A Q O
• o
- 0,3
- 0.2
- 0.1
- 0.0
13579 11 1 3579 11 1 3579 11 1 3579 11
Months
Figure 1.2-26. 2002-2005 domain-wide monthly aggregate model performance statistics
for NH4+ deposition based on NADP sites in the eastern domain.
Final Risk and Exposure Assessment
Appendix 1-20
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
1.2.4 Extended Evaluation of CMAQv4.6 2002 Predictions versus Observations
In addition to the preceding evaluation of annual and monthly average concentrations and
deposition for nitrogen and sulfur species, model performance statistics are also calculated and
predicted for CMAQv4.6 for a more comprehensive set of species, with comparisons of
observations and predictions paired for the native averaging times of the observations (i.e.,
hourly or weekly)*. These statistics are presented separately for each of four subregionst:
Northeast, Midwest, Southeast, Central, and West for the following species*: PM2.5 total mass,
selected PM2.5 component species (i.e., SC>42", N(V, NH4+, elemental carbon, and organic
carbon), HNO3, total nitrate, SO2, NO2*, ozonet, and wet deposition of SO42", NO3", and NH4+.
The performance statistics are provided in Tables 1.2-3 through 1.2-7. As one way of
comparing performance for sulfur and nitrogen containing species across the five subregions, the
species/subregion combinations are identified, which had NMB estimates within + 15%, between
+ 15% and + 30%, and beyond + 30%. The results of categorizing the statistics in this manner
lead to the following findings:
1. Sulfate Concentration
• NMB is within + 15%, except for the Central subregion and the West subregion,
where NMB is between -15% and ^30%.
2. Sulfur Dioxide Concentration
• NMB is between +15% and + 30% in the Northeast and West subregions and
greater than + 30% in the Midwest, Southeast, and Central subregions.
3. Sulfate Wet Deposition
• NMB is within + 15%, except for the Midwest subregion, where NMD is between
+ 15% and+ 30%.
* Measurement sampling periods used in this analysis are as follows: Speciation Trends Network (STN) and
IMPROVE - 24-hour average concentrations; CASTNet and NADP - weekly average concentrations and weekly
total deposition, respectively; and Air Quality System (AQS) - hourly average concentrations.
t The subregions are defined as groups of states: Midwest includes IL, IN, MI, OH, and WI; Northeast CT, DE, MA,
MD, ME, NH, NJ, NY, PA, RI, and VT; Southeast includes AL, FL, GA, KY, MS, NC, SC, TN, VA, and WV;
Central includes AR, IA, KS, LA, MN, MO, ME, OK, and TX; and West includes AZ, CA, CO, ID, MT, ND,
MM, NV, OR, SD, UT, WY, and WY.
t Measurements of NO2 and NOy from sites in the Southeastern Aerosol Research and Characterization (SEARCH)
network were not available for use in this assessment.
* The model performance statistics for NO2 were not calculated in the West subregion because the results would be
disproportionately skewed to California where most of the NO2 AQS sites in this subregion are located.
t Note that performance statistics for ozone excluded any observed or predicted hourly ozone concentrations less
than a threshold of 40ppb.
Final Risk and Exposure Assessment September 2009
Appendix 1-21
-------
Community Multiscale Air Quality (CMAQ) Model
4. Nitrate Concentration
• NMB is within + 15%, except for the Northeast subregion, where NMB exceeds +
30% and the West subregion where NMB is less than - 30%.
5. Nitric Acid Concentration
• NMB is within + 15%, except for the Southeast subregion, where NMB is
between +15% and + 30%.
6. Nitrogen Dioxide Concentration
• NMB is within + 15%, except for the Central subregion, where NMB exceeds +
30%.
7. Nitrate Wet Deposition
• NMB is within + 15%, except for the Central subregion, where NMB is between -
15% and - 30%, and the West subregion, where NMB is less than - 30%.
8. Ammonium Concentration
• NMB is within + 15%, except for the West subregion, where NMB exceeds +
30%.
9. Ammonium Wet Deposition
• NMB is within + 15%, except for the Central subregion, where NMB is between
15% and - 30%, and the West subregion, where NMB is less than - 30%.
Table 1.2-3. CMAQv4.6 Model Performance Statistics for PM2.5 and Selected Component
Species Concentrations
CMAQ 2002
PM25
Total Mass
STN
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
10307
3000
2780
1516
2554
2738
2487
NMB
(%)
2.4
-5.8
10.7
8.4
-9.4
4.4
-7.4
NME
(%)
39.5
46.9
41.6
32.8
35.6
46.5
46.8
FB
(%)
-3.2
-3.1
7.5
6.2
-17.2
-3.4
-4.5
FE
(%)
42.1
45.0
39.3
33.9
41.4
50.2
44.8
Final Risk and Exposure Assessment
Appendix 1-22
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002
PM25
Total Mass
(continued)
Sulfate
Sulfate
(continued)
Nitrate
IMPROVE
STN
IMPROVE
CASTNet
STN
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
8436
10123
2060
592
1803
1624
9543
10157
2926
2730
1487
2541
2686
2446
8532
10232
2070
597
1805
1671
9645
3173
1158
839
663
1085
229
1118
8770
2726
2731
1488
2540
1298
2446
NMB
(%)
-12.0
-26.4
4.6
-0.8
-19.7
-20.1
-27.8
-5.3
-20.6
-4.1
7.1
-9.8
-9.2
-26.1
-11.7
-7.5
-12.7
-3.1
-10.8
-18.5
-5.5
-12.3
-21.3
-13.3
-8.2
-12.9
-22.9
-20.4
0.1
-45.0
9.8
2.7
-5.8
-3.8
-47.5
NME
(%)
44.8
53.5
47.2
36.1
40.7
48.9
53.1
34.1
41.9
29.6
36.3
33.7
39.4
44.9
33.8
42.4
30.8
30.6
33.2
37.0
43.5
20.3
34.6
18.7
17.9
21.4
28.6
35.3
61.1
63.1
60.6
55.9
78.2
53.7
62.8
FB
(%)
-15.1
-26.3
3.5
-6.0
-28.8
-25.3
-27.1
-11.3
-12.2
-8.4
0.9
-18.7
-13.0
-15.8
-8.8
7.6
-11.4
-6.9
-18.8
-19.3
8.6
-17.2
-11.2
-16.9
-14.4
-19.9
-29.9
-10.7
-39.8
-70.6
-26.2
-14.5
-70.0
-21.5
-73.8
FE
(%)
50.5
57.5
45.6
39.2
51.8
58.6
57.2
39.3
43.5
33.5
36.9
39.8
46.4
44.8
41.8
45.7
37.7
35.8
41.3
44.6
45.9
25.7
35.9
22.2
21.6
27.4
35.8
36.1
85.8
95.0
77.5
70.4
106.9
73.8
95.4
Final Risk and Exposure Assessment
Appendix 1-23
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002
Nitrate
(continued)
Total Nitrate
(NO3+HNO3)
Ammonium
Elemental
Carbon
IMPROVE
CASTNet
STN
CASTNet
STN
12-km BUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
8514
10110
2069
597
1803
1672
9522
3171
1157
839
662
1085
229
1117
10157
2926
2731
1488
2540
2685
2446
3166
1156
837
661
1085
229
1116
10031
2975
2744
1498
2506
2570
2475
NMB
(%)
13.3
-34.8
56.3
11.5
3.8
0.8
-39.6
11.1
-19.5
18.1
10.2
12.4
-2.9
-20.4
5.0
-23.6
5.4
11.8
1.9
5.1
-30.6
28.2
-16.8
-3.2
7.1
-13.7
0.5
-21.1
33.0
43.1
39.1
25.9
10.1
76.4
49.0
NME
(%)
85.1
80.6
99.1
72.04
94.8
77.3
81.1
30.5
44.2
30.4
27.9
33.5
33.2
45.8
39.5
55.7
35.3
39.1
38.6
46.9
56.7
-3.1
42.5
27.8
26.5
29.9
31.5
43.5
72.4
82.6
68.7
53.1
63.8
107.5
86.2
FB
(%)
-66.5
-101.0
-19.6
-50.4
-91.4
-64.7
-104.0
7.2
-12.0
15.4
7.7
8.3
-6.5
-12.1
7.0
7.2
12.1
17.3
-0.7
4.7
2.9
30.8
-13.0
-0.8
5.6
-17.6
-4.2
-14.4
12.6
18.2
16.1
14.0
-0.1
31.5
17.1
FE
(%)
115.4
130.0
97.6
101.5
132.6
115.8
131.1
31.3
46.0
29.7
25.6
33.9
35.3
46.6
45.7
58.1
40.1
43.4
43.4
54.8
59.7
0.8
41.1
28.4
26.1
35.1
35.9
41.4
56.1
61.3
53.0
47.5
52.2
66.1
62.7
Final Risk and Exposure Assessment
Appendix 1-24
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002
Elemental
Carbon
(continued)
Organic Carbon
IMPROVE
STN
IMPROVE
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
8282
10069
2056
599
1795
1532
9493
9726
2903
2641
1447
2474
2504
2408
8287
10082
2057
598
1800
1531
9508
NMB
(%)
48.8
-14.1
-4.1
-31.2
-39.9
-31.1
-15.5
-45.8
-37.6
-35.2
-50.5
-50.6
-49.1
-36.3
-41.3
-34.8
-21.8
-49.6
-51.9
-53.6
-34.5
NME
(%)
-23.5
67.2
51.8
39.4
47.0
50.4
67.8
57.5
60.3
58.2
60.6
56.0
56.2
61.4
59.0
60.0
58.8
54.9
59.8
63.3
59.6
FB
(%)
-34.1
-29.5
-15.0
-39.3
-53.6
-38.8
-31.3
-50.4
-40.4
-30.3
-51.1
-60.7
-61.2
-37.9
-49.0
-31.2
-16.9
-52.5
-80.9
-71.0
-29.7
FE
(%)
56.1
62.1
50.6
51.3
62.2
61.2
62.7
72.8
69.3
66.8
75.1
73.4
75.6
70.2
70.6
63.0
57.4
66.4
87.8
85.1
61.9
Table 1.2-4. CMAQ v4.6 2002 Model Performance Statistics for Nitric Acid and Sulfur Dioxide
CMAQ 2002
Nitric Acid
Sulfur Dioxide
CASTNet
CASTNet
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
3172
1157
839
662
1086
229
1117
3174
1157
839
663
1086
229
1117
NMB
(%)
10.0
6.1
3.8
9.2
17.7
2.8
5.5
33.08
26.5
24.7
35.3
39.0
91.6
19.4
NME
(%)
39.1
47.7
34.2
38.8
44.0
31.3
48.5
46.61
71.4
39.4
47.9
51.6
99.9
68.1
FB
(%)
9.0
17.0
4.4
3.0
20.5
2.4
17.2
31.6
14.9
29.7
38.7
29.8
48.5
12.9
FE
(%)
40.3
49.0
35.4
40.4
44.0
31.3
49.5
46.4
57.7
40.8
45.3
51.4
61.0
57.2
Final Risk and Exposure Assessment
Appendix 1-25
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002
Sulfur Dioxide
(continued)
Nitrogen
Dioxide
SEARCH
AQS
12-kmEUS
12-kmEUS
Northeast
Midwest
Southeast
Central
No. of
Obs.
61760
921032
620014
14913
342422
625924
NMB
(%)
28.1
13.6
-2.9
12.3
7.3
34.3
NME
(%)
98.4
54.7
45.2
44.0
57.2
70.8
FB
(%)
24.4
4.9
-13.0
5.8
2.0
10.9
FE
(%)
87.1
56.5
52.5
43.8
60.8
61.1
Table 1.2-5. CMAQv4.6 2002 Model Performance Statistics for Annual Sulfate and Nitrate Wet
Deposition Model Performance Statistics.
CMAQ 2002 Annual
Sulfate
Nitrate
Ammonium
NADP
NADP
NADP
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
6896
2354
1405
1413
1856
1246
2058
6896
2354
1413
1405
1856
1246
2058
6896
2354
1413
1405
1856
1246
2058
NMB
(%)
5.53
-18.0
8.2
17.0
4.8
-10.2
-10.5
-16.2
-42.4
-7.1
-9.5
-12.5
-25.8
-45.4
-11.5
-40.4
-8.0
-2.0
1.1
-24.4
-41.1
NME
(%)
63.7
73.5
57.6
62.9
68.6
62.3
76.6
59.2
72.3
56.8
52.7
64.3
58.9
72.0
65.2
73.1
61.9
63.1
75.2
60.9
74.2
FB
(%)
1.0
-23.5
3.0
16.4
1.4
-2.5
-20.5
-22.2
-59.4
-6.0
-13.6
-14.8
-24.1
-59.8
-10.2
-55.5
0.4
1.3
-0.2
-16.5
-54.8
FE
(%)
70.9
94.7
59.7
64.0
73.2
71.6
94.7
73.3
99.2
66.0
60.7
74.2
72.8
99.3
76.0
100.6
68.5
67.3
77.9
75.0
100.6
Final Risk and Exposure Assessment
Appendix 1-26
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
Table 1.2-6. CMAQv4.6 Model Performance Statistics for Hourly Ozone Concentrations
CMAQ 2002 Hourly Ozone:
Threshold of 40 ppb
May
June
July
August
September
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of Obs.
241185
124931
51055
55859
69073
41728
111385
256263
125662
61354
54515
67867
46026
109157
257076
116785
66774
59360
68619
36021
104321
235090
125575
53837
54179
62506
41456
110225
179156
99710
44678
34285
41627
41549
83921
NMB (%)
-0.7
-3.7
1.2
3.3
-2.5
-6.4
-3.9
-7.5
-8.37
-8.46
-7.19
-7.2
-10.0
-8.8
-5.3
-12.0
-3.9
-10.5
-3.6
-3.6
-13.6
-8.7
-7.91
-6.4
-10.8
-9.4
-9.3
-8.5
-9.9
-10.7
-8.7
-11.4
-8.2
-12.8
-11.7
NME (%)
15.9
15.9
17.1
16.2
14.1
17.3
16.1
16.8
17.7
17.3
17.9
15.3
17.5
18.2
17.7
21.5
17.0
19.4
16.5
18.7
21.8
17.8
20.1
16.7
19.1
17.3
18.7
20.6
17.2
19.0
16.3
18.5
16.5
18.8
20.0
FB (%)
-2.0
-5.0
-0.3
2.4
-3.1
-9.2
-5.2
-9.0
-9.3
-9.9
-8.3
-7.6
-13.5
-9.9
-6.6
-14.9
-4.8
-12.3
-3.9
-6.3
-16.8
-10.2
-10.2
-7.4
-12.4
-9.9
-12.8
-11.1
-11.8
-12.7
-10.6
-12.9
-9.0
-16.6
-13.8
FE (%)
17.1
17.3
18.2
16.9
14.8
20.3
17.6
18.6
19.1
19.1
19.6
16.3
21.2
19.7
19.2
24.3
18.0
21.7
17.2
21.1
24.9
19.7
22.1
18.0
21.4
18.5
22.4
22.8
19.5
21.1
18.4
20.4
17.8
22.8
22.1
Final Risk and Exposure Assessment
Appendix 1-27
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002 Hourly Ozone:
Threshold of 40 ppb
Seasonal Aggregate
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
1168770
592663
277698
258198
309692
206780
519009
NMB
(%)
-6.4
-8.4
-5.4
-7.3
-6.0
-8.6
-9.2
NME
(%)
17.1
18.8
16.9
18.3
15.9
18.2
19.3
FB
(%)
-7.7
-10.3
-6.5
-8.4
-6.4
-11.9
-11.2
FE
(%)
18.8
20.7
18.4
20.0
16.8
21.6
21.3
Table 1.2-7. CMAQv4.6 Model Performance Statistics for 8-hour Daily Maximum Ozone
Concentrations
CMAQ 2002 8-Hour
Maximum Ozone:
Threshold of 40 ppb
May
June
July
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
18115
9188
4123
5596
5257
2921
1199
18341
8788
4038
5388
4901
3137
7789
19499
8809
4300
5612
5397
2799
1242
NMB (%)
4.0
0.1
7.1
13.9
0.5
0.4
-1.7
-4.3
-4.9
-3.0
-0.8
-5.5
-6.1
-5.4
-0.9
-7.5
-5.5
2.0
-0.6
2.1
-14.1
NME (%)
12.7
12.6
13.1
19.7
11.1
12.3
11.7
12.3
14.0
12.7
14.7
11.9
12.4
14.5
13.5
17.0
14.0
13.6
13.2
14.3
20.3
FB (%)
4.4
0.5
7.4
14.6
0.8
0.8
-1.6
-3.8
-4.3
-2.0
2.6
-5.0
-6.2
-4.8
-0.5
-8.2
-4.9
3.1
-0.1
1.9
-16.1
FE (%)
12.7
12.8
12.9
19.7
11.2
12.5
11.8
12.5
14.2
12.8
16.4
12.1
13.0
14.7
13.6
18.0
14.2
13.7
13.2
14.6
22.2
Final Risk and Exposure Assessment
Appendix 1-28
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
CMAQ 2002 8-Hour
Maximum Ozone:
Threshold of 40 ppb
August
September
Seasonal Aggregate
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
No. of
Obs.
18204
9516
3956
5620
4882
3104
8283
14921
8138
3075
5439
3540
3026
6864
89080
44626
19492
27655
23977
14987
38823
NMB
(%)
-4.4
-2.9
-7.3
2.6
-5.8
-3.6
-3.2
-5.7
-6.7
-7.0
-0.1
-3.7
-8.4
-7.4
-2.3
-4.3
o o
-J.J
3.2
-3.0
o o
-J.J
-5.0
NME
(%)
13.0
15.8
13.5
14.3
13.5
13.0
16.1
12.6
15.0
13.4
14.2
12.6
13.7
15.9
12.8
14.9
13.3
15.1
12.5
13.2
15.3
FB
(%)
-3.8
-3.2
-6.7
4.7
-4.9
-3.7
-3.7
-5.5
-6.9
-6.4
2.6
-3.1
-8.8
-7.7
-1.7
-4.3
-2.3
5.6
-2.3
-3.3
-5.0
FE
(%)
13.1
16.1
13.7
14.8
13.5
13.6
16.4
13.0
15.5
13.4
15.6
12.9
14.5
16.4
13.0
15.3
13.4
16.1
12.5
13.6
15.7
2.0 REFERENCES
Aiyyer, A, Cohan, D., Russell, A., Stockwell, W., Tanrikulu, S., Vizuete, W., and Wilczak, J.
2007. Final Report: Third Peer Review of the CMAQ Model.
Byun, D.W., and Schere, K.L. 2006. Review of the Governing Equations, Computational
Algorithms, and Other Components of the Models-3 Community Multiscale Air Quality
(CMAQ) Modeling System. J. Applied Mechanics Reviews, 59 (2), 51-77.
Dennis, R.L., Byun, D.W., Novak, J.H., Galluppi, K.J., Coats, C.J., and Vouk, M.A. 1996. The
next generation of integrated air quality modeling: EPA's Models-3. Atmospheric
Environment, 30, 1925-1938.
Final Risk and Exposure Assessment
Appendix 1-29
September 2009
-------
Community Multiscale Air Quality (CMAQ) Model
Dentener F. J., and Crutzen PJ. 1993. Reaction of ^Os on tropospheric aerosols: impact on the
global distributions of NOX, O3, and OH. JGeophysRes, 98, 7149-7163.
U.S. EPA (Environmental Protection Agency). 1999. Science Algorithms of EPA Models-3
Community Multiscale Air Quality (CMAQ) Modeling System. Byun, D.W., and Ching,
J.K.S., Eds. EPA/600/R-99/030. Office of Research and Development.
U.S. EPA (Environmental Protection Agency). 2008. Technical Support Document for the Final
Locomotive/Marine Rule: Air Quality Modeling Analyses. EPA 454/R-08-002. Office of
Air Quality Planning and Standards, Research Triangle Park, NC. January 2008.
U.S. EPA (Environmental Protection Agency). 2006a. Technical Support Document for the Final
Clean Air Interstate Rule: Air Quality Modeling. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. March 2005 (CAIR Docket OAR-2005-0053-
2149).
U.S. EPA (Environmental Protection Agency). 2006b. Technical Support Document for the
Final PMNAAQS Rule. Office of Air Quality Planning and Standards, Research Triangle
Park, NC.
Final Risk and Exposure Assessment September 2009
Appendix 1-30
-------
September 2009
Appendix 2
Supplemental Deposition Information
Wet Deposition Trends in or Near Case Study Areas and
Supplemental Nationwide Maps Depicting the Ratio of
Deposition to Concentration and Deposition to Emissions
Final
Prepared by
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27709
-------
-------
Trends in Wet Deposition at NADP Sites
TABLE OF CONTENTS
1.0 Trends in Wet Deposition of Inorganic Nitrogen and Sulfate at National
Atmospheric Deposition Program Sites in or Near Case Study Areas 1
2.0 Supplemental Nationwide Maps Depicting the Ratio of Deposition to
Concentration and Deposition to Emissions 17
LIST OF FIGURES
Figure 1-1. Adirondack Case Study Area; Site: Huntington Wildlife Forest, NY
NTN = National Trends Network 1
Figure 1-2. Adirondack Case Study Area; Site: Whiteface Mountain, NY 2
Figure 1-3. Hubbard Brook Experimental Forest Case Study Area; Site: Hubbard
Brook, NH 2
Figure 1-4. Kane Experimental Forest Case Study Area; Site: Kane Experimental
Forest, PA 3
Figure 1-5. Potomac River/Potomac Estuary Case Study Area; Site: Arendtsville, PA 3
Figure 1-6. Potomac River/Potomac Estuary Case Study Area; Site: Parsons, WV 4
Figure 1-7. Potomac River/Potomac Estuary Case Study Area; Site: Wye, MD 4
Figure 1-8. Shenandoah Case Study Area; Site: Shenandoah National Park, VA 5
Figure 1-9. Shenandoah Case Study Area; Site: Horton's Station, VA 5
Figure 1-10. Neuse River/Neuse River Estuary Case Study Area; Site: Finley Farm, NC 6
Figure 1-11. Neuse River/Neuse River Estuary Case Study Area; Site: Beaufort, NC 6
Figure 1-12. Rocky Mountain National Park Supplemental Area; Site: Beaver
Meadows, CO 7
Figure 1-13. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area; Site:
Yosemite National Park, CA 7
Figure 1-14. Mixed Conifer Forest (Transverse Range) Case Study Area; Site: Joshua
Tree National Park, CA 8
Figure 1-15. Mixed Conifer Forest (Transverse Range) Case Study Area; Site: Tanbark
Flat, CA 8
Figure 1-16. Adirondack Case Study Area; Site: Huntington Wildlife Forest, NY 9
Figure 1-17. Adirondack Case Study Area; Site: Whiteface Mountain, NY 9
Figure 1-18. Hubbard Brook Experimental Forest Case Study Area; Site: Hubbard
Brook, NH 10
Figure 1-19. Kane Experimental Forest Case Study Area; Site: Kane Experimental
Forest, PA 10
Figure 1-20. Potomac River/Potomac Estuary Case Study Area; Site: Arendtsville, PA 11
Figure 1-21. Potomac River/Potomac Estuary Case Study Area; Site: Parsons, WV 11
Figure 1-22. Potomac River/Potomac Estuary Case Study Area; Site: Wye, MD 12
Figure 1-23. Shenandoah Case Study Area; Site: Shenandoah National Park, VA 12
Figure 1-24. Shenandoah Case Study Area; Site: Horton's Station, VA 13
Figure 1-25. Neuse River/Neuse River Estuary Case Study Area; Site: Finley Farm, NC 13
Figure 1-26. Neuse River/Neuse River Estuary Case Study Area; Site: Beaufort, NC 14
Final Risk and Exposure Assessment September 2009
Appendix 2 - i
-------
Trends in Wet Deposition at NADP Sites
Figure 1-27. Rocky Mountain National Park (Supplemental Area); Site: Beaver
Meadows, CO 14
Figure 1-28. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area; Site:
Yosemite National Park, CA 15
Figure 1-29. Mixed Conifer Forest (Transverse Range) Case Study Area; Site: Joshua
Tree National Park, CA 15
Figure 1-30. Mixed Conifer Forest (Transverse Range) Case Study Area; Site: Tanbark
Flat, CA 16
Figure 2-1. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual average
sulfur dioxide concentrations (|ig/m3) 18
Figure 2-2. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual average
sulfur dioxide concentrations (|ig/m3) 19
Figure 2-3. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual average
sulfur dioxide concentrations (|ig/m3) 20
Figure 2-4. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual total sulfur
dioxide emissions (tons/yr) 21
Figure 2-5. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual total sulfur
dioxide emissions (tons/yr) 22
Figure 2-6. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual total
sulfur dioxide emissions (tons/yr) 23
Figure 2-7. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to
annual average nitrogen dioxide concentrations (ppb) 24
Figure 2-8. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to
annual average nitrogen dioxide concentrations (ppb) 25
Figure 2-9. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to
annual average nitrogen dioxide concentrations (ppb) 26
Figure 2-10. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to
annual total nitrogen dioxide emissions (tons/yr) 27
Figure 2-11. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to
annual total nitrogen dioxide emissions (tons/yr) 28
Figure 2-12. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to
annual total nitrogen dioxide emissions (tons/yr) 29
Figure 2-13. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb) 30
Figure 2-14. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb) 31
Figure 2-15. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual
average nitrogen dioxide concentrations (ppb) 32
Figure 2-16. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual total
nitrogen dioxide emissions (tons/yr) 33
Figure 2-17. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual total
nitrogen dioxide emissions (tons/yr) 34
Figure 2-18. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual
total nitrogen dioxide emissions (tons/yr) 35
Final Risk and Exposure Assessment September 2009
Appendix 2 - ii
-------
Trends in Wet Deposition at NADP Sites
1.0 TRENDS IN WET DEPOSITION OF INORGANIC
NITROGEN AND SULFATE AT NATIONAL ATMOSPHERIC
DEPOSITION PROGRAM SITES IN OR NEAR CASE STUDY
AREAS
Trends in Wet Deposition of Inorganic Nitrogen
(Arrow on plots indicate the year 2002)
NADP/NTN Site NY20
Annual inorganic N wet depositions, 1978-2007
ra
€
OJ
0 — 4—<—>~f—I—t—r—*-
197719791981 1983 (935 198? 1989 1S91 1993 1395 1«? IflM 2QOt 20032005200?
* Ms! criteria A Did nol meal ci-ilens / Trend line
Figure 1-1. Adirondack Case Study Area; Site: Huntington Wildlife Forest, NY
NTN = National Trends Network
Final Risk and Exposure Assessment
Appendix 2-1
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site NY98
8
6
re
JC 4
D)
2
0
19
Annual inorganic N wet depositions, 1984-2007
* * * ;
« X ^~~**\ * v%^ »
* ^v*^ \*»«^-»^^*1%—*^^^^**\. /"**" ^*
* *
A
•
! * i
83 1^35 t9S7 liSS 1991 1993 1115 1997 1899 2001 2003 2005 2007
* Mel criteria A Did noi meel tniena /Trend line
Figure 1-2. Adirondack Case Study Area; Site: Whiteface Mountain, NY
Fi
NADP/NTN Site NH02
Annual inorganic N wet depositions, 1978-2007
8
6
§ 4
1
0
19
r
* *
— x» /**\^jf^v * ^
*
, A
I { I 1 ) ' I 1 i I I 1 [ I L I I 1 j 1
7719791931 1 983 1 985 1 987 1 989 1 331 1 993 1 995 1 997 1 999 3001
* Met criteria 4 Did nol meet criteria / Trend line
gure 1-3. Hubbard Brook Experimental Forest Case Study Area; Site
r
» *
*
I ! I E I
2003 2007
Hubbard Brook, NH.
Final Risk and Exposure Assessment
Appendix 2-2
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site PA29
Annual inorganic N wet depositions, 1978-2007
A A
1977197919811983198519871989199119931 f 35 199? 1919 2001 3003 2005 200?
* Met criteria 4 Did nQtmeentteria /Trend line
Figure 1-4. Kane Experimental Forest Case Study Area; Site: Kane Experimental Forest, PA.
NADP/NTN Site PAOO
i
i
m
-E 1
O»
2
0
19
Annual inorganic N wet depositions, 1999-2007
*
. i
I I I I I I I I ;
98 2000 2002 2004 2008
# Met criteria /Trend line
Figure 1-5. Potomac River/Potomac Estuary Case Study Area; Site: Arendtsville, PA
Final Risk and Exposure Assessment
Appendix 2-3
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTNSiteWV18
Annual inorganic N wet depositions, 1978-2007
in
a
i
fW
J&£ 4
2
0
H
r
*
.K /\ * >;
• v*^^^^-
* 1 *
4
i i i I i I i I i i i I i i i i i i i i s i i i t I
77 1 979 1 981 1 983 1 985 1 987 1 989 1 991 1 993 1 995 1 997 1 993 2001 2007
» Met criteria 4 Did not meet criteria /Trend line
Figure 1-6. Potomac River/Potomac Estuary Case Study Area; Site: Parsons, WV
NADP/NTN SiteMD13
Annual inorganic N wet depositions, 1983-2007
8 r
6
fS
JC 4
2
0
HE
*
* 9
* * * * *
I
32 liS4 1986 1988 1990 19S2 1994 1996 1*3% 2000 2002 2004 2006 2008
* Met criteria 4 Did not meet tntena /Trend line
Figure 1-7. Potomac River/Potomac Estuary Case Study Area; Site: Wye, MD
Final Risk and Exposure Assessment
Appendix 2-4
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site VA28
Annual inorganic N wet depositions, 1981-2007
8
i
Of
I4
2
0
1i
r
* *
A
A A -A *X^^\ *
•
A A
.
4
L [ 1 i | ^ _j J j J , j j | j j | I
80 19B2 1984 1986 1988 1990 1992 1994 1996 1998 2004 2BOB
* Met criteria A Did not meet trtefia /Trend line
Figure 1-8. Shenandoah Case Study Area; Site: Shenandoah National Park, VA
NADP/NTN Site VA1 3
Annual inorganic N wet depositions, 1978-2007
E
5
4
m
=E 3
OS
2
1:
0
19
-
*
t
* . ** ' * •' f ^
* 4 1
A
A
| - l | t .._ ,.| ..| t f.-^---) ^ .. .) ..-.,.. j j ..j. _.j...i. ... | _,. j... . .\
??1t?&liS1 198319851387 19891931 1933 1395 1997 1999 2001
* Met criteria 4 Did tint rnesi . tritenfl /Trend line
Figure 1-9. Shenandoah Case Study Area; Site: Horton's Station, VA
Final Risk and Exposure Assessment
Appendix 2-5
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site NC41
Annual inorganic N wet depositions, 1978-2007
10
8
B
IS
JjjH ^
2
0
n
r
*
* * •
* >"X, »V jf*~**~^-^^*
. ^ * * * t
t 1 III 1 1 j 1 I i i t i 1 * i i ' ' 1 1 i- t
:?7 1 979 1 981 1 983 1 985 1 987 1 989 1 991 1 993 ! 395 1 997 1 398 2001 2003 2007
» Met criteria A Did not meet criteria /Trend line
Figure 1-10. Neuse River/Neuse River Estuary Case Study Area; Site: Finley Farm, NC
NADP/NTN Site NC06
™
4 .
3
«
JC 2
1 -
0
Annual inorganic N wet depositions, 1999-2007
* ^
. ' ' f ' * .
IS 2000 2002 3004 2008 2008
-------
Trends in Wet Deposition at NADP Sites
F
NADP/NTNSiteCO19
Annual inorganic N wet depositions, 1980-2007
4
3
«
-t 1
i
0
fi
-
-
«
- * * * — _/^"*
^** * y^-*-^^** '
-A *
4
79 t98t 1983 1985 1987 1989 1991 1993 1995 1997
» Met criteria 4 Did not meet criteria /Trend
igure 1-12. Rocky Mountain National Park Supplemental Area;
NADP/NTN Site CA99
#
/ * *
>t/* *
!
2001 2003 20D5 2007
ine
Site: Beaver Meadows, CO
Annual inorganic N wet depositions, 1981-2007
3 r *
4
3
_C
J** 2
1
0
19
A * A
-
" '- \ -A
• .• . \. -v
4 *
' [ ' 1 ' ' j J J 1 i 1
80 1S82 1984 1986 1988 1990 1992 1994 1996 1998 2000
9 Mel criteria A Did not meet criteria /Trend
*
/-^
i * •
(ill ' t '
2002 2004
line
Figure 1-13. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area;
Site: Yosemite National Park, CA
Final Risk and Exposure Assessment
Appendix 2-7
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site CA67
Annual inorganic N wet depositions, 2000-2007
wr
=E
o>
2.5 r
20
1 5
10
0.5
00
1999 2001 2003 2005
• Met criteria A Did noirneelrateia /Trend line
200?
Figure 1-14. Mixed Conifer Forest (Transverse Range) Case Study Area;
Site: Joshua Tree National Park, CA
NADP/NTN Site CA42
Annual inorganic N wet depositions, 1982-2007
^f
I
^-4—^-4-+.
1981 1B83 1386 1187 1989 1191 1§§3 1995 1997 1999 2001 2003 3005 2001
» Met criteria A Dies not meet criteria /Trend line
Figure 1-15. Mixed Conifer Forest (Transverse Range) Case Study Area;
Site: Tanbark Flat, CA
Final Risk and Exposure Assessment
Appendix 2-8
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site NY20
Annual S04 wet depositions, 1978-2007
30 r
25
20
15
10
* »
"7^* I
( ) ' 1 \
f r f 1 i- 1
M t
0
101"? 19-731981 18831985196? 18891981 199319961997 19992001 20032005200?
• Met criteria A Did not meet criteria /Trend line
Figure 1-16. Adirondack Case Study Area; Site: Huntington Wildlife Forest, NY
NADP/NTN Site NY98
Annual S04 wet depositions, 1984-2007
o>
30 r
25
20
15
10
* »
• *
1983 1985 198? 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007
* Met criteria A Did nol meet criteria / Trend line
Figure 1-17. Adirondack Case Study Area; Site: Whiteface Mountain, NY
Final Risk and Exposure Assessment
Appendix 2-9
September 2009
-------
Trends in Wet Deposition at NADP Sites
m
NADP/NTN Site NH02
Annual SO4 wet depositions, 1978-2007
30'
20
10
o' i. i I i I i ' i I i I i I i I i I I i i i i i i i i I t i
1377 1 979 19B1 1 983 1 985 1 987 1 989 1 991 1993 1995 1997 1993 2001 2007
» Met criteria
A Did not meet criteria / Trend line
Figure 1-18. Hubbard Brook Experimental Forest Case Study Area;
Site: Hubbard Brook, NH.
NADP/NTN Site PA29
Annual SO4 wet depositions, 1978-2007
60 r
40
W
JC 3®
era
20
10
4-
1077 1 973 1981 1 983 1 985 1 987 1 989 1 991 1993 1995 1997 1999 2001 2005 2007
» Met criteria A. Did not meet criteria /Trend line
Figure 1-19. Kane Experimental Forest Case Study Area; Site: Kane Experimental Forest, PA.
Final Risk and Exposure Assessment
Appendix 2 - 10
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site PAOO
Annual SO4 wet depositions, 1999-2007
3D
25
20
«f
-C 15
10
5
0
19
„
*
*
-
i I I I I i I i *
i& 2000 2002 2004 2008
* Met criteria /Trend line
Figure 1-20. Potomac River/Potomac Estuary Case Study Area; Site: Arendtsville, PA
NADP/NTN SiteWV18
Annual SO4 wet depositions, 1978-2007
60
50
40
30
20
10
i
* 4
197719791981 13831986 188? 1989 1 if1 19931995199719992001 200320052001
* Met criteria 4 Old not /Trend line
Figure 1-21. Potomac River/Potomac Estuary Case Study Area; Site: Parsons, WV
Final Risk and Exposure Assessment
Appendix 2-11
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTNSiteMD13
Annual SO4 wet depositions, 1983-2007
40 r
ns
10
•
I I I I I I I I I I I i
I I I I I i I
1384 1986 1986 1990 1392 1194 1998 2000 2002 2004 2006 2008
» Met criteria 4 CM dot mirM enter ia /Trend line
Figure 1-22. Potomac River/Potomac Estuary Case Study Area; Site: Wye, MD
NADP/NTN Site VA28
Annual SO4 wet depositions, 1981-2007
3D
25
20
m
Jn*
ID
5
0
li
A »
*AI A • *
* . | .
-
A
80 1SS2 1934 1 986 1188 1990 1312 1394 1998 1998 2000 2002 "MM MOB 2008
* Met criteria A. Did not meet criteria / Trend line
Figure 1-23. Shenandoah Case Study Area; Site: Shenandoah National Park, VA
Final Risk and Exposure Assessment
Appendix 2-12
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN SiteVA13
Annual SO4 wet depositions, 1978-2007
25
20
15
Itf
JseJ 10
5
0
19
* A *
A * * * A
4 * * * T
-
i ' i ' i i I i i i i i i i i i i i i ' '
77 1 979 1 981 1 983 1 985 1 987 1 989 1 991 1 993 1 995 1 997 1 993 2001 2007
» Met criteria 4 Did not meet criteria /Trend line
Figure 1-24. Shenandoah Case Study Area; Site: Horton's Station, VA
NADP/NTN Site NC41
Annual SO4 wet depositions, 1978-2007
3D
25
20
w
JE 15
O1
10
5
0
11
*
• 4
• t— /\ • 'N
\^A *\ * * * * *
: ' ' " ' i' ' .
" A
1 .,. |_ j_. | j. .| f--i— 1 j-- - t -1 -} ! -3— ' "3 • •'- - 1- '!• i l-t--1
11 1 ITS 1 BS1 1 383 1 9S6 1 187 1 9S9 1 391 1913 1 395 1 9i7 (9ii 2001 3003 2005 2007
• Met criteria 4 Did nrt meet criteria / Trend line
Figure 1-25. Neuse River/Neuse River Estuary Case Study Area; Site: Finley Farm, NC
Final Risk and Exposure Assessment
Appendix 2 - 13
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site NC06
Annual SO4 wet depositions, 1999-2007
20
15
to.
JC 10
5
0
19
-
A
* ^ . *
- ^ ' I
III ill1
98 2000 2002 2004
» Met criteria A Did not meet criteria /Trend line
Figure 1-26. Neuse River/Neuse River Estuary Case Study Area; Site: Beaufort, NC
NADP/NTN Site CO19
Annual SO4 wet depositions, 1980-2007
6 r
1
-h
-+•
0
11?i 1981 1983 1985 198? 1989 1991 1033 1395 19iF 1999 2001 20052007
• Met criteria 4 DidnDtrneeltTitena /Trend line
Figure 1-27. Rocky Mountain National Park (Supplemental Area); Site: Beaver Meadows, CO
Final Risk and Exposure Assessment
Appendix 2-14
September 2009
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site CA99
Annual SO4 wet depositions, 1981-2007
8 r
CO
0
\
* * * • f *
H—f-
i i . i
H—f- ,!,!,,...
4-
1380 1iB2 1984 1986 1988 1990 1992 1994 1996 1998 2000 2004 3006 100B
• Met criteria A Did not meet crt en a /Trend line
Figure 1-28. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area;
Site: Yosemite National Park, CA
NADP/NTN Site CA67
Annual SO4 wet depositions, 2000-2007
3,0
2.5
20
fa
•C 15
1.0
0.5
0.0
1i
-
*
-
. /^s.
/
_ — * *
* I
. Jt_ __l_ j | .., j A... J
0s 2001 2003 mm imi
-------
Trends in Wet Deposition at NADP Sites
NADP/NTN Site CA42
Annual SO4 wet depositions, 1982-2007
o '—i
Jill i !
-H-h
-I—I—J—1—P—t—f-
1383 1985 1987 1989 1991 1993 1995 1937 2001 2003 2005 2007
• Met criteria A Did not meet criteria /Trend line
Figure 1-30. Mixed Conifer Forest (Transverse Range) Case Study Area;
Site: Tanbark Flat, CA
Final Risk and Exposure Assessment
Appendix 2 - 16
September 2009
-------
Trends in Wet Deposition at NADP Sites
2.0 SUPPLEMENTAL NATIONWIDE MAPS DEPICTING THE
RATIO OF DEPOSITION TO CONCENTRATION AND
DEPOSITION TO EMISSIONS
This section of Appendix 2 contains the following maps (shown by figure number):
2-1. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (|ig/m3)
2-2. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (|ig/m3)
2-3. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (|ig/m3)
2-4. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual total sulfur dioxide
emissions (tons/yr).
2-5. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual total sulfur dioxide
emissions (tons/yr)
2-6. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual total sulfur
dioxide emissions (tons/yr)
2-7. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb)
2-8. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb)
2-9. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb)
2-10. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual total
nitrogen dioxide emissions (tons/yr)
2-11. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to annual
total nitrogen dioxide emissions (tons/yr)
2-12. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
total nitrogen dioxide emissions (tons/yr)
2-13. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb)
2-14. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb)
2-15. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb)
2-16. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual total nitrogen
dioxide emissions (tons/yr)
2-17. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual total nitrogen
dioxide emissions (tons/yr)
2-18. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual total
nitrogen dioxide emissions (tons/yr).
Final Risk and Exposure Assessment September 2009
Appendix 2 - 17
-------
Trends in Wet Deposition at NADP Sites
These maps were created using CMAQ-predicted 12 x 12 km concentrations and
deposition, and CMAQ gridded emissions inputs data.
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio ol annual total dry sulfur tteposfton to annual average SO2 concentration
Figure 2-1. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (|ig/m3).
Final Risk and Exposure Assessment
Appendix 2-18
September 2009
-------
Trends in Wet Deposition at NADP Sites
.
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River/NeuseEstuary
B Kane Experimental Fores!
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Fores! (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio ofanmat tolat wef suffw deposition to annual a^e/a^e $O2 concentration
Figure 2-2. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (jag/m3).
Final Risk and Exposure Assessment
Appendix 2 - 19
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Polomac River /Potomac Estuary
4 Neuse River/Neuse Estuary
5 Kane Experimental Fores)
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Fores! (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio QfiBVBad tola! wef+cfty sulfur eftprafitan to annual average SO2 concenrraftw
Figure 2-3. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual average sulfur
dioxide concentrations (jag/m3).
Final Risk and Exposure Assessment
Appendix 2 - 20
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
*i.o
» 1.0to<30
'_ >=3 Qto <5G
! >=50to<100
>= 10 Ota < 25 0
! =•= 25.0 to < 50 0
>= 50.0 lo* 750
>= 75.0 10 < 100.0
>= 100.0
1 Adirondack
2 Shenandoah
3 Polomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual iota! dry sutfw deposition to annual total SQ2 emissions
Figure 2-4. Ratio of annual total dry sulfur deposition (kg S/ha/yr) to annual total sulfur dioxide
emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2 -21
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
BB - i.o
• >= 1 Oto<30
'_ >=3 Qto <5G
I I >= 5 0 to < 10 0
>= 10 Ota < 25 0
I 1 « 25.0 to < 50 0
•I >= 50.0 lo* 750
H >= 75.010 < 100.0
• >= 100.0
1 Adirondack
2 Shenandoah
3 Polomac River/Potomac Estuary
4 Neuse River / Neuse Esluary
6 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annuat total wet sulfur deposition to annual toiat SO2 emissions
Figure 2-5. Ratio of annual total wet sulfur deposition (kg S/ha/yr) to annual total sulfur dioxide
emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2 - 22
September 2009
-------
Trends in Wet Deposition at NADP Sites
<1.0
» 1.0 to < 3.0
! >=3 Oto<5.0
' >= 5 0 to < 10 0
>= 10.010 < 25 0
! =•= 25.010 < 50 0
| -•>= 50.010 * 75 0
| »= 75.Olos 100.0
>= 1000
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River / Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual total wef+ctry sutfur opposition to annual total SO2 emissions
Figure 2-6. Ratio of annual total wet+dry sulfur deposition (kg S/ha/yr) to annual total sulfur
dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2-23
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Potomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
annual total wet oxidized nitrogen deposition fo annual average NO2 concentrations
Figure 2-7. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2 - 24
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
05
= 0.5 to < 1 0
= 1.0 to < 2.0
= 2.0 to < 3.0
= 3.0 to < A 0
= 4.0 lo < 5 0
= 5 0 to < 6 0
= 6 0 (o < 7 0
= 70
1 Adirondack
2 Shenandoah
3 Potomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
annual total dry oxidized nitrogen cteposttjon to annual average A/02 concenttations
Figure 2-8. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2-25
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Potomac River/ Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual tota! wef-Kfry owtftzert nitrogen deposition to antiual av&rage NO2 concentrations
Figure 2-9. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
average nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2-26
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
I <1 0
jul.DtoOO
; >=3 Oto<5.0
' >= 5 0 to < 10.0
>= 10.010 < 25 0
! =•= 25.010 < 50 0
| >= 50.010 * 75 0
| »= 75 010*1000
>= 1000
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River / Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual total dry oxKfoetf mtfogsn tfetxisition to annual totat A/O2 emissions
Figure 2-10. Ratio of annual total dry oxidized nitrogen deposition (kg NOx/ha/yr) to annual
total nitrogen dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2-27
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
I *ro
| »• 1.0 to < 3.0
I >=3 Oto<5.0
j >= 5 0 to < 10.0
>= 10.010 < 25 0
} « 25.0 to * 50 0
| *= 50.0 lo* 750
| >= 75.0 to < 100.0
>= 1000
1 Adirondack
2 Shenandoah
3 Polomac River/Potomac Estuary
4 Neuse River/Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annwa/ tots! wet cwcteefl nitrogen deposition to annual tola! NO2 emissions
Figure 2-11. Ratio of annual total wet oxidized nitrogen deposition (kg NOx/ha/yr) to annual
total nitrogen dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2 - 28
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
I *1.0
j >= 1.0 to < 3.0
I >=3 Oto<5.0
j >= 5 0 to < 10.0
>= 10.010 < 25 0
} « 25.0 to * 50 0
| *= 50.0 lo* 750
| >= 75.0 to < 100.0
>= 1000
1 Adirondack
2 Shenandoah
3 Polomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual tots! wef+ctry owtfrzetf nitrogen deposition lo annual total NQ2 emissions
Figure 2-12. Ratio of annual total wet+dry oxidized nitrogen deposition (kg NOx/ha/yr) to
annual total nitrogen dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2 - 29
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Polomac River/ Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Foresl
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of aftnual total tfry nitrogen deposition to annual average WO2 concentrations
Figure 2-13. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2-30
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Potomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annua! tota) wet nitrogen deposition to annual average NO2 concentrations
Figure 2-14. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2-31
September 2009
-------
Trends in Wet Deposition at NADP Sites
1 Adirondack
2 Shenandoah
3 Potomac River/Potomac Estuary
4 Neuse River /Neuse Estuary
5 Kane Experimental Foresl
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual total wet+dry nitrogen deposition lo amttiaf average NO2 concentrations
Figure 2-15. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual average
nitrogen dioxide concentrations (ppb).
Final Risk and Exposure Assessment
Appendix 2-32
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
—
10
= 1.0to<3.0
= 3 Oto<5.0
: 5 0 10 < 10 0
= 10.010*250
= 25.010 « 50 0
= 50.010 * 75 0
= 75.0 los 100.0
= 100.0
1 Adirondack
2 Shenandoan
3 Potomac River / Potomac Estuary
4 Neuse River / Neuse Estuary
5 Kane Experimental Forest
6 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual total (try nitrogen deoosriton to annual total NO2 emissions
Figure 2-16. Ratio of annual total dry nitrogen deposition (kg N/ha/yr) to annual total nitrogen
dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2-33
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
^H'- vo
! >=3 Oto<5.0
; >= 5 0 to < 10 0
>= 10.010 < 25 0
1 =•= 25.010 < 50 0
| >= 50.010 * 75 0
| »= 75 Olos 1000
>= 1000
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River / Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
fatio of annual total wef nitfogan imposition to annual total NO2 emissions
Figure 2-17. Ratio of annual total wet nitrogen deposition (kg N/ha/yr) to annual total nitrogen
dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2-34
September 2009
-------
Trends in Wet Deposition at NADP Sites
Legend
|<1 0
jul.DtoOO
; >=3 Oto<5.0
' >= 5 0 to < 10 0
>= 10.010 < 25 0
! =•= 25.010 < 50 0
| >= 50.010 * 75 0
| »= 75 010*1000
>= 1000
1 Adirondack
2 Shenandoah
3 Potomac River / Potomac Estuary
4 Neuse River / Neuse Estuary
5 Kane Experimental Forest
5 Hubbard Brook Experimental Forest
7 Mixed Conifer Forest (Transverse Range)
8 Mixed Conifer Forest (Sierra Nevada Range)
9 Rocky Mountain National Park
ratio of annual total v/er+cfly nitrogen deposition to annual total N02 emissions
Figure 2-18. Ratio of annual total wet+dry nitrogen deposition (kg N/ha/yr) to annual total
nitrogen dioxide emissions (tons/yr).
Final Risk and Exposure Assessment
Appendix 2 - 35
September 2009
-------
-------
September, 2009
Appendix 3
Components of Reactive Nitrogen Deposition Based
on Average Deposition Over the Period 2002-2005
Dry Deposition from CMAQ/Wet Deposition from NADP
Final
Prepared by
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27709
-------
-------
Components of Reactive Nitrogen Deposition: 2002-2005
LIST OF FIGURES
Figure 1. Adirondack Case Study Area: 2002-2005 1
Figure 2. Hubbard Brook Experimental Forest Case Study Area: 2002-2005 1
Figure 3. Kane Experimental Forest Case Study Area: 2002—2005 2
Figure 4. Neuse River/Neuse River Estuary Case Study Area: 2002-2005 2
Figure 5. Potomac River/Potomac Estuary Case Study Area: 2002-2005 3
Figure 6. Shenandoah Case Study Area: 2002-2005 3
Figure 7. Rocky Mountain National Park (Supplemental Area): 2002-2005 4
Figure 8. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area: 2002-2005 4
Figure 9. Mixed Conifer Forest (Transverse Range) Case Study Area: 2002—2005 5
Final Risk and Exposure Assessment September 2009
Appendix 3 - i
-------
Components of Reactive Nitrogen Deposition: 2002-2005
[This page intentionally left blank.]
Final Risk and Exposure Assessment September 2009
Appendix 3 - ii
-------
Components of Reactive Nitrogen Deposition: 2002-2005
Re N - Dry
7%
Re N - Wet
23%
Ox N - Dry
35%
Ox N - Wet
35%
Figure 1. Adirondack Case Study Area: 2002-2005.
Re N - Dry
6%
Re N - Wet
19%
Ox N - Dry
40%
Ox N - Wet
35%
Figure 2. Hubbard Brook Experimental Forest Case Study Area: 2002-2005.
Final Risk and Exposure Assessment
Appendix 3-1
September 2009
-------
Components of Reactive Nitrogen Deposition: 2002-2005
Re N - Dry
4%
Re N - Wet
22%
Ox N - Dry
41%
Ox N - Wet
33%
Figure 3. Kane Experimental Forest Case Study Area: 2002—2005.
Re N - Dry
37%
Ox N - Dry
30%
Ox N - Wet
15%
Re N - Wet
18%
Figure 4. Neuse River/Neuse River Estuary Case Study Area: 2002-2005.
Final Risk and Exposure Assessment
Appendix 3-2
September 2009
-------
Components of Reactive Nitrogen Deposition: 2002-2005
Re N - Dry
15%
Re N - Wet
20%
Ox N - Dry
42%
Ox N - Wet
23%
Figure 5. Potomac River/Potomac Estuary Case Study Area: 2002-2005.
Re N - Dry
16%
Re N - Wet
18%
Ox N - Dry
42%
Ox N - Wet
24%
Figure 6. Shenandoah Case Study Area: 2002-2005.
Final Risk and Exposure Assessment
Appendix 3-3
September 2009
-------
Components of Reactive Nitrogen Deposition: 2002-2005
Re N - Dry
7%
Re N - Wet
36%
Ox N - Dry
32%
Ox N - Wet
25%
Figure 7. Rocky Mountain National Park (Supplemental Area): 2002-2005.
Re N - Dry
23%
Re N - Wet
18%
Ox N - Dry
46%
Ox N - Wet
13%
Figure 8. Mixed Conifer Forest (Sierra Nevada Range) Case Study Area: 2002-2005.
Final Risk and Exposure Assessment
Appendix 3-4
September 2009
-------
Components of Reactive Nitrogen Deposition: 2002-2005
Re N - Dry
14%
Re N - Wet
6%
Ox N - Wet
6%
Ox N - Dry
74%
Figure 9. Mixed Conifer Forest (Transverse Range) Case Study Area: 2002—2005.
Final Risk and Exposure Assessment
Appendix 3-5
September 2009
-------
-------
September 2009
Appendix 4
Aquatic Acidification Case Study
Final
Prepared by
U.S. Environmental Protection Agency
Office of Atmospheric Programs
Washington, DC
-------
-------
Aquatic Acidification Case Study
TABLE OF CONTENTS
Acronyms and Abbreviations vii
1.0 Purpose 1
2.0 Background 1
2.1 Acidification 1
2.2 Indicators of Acidification 2
2.3 Biological Response to Acidification and Acid Neutralizing Capacity 4
3.0 Case Studies 6
3.1 Surface Waters Acidification in the Eastern United States 6
3.2 Objectives 8
3.3 Adirondack Case Study Area 9
3.3.1 General Description 9
3.3.2 Levels of Air Pollution and Acidifying Deposition 10
3.3.3 Levels of Sulfate,Nitrate, and ANC Concentrations in Surface Water 10
3.4 Shenandoah Case Study Area 14
3.4.1 General Description 14
3.4.2 Levels of Air Pollution and Acidifying Deposition 15
3.4.3 Levels of Sulfate, Nitrate, and ANC Concentrations in Surface Water 15
4.0 Methods 19
4.1 Biological Response to Acidification 19
4.2 Past, Present, and Future Surface Water Chemistry—the MAGIC Modeling
Approach 23
4.3 Connecting Current Nitrogen and Sulfur Deposition to Acid-Base Conditions of
Lakes and Streams: The Critical Load Approach 24
4.3.1 Regional Assessment of Adirondack Case Study Area Lakes and
Shenandoah Case Study Area Trout Streams 27
4.3.1.1 Adirondack Case Study Area 27
4.3.1.2 Shenandoah Case Study Area 27
5.0 Re suits 29
5.1 Adirondack Case Study Area 29
5.1.1 Current and Preacidification Conditions of Surface Waters 29
5.1.2 ANC Inferred Condition—Aquatic Status Categories 31
5.1.3 The Biological Risk from Current Nitrogen and Sulfur Deposition:
Critical Load Assessment 33
5.1.4 Representative Sample of Lakes in the Adirondack Case Study Area 37
5.1.5 Recovery from Acidification Given Current Emission Reductions 38
5.2 Shenandoah Case Study Area 39
5.2.1 Current and Preacidification Conditions of Surface Waters 39
5.2.2 ANC Inferred Condition—Aquatic Status Categories 41
5.2.3 The Biological Risk from Current Nitrogen and Sulfur Deposition:
Critical Load Assessment 43
5.2.4 Regional Assessment of Trout Streams in the Shenandoah Case Study
Area 45
Final Risk and Exposure Assessment September 2009
Appendix 4 - i
-------
Aquatic Acidification Case Study
5.2.5 Recovery from Acidification Given Current Emission Reductions 45
6. References 47
ATTACHMENT A 1
1. Modeling Descriptions 1
1.1 MAGIC 1
1.1.1 Input Data and Calibration 2
1.1.2 Lake, Stream, and Soil Data for Calibration 2
1.1.3 Wet Deposition and Meteorology Data for Calibration 3
1.1.4 Wet Deposition Data (Reference Year and Calibration Values) 5
1.1.5 Dry, Cloud, and Fog Deposition Data and Historical
Deposition Sequences 6
1.1.6 Protocol for MAGIC Calibration and Simulation at Individual
Sites 7
1.1.7 Combined Model Calibration and Simulation Uncertainty 9
1.1.8 Results of the Uncertainty Analysis 10
1.2 Critical Loads: Steady-State Water Chemistry Models 15
1.2.1 Preindustrial Base Cation Concentration 18
1.2.2 F-factor 19
1.2.3 ANC Limits 20
1.2.4 Sea Salt Corrections 20
1.2.5 Critical load exceedance 21
1.2.6 Lake-to-Catchment-Area Ratios 23
1.2.7 Denitrification and N Immobilization in Soils 23
1.2.8 Uncertainty and Variability 23
1.2.8.1 Results of the uncertainty analysis 26
ATTACHMENTS 1
1. EMAP/TIME/LTM Programs 1
2. Temporally Integrated Monitoring of Ecosystems and Long-Term Monitoring
Programs 2
2.1 TIME Program 2
2.2 LTM Program 3
LIST OF FIGURES
Figure 2.3-1. (a) Number offish species per lake or stream versus acidity, expressed as
ANC for Adirondack Case Study Area lakes (Sullivan et al., 2006). (b)
Number offish species among 13 streams as a function of ANC in the
Shenandoah Case Study Area. Values of ANC are means based on
quarterly measurements from 1987 to 1994. The regression analysis shows
a highly significant relationship (p <.0001) between mean stream ANC
and the number offish species
Figure 3.1-1. Regions containing ecosystems sensitive to acidifying deposition in the
eastern United States (U.S. EPA, based onNAPAP, 2005)
Final Risk and Exposure Assessment September 2009
Appendix 4 - ii
-------
Aquatic Acidification Case Study
Figure 3.3-1. Annual average total wet deposition (kg/ha/yr) for the period 1990 to 2006
in SO42" (blue) and NO3" (green) from eight NADP/NTN sites in the
Adirondack Case Study Area (shown with linear trend regression lines) 10
Figure 3.3-2. Trends over time for annual average SO42- (blue) and NO3" (green)
concentrations and ANC (red) in 50 Adirondacks Long-Term Monitoring
monitored lakes in the Adirondack Case Study Area (shown with linear
trend regression lines). Both SC>42" and NCV concentrations have
decreased in surface waters by approximately 26% and 13%, respectively 11
Figure 3.3-3. Current yearly average for 2005 to 2006 (a) SO42"concentrations (ueq/L),
(b) NCV concentrations (ueq/L), and (c) ANC (ueq/L) in surface waters
from 88 monitored lakes in the Temporally Integrated Monitoring of
Ecosystems (38 Lakes) and Adirondacks Long-Term Monitoring (50
Lakes) networks in the Adirondack Case Study Area 13
Figure 3.4-1. Air pollution concentrations and wet deposition for the period 1990 to 2006
using a CastNET (SHN418) and seven NADP/NTN sites in the
Shenandoah Case Study Area, (a) Annual average atmospheric
concentrations (ug/m3) of 862 (blue), oxidized nitrogen (red), SO42"
(green), and reduced nitrogen (black), (b) Annual average total wet
deposition (kg/ha/yr) of SO42" (green) and NCV (blue) 15
Figure 3.4-2. Trends over time for annual average SO42" (blue) and NCV (green)
concentrations, and ANC (red) in 67 streams in the Surface Water
Acidification Study, Virginia Trout Stream Sensitivity Survey, and Long-
Term Monitoring programs in the Shenandoah Case Study Area (shown
with linear trend regression lines) 16
Figure 3.4-3. Current yearly average for 2005 to 2006 (a) SC>42" concentration, (b) NCV
concentration, and (c) ANC (ueq/L) in surface waters from 67 monitored
streams in the Surface Water Acidification Study, Virginia Trout Stream
Sensitivity, and Long-Term Monitoring network in the Shenandoah Case
Study Area 18
Figure 4.1-1. Relationship between summer and spring ANC values at LTM sites in New
England, the Adirondack Mountains, and the Northern Appalachian
Plateau. Values are mean summer values for each site for the period 1990
to 2000 (horizontal axis) and mean spring minima for each site for the
same time period. On average, spring ANC values are at least 30 ueq/L
lower than summer values 22
Figure 4.1-2. Number offish species per lake or stream versus ANC level and aquatic
status category (represented by color) for lakes in the Adirondack Case
Study Area (Sullivan et al., 2006). The five aquatic status categories are
described in Table 4.1-1 22
Figure 4.3-1. (a) The location of lakes in the Adirondack Case Study Area used for
MAGIC (red circles) and critical load (green circles) modeling, (b) The
location of streams in the Shenandoah Case Study Area used for both
MAGIC and critical load modeling 25
Figure 5.1-1. Average NCV concentrations (orange), SC>42" concentrations (red), and ANC
(blue) for the 44 lakes in the Adirondack Case Study Area modeled using
MAGIC for the period 1850 to 2050 29
Final Risk and Exposure Assessment September 2009
Appendix 4 - iii
-------
Aquatic Acidification Case Study
Figure 5.1-2. (a) N(V and (b) SC>42" concentrations (ueq/L) of preacidification (1860) and
current (2006) conditions based on hindcasts of 44 lakes in the
Adirondack Case Study Area modeled using MAGIC 31
Figure 5.1-3. Percentage of lakes in the five classes (Acute, Severe, Elevated, Moderate,
Low) for years 1860 (preacidification) and 2006 (current) conditions for
44 lakes modeled using MAGIC. Error bars indicate the 95% confidence
interval 32
Figure 5.1-4. ANC concentrations of preacidification (1860) and current (2006)
conditions based on hindcasts of 44 lakes in the Adirondack Case Study
Area modeled using MAGIC 33
Figure 5.1-5. Critical loads of acidifying deposition that each surface water location can
receive in the Adirondack Case Study Area while maintaining or
exceeding an ANC concentration of 50 ueq/L based on 2002 data.
Watersheds with critical load values <100 meq/m2/yr (red and orange
circles) are most sensitive to surface water acidification, whereas
watersheds with values >100 meq/m2/yr (yellow and green circles) are the
least sensitive sites 34
Figure 5.1-6. Critical load exceedances (red circles) in the Adirondack Case Study Area
based on the year 2002 deposition magnitudes for waterbodies where the
critical limit ANC concentration is 0, 20, 50, and 100 ueq/L, respectively.
Green circles represent lakes where deposition is below the critical load.
See Table 5.1-3 36
Figure 5.1-7. Percentage of lakes in each of the five classes of acidification (Acute,
Severe, Elevated, Moderate, Low) for the years 2006, 2020, and 2050 for
44 lakes modeled using MAGIC, where current emissions are held
constant. Error bars indicate the 95% confidence interval 38
Figure 5.2-1. Average N(V concentrations orange), SO42"concentrations (red), and ANC
(blue) levels for the 60 streams in the Shenandoah Case Study Area
modeled using MAGIC for the period 1850 to 2050 39
Figure 5.2-2. (a) N(V and (b) SC>42" concentrations (ueq/L) of 1860 (preacidification) and
2006 (current) conditions based on hindcasts of 60 streams in the
Shenandoah Case Study Area modeled using MAGIC 40
Figure 5.2-3. Percentage of streams in the five classes of acidification (Acute, Severe,
Elevated, Moderate, Low) for years 1860 (preacidification) and 2006
(current) conditions for 60 streams modeled using MAGIC. The number
of streams in each category is above the bar. Error bars indicate the 95%
confidence interval 42
Figure 5.2-4. ANC levels of 1860 (preacidification) and 2006 (current) conditions based
on hindcasts of 60 streams in the Shenandoah Case Study Area modeled
using MAGIC 42
Figure 5.2-5. Critical loads of surface water acidity for an ANC of 50 ueq/L for
Shenandoah Case Study Area streams. Each dot represents an estimated
amount of acidifying deposition (i.e., critical load) that each stream's
watershed can receive and still maintain a surface water ANC >50 ueq/L.
Watersheds with critical load values <100 meq/m2/yr (red and orange
circles) are most sensitive to surface water acidification, whereas
Final Risk and Exposure Assessment September 2009
Appendix 4 - iv
-------
Aquatic Acidification Case Study
watersheds with values >100 meq/m2/yr (yellow and green circles) are the
least sensitive sites 43
Figure 5.2-6. Critical load exceedances for ANC levels of 0, 20, 50, and 100 ueq/L for
Shenandoah Case Study Area streams. Green circles represent streams
where current nitrogen and sulfur deposition is below the critical load and
that maintain an ANC level of 0, 20, 50, and 100 ueq/L, respectively. Red
circles represent streams where current nitrogen and sulfur deposition
exceeds the critical load, indicating they are currently impacted by
acidifying deposition. See Table 5.2-3 44
Figure 5.2-7. Percentage of streams in the five categories of acidification (Acute, Severe,
Elevated, Moderate, Low) for the years 2006, 2020, and 2050 for 60
streams in the Shenandoah Case Study Area modeled using MAGIC. The
number of streams in each category is above the bar. Error bars indicate
the 95% confidence interval 46
Figure 1.1-1. Simulated versus observed annual average surface water SC>42", N(V. ANC,
and pH during the model calibration period for each of the 44 lakes in the
Adirondacks Case Study Area. The black line is the 1:1 line 12
Figure 1.1-2. Simulated versus observed annual average surface water SC>42", NOs". ANC,
and pH during the model calibration period for each of the 60 streams in
the Shenandoah Case Study Area. The black line is the 1:1 line 13
Figure 1.1-3. MAGIC simulated and observed values of ANC for two lakes in the
Shenandoah Case Study Area. Red points are observed data and the
simulated values are the line. The Root Mean Squared Error (RMSE) for
ANC was 11.8 ueq/L for Helton Creek and 4.0 ueq/L for Nobusiness
Creek 14
Figure 1.1-4. MAGIC simulated and observed values of ANC for two lakes in the
Shenandoah Case Study Area. Red points are observed data and the
simulated values are the line. The Root Mean Squared Error (RMSE) for
ANC was 11.8 ueq/L for Helton Creek and 4.0 ueq/L for Nobusiness
Creek 15
Figure 1.2-1. The depositional load function defined by the model 22
Figure 1.2-2. The inverse cumulative frequency distribution for Little Hope Pond. The x-
axis shows critical load exceedance in meq/ha/yr and y-axis is the
probability. The dashed lines represent zero exceedance. In the case of
Little Hope Pond, the dash line divides mostly the probability distribution
on the left hand side, indicating Little Hope Pond has a relative low
probability of being exceeded (0.3). Critical load and exceedances values
were based on a critical level of protection of ANC = 50 ueq/L 25
Figure 1.2-3. Coefficients of variation of surface water critical load for acidity CL(A) and
exceedances (EX(A)). Critical load and exceedances values were based on
a critical level of protection of ANC = 50 ueq/L 27
Figure 1.2-4. Probability of exceedance of critical load for acidity for 2002 28
LIST OF TABLES
Table 4.1-1. Aquatic Status Categories 20
Final Risk and Exposure Assessment September 2009
Appendix 4 - v
-------
Aquatic Acidification Case Study
Table 5.1-1. Estimated Average Concentrations of Surface Water Chemistry for 44 Lakes
in the Adirondack Case Study Area Modeled Using MAGIC for
Preacidification (1860) and Current (2006) Conditions 30
Table 5.1-2. Percentage of Lakes in the Five Aquatic Status Categories Based on Their
Surface Water ANC Levels for 44 Lakes Modeled Using MAGIC and 88
Lakes in the TIME/LTM Monitoring Network. Results Are for the
Adirondack Case Study Area for the Years 2005 to 2006 32
Table 5.1-3. Adirondack Case Study Area Critical Load Exceedances, Where Nitrogen
and Sulfur Deposition Is Larger Than the Critical Load for Four Different
ANC Critical Limit Thresholds, for 169 Modeled Lakes within the
TEVIE/LTM and EMAP Monitoring and Survey Programs. "No.
Exceedances" Indicates the Number of Lakes at or below the Given ANC
Critical Limit, and "% Lakes" Indicates the Total Percentage of Lakes at
or below the Given ANC Critical Limit 35
Table 5.1-4. Critical Load Exceedances for the Regional Population of 1,842 Lakes in the
Adirondack Case Study Area That Are from 0.5 to 2,000 ha in Size and at
Least 1 m in Depth for the Four Critical Limit ANC Levels (0, 20, 50, and
100 ueq/L). Estimates Use the Exceedances for the Subset of 169 Lakes
Using 2002 Deposition Magnitudes (Table 5.1-3) and Are Extrapolated to
the Full Population Based on the EMAP Lake Probability Survey of 1991
to 1994. "No. Exceedances" Indicates the Number of Lakes at or Below
the Given ANC Critical Limit; "% Lakes" Indicates the Total Percentage
of Lakes at or Below the Given ANC Critical Limit 37
Table 5.2-1. Estimated Average Concentrations of Surface Water Chemistry for 60
Streams in the Shenandoah Case Study Area Modeled Using MAGIC for
Preacidification (1860) and Current (2006) Conditions 41
Table 5.2-2. Percentage of Streams in the Five Aquatic Status Categories Based on Their
Surface Water ANC Concentrations for 60 Streams Modeled Using
MAGIC and 67 Streams in the SWAS-VTSSS LTM Network. Results are
for the Shenandoah Case Study Area for the Year 2006 41
Table 5.2-3. Critical Load Exceedances (Nitrogen + Sulfur Deposition > Critical Load)
for 60 Modeled Streams within the VTSSS LTM Program in the
Shenandoah Case Study Area. "No. Exceedances" Indicates the Number
of Streams at the Given ANC Limit; "% Streams" Indicates the Total
Percentage of Streams at the Given ANC Limit 45
Table 1.2-1. Parameters Used and their Uncertainty Range. The Range of Surface Water
Parameters (e.g., CA, MG, CL, NA, NOs, 864) were Determined from
Surface Water Chemistry Data for the Period from 1992 to 2006 from the
LTM-TEVIE Monitoring Network. Runoff(Q) and Acidic Deposition were
Set at 50% and 25% 24
Table 1.2-2. Means and Coefficients of Variation of Critical Loads and Exceedances for
Surface Water 26
Final Risk and Exposure Assessment September 2009
Appendix 4 - vi
-------
Aquatic Acidification Case Study
ACRONYMS AND ABBREVIATIONS
A13+
A1(OH)3
ALTM
ANC
ASTRAP
Ca2+
cr
CL(A)
CO2
DDF
EMAP
eq/ha/yr
F-
FAB
H+
H4SiO4
ha
ISA
K+
kg/ha/yr
km
LTM
m
m/yr
MAGIC
MAHA
meq/m2yr)
Mg2+
Na+
NADP
NH4+
NOX
NSWS
NTN
03
PnET-BGC
Si
SO2
SO42-
sox
SWAS
aluminum
aluminum hydroxide
Adirondack Long-Term Monitoring
acid neutralizing capacity
Advanced Statistical Trajectory Regional Air Pollution
calcium
chloride
critical loads of acidity
carbon dioxide
dry and occult deposition factor
Environmental Monitoring and Assessment Program
equivalents per hectare per year
fluoride
First-Order Acidity Balance
hydrogen ion
silicic acid
hectare
Integrated Science Assessment
potassium
kilograms/hectare/year
kilometer
Long-Term Monitoring
meter
meters/year
Model of Acidification of Groundwater in Catchments
Mid-Atlantic Highlands Assessment
milliequivalents per square meter per year
magnesium
sodium
National Atmospheric Deposition Program
ammonium
nitrate
nitrogen oxides
National Lake/Stream Surveys
National Trends Network
ozone
biogeochemical model
silicon
sulfur dioxide
sulfate
sulfur oxides
Shenandoah Nation Park Surface Water Acidification Study
Final Risk and Exposure Assessment
Appendix 4 - vii
September 2009
-------
Aquatic Acidification Case Study
SSWC Steady-State Water Chemistry
TIME Temporally Integrated Monitoring of Ecosystems
ueq/L microequivalents per liter
uM micrometer
VTSSS Virginia Trout Stream Sensitivity Survey
Final Risk and Exposure Assessment September 2009
Appendix 4 - viii
-------
Aquatic Acidification Case Study
1.0 PURPOSE
This case study is intended to estimate the ecological exposure and risk to aquatic
ecosystems from acidification associated with the deposition of nitrogen and sulfur for two
sensitive regions of eastern United States: the Adirondack Mountains in New York (hereafter
referred to as the Adirondack Case Study Area) and Shenandoah National Park and the
surrounding areas of Virginia (hereafter referred to as the Shenandoah Case Study Area).
2.0 BACKGROUND
2.1 ACIDIFICATION
Sulfur oxides (SOX) and nitrogen oxides (NOX) compounds in the atmosphere undergo a
complex mix of reactions and thermodynamic processes in gaseous, liquid, and solid phases to
form various acidic compounds. These acidic compounds are removed from the atmosphere
through deposition: either wet (e.g., rain, snow), fog or cloud, or dry (e.g., gases, particles).
Deposition of these acidic compounds leads to ecosystem exposure and effects on ecosystem
structure and function. Following deposition, these compounds can, in some instances, leach out
of the soils in the form of sulfate (SC>42") and nitrate (N(V), leading to the acidification of surface
waters. The effects on ecosystems depend on the magnitude of deposition, as well as a host of
biogeochemical processes occurring in the soils and waterbodies.
When sulfur or nitrogen leaches from soils to surface waters in the form of SO42" or NO3",
an equivalent amount of positive cations, or countercharge, is also transported. This maintains
electroneutrality. If the countercharge is provided by base cations, such as calcium (Ca2+),
magnesium (Mg2+), sodium (Na+), or potassium (K+), rather than hydrogen (H+) and dissolved
inorganic aluminum, the acidity of the soil water is neutralized, but the base saturation of the soil
is reduced. Continued SC>42" or N(V leaching can deplete the base cation supply of the soil. As
the base cations are removed, continued deposition and leaching of SC>42" and/or N(V (with H+
and A13+) leads to acidification of soil water, and by connection, surface water. Loss of soil base
saturation is a cumulative effect that increases the sensitivity of the watershed to further
acidifying deposition. Base cations are replenished through the natural weathering of the rocks
and soils, but weathering is a slow process, which results in the depletion of cations in the soil in
Final Risk and Exposure Assessment September 2009
Appendix 4-1
-------
Aquatic Acidification Case Study
the presence of SC>42" and/or N(V pollution. A watershed's ability to neutralize acidic deposition
is determined by a host of biogeophysical factors, including base cation concentrations,
weathering rates, uptake by vegetation, rate of surface water flow, soil depth, and bedrock.
Following deposition, SC>42" can absorb to, or bind with, soil particles, a process that
removes it from the aqueous soil solution, and therefore, prevents the leaching of base cations (at
least temporarily) and further acidifying of the soil water. This process results in an
accumulation of sulfur in the soil. This process is potentially reversible and can contribute to soil
acidification if, and when, the SC>42" is desorbed and released back into the soil water solution.
The degree to which SC>42" adsorbs on soil is dependent on soil characteristics. The locations of
soils in the United States that most effectively adsorb SC>42" are found south of the areas that
experienced glaciation during the most recent ice age (Rochelle and Church, 1987; Rochelle et
al., 1987). SC>42" adsorption is strongly pH-dependent, and a decrease in soil pH resulting from
acidifying deposition can enhance the ability of soil to adsorb SC>42". Consequently, as deposition
increases, the soil potentially stores a disproportionate amount of SC>42". When deposition
decreases, this stored SC>42" is slowly, but continually, released, keeping soil water acidified
and/or depleting the base cation supply.
2.2 INDICATORS OF ACIDIFICATION
The chemistry of the surface water is directly related to the biotic integrity of freshwater
ecosystems. There are numerous chemical constituents in surface water that can be used to
indicate the acidification condition of lakes and streams and to assess the effects of acidifying
deposition on ecosystem components. These include surface water pH (log[H+]) and
concentrations of SC>42", N(V, A13+, and Ca2+; the sum of base cations; the recently developed
base cation surplus; and the acid neutralizing capacity (ANC). Each of these chemical indicators
provides direct links to the health of individual biota and the overall health and integrity of
aquatic ecosystems as a result of surface water acidification.
Although ANC does not directly affect the health of biotic communities, it is calculated
(or measured) based on the concentrations of chemical constituents that directly contribute to or
ameliorate acidity-related stress, in particular, pH, Ca2+, and dissolved inorganic aluminum.
Furthermore, numerical models of surface water acidification can more accurately estimate ANC
than all of the individual constituents that comprise it. Consequently, for the purpose of this case
Final Risk and Exposure Assessment September 2009
Appendix 4-2
-------
Aquatic Acidification Case Study
study, annual average ANC of surface waters was used as the primary metric to quantify the
current acidic conditions and biological impacts for a subset of waterbodies in the study areas.
The remainder of this section focuses on a description of ANC.
ANC reflects the relative balance between base cations and strong acid anions. It
accounts for the cumulative effects of all of the ionic interactions that occur as the acidic
compounds are removed from the atmosphere to the catchment and drainage water to emerge in
a stream or lake. ANC of surface waters is defined (in this study) as the total amount of strong
base ions minus the total amount of strong acid anions:
ANC = (Ca2+ + Mg2++ K+ + Na+ + NH4+) - (SO42" + NCV + Cl') (1)
The unit of ANC is microequivalents per liter (ueq/L). If the sum of the equivalent
concentrations of the base cations exceeds those of the strong acid anions, then the ANC of a
waterbody will be positive. To the extent that the base cation sum exceeds the strong acid anion
sum, the ANC will be higher. Higher ANC is generally associated with high pH and Ca2+
concentrations, and lower ANC is generally associated with low pH and high dissolved inorganic
aluminum concentrations and a greater likelihood of toxicity to biota.
ANC samples from waterbodies are typically measured using the Gran titration approach.
Process-based numerical models, such as Model of Acidification of Groundwater in Catchments
(MAGIC) utilize the ANC calculated from the charge balance. For assessment purposes,
including resource characterization and Long-Term Monitoring (LTM) programs, it is always
best to use both directly measured and numerically estimated ANC values. The difference
between the two can be used to quantify uncertainty and reveal the influences of natural organic
acidity and/or dissolved inorganic aluminum on the overall acid-base chemistry of the water.
Relative to some individual chemical parameters, such as pH, ANC reflects sensitivity to
acidifying deposition input and effects on surface water chemistry in a linear fashion across the
full range of ANC values. Consequently, ANC is a preferred indicator variable for surface water
acidification. Other parameters, such as surface water pH, can complement the assessment of
surface water acidification; however, the response of this parameter to inputs is not necessarily
linear throughout its range. For example, at pH values >6.0, pH is not a good indicator of either
sensitivity to acidification or level of biological effect. In addition, pH measurements (especially
at these higher values) are sensitive to and can be confounded by the level of dissolved carbon
dioxide (CO2) in the water.
Final Risk and Exposure Assessment September 2009
Appendix 4-3
-------
Aquatic Acidification Case Study
2.3 BIOLOGICAL RESPONSE TO ACIDIFICATION AND ACID
NEUTRALIZING CAPACITY
Ecological effects occur at four levels of biological organization: (1) the individual; (2)
the population, which is composed of a single species of individuals; (3) the biological
community, which is composed of many species; and (4) the ecosystem. Low ANC
concentrations are linked with negative effects on aquatic systems at all four of these biological
levels. For the individual level, impacts are assessed in terms of fitness (i.e., growth,
development, and reproduction) or sublethal effects on condition. Surface water with low ANC
concentrations can directly influence aquatic organism fitness or mortality by disrupting ion
regulation and can mobilize dissolved inorganic aluminum, which is highly toxic to fish under
acidic conditions (i.e., pH <6 and ANC <50 ueq/L). For example, research showed that as the pH
of surface waters decreased to <6, many aquatic species, including fish, invertebrates,
zooplankton, and diatoms, tended to decline sharply causing species richness to decline
(Schindler, 1988). Van Sickle et al. (1996) also found that blacknose dace (Rhinichthy spp.) were
highly sensitive to low pH and could not tolerate inorganic Al concentrations greater than about
3.7 micromolar (uM) for extended periods of time. For example, they found that after 6 days of
exposure to high inorganic Al, blacknose dace mortality increased rapidly to nearly 100%.
At the community level, species richness and community structure can be used to
evaluate the effects of acidification. Species composition refers to the mix of species that are
represented in a particular ecosystem, whereas species richness refers to the total number of
species in a stream or lake. Acidification alters species composition and richness in aquatic
ecosystems. There are a number of species common to many oligotrophic waterbodies that are
sensitive to acidification and cannot survive, compete, or reproduce in acidic waters. In response
to small to moderate changes in acidity, acid-sensitive species are often replaced by other more
acid-tolerant species, resulting in changes in community composition and richness, but with little
or no change in total community biomass. The effects of acidification are continuous, with more
species being affected at higher degrees of acidification. At a point, typically a pH <4.5 and an
ANC <0 ueq/L, complete to near-complete loss of many classes of organisms occur, including
fish and aquatic insect populations, whereas others are reduced to only a few acidophilic forms.
These changes in species integrity are because energy cost in maintaining physiological
Final Risk and Exposure Assessment September 2009
Appendix 4-4
-------
Aquatic Acidification Case Study
homeostasis, growth, and reproduction is high at low ANC levels (Schreck, 1981, 1982;
Wedemeger et al., 1990).
Decreases in species richness related to acidification have been observed in the
Adirondack Mountains and Catskill Mountains of New York (Baker et al., 1993), the Upper
Midwest of the United States (Schindler et al., 1989), New England and Pennsylvania (Haines
and Baker, 1986), and Virginia (Bulger et al., 2000). Studies on fish species richness in the
Adirondack Case Study Area demonstrated the effect of acidification; of the 53 fish species
recorded in Adirondack Case Study Area lakes, only 27 species were found in lakes with a pH
<6.0. The 26 species missing from lakes with a pH <6.0 include important recreational species,
such as Atlantic salmon, tiger trout (Salmo trutta X Salvelinus fontinalis)., redbreast sunfish
(Lepomis auritus)., bluegill (Lepomis macrochirus\ tiger musky (Esox masquinongy X lucius\
walleye (Sander vitreus), alewife (Alosapseudoharengus), and kokanee (Oncorhynchus nerkd)
(Kretser et al., 1989), as well as ecologically important minnows that are commonly eaten by
sport fish. A survey of 1,469 lakes in the late 1980s found 346 lakes to be devoid offish. Among
lakes with fish, there was a relationship between the number offish species and lake pH, ranging
from about one species per lake for lakes having a pH <4.5 to about six species per lake for lakes
having a pH >6.5 (Driscoll et al., 2001; Kretser et al., 1989).
These decreases in species richness due to acidifying deposition are positively correlated
with ANC concentrations (Kretser et al., 1989; Rago and Wiener, 1986). Most notably, Sullivan
et al. (2006) found a logistic relationship between fish species richness and ANC category for
Adirondack Case Study Area lakes (Figure 2.3-1, a), which indicates the probability of
occurrence of an organism for a given value of ANC. In addition, a similar relationship has been
found for the Shenandoah Case Study Area, where a statistically robust relationship between
ANC and fish species richness was documented (Figure 2.3-1, b). In fact, ANC has been found
in studies to be the best single indicator of the biological response and health of aquatic
communities in acid-sensitive systems (Lien et al., 1992; Sullivan et al., 2006).
Final Risk and Exposure Assessment September 2009
Appendix 4-5
-------
Aquatic Acidification Case Study
(a)
(b)
14
I 12
£
200 ueq/L and are not sensitive to the acidifying deposition of
NOX and SOX air pollution at their existing ambient concentration levels. Figure 3.1-1 shows the
acid-sensitive regions of the eastern United States with the potential for low surface water ANC,
as determined by geology and surface water chemistry.
Freshwater surveys and monitoring in the eastern United States have been conducted by
many programs since the mid-1980s, including the National Lake/Stream Surveys (NSWS),
EPA's Environmental Monitoring and Assessment Program (EMAP), the Temporally Integrated
Monitoring of Ecosystems (TIME) monitoring program (Stoddard, 1990), and LTM project
(Ford et al., 1993; Stoddard et al., 1998) (Appendix Attachment B). The purpose of these
programs is to determine the current state and document the trends over time in surface water
Final Risk and Exposure Assessment
Appendix 4-6
September 2009
-------
Aquatic Acidification Case Study
chemistry for regional populations of lakes or streams impacted by acidifying deposition. Based
on extensive surveys and surface water data from these programs, it was determined that the
most sensitive lakes and streams (i.e., ANC less than about 50 ueq/L) in the eastern United States
are found in New England, the Adirondack Mountains, the Appalachian Mountains (northern
Appalachian Plateau and Ridge/Blue Ridge region), northern Florida, and the Upper Midwest.
These areas are estimated to contain 95% of the lakes and 84% of the streams in the United
States that have been anthropogenically acidified through deposition (see Annex 4.3.3.2 of the
Integrated Science Assessment (ISA) for Oxides of Nitrogen and Sulfur-Ecological Criteria
(FinalReport) (ISA) (U.S. EPA, 2008).
Figure 3.1-1. Regions containing ecosystems sensitive to acidifying deposition in
the eastern United States (U.S. EPA, based on NAPAP, 2005).
The number and proportion of acidic waterbodies in these regions, defined as having a
pH <5.0 and ANC <0 ueq/L, are substantial. The Adirondack Case Study Area had a large
proportion of acidic surface waters (14%) in the NSWS; from 1984 to 1987, the Adirondack
Lakes Survey Corporation sampled 1,469 Adirondack Case Study Area lakes >0.5 hectares (ha)
in size and estimated that many more (26%) were acidic (Driscoll et al., 1991). The proportion of
lakes estimated by NSWS to be acidic was smaller in New England and the Upper Midwest (5%
Final Risk and Exposure Assessment
Appendix 4-7
September 2009
-------
Aquatic Acidification Case Study
and 3%, respectively), but because of the large number of lakes in these regions, there were
several hundred acidic waters in each of these two regions. The Shenandoah Case Study Area
had 5.5% and 6% acidic sites, respectively, based on data from the early 1990s. Portions of
northern Florida also contain many acidic and low-ANC lakes and streams, although the role of
acidifying deposition in these areas is less clear. In 2002, Stoddard et al. (2003) suggested that
-8% of lakes in the Adirondack Mountains and from 6% to 8% of streams in the northern
Appalachian Plateau and Ridge/Blue Ridge region were acidic at base-flow conditions.
The Adirondack Case Study Area and the Shenandoah Case Study Area provide ideal
areas to assess the risk to aquatic ecosystems from NOX and SOX acidifying deposition. Four
main reasons support the selection of these two areas. First, both regions fall within the areas of
the United States known to be sensitive to acidifying deposition because of a host of
environmental factors that make these regions predisposed to acidification. Second, these areas
are representative of other areas sensitive to acidification, which will allow the results of this
case study to be generalized. Third, these regions have in the past and continue to experience
substantial exposure to NOX and SOX air pollution. Fourth, these areas have been extensively
studied (e.g., from atmospheric concentrations, soil characteristics, surface water chemistry, to
the changes in biological communities in response to aquatic acidification) over the last 3
decades (see Section 4 of the ISA Report) (U.S. EPA, 2008). For example, extensive water
quality data exists from monitoring networks in operation since the 1980s, along with numerous
research studies that directly link the biological harm of individuals, populations, communities,
and ecosystems to aquatic acidification. The sections below describe each of the case studies
areas, in turn, indicating past impacts of acidifying deposition, and identifying research linking
biological and acidic conditions for each region.
3.2 OBJECTIVES
For the two case study areas, the Adirondack and the Shenandoah, conditions of the
aquatic ecosystems and responses to nitrogen and sulfur deposition were evaluated by using
multiple approaches that rely on monitoring data and modeled output. Current conditions were
evaluated by a three-step process:
Final Risk and Exposure Assessment September 2009
Appendix 4-8
-------
Aquatic Acidification Case Study
• By evaluating the status and trends of surface water chemistry data to establish linkages
between current ambient air pollution levels of nitrogen and sulfur and the total amount
of deposition
• By evaluating the biological risk to individuals, populations, and communities from
acidification
• By evaluating the response of the aquatic ecosystem to current and future deposition
compared with the likelihood for recovery of currently impacted aquatic waterbodies.
In evaluating these conditions, this case study addresses the welfare effects of
acidification by building linkages between ambient pollutant levels, deposition, surface water
chemistry, and the resulting response in the biological communities.
3.3 ADIRONDACK CASE STUDY AREA
3.3.1 General Description
The Adirondack Case Study Area is situated in northeastern New York and is
characterized by dense forest cover of evergreen and deciduous trees and abundant surface
waters, with 46 peaks that extend up to 1,600 meters. The Adirondack Case Study Area has long
been a nationally important recreation area for fishing, hiking, boating, and other outdoor
activities. The area includes the headlands of five major drainage basins: Lake Champlain and
the Hudson, Black, St. Lawrence, and Mohawk rivers, which all draw water from the preserve.
There are more than 2,800 lakes and ponds, and more than 1,500 miles of rivers that are fed by
an estimated 30,000 miles of brooks and streams. The Adirondack Case Study Area, particularly
its southwestern section, is sensitive to acidifying deposition because it receives high
precipitation amounts with high concentrations of pollutants, has shallow base-poor soils, and is
underlain by igneous bedrock with low weathering rates and buffering ability (Driscoll et al.,
1991; Sullivan et al., 2006). This case study area is among the most severely acid-impacted
regions in North America (Driscoll et al., 2003; Landers et al., 1988; Stoddard et al., 2003). It
has long been used as an indicator of the response of forest and aquatic ecosystems to changes in
emissions of sulfur dioxide (802) and NOX resulting, in part, from the Clean Air Act
Amendments of 1990 (NAPAP, 1998; U.S. EPA, 1995).
Final Risk and Exposure Assessment September 2009
Appendix 4-9
-------
Aquatic Acidification Case Study
3.3.2 Levels of Air Pollution and Acidifying Deposition
Wet deposition in the Adirondack Case Study Area has been monitored by the National
Atmospheric Deposition Program/National Trends Network (NADP/NTN) since 1978 at two
sites (i.e., Huntington Forest and Whiteface Mountain) and at seven other sites since the 1980s.
Since 1990, wet SO42" and NO3" deposition at these NADP/NTN sites in the Adirondack Case
Study Area has declined by about 45% and 40%, respectively (Figure 3.3-1). However, annual
total wet deposition is still >15 and 10 kilograms/hectare/year (kg/ha/yr) of SC>42" and NCV,
respectively.
re
I
1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: NADP
NADP Sites: NY08,NY20,NY52,NY68,NY98,NY99,VT01VT99
Figure 3.3-1. Annual average total wet deposition (kg/ha/yr) for the period 1990 to 2006 in
SC>42" (blue) and NCV (green) from eight NADP/NTN sites in the Adirondack Case Study Area
(shown with linear trend regression lines).
3.3.3 Levels of Sulfate, Nitrate, and ANC Concentrations in Surface Water
Environmental monitoring data reported above demonstrate decreasing trends in
depositional loading, reflecting decreases in air pollution. Figure 3.3-2 shows annual average
trends since 1990 in SC>42", NCV, and ANC in surface water from 50 lakes in the Adirondack
Final Risk and Exposure Assessment
Appendix 4-10
September 2009
-------
Aquatic Acidification Case Study
Case Study Area monitored through the Adirondack Long-Term Monitoring (ALTM) program.
As a result of decreases in air pollution and deposit!onal loading, regional annual average SC>42"
concentrations in these lakes has dropped by approximately 26% since the mid-1990s. While
inter-annual variability in annual average N(V concentrations is evident in the Adirondack Case
Study Area monitored lakes, the overall trend is modestly downward (13% over the entire
period). An increase in long-term annual average ANC of+0.8 ueq/L/yr has corresponded to the
declines in NCV and SC>42" However, this increase in ANC also correlates with reductions in
annual average base cations of calcium (Ca2+) and magnesium (Mg2+) during the same period of
time (data not shown). In the Adirondack Case Study Area, toxic levels of dissolved inorganic
aluminum also declined slightly (data not shown).
0)
120.0
100.0
80.0
60.0
40.0
20.0
0.0
Annual Average Suface Water Trends 1990-2006
(Adirondack LTM Lakes)
1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: ALTM -50 Lakes
Figure 3.3-2. Trends over time for annual average SC>4 - (blue) and N(V (green) concentrations
and ANC (red) in 50 Adirondacks Long-Term Monitoring monitored lakes in the Adirondack
Case Study Area (shown with linear trend regression lines). Both SC>42~ and NCV concentrations
have decreased in surface waters by approximately 26% and 13%, respectively.
Final Risk and Exposure Assessment
Appendix 4-11
September 2009
-------
Aquatic Acidification Case Study
Despite decreases in deposition and surface water concentrations of SC>42" and
levels remain elevated in monitored lakes. Figure 3.3-3 shows current yearly average (2005 to
2006) SO42", NO3", and ANC for 88 Adirondack Case Study Area lakes monitored through the
ALTM (50 Lakes) and TIME (38 lakes) programs. The yearly averages for the period from 2005
to 2006 of SO42", NO3", and ANC are 70.88 ± 19.87, 9.18 ± 10.51, and 33.84 ± 44.40 ueq/L,
respectively.
There is still a substantial number of lakes in the Adirondack Case Study Area that have
low ANC values (<50 ueq/L) based on the observed yearly average of ANC from the years 2005
and 2006 for the waterbodies in the TIME/ALTM monitoring networks. Of the 88 monitored
lakes, 24% have ANC values >50 ueq/L, whereas 76% of the monitored lakes have ANC values
<50 ueq/L. Of the 73 monitored lakes with >50 ueq/L, 17 are chronically acidic (ANC <0
ueq/L). Twenty-seven of the lakes have <20 ueq/L, making their biological communities
susceptible to episodic acidification.
Final Risk and Exposure Assessment September 2009
Appendix 4-12
-------
Aquatic Acidification Case Study
(a)
TIME/ALTM: 2005-2006
Sulfate (SO42')
Sulfate (Meq/L)
I Adirani
Source TIME/ALTM 200E
(b)
TIME/ALTM: 2005-2006
Nitrate (NO3)
| | Adirondack Boundary
Source TIME/ALTM 20QS
(c)
TIME/LTM: 2005-2006
ANC
ANC (ueq/L)
^] Adirondack Boundary
Source TIME/ALTM 2009
2-
Figure 3.3-3. Current yearly average for 2005 to 2006 (a) SC>4 "concentrations
(ueq/L), (b) N(V concentrations ((j,eq/L), and (c) ANC ((j,eq/L) in surface waters
from 88 monitored lakes in the Temporally Integrated Monitoring of Ecosystems
(38 Lakes) and Adirondacks Long-Term Monitoring (50 Lakes) networks in the
Adirondack Case Study Area.
Final Risk and Exposure Assessment
Appendix 4 - 13
September 2009
-------
Aquatic Acidification Case Study
3.4 SHENANDOAH CASE STUDY AREA
3.4.1 General Description
The Shenandoah Case Study Area straddles the crest of the Blue Ridge Mountains in
western Virginia, on the eastern edge of the central Appalachian Mountain region. Several areas
in Shenandoah National Park have been designated Class 1 Wilderness areas. Shenandoah
National Park is known for its scenic beauty, outstanding natural features, and biota. The Skyline
Drive, a scenic 165-kilometer (km) parkway, provides the opportunity for views of the Blue
Ridge Mountains and surrounding areas. The Appalachian National Scenic Trail is the backbone
of the park's trail system. The natural features and biota of the park include the well-exposed
rock strata of the Appalachians, which is one of the oldest mountain ranges in the world. The
park comprises one of the nation's most diverse botanical reserves and wildlife habitats. A
congressionally designated wilderness area within the park is the largest in the mid-Atlantic
states and provides a comparatively accessible opportunity for solitude, study, and experience in
a natural area.
Air pollution within the Shenandoah Case Study Area, including concentrations of sulfur,
nitrogen, and ozone (O3), is higher than in most other national parks in the United States. This
area is sensitive to acidifying deposition because of the noncarbonate composition and
weathering-resistant characteristics of much of the underlying bedrock, which result in base-poor
soils with low weathering rates and poor buffering capacity. At base flow conditions, Lynch and
Dise (1985) determined that stream water ANC, pH, and base cation concentrations in this region
are strongly correlated with bedrock geology. This landscape includes three major bedrock types:
siliceous (e.g., quartzite and sandstone), felsic (e.g., granitic), and mafic (e.g., basaltic). Each of
these bedrock types influence about one-third of the stream miles in this region. ANC
concentrations for streams associated with siliceous bedrock are extremely low. Almost half of
the sampled streams had ANC in the chronically acidic range (<0 (j,eq/L). The balance of the
streams associated with siliceous bedrock had ANC in the episodically acidic range (0 to 20
ueq/L). Consequently, this region is among the most severely acid-impacted areas in North
America (Stoddard et al., 2003; Webb et al., 2004).
Final Risk and Exposure Assessment September 2009
Appendix 4-14
-------
Aquatic Acidification Case Study
2-
3.4.2 Levels of Air Pollution and Acidifying Deposition
Annual average atmospheric concentrations of sulfur dioxide (802), NOX, SC>4Z", and
reduced nitrogen at the Big Meadow air monitoring location within the Shenandoah National
Park have all decreased since the 1990s, with the exception of reduced nitrogen (Figure 3.4-1,
a). As a result, wet deposition in the Shenandoah National Park monitored at 7 sites by the
NADP/NTN since the 1980s shows wet SC>42" and N(V deposition declining by about 28% and
20%, respectively (Figure 3.4-1, b). However, annual total deposition is still 15 and 10 kg/ha/yr
of SC>42" and N(V, respectively.
1990 1992 1994 1996 1998 2000 2002 2004 2006
Source:CASTNET
CASTNET Sites: SHN418
1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: NADP
NADP Sites: NY08,NY20,NY52,NY68,NY98,NY99,VT01VT99
Figure 3.4-1. Air pollution concentrations and wet deposition for the period 1990
to 2006 using a CastNET (SHN418) and seven NADP/NTN sites in the
Shenandoah Case Study Area, (a) Annual average atmospheric concentrations
(ug/m3) of SC>2 (blue), oxidized nitrogen (red), SC>42" (green), and reduced
nitrogen (black), (b) Annual average total wet deposition (kg/ha/yr) of SC>42"
(green) and N(V (blue).
3.4.3 Levels of Sulfate, Nitrate, and ANC Concentrations in Surface Water
Figure 3.4-2 shows trends in annual average surface water concentrations of SC>42" and
and ANC values for streams in the Shenandoah Case Study Area monitored through the
Shenandoah Nation Park Surface Water Acidification Study (SWAS), Virginia Trout Stream
Sensitivity Survey (VTSSS), and LTM programs. The annual average for 67 streams for the
period 2005 to 2006 of SO42", NO3", and ANC are 59.29 ± 28.2, 3.74 ± 7.0, and 60.60 ± 72.7
Final Risk and Exposure Assessment
Appendix 4-15
September 2009
-------
Aquatic Acidification Case Study
ueq/L, respectively (Figure 3.4-3). Despite the decreases in pollution and regional acidifying
deposition, SC>42" and N(V concentrations in these streams have seen slight improvements since
the mid-1990s. There is a slight decline in SC>42" concentrations (-0.04 ueq/L/yr) in surface
waters, whereas N(V declined by only -0.3 ueq/L/yr. On the other hand, average ANC values of
the 67 streams increased to 79 ueq/L until the year 2002, from about 50 ueq/L in the early 1990s.
However, since 2002, ANC levels have fluctuated by first declining back to early 1990s levels
and then increasing to 67 ueq/L in 2006. Despite improvement in deposition, surface water
concentrations of SC>42" and N(V levels remain elevated in monitored streams in the Shenandoah
Case Study Area.
120.0
100.0
Annual Average Suface Water Trends 1990-2006
(SWAS-VTSSS-LTM)
0.0
1990 1992 1994 1996
Source: SWAS-VTSSS-LTM-67 Streams
1998
2000
2002
2004
2006
Figure 3.4-2. Trends over time for annual average SC>42" (blue) and N(V (green)
concentrations, and ANC (red) in 67 streams in the Surface Water Acidification Study,
Virginia Trout Stream Sensitivity Survey, and Long-Term Monitoring programs in the
Shenandoah Case Study Area (shown with linear trend regression lines).
Final Risk and Exposure Assessment
Appendix 4 - 16
September 2009
-------
Aquatic Acidification Case Study
There are a significant number of the 67 streams in SWAS-VTSSS and LTM programs
that currently have ANC <50 ueq/L based on the observed annual average ANC concentrations
(Figure 3.4-3). Forty-five percent of all monitored streams have ANC values >50 ueq/L,
whereas 55% have <50 ueq/L. Of the 55% <50 ueq/L, 18% experience episodic acidification
(<20 ueq/L) and 12% are chronically acidic (<0 ueq/L) at the current level of acidifying
deposition and ambient concentrations of NOX and SC>2.
Final Risk and Exposure Assessment September 2009
Appendix 4 - 17
-------
Aquatic Acidification Case Study
SWAS-VTSSS/LTM: 2005 - 2006
Sulfate (SO42')
Source: SWAS-VTSSS/LTM 2009
Sulfate (|ioq/L)
0-25
• 25-50
50-75
75-100
• >100
Case Study Boundary
SWAS-VTSSS/LTM: 2005-2006 /^
Nitrate (NO3") /£
• X
Source SWAS-VTSSS/LTM 2009
SWAS-VTSSS/LTM: 2005 - 2006
ANC
Source. SWAS-VTSSS;LTM 2009
Figure 3.4-3. Current yearly average for 2005 to 2006 (a) SC>42" concentration, (b) N(V concentration, and (c) ANC (ueq/L) in
surface waters from 67 monitored streams in the Surface Water Acidification Study, Virginia Trout Stream Sensitivity, and
Long-Term Monitoring network in the Shenandoah Case Study Area.
Final Risk and Exposure Assessment
Appendix 4-18
September 2009
-------
Aquatic Acidification Case Study
4.0 METHODS
4.1 BIOLOGICAL RESPONSE TO ACIDIFICATION
Because there is a continuum in the relationship between ANC concentrations and
resulting biological effects, a range of ANC values related to specific biological effects is needed
for the following reasons:
(1) ANC values within a waterbody are not constant; there is variation with time and
season. For example, during spring, following snowmelt and the resulting influx of
acidifying compounds, surface water ANC levels can substantially drop. There is also
spatial uncertainty. Consequently, the length of exposure (i.e., chronic vs. episodic) can
affect biological responses.
(2) The biological effects of particular ANC values vary between individual organisms
because of differences in developmental stage and size, innate differences between
different species of the same general types of organisms, and differences between
different kingdoms, phyla, and classes of organisms.
Therefore, five categories of ANC values were used that link specific biological health
conditions to the effects of aquatic communities, ranging from no impacts to complete loss of
populations. These five classes are based on the relationships among ANC and ecological
attributes, including richness, diversity, community structure, and individual fitness of
organisms. The following paragraphs describe the biological impacts, given a range of ANC
values and the scientific research that supports the grouping. Section AX4 of the Annexes to the
ISA (U.S. EPA, 2008) presents a more in-depth description of the biological relationship used in
this case study.
For freshwater systems, ANC values are grouped into five major categories: Acute
Concern (<0 ueq/L; acidic), Severe Concern (0 to 20 ueq/L), Elevated Concern (20 to 50 ueq/L),
Moderate Concern (50 to 100 ueq/L), and Low Concern (>100 ueq/L), with each range
representing a probability of ecological damage to the community (Table 4.1-1).
Final Risk and Exposure Assessment September 2009
Appendix 4-19
-------
Aquatic Acidification Case Study
Table 4.1-1. Aquatic Status Categories
Category Label ANC Levels* Expected Ecological Effects
Acute
Concern
<0 ueq/L
(Acidic)
Near complete loss offish populations is expected. Planktonic
communities have extremely low diversity and are dominated by
acidophilic forms. The number of individuals in plankton species that
are present is greatly reduced.
Severe
Concern
0-20 ueq/L
Highly sensitive to episodic acidification. During episodes of high
acidifying deposition, brook trout populations may experience lethal
effects. Diversity and distribution of zooplankton communities decline
sharply.
Elevated
Concern
20-50 ueq/L
Fish species richness is greatly reduced (i.e., more than half of expected
species can be missing). On average, brook trout populations
experience sublethal effects, including loss of health, reproduction
capacity, and fitness. Diversity and distribution of zooplankton
communities decline.
Moderate
Concern
50-100
ueq/L
Fish species richness begins to decline (i.e., sensitive species are lost
from lakes). Brook trout populations are sensitive and variable, with
possible sublethal effects. Diversity and distribution of zooplankton
communities also begin to decline as species that are sensitive to
acidifying deposition are affected.
Low
Concern
>100 ueq/L
Fish species richness may be unaffected. Reproducing brook trout
populations are expected where habitat is suitable. Zooplankton
communities are unaffected and exhibit expected diversity and
distribution.
Low Concern - Biota is generally not harmed when ANC values are >100 ueq/L. For
example, the number offish species tend to peak at ANC values >100 ueq/L (Bulger et al., 1999;
Driscoll et al., 2001; Kretser et al., 1989; Sullivan et al., 2006). Typically, with ANC
concentrations >100 ueq/L, the diversity of the aquatic community is more influenced by other
environmental factors, such as habitat availability, than the acid-base balance of the surface
water.
Moderate Concern - At ANC levels 50 to 100 ueq/L, declines in the fitness and
recruitment of species sensitive to acidity (e.g., some fish and invertebrate organisms) have been
demonstrated and may result in decreases in community-level diversity as the few highly acid-
sensitive species are lost (see Figure 2.3-1). However, minimal (no measurable) change in total
community abundance or production generally occurs, resulting in good overall health of the
community.
Final Risk and Exposure Assessment
Appendix 4-20
September 2009
-------
Aquatic Acidification Case Study
Elevated Concern - When ANC values drop between 20 and 50 ueq/L, they are generally
associated with negative effects on the fitness and recruitment of aquatic biota. Kretser et al.
(1989) showed that a 50% reduction in the number offish species occurred when ANC values
dropped to <50 ueq/L in lakes that were surveyed. Furthermore, Dennis and Bulger (1995)
showed that when ANC values drop to between 20 and 50 ueq/L, the overall fitness of most fish
species are reduced, such as sensitive species of minnows and daces (e.g., fathead minnow and
blacknose dace), and recreation fish species (e.g., lake trout and walleye). In addition to the
changes in the fish community, a drop in ANC values can cause some loss of common
invertebrate species from zooplankton and benthic communities, which include many species of
snails, clams, mayflies, and amphipods. These losses of sensitive species often result in distinct
decreases in species richness and changes in species composition of the biota. However, the total
community abundance or production remains high, with little if any change.
Severe Concern - When ANC levels drop <20 ueq/L, almost all biota exhibit some level
of negative effects. Fish and plankton diversity and the structure of the communities continue to
decline sharply to levels where acid-tolerant species begin to outnumber all other species
(Driscoll et al., 2001; Matuszek and Beggs, 1988). Loss of several important sport fish species is
possible, including lake trout, walleye, and rainbow trout, and losses of additional nongame
species, such as creek chub, occur. In addition, several other invertebrate species, including all
snails, most slams, and many species of mayflies, stoneflies, and other benthic invertebrates, are
lost or greatly reduced in population size, which further depresses species composition and
community richness. Also, at <20 ueq/L, surface waters are susceptible to episodic acidification,
and a total loss of biota can occur when ANC concentration goes to <0 ueq/L for a short period
of time. Stoddard et al. (2003) showed that to protect biota from episodic acidification in the
spring, base flow (i.e., summer nonstorm event) ANC levels need to have an ANC of at least 30
to 40 ueq/L (Figure 4.1-1).
Acute Concern-Near complete loss offish populations and extremely low diversity of
planktonic communities occur with ANC levels <0 ueq/L. Only acidophilic species are present,
but their population numbers are sharply reduced. For example, lakes in the Adirondack Case
Study Area have been shown to be fishless when the average ANC is <0 ueq/L (Sullivan et al.,
2006). A summary of the five categories of ANC and expected ecological effects can be found in
Figure 4.1-2 and Table 4.1-1
Final Risk and Exposure Assessment September 2009
Appendix 4 -21
-------
Aquatic Acidification Case Study
200
o
New England Lakes
Adirondack Lakes
Appalachian Streams
-50 0 50 100 150
Mean Summer ANC (peq/L)
200
Figure 4.1-1. Relationship between summer and spring ANC values at LTM sites
in New England, the Adirondack Mountains, and the Northern Appalachian
Plateau. Values are mean summer values for each site for the period 1990 to 2000
(horizontal axis) and mean spring minima for each site for the same time period.
On average, spring ANC values are at least 30 ueq/L lower than summer values.
Severe Elevated Moderate
14 -,
8 12 -
8 10 -
0.
^*F
/. '
^MM-
T
.
•'.:
:j-
"
-200 -100 0 100 200
Low
'
* "
i
^^ ^^ "
T
•
T
3OO 400 500
ANC(peq/L)
Figure 4.1-2. Number offish species per lake or stream versus ANC level and aquatic
status category (represented by color) for lakes in the Adirondack Case Study Area
(Sullivan et al., 2006). The five aquatic status categories are described in Table 4.1-1.
Final Risk and Exposure Assessment
Appendix 4 - 22
September 2009
-------
Aquatic Acidification Case Study
4.2 PAST, PRESENT, AND FUTURE SURFACE WATER
CHEMISTRY—THE MAGIC MODELING APPROACH
The preacidification condition of a waterbody is rarely known because historical
measurements are not available. Likewise, it is also difficult to determine if a waterbody has or
will recover from acidification as acidifying deposition inputs decline, because recovery may
take many years to decades to occur. For these reasons, biogeochemical models, such as
MAGIC, enable estimates of past, present, and future water chemistry levels that can be used to
evaluate the associated risk and uncertainty of the current levels of acidification compared with
estimated preacidification conditions and to evaluate whether a system will recover as a result of
reduction in acidifying deposition.
Dynamic hydrological models use surface water measurements of multiple parameters
from the long-term record, information about the current exposure (i.e., ambient pollutant
concentrations, deposition estimates), and known/measurable biogeochemical factors to
characterize a watershed and estimate its preindustrial (i.e., preacidification) state and to estimate
its response to changes in deposition in the future.
In both case study areas, MAGIC was used to estimate the past (i.e., preacidification,
1860), present (i.e., the years 2002 and 2006), and future (i.e., the years 2020 and 2050) acidic
conditions of 44 lakes in the Adirondack Case Study Area and 60 streams in the Shenandoah
Case Study Area (Figure 4.3-1). Fewer lakes and streams were modeled with MAGIC because
calibration data are not available for all sites in the ALTM and SWAS-VTSSS/LTM monitoring
networks. Furthermore, MAGIC was used to quantify the associated uncertainty in these
estimates, as well as in input parameters used in MAGIC. The MAGIC model output for each
waterbody was summarized into five ANC levels that correspond to the aquatic status categories
in Table 4.1-1. This grouping permits an assessment of the risk to the biological communities for
each of the conditions. The hydrological model, MAGIC, along with all the necessary inputs and
calibration procedure, is described in detail in Appendix Attachment A.
Final Risk and Exposure Assessment September 2009
Appendix 4-23
-------
Aquatic Acidification Case Study
4.3 CONNECTING CURRENT NITROGEN AND SULFUR
DEPOSITION TO ACID-BASE CONDITIONS OF LAKES AND
STREAMS: THE CRITICAL LOAD APPROACH
Critical loads were calculated for lakes and streams in both case study areas. The critical
load for a lake or stream provides a means to gauge the extent to which a waterbody has
recovered from past acidifying deposition or is potentially at risk because of current deposition
levels. The critical load approach provides a quantitative estimate of the level of exposure to one
or more pollutants, below which significant harmful effects on specific sensitive elements of the
environment do not occur, according to present knowledge.
The critical load approach relates specific amounts of deposition to particular ANC levels
for individual waterbodies, using the relationships established between the biogeochemical state
of the environment, current pollutant deposition, the surface water chemistry, and the response of
the biological communities to deposition. Conversely, it is possible to specify a "critical limit"
ANC level and to estimate the "critical load" of deposition required to cause the stream to have
that specified ANC level. The past, current, or estimated future levels of deposition can be
compared with the critical load estimate. For example, a critical limit ANC value of 50 ueq/L
could be specified for a particular stream or lake. The amount of deposition that the stream or
lake could take and maintain an ANC of 50 ueq/L would be its critical load. Clearly, if the
critical limit ANC value is lower (20 ueq/L), the critical load would increase—it would take
more deposition to lower the stream's ANC to that new value.
A critical load estimate is analogous to a "susceptibility" estimate, relating the sensitivity
of the waterbody to become acidified from the deposition of nitrogen and sulfur to the critical
limit ANC concentration. Low critical load values (e.g., less than 50 milliequivalents per square
meter per year (meq/m2/yr)) mean that the watershed has a limited ability to neutralize the
addition of acidic anions, and hence, it is at risk or susceptible to acidification and the resulting
deleterious effects. The greater the critical load value, the greater the ability of the watershed to
neutralize the additional acidic anions and resist acidification, thereby protecting the aquatic
ecosystem.
Final Risk and Exposure Assessment September 2009
Appendix 4-24
-------
Aquatic Acidification Case Study
(a)
Adirondack Case Study Area
• MAGIC Locations
* Crritical Locations
Adirondack Boundary
Watershed Boundary
Source: EPA 2009
(b)
Shenandoah Case Study Area
• MAGIC & Critical Load Locations
| | Case Study Boundary
Watersheds
Source: EPA 2009
Figure 4.3-1. (a) The location of lakes in the Adirondack Case Study Area used for
MAGIC (red circles) and critical load (green circles) modeling, (b) The location of
streams in the Shenandoah Case Study Area used for both MAGIC and critical load
modeling.
Final Risk and Exposure Assessment
Appendix 4-25
September 2009
-------
Aquatic Acidification Case Study
Applied at locations over a region, the critical load approach provides a method to
quantify the number of lakes or streams in a given area that receive harmful levels of deposition.
The magnitude of the biological harm is defined by the critical limit ANC level (e.g., ANC of
50 ueq/L) (see Table 4.1-1). Critical load exceedance (i.e., the amount of actual deposition
above the critical load, if any) can be calculated for each waterbody in the region. Lakes and
streams with positive exceedance values, where actual deposition was above its critical load, are
not protected at that critical limit, whereas negative exceedance values, where deposition of
nitrogen and sulfur was below its critical load, are protected.
Critical loads and their exceedances were calculated for four critical limit thresholds (i.e.,
ANC of 0, 20, 50, and 100 ueq/L), separating the five ANC categories of biological protections
(Table 4.1-1) for 169 lakes in the Adirondack Case Study Area and 60 streams in the
Shenandoah Case Study Area. There are numerous methods and models that can be used to
calculate critical loads for acidity. Drawing on the peer-reviewed scientific literature (Dupont et
al., 2005), this case study used a steady-state critical load model that uses surface water
chemistry as the base for calculating the critical load. A combination of the Steady-State Surface
Water Chemistry (SSWC) and First-Order Acidity Balance (FAB) models were used to calculate
the critical load. This analysis uses water chemistry data from the TIME and LTM programs that
are part of the Environmental Monitoring and Assessment Program (see discussion of surface
water trends in Section 5.1.1). This case study focuses on the combined load of sulfur and
nitrogen deposition, below which the ANC level would still support healthy aquatic ecosystems.
For each waterbody, the actual current total deposition in the year 2002 was compared with the
estimated critical loads for the four critical limit thresholds to determine which sites exceed their
critical limit of deposition and biological protection level. Estimates of actual current deposition
were based on the sum of measured wet deposition values from the year 2002 NADP network
and modeled dry deposition values based on the year 2002 emissions and meteorology using the
Community Multiscale Air Quality (CMAQ) model, respectively.
The actual deposition was compared with the critical load for each of the waterbodies
within the case study areas for each of the critical limit levels, and exceedances were determined.
Results for an individual lake were grouped by whether or not the lake exceeded its critical load.
For each of two case study areas, the number and percentage of lakes that receive acidifying
deposition above their critical load for each of the ANC critical limits of 0, 20, 50, and 100 ueq/L
Final Risk and Exposure Assessment September 2009
Appendix 4-26
-------
Aquatic Acidification Case Study
were determined. The critical load models and their inputs are described in detail in Appendix
Attachment A.
4.3.1 Regional Assessment of Adirondack Case Study Area Lakes and Shenandoah
Case Study Area Trout Streams
4.3.1.1 Adirondack Case Study Area
In the Adirondack Case Study Area, critical load exceedances were extrapolated to lakes
defined by the New England EMAP probability survey. The EMAP probability survey was
designed to estimate, with known confidence, the status, extent, change, and trends in condition
of the nation's ecological resources, such as surface water quality. In probability sampling, the
inclusion probability for each sampled lake represents a proportion of the target population.
Lakes selected with relatively high probability represent relatively few lakes in the population;
therefore, they carry relatively low weight and influence the final inferences less than lakes
selected with low probability. These inclusion probabilities (i.e., weighting or expansion factors)
are used to infer or estimate population frequency distributions and to evaluate sampling
uncertainty.
For the Adirondack Case Study Area, the regional EMAP probability survey of 117 lakes
(i.e., weighting factors) were used to infer the number of lakes and percentage of lakes that
receive acidifying deposition above their critical load of a target population of 1,842 lakes. The
target population of 1,842 lakes represents all lakes from 0.5 to 2,000 ha with a depth of
>1 meter (m) and > 1,000 m2 of open water in the Adirondack Case Study Area. ANC limits of
20, 50, and 100 ueq/L were examined.
The 117 lakes in the regional Adirondack probability survey represent a subset of 344
sampled lakes throughout New England (e.g., lakes in Maine, New Hampshire, Vermont, Rhode
Island, Massachusetts, Connecticut, New York, New Jersey) from 1991 through 1994. For New
England, 11,076 lakes are represented in the target population (Larsen et al., 1994).
4.3.1.2 Shenandoah Case Study Area
In the Shenandoah Case Study Area, critical load exceedances were extrapolated using
the SWAS-VTSSS LTM quarterly monitored sites to the population of brook trout streams that
do not lie on limestone bedrock and/or are not significantly affected by human activity within the
Final Risk and Exposure Assessment September 2009
Appendix 4-27
-------
Aquatic Acidification Case Study
watershed. The total number of brook trout streams represented by the SWAS-VTSSS LTM
quarterly monitored sites is approximately 310 streams out of 440 mountain headwater streams
known to support reproducing brook trout in the Shenandoah Case Study Area.
The SWAS-VTSSS LTM programs were designed to track the effects of acidifying
deposition and other factors that determine water quality and related ecological conditions in the
Shenandoah Case Study Area's native trout streams. The SWAS-VTSSS LTM began in spring
1987, when water samples were collected from 440 streams known to have brook trout.
Following the 1987 survey, a representative subset of 69 streams was selected for long-term
quarterly monitoring of water quality, mostly located on National Forest lands or within the
Shenandoah National Park Case Study Area (14 SWAS and 55 VTSSS streams). These streams
were selected to achieve geographic distribution and representation of major bedrock types
(Webb et al., 1994), allowing the streams to be stratified into bedrock type. This enabled results
from the monitored streams (n=69) to be extrapolated to the entire regional population of trout
streams (440). Webb et al. (1994) identified six bedrock classes that account for much of the
spatial variation in ANC among the SWAS-VTSSS LTM quarterly sampled streams. The
landscape classes adopted for this study and the number of selected stream sites within each of
the classes include Blue Ridge siliciclastic (16 streams), Blue Ridge granitic (18 streams), Blue
Ridge basaltic (four streams), and Valley and Ridge siliciclastic (22 streams). Streams in the
carbonate classes (13) were not included because they are not considered susceptible to
acidification. A weighting scheme based on the number of monitoring streams in each of the
bedrock classes was used to extrapolate to the regional population of trout streams. For example,
103 VTSSS streams lie on granitic bedrock, of which 18 were monitored quarterly, resulting in a
weighting factor of 5.7 (=103/18). The weights for streams in the other bedrock classes are 6.25
(= 25/4) for basaltic, 4.3 (= 69/16) for Blue Ridge siliciclastic, and 4.86 (=107/22) for Valley and
Ridge siliciclastic. Thus, the total number of brook trout streams represented in the Shenandoah
Case Study Area is approximately 310; these are all brook trout streams that do not lie on
limestone (10% of 440 streams) and/or have not been significantly affected by human activity
within the watersheds (20% of the streams).
Final Risk and Exposure Assessment September 2009
Appendix 4 - 28
-------
Aquatic Acidification Case Study
5.0 RESULTS
5.1 ADIRONDACK CASE STUDY AREA
5.1.1 Current and Preacidification Conditions of Surface Waters
Since the mid-1990s, lakes in the Adirondack Case Study Area have shown signs of
improvement in ANC, NCV, and SC>42" levels in surface waters, as shown in Figure 3.3-2.
However, current average surface water concentrations of NCV and SC>42" are still well above
preacidification conditions based on MAGIC model simulations of 44 lakes (Figure 5.1-1),
resulting in lower than average ANC surface water chemistry.
140
120
c
o
c of 80
23: 60
O 40
< 20
0
1850
1900
1950
Years
2000
2050
Figure 5.1-1. Average NCV concentrations (orange), SC>4 " concentrations (red), and ANC (blue)
for the 44 lakes in the Adirondack Case Study Area modeled using MAGIC for the period 1850
to 2050.
2-
On average, simulated SC>4 " and NCV surface water concentrations are 5- and 17-fold
higher today, respectively, compared with 1860 acidification levels, whereas ANC has dropped
by a factor of 2 (Table 5.1-1, Figure 5.1-2). Although NCV deposition can be an important
factor in acid precipitation and waterbody acidification, the strong inverse relationship between
SC>42" and ANC, in combination with the low levels of NCV in surface waters, suggests that
acidification in the Adirondack Case Study Area has and continues to be driven by SC>42"
deposition.
Final Risk and Exposure Assessment
Appendix 4-29
September 2009
-------
Aquatic Acidification Case Study
Table 5.1-1. Estimated Average Concentrations of Surface Water Chemistry for 44 Lakes in the
Adirondack Case Study Area Modeled Using MAGIC for Preacidification (1860) and Current
(2006) Conditions
u«q/L
ANC
SO42
NO3
NH4+
Preacidification
Ave.
120.3
12.4
0.2
0.0
(±)
13.6
2.1
1.7
0.0
Current
Ave.
62.1
66.1
3.4
0.1
(±)
15.7
1.24
14.8
0.1
An estimate of how much of this current condition is attributed to the effects of
industrially generated acidifying deposition can be made by examining the hindcast conditions of
the streams. Based on the MAGIC model simulations, the preacidification average ANC
concentration of 44 modeled lakes is 120.3 ueq/L (95% CI 106.8 to 134.0) compared with 62.1
ueq/L (95% CI 50.5 to 81.8) for today (Table 5.1-1).
(a)
Nitrate Pre-acidification (1860) and Current condition (2006)
Nitrate (|jeq/L)
« 0-3
• 3-6
6-9
* 9 -12
Source: EPA 2009
| | Adirondack Boundary
Final Risk and Exposure Assessment
Appendix 4-30
September 2009
-------
Aquatic Acidification Case Study
(b)
Sulfate Pre-acidification (1860) and Current condition (2006)
Sulfate (Meq/L)
• 0-25
• 25-50
50-75
• >75
| | Adirondack Boundary
Source: EPA 2009
Figure 5.1-2. (a) N(V and (b) SC>42" concentrations (ueq/L) of
preacidification (1860) and current (2006) conditions based on hindcasts
of 44 lakes in the Adirondack Case Study Area modeled using MAGIC.
5.1.2 ANC Inferred Condition—Aquatic Status Categories.
The deposition of sulfur and nitrogen and resulting changes in water quality has effects
on ANC values, and by consequence, the biological integrity of the water ecosystem. By
comparing their current surface water condition with their preindustrial (i.e., preacidification or
1860) condition through the MAGIC model simulations of 44 lakes, it is possible to estimate
how the distribution of affected lakes within the Adirondack Case Study Area has changed over
time. Furthermore, grouping each lake into aquatic status categories provides the range of
biological condition. The percentage of lakes in each of the five aquatic status categories is
shown in Table 5.1-2. Eighty-nine percent of the modeled lakes were likely Low Concern or
Moderate Concern (i.e., ANC levels >50 ueq/L) prior to the onset of acidifying deposition
(Figure 5.1-3), whereas the remaining 11% of lakes have ANC >20 ueq/L.
Final Risk and Exposure Assessment
Appendix 4-31
September 2009
-------
Aquatic Acidification Case Study
Table 5.1-2. Percentage of Lakes in the Five Aquatic Status Categories Based on Their Surface
Water ANC Levels for 44 Lakes Modeled Using MAGIC and 88 Lakes in the TEVIE/LTM
Monitoring Network. Results Are for the Adirondack Case Study Area for the Years 2005 to 2006.
Concern
Low
Moderate
Elevated
Severe
Acute
ANC
(jieq/L)
>100
50-100
20-50
0-20
<0
Modeled Pre-
Acidification
(% of Lakes)
50
39
11
0
0
Modeled Current
Condition
(% of Lakes)
20
36
23
9
11
Measured Current
Condition
(% of Lakes)
6
16
32
29
17
In contrast, the current simulated condition of the lakes has shifted toward the chronically
acidified categories. Only 20% and 36% of the lakes currently experience the Low Concern or
Moderate Concern condition, having ANC concentrations >50 ueq/L, whereas 43% of the lakes
are in either the Acute Concern, Severe Concern, or Elevated Concern condition. The result that
the hindcast simulations produced no lakes with Acute Concern or Severe Concern
preacidification suggests that current and recent historical ambient concentrations of NOX and
SOX and their associated levels of N(V and SC>42" deposition have substantially contributed to
acidification (<50 ueq/L) to approximately 32% of modeled lakes. Figure 5.1-4 shows the
spatial extent of preacidification and the current annual average ANC levels in the Adirondack
Case Study Area.
70
60
50
40
30
20
10
0
T
Acute (Below 0 |Jeq/L)
Severe (0-20 ueq/L)
Elevated (20-50 ueq/L)
Moderate (50-100 ueq/L)
Low (Above 100 ueq/L)
1860
2006
Figure 5.1-3. Percentage of lakes in the five classes (Acute, Severe, Elevated,
Moderate, Low) for years 1860 (preacidification) and 2006 (current) conditions
for 44 lakes modeled using MAGIC. Error bars indicate the 95% confidence
interval.
Final Risk and Exposure Assessment
Appendix 4-32
September 2009
-------
Aquatic Acidification Case Study
ANC Preacidification (1860) and Current Condition (2006)
Preacidification (1860)
Current (2006)
ANC
•
Source: EPA 2009
>0
0-20
20-50
50-100
>100
Figure 5.1-4. ANC concentrations of preacidification (1860) and current (2006)
conditions based on hindcasts of 44 lakes in the Adirondack Case Study Area
modeled using MAGIC.
5.1.3 The Biological Risk from Current Nitrogen and Sulfur Deposition: Critical
Load Assessment
The amount of acidifying deposition of sulfur and nitrogen that a watershed can receive
and effectively neutralize varies with location, depending on the biogeochemical properties of
the watershed. A critical load of deposition approach, where the amount of deposition that a
watershed can receive and maintain an ANC critical limit level, can provide insight into the
sensitivity of the waterbody to deposition and can allow an assessment of what the current
condition of the lake might be under current deposition loads.
A critical load of deposition analysis for a critical limit ANC threshold level of 50 ueq/L
was done for 169 waterbodies in the Adirondack Case Study Area. Sites that are unable to
maintain the critical limit ANC level of 50 ueq/L while experiencing 100 meq/m2/yr or less of
deposition are classified as "highly" or "moderately sensitive," indicating that they have a
limited ANC ability and could shift toward acidic aquatic status levels with modest acidifying
deposition inputs. Figure 5.1-5 shows the locations and relative sensitivity of the 169
waterbodies for the critical load analysis (with the 50 ueq/L ANC critical limit). Sites labeled by
red or orange circles have less neutralizing ability than lakes labeled with yellow and green
Final Risk and Exposure Assessment
Appendix 4-33
September 2009
-------
Aquatic Acidification Case Study
circles, and therefore, are those lakes most sensitive to acidifying deposition. Approximately
50% of the 169 lakes modeled in the Adirondack Case Study Area are sensitive, or at risk, to
acidifying deposition.
In Figure 5.1-6, a critical load exceedance "value" indicates combined sulfur and
nitrogen deposition for the year 2002 is greater than the amount of deposition the lake could
neutralize and still maintain an ANC level at or above the critical limit threshold. For the
deposition load for the year 2002, 18%, 28%, 44%, and 58% of the 169 lakes modeled received
levels of combined sulfur and nitrogen deposition that exceeded their critical load for the critical
limit ANC values of 20, 50, and 100 ueq/L, respectively (Table 5.1-3).
Current Condition of Acidity
and Sensitivity
Criticial Load
meq/m2/yr
• Highly Sensitive: < 50
Moderately Sensitive: 51 -100
Low Sensitivity: 101 - 200
a Not Sensitive: > 201
I | Adirondack Boundary
Source: EPA 2009
Figure 5.1-5. Critical loads of acidifying deposition that each surface water
location can receive in the Adirondack Case Study Area while maintaining or
exceeding an ANC concentration of 50 ueq/L based on 2002 data. Watersheds
with critical load values <100 meq/m2/yr (red and orange circles) are most
sensitive to surface water acidification, whereas watersheds with values >100
meq/m2/yr (yellow and green circles) are the least sensitive sites.
Final Risk and Exposure Assessment
Appendix 4-34
September 2009
-------
Aquatic Acidification Case Study
Table 5.1-3. Adirondack Case Study Area Critical Load Exceedances, Where Nitrogen and
Sulfur Deposition Is Larger Than the Critical Load for Four Different ANC Critical Limit
Thresholds, for 169 Modeled Lakes within the TIME/LTM and EMAP Monitoring and Survey
Programs. "No. Exceedances" Indicates the Number of Lakes at or below the Given ANC
Critical Limit, and "% Lakes" Indicates the Total Percentage of Lakes at or below the Given
ANC Critical Limit.
ANC
Critical Limit
100 (ieq/L
50 (ieq/L
20 neq/L
0 (j,eq/L
No. Exceedances
(out of 169 Lakes)
98
74
47
30
%
Lakes
58
44
28
18
Final Risk and Exposure Assessment
Appendix 4 - 35
September 2009
-------
Aquatic Acidification Case Study
Critical Load Exceedances
(100|jeq/L)
Critical Load Exceedences
( > ANC of 100 |ieq/L)
Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source EPA 2009
Critical Load Exceedances
(50 |jeq/L)
Critical Load Exceedences
( > ANC of 50 (jeq/L)
Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2009
Critical Load Exceedances
(20
Critical Load Exceedences
( > ANC of 20 peq/L)
Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2009
Critical Load Exceedances
(0 Meq/L)
Critical Load Exceedences
( > ANC of 0 (jeq/L)
« Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2008
Figure 5.1-6. Critical load exceedances (red circles) in the Adirondack Case
Study Area based on the year 2002 deposition magnitudes for waterbodies where
the critical limit ANC concentration is 0, 20, 50, and 100 ueq/L, respectively.
Green circles represent lakes where deposition is below the critical load. See
Table 5.1-3
Final Risk and Exposure Assessment
Appendix 4-36
September 2009
-------
Aquatic Acidification Case Study
5.1.4 Representative Sample of Lakes in the Adirondack Case Study Area
Estimating the acidification risk to the entire population of lakes in the Adirondack Case
Study Area from the current the levels of NOX and SOX ambient concentrations and deposition
requires extrapolating from the 169 modeled lakes to the broader population of lakes in the
Adirondack Case Study Area. One hundred seventeen lakes of the 169 lakes modeled for critical
loads are part of a subset of 1,842 lakes in the Adirondack Case Study Area, which include all
lakes from 0.5 to 2000 ha in size and at least 1 m in depth. Using weighting factors for frequency
of occurrence derived from the EMAP probability survey and critical load calculations from the
117 lakes, exceedances estimates were derived for the 1,842 lakes in the Adirondack Case Study
Area. Based on this approach, 945, 666, 242, and 135 lakes exceed their critical load for the year
2002 deposition with critical limits of 100, 50, 20, and 0 ueq/L, respectively (Table 5.1-4). The
current effect level for a moderate protective (i.e., ANC limit of 20 and 50 ueq/L) is 13% of the
total population and 36% of the total population, whereas it is 51% of the total population for the
most protective level (i.e., ANC limit of 100 ueq/L).
Table 5.1-4. Critical Load Exceedances for the Regional Population of 1,842 Lakes in the
Adirondack Case Study Area That Are from 0.5 to 2,000 ha in Size and at Least 1 m in Depth for
the Four Critical Limit ANC Levels (0, 20, 50, and 100 ueq/L). Estimates Use the Exceedances
for the Subset of 169 Lakes Using 2002 Deposition Magnitudes (Table 5.1-3) and Are
Extrapolated to the Full Population Based on the EMAP Lake Probability Survey of 1991 to
1994. "No. Exceedances" Indicates the Number of Lakes at or Below the Given ANC Critical
Limit; "% Lakes" Indicates the Total Percentage of Lakes at or Below the Given ANC Critical
Limit.
ANC Critical Limit
100 neq/L
50 neq/L
20 (ieq/L
0 (j,eq/L
No. Exceedances
(out of 1842)
945
666
242
13
% Lakes
51
36
13
7
Because some lakes in the Adirondack Case Study Area have natural sources of acidity,
they would never have ANC levels >50 or 100 ueq/L, even in the absence of all anthropogenic-
derived acidifying deposition. Based on the hindcast simulations of 44 lakes using the MAGIC
model, no modeled lakes have ANC levels <20 ueq/L. However, 5 modeled lakes, or 11% of
modeled lakes, have ANC concentrations between 22 and 47 ueq/L. This equates to
Final Risk and Exposure Assessment
Appendix 4-37
September 2009
-------
Aquatic Acidification Case Study
approximately 300 lakes, or 16%, of the representative population of lakes in the Adirondack
Case Study Area that likely had preacidification ANC levels <50 ueq/L. On the other hand,
potentially >52% of lakes likely had preacidification ANC levels <100 ueq/L. The higher
percentage of lakes in the regional population compared with the modeled population is because
the lake classes or sizes that are likely to have preacidification ANC levels <50 or 100 ueq/L are
more abundant in the Adirondack Case Study Area than lakes with preacidification ANC levels
>50 or 100 ueq/L.
5.1.5 Recovery from Acidification Given Current Emission Reductions
The question is whether lakes can recover to healthy systems (i.e., ANC >50, or 100
ueq/L) given the current surface water and ecosystem conditions of the lakes and current
emission and deposition levels in the Adirondack Case Study Area. The forecast model runs of
44 lakes using MAGIC were used to determine if current deposition could lead to recovery of the
acidified lakes. Based on a deposition scenario that maintains current emission levels up to years
2020 and 2050, the simulation forecast indicates no improvement in water quality over either of
these periods. The percentage of lakes within the Elevated Concern to Acute Concern categories
remains the same in the years 2020 and 2050 (Figure 5.1-7). Moreover, the percentage of
modeled lakes classified as not acidic remains the same, suggesting that current emission levels
will not likely improve their recovery from acidification in the Adirondack Case Study Area.
2006
2020
2050
• Acute (Below 0 ueq/L)
D Elevated (20-50 ueq/L)
• Low (Above 100 ueq/L)
D Severe (0-20 ueq/L)
• Moderate (50-100 ueq/L)
Figure 5.1-7. Percentage of lakes in each of the five classes of acidification (Acute,
Severe, Elevated, Moderate, Low) for the years 2006, 2020, and 2050 for 44 lakes
modeled using MAGIC, where current emissions are held constant. Error bars
indicate the 95% confidence interval.
Final Risk and Exposure Assessment
Appendix 4 - 38
September 2009
-------
Aquatic Acidification Case Study
5.2 SHENANDOAH CASE STUDY AREA
5.2.1 Current and Preacidification Conditions of Surface Waters
Since the mid-1990s, streams in the Shenandoah Case Study Area have shown slight
signs of improvement in NCV and SC>42" concentrations in surface waters (see Figure 3.4-2).
However, current concentrations of NCV and SC>42" are still well above preacidification
conditions based on MAGIC model simulations, resulting in lower ANC levels of surface water
(Figure 5.2-1)
120
1850
1900
1950
Years
2000
2050
Figure 5.2-1. Average NCV concentrations orange), SO42"concentrations (red),
and ANC (blue) levels for the 60 streams in the Shenandoah Case Study Area
modeled using MAGIC for the period 1850 to 2050.
Figure 5.2-2 shows the condition of the streams in the year 1860 (i.e., preacidification)
and in the year 2006 (i.e., current) conditions. On average, NCV and SC>42" concentrations are 32-
and 10-fold higher today (Table 5.2-1). These results also demonstrate that acidification in most
streams in the Shenandoah Case Study Area are currently being driven by SO42" deposition
because the current average SC>42" concentration is 11-folder greater than NCV concentrations in
surface waters.
An estimate of how much of this current condition is attributed to the effects of
industrially generated acidifying deposition can be made by examining the hindcast conditions of
the streams. Based on the MAGIC model simulations, preacidification average ANC level of the
60 modeled streams is 101.4 (95% CI 91.9. to 110.9) ueq/L compared with 57.9 (95% CI. 53.4 to
62.4) ueq/L for today (Table 5.2-1).
Final Risk and Exposure Assessment
Appendix 4-39
September 2009
-------
Aquatic Acidification Case Study
(a) Nitrate Preacidification (1860) and Current Conditions (2006)
Preacidification (1860) Current (2006)
Source: EPA 2009
Nitrate (ueq/L)
• 0-5
• 5-10
10-15
• 15-20
• >20
(b) Sulfate Preacidification (1860) and Current Conditions (2006)
Preacidification (1860) Current (2006)
Source: EPA 2009
Sulfate (|Jeq/L)
• 0-25
• 25-50
• 50-75
75 -100
• >100
Figure 5.2-2. (a) NO3" and (b) SO42" concentrations (ueq/L) of 1860
(preacidification) and 2006 (current) conditions based on hindcasts of 60 streams
in the Shenandoah Case Study Area modeled using MAGIC.
Final Risk and Exposure Assessment
Appendix 4-40
September 2009
-------
Aquatic Acidification Case Study
Table 5.2-1. Estimated Average Concentrations of Surface Water
Chemistry for 60 Streams in the Shenandoah Case Study Area Modeled
Using MAGIC for Preacidification (1860) and Current (2006) Conditions
jieq/L
ANC
SO42
NO3
NH4+
Preacidification
Avg.
101.4
2.1
0.6
n/a
(±)
9.5
0.1
0.01
n/a
Current
Avg.
57.9
68.0
6.2
n/a
(±)
4.5
8.4
0.1
n/a
5.2.2 ANC Inferred Condition—Aquatic Status Categories
The percentage of streams in each of the five aquatic status categories is shown in Table
5.2-2 and in the graph in Figure 5.2-3. Ninety-two percent of the modeled streams likely were at
the Low Concern or Moderate Concern conditions prior to the onset of acidifying deposition.
The other 8% of streams have ANC of >27 ueq/L. The hindcast simulations produced no streams
with Acute Concern or Severe Concern conditions.
Table 5.2-2. Percentage of Streams in the Five Aquatic Status Categories Based on Their
Surface Water ANC Concentrations for 60 Streams Modeled Using MAGIC and 67 Streams in
the SWAS-VTSSS LTM Network. Results are for the Shenandoah Case Study Area for the Year
2006.
Concern
Low
Moderate
Elevated
Severe
Acute
ANC
(ueq/L)
>100
50-100
20-50
0-20
<0
Modeled
Preacidification
Condition
(% of Streams)
42
50
8
0
0
Modeled Current
Condition
(% of Streams)
27
20
28
12
13
Measured Current
Condition
(% of Streams)
15
30
25
18
12
In contrast, the current simulated condition of the streams has shifted toward the
chronically acidified categories. Only 47% of the streams currently experience the Low Concern
or Moderate Concern conditions, whereas 53% of the streams experience the Acute Concern,
Severe Concern, or Elevated Concern conditions. These results based on model reconstructions
suggest that anthropogenic acidifying deposition is responsible for acidifying (ANC <50 ueq/L)
Final Risk and Exposure Assessment
Appendix 4-41
September 2009
-------
Aquatic Acidification Case Study
approximately 45% of streams modeled in the Shenandoah Case Study Area. Figure 5.2-4 shows
the spatial extent of preacidification and current annual average ANC levels in the Shenandoah
Case Study Area.
• Acute (Below 0 ueq/L)
• Severe (0-20 ueq/L)
D Elevated (20-50 ueq/L)
• Moderate (50-100 ueq/L)
• Low (Above 100 ueq/L)
1860
2006
Figure 5.2-3. Percentage of streams in the five classes of acidification (Acute,
Severe, Elevated, Moderate, Low) for years 1860 (preacidification) and 2006
(current) conditions for 60 streams modeled using MAGIC. The number of
streams in each category is above the bar. Error bars indicate the 95% confidence
interval.
ANC Preacidification (1860) and Current Condition (2006)
Pre-acidification (1860) Current (2006)
Source: EPA 2009
ANC (yeq/L)
• <0
0-20
20-50
• 50-100
• >100
Figure 5.2-4. ANC levels of 1860 (preacidification) and 2006 (current) conditions
based on hindcasts of 60 streams in the Shenandoah Case Study Area modeled
using MAGIC.
Final Risk and Exposure Assessment
Appendix 4-42
September 2009
-------
Aquatic Acidification Case Study
5.2.3 The Biological Risk from Current Nitrogen and Sulfur Deposition: Critical
Load Assessment
In Figure 5.2-5, sites labeled by red or orange circles have less ability to neutralize acid
inputs than streams labeled with yellow and green circles, and therefore, are streams most
sensitive to acidifying deposition. Approximately 75% of the 60 streams modeled in the
Shenandoah Case Study Area are sensitive or at risk to acidifying deposition.
Current Condition of Acidity
and Sensitivity
Criticial Load
meq/m2/yr
• Highly Sensitive: < 50
Moderately Sensitive: 51-100
Low Sensitivity: 101 -200
• Not Sensitive: > 201
Source: EPA 2009
Figure 5.2-5. Critical loads of surface water acidity for an ANC of 50 ueq/L for
Shenandoah Case Study Area streams. Each dot represents an estimated amount
of acidifying deposition (i.e., critical load) that each stream's watershed can
receive and still maintain a surface water ANC >50 ueq/L. Watersheds with
critical load values <100 meq/m2/yr (red and orange circles) are most sensitive to
surface water acidification, whereas watersheds with values >100 meq/m2/yr
(yellow and green circles) are the least sensitive sites.
In Figure 5.2-6, a critical load exceedance "value" indicates combined sulfur and
nitrogen deposition in the year 2002 that is greater than the amount of deposition the stream
could buffer and still maintain the ANC level of above each of the four different ANC limits of
0, 20, 50, and 100 ueq/L. For the year 2002, 52%, 72%, 85%, and 92% of the 60 streams
Final Risk and Exposure Assessment
Appendix 4-43
September 2009
-------
Aquatic Acidification Case Study
modeled receive levels of combined sulfur and nitrogen deposition that exceeded their critical
load with critical limits of 0, 20, 50, and 100 ueq/L, respectively (Table 5.2-3).
Critical Load Exceedances
(100
Critical Load Exceedences
(>ANCof 100(jeq/L)
• Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2009
Critical Load Exceedances
(20
Critical Load Exceedences
( > ANC of 20 Meq/L)
• Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA Z009
Critical Load Exceedances
(50 neq/L)
Critical Load Exceedences
( > ANC of 50 peq/L)
• Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2009
Critical Load Exceedances
(0 Meq/L)
Critical Load Exceedences
( > ANC of 0 peq/L)
• Deposition does not Exceed Critical Load
• Deposition Exceeds Critical Load
Source: EPA 2009
Figure 5.2-6. Critical load exceedances for ANC levels of 0, 20, 50, and 100 ueq/L
for Shenandoah Case Study Area streams. Green circles represent streams where
current nitrogen and sulfur deposition is below the critical load and that maintain an
ANC level of 0, 20, 50, and 100 ueq/L, respectively. Red circles represent streams
where current nitrogen and sulfur deposition exceeds the critical load, indicating they
are currently impacted by acidifying deposition. See Table 5.2-3.
Final Risk and Exposure Assessment
Appendix 4-44
September 2009
-------
Aquatic Acidification Case Study
Table 5.2-3. Critical Load Exceedances (Nitrogen + Sulfur Deposition > Critical Load) for 60
Modeled Streams within the VTSSS LTM Program in the Shenandoah Case Study Area. "No.
Exceedances" Indicates the Number of Streams at the Given ANC Limit; "% Streams" Indicates
the Total Percentage of Streams at the Given ANC Limit.
ANC Critical Limit
No. Exceedances (out
of 60)
% Streams
100
jieq/L
55
92
50
jieq/L
51
85
20 ueq/L
43
72
0 ueq/L
31
52
5.2.4 Regional Assessment of Trout Streams in the Shenandoah Case Study Area
The 60 trout streams modeled are characteristic of first- and second-order streams on
nonlimestone bedrock in the Blue Ridge Mountains of Virginia. Because of the strong
relationship between bedrock geology and ANC in this region, it is possible to consider the
results in the context of similar trout streams in the Southern Appalachians that have the same
bedrock geology and size. In addition, the 60 streams are a subset of 344 streams sampled by the
Virginia Trout Stream Sensitivity Study, which can be applied to 304 of the original 344 streams.
Using the 304 streams to which the analysis applies directly as the total, 279, 258, 218, and 157
streams exceed their critical load for the year 2002 deposition with critical limits of 100, 50, 20,
and 0 ueq/L, respectively. However, it is likely that many more of the -12,000 trout streams in
Virginia would exceed their critical load, given the extent of similar bedrock geology outside of
the case study area in the southern Appalachian Mountains.
5.2.5 Recovery from Acidification Given Current Emission Reductions
Based on a deposition scenario that maintains current emission levels to the years 2020
and 2050, there will still be a large number of streams in Virginia that have Elevated Concern to
Acute Concern problems with acidity (Figure 5.2-7). In the short term (i.e., by the year 2020)
and in the long term (i.e., by the year 2050), the response of the 60 modeled streams shows no
improvement in the number of streams that have Moderate Concern conditions. In fact, under
current emission levels, the modeling suggests that conditions may degrade by the year 2050. In
Figure 5.2-7 the percentage of streams in Acute Concern condition increases by 5%, whereas
streams in Moderate Concern condition decreases by 5%.
Final Risk and Exposure Assessment
Appendix 4-45
September 2009
-------
Aquatic Acidification Case Study
iL
2006
2020
2050
• Acute (Below 0 |jeq/L)
D Elevated (20-50 |jeq/L)
• Low (Above 100 |jeq/L)
D Severe (0-20 |jeq/L)
• Moderate (50-100 |jeq/L)
Figure 5.2-7. Percentage of streams in the five categories of acidification (Acute,
Severe, Elevated, Moderate, Low) for the years 2006, 2020, and 2050 for 60 streams in
the Shenandoah Case Study Area modeled using MAGIC. The number of streams in
each category is above the bar. Error bars indicate the 95% confidence interval.
Final Risk and Exposure Assessment
Appendix 4-46
September 2009
-------
Aquatic Acidification Case Study
6. REFERENCES
Baker, J.P., WJ. Warren-Hicks, J. Gallagher, and S.W. Christensen. 1993. Fish population losses
from Adirondack lakes: The role surface water acidity and acidification. Water Resources
Research 29:861-874.
Brakke, D.F., A. Henriksen, and S.A. Norton. 1989. Estimated background concentrations of
sulfate in dilute lakes. Water Resources Bulletin 25(2):247-253.
Brakke, D.F., A. Henriksen, and S.A. Norton. 1990. A variable F-factor to explain changes in
base cation concentrations as a function of strong acid deposition. International
Association of Theoretical and Applied Limnology, Proceedings 24:146-149.
Bulger, A.J., BJ. Cosby, C.A. Dolloff, K.N. Eshleman, J.R. Webb, and J.N. Galloway. 1999.
SNP:FISH, Shenandoah National Park: Fish in Sensitive Habitats, Volumes 1 through
IV. Project final report. Project Completion Report to the National Park Service.
Cooperative Agreement CA-4000-2-1007, Supplemental Agreement #2. University of
Virginia, Department of Environmental Sciences, Charlottesville, VA.
Bulger, A.J., BJ. Cosby, and J.R. Webb. 2000. Current, reconstructed past, and projected future
status of brook trout (Salvelinus fontinalis) streams in Virginia. Canadian Journal of
Fisheries and Aquatic Sciences 57:1515-1523.
Cosby, B.J., R.F. Wright, G.M. Hornberger, and J.N. Galloway. 1985a. Modelling the effects of
acid deposition: Assessment of a lumped parameter model of soil water and streamwater
chemistry. Water Resources Research 27:51-63.
Cosby, B.J., R.F. Wright, G.M. Hornberger, and J.N. Galloway. 1985b. Modelling the effects of
acid deposition: Estimation of long-term water quality responses in a small forested
catchment. Water Resources Research 27:1591-1601.
Cosby, B.J., G.M. Hornberger, P.F. Ryan, and D.M. Wolock. 1989. MAGIC/DDRP Final
Report. Project Completion Report. U.S. Environmental Protection Agency,
Direct/Delayed Response Project, Corvallis, OR.
Final Risk and Exposure Assessment September 2009
Appendix 4-47
-------
Aquatic Acidification Case Study
Cosby, B.J., A. Jenkins, R.C. Ferrier, J.D. Miller, and T.A.B. Walker. 1990. Modelling stream
acidification in afforested catchments: Long-term reconstructions at two sites in central
Scotland. Journal of Hydrology 720:143-162.
Cosby, B.J., S.A. Norton, and J.S. Kahl. 1996. Using a paired-catchment manipulation
experiment to evaluate a catchment-scale biogeochemical model. Science of the Total
Environment 183:49-66.
Dennis, T.E., and AJ. Bulger. 1995. Condition factor and whole-body sodium concentrations in
a freshwater fish: Evidence for acidification stress and possible ionoregulatory
overcompensation. Water, Air, and Soil Pollution 55:377-382.
De Vries, W., G. J. Reinds, and M. Posch. 1994. Assessment of critical loads and their
exceedance on European forests using a one-layer steady-state model. Water, Air, and
Soil Pollution 72:357-394.
DeWalle, D. R. Swistock, B. R. 1994. Causes of episodic acidification in five Pennsylvania
streams on the northern Appalachian Plateau. Water Resources Research 30: 1955-1963.
Driscoll C. T. Van Dreason, R. 1993. Seasonal and long-term temporal patterns in the chemistry
of Adirondack lakes. Water Air Soil Pollution 67: 319-344
Driscoll, C.T., R.M. Newton, C.P. Gubala, J.P. Baker, and S.W. Christensen. 1991. Adirondack
mountains. Pp. 133-202 in Acidic Deposition and Aquatic Ecosystems: Regional Case
Studies. Edited by D.F. Charles. New York: Springer-Verlag.
Driscoll, C. T. Postek, K. M., Kretser, E, Raynal, D. J. 1995. Long-term trends in the chemistry
of precipitation and lake water in the Adirondack region of New York, USA. Water Air
Soil Pollution 85: 583-588.
Driscoll, C.T., G.B. Lawrence, AJ. Bulger, TJ. Butler, C.S. Cronan, C. Eagar, K.F. Lambert,
G.E. Likens, J.L. Stoddard, and K.C. Weathers. 2001. Acidic deposition in the
northeastern United States: sources and inputs, ecosystem effects, and management
strategies. BioScience 57:180-198.
Final Risk and Exposure Assessment September 2009
Appendix 4-48
-------
Aquatic Acidification Case Study
Driscoll, C.T., K.M. Driscoll, K.M. Roy, and MJ. Mitchell. 2003. Chemical response of lakes in
the Adirondacks of New York to declines in acidic deposition. Environmental Science
and Technology 37:2036-2042.
Dupont, 1, T.A. Clair, C. Gagnon, D.S. Jeffries, IS. Kahl, SJ. Nelson, and J.M. Peckenham.
2005. Estimation of critical loads of acidity for lakes in northeastern United States and
eastern Canada. Environmental Monitoring and Assessment 109: 275-291
Ford, J., J.L. Stoddard, and C.F. Powers. 1993. Perspectives in environmental monitoring: an
introduction to the U.S. EPA Long-Term Monitoring (LTM) project. Water, Air, and Soil
Pollution 67:247-255.
Gran G. Determination of the equivalence point in potentiometric titrations. Part 11. Analyst
1952; 77:661-671.
Grimm, J.W., and J. A. Lynch. 2004. Enhanced wet deposition estimates using modeled
precipitation inputs. Environmental Monitoring and Assessment 90:243-268.
Haines, T.A., and J.P Baker. 1986. Evidence offish population responses to acidification in the
Eastern United States. Water, Air, and Soil Pollution, 37:605-629
Henriksen, A. 1984. Changes in base cation concentrations due to freshwater acidification.
International Association of Theoretical and Applied Limnology, Proceedings 22:692-
698.
Henriksen, A., and M. Posch. 2001. Steady-state models for calculating critical loads of acidity
for surface waters. Water, Air, and Soil Pollution: Focus 7:375-398.
Henriksen, A., J. Kamari, M. Posch, and A. Wilander. 1992. Critical loads of acidity: Nordic
surface waters. Ambio 27:356-363.
Herlihy, A.T., Larsen, D.P., Paulsen, S.G., Urquhart, N.S., Rosenbaum, B.J., 2000. Designing a
spatially balanced, randomized site selection process for regional stream surveys: the
EMAP Mid-Atlantic Pilot Study. Environmental Monitoring and Assessment 63: 95-113.
Final Risk and Exposure Assessment September 2009
Appendix 4-49
-------
Aquatic Acidification Case Study
Hornberger, G.M., BJ. Cosby, and R.F. Wright. 1989. Historical reconstructions and future
forecasts of regional surface water acidification in southernmost Norway. Water
Resources Research 25:2009-2018.
Jenkins, A., BJ. Cosby, R.C. Ferrier, T.A.B. Walker, and J.D. Miller. 1990a. Modelling stream
acidification in afforested catchments: An assessment of the relative effects of acid
deposition and afforestation. Journal of Hydrology 720:163-181.
Jenkins, A., P.O. Whitehead, BJ. Cosby, and H.J.B. Birks. 1990b. Modelling long-term
acidification: A comparison with diatom reconstructions and the implication for
reversibility. Philosophical Transactions of the Royal Society of London, B 327:435-440.
Jenkins, A., P.O. Whitehead, TJ. Musgrove, and B J. Cosby, 1990c. A regional model of
acidification in Wales. Journal of Hydrology 77(5:403-416.
Kahl, J. S., Norton, S. A., Cornan, C. S. Fernandez, I. J. Bacon, K. C., Haines, T. A. (1991)
Maine. In Charles D. F. ed. Acidic deposition and aquatic ecosystems: regional case
studies, New York, NY: Spring-Verlag: pp. 203-235.
Kahl, J. S., Haines, T. A., Norton, S. A., Davis R. B. (1993) Recent trends in the acid-base status
of surface waters in Maine, USA. Water Air Soil Pollution 67: 281-300
Klemedtsson, L., and B. H. Svensson. 1988. Effects of acidic deposition on denitrification and
N2O emission from forest soils. Pages 343-362 in J. Nilsson and P. Grennfelt (eds.),
Critical loads for sulphur and nitrogen, Nord 1988:97. Nordic Council of Ministers,
Copenhagen, Denmark.
Kretser, W., J. Gallagher, and J. Nicolette. 1989. Adirondack Lakes Study, 1984-1987: An
Evaluation of Fish Communities and Water Chemistry. Ray Brook, NY: Adirondack
Lakes Survey Corporation.
Landers, D.H., W.S. Overton, R.A. Linthurst, andD.F. Brakke. 1988. Eastern lake survey: Regional
estimates of lake chemistry. Environmental Science and Technology 22:128-135.
Final Risk and Exposure Assessment September 2009
Appendix 4-50
-------
Aquatic Acidification Case Study
Larsen B. P., Urquhart, N. S. (1993) A framework for assessing the sensitivity of the EMAP
design. In Larsen, D. P., Christie, S. J. eds. EMAP-surf ace waters 1991 pilot report.
Corvallis, OR: U.S. Environmental Protection Agency: pp. 4.1-4.37
Larsen, D.P., K.W. Thornton, N.S. Urquhart, and S.G. Paulsen. 1994. The role of sample surveys
for monitoring the condition of the nation's lakes. Environmental Monitoring and
Assessment 32:101-134.
Lepisto, A., P.O. Whitehead, C. Neal, and BJ. Cosby. 1988. Modelling the effects of acid
deposition: Estimation of longterm water quality responses in forested catchments in
Finland. Nordic Hydrology 79:99-120.
Lien, L., G.G. Raddum, and A. Fjellheim. 1992. Critical Loads of Acidity to Freshwater: Fish
and Invertebrates. The Environmental Tolerence Levels Programme. Rep. No. 23/1992.
Norwegian Ministry of Environment, Oslo, Norway.
Lynch, D.D. and N.B. Dise. 1985. Sensitivity of stream basins in Shenandoah National Park to
acid deposition. Water Resources Investigations Report 85-4115 1985:1]61. U.S.
Department of Interior, U.S. Geological Survey, Reston, VA.
Matuszek, I.E., and G.L. Beggs. 1988. Fish species richness in relation to lake area, pH, and
other abiotic factors in Ontario lakes. Canadian Journal of Fisheries and Aquatic
Sciences 45:1931-1941.
Murdoch, P. S. Stoddard, J. L. (1993) Chemical characteristics and temporal trends in eight
streams of the Catskill Mountains, New York. Water Air Soil Pollution 67: 367-395.
NAPAP (National Acid Precipitation Assessment Program). 2005. National Acid Precipitation
Assessment Program Report to Congress: An Integrated Assessment. National Acid
Precipitation Assessment Program, Washington, DC.
NAPAP (National Acid Precipitation Assessment Program). 1998. NAPAP Biennial Report to
Congress: An Integrated Assessment. National Science and Technology Council,
Committee on Environment and Natural Resources, Silver Spring, MD.
Final Risk and Exposure Assessment September 2009
Appendix 4-51
-------
Aquatic Acidification Case Study
Norton, S.A., R.F. Wright, J.S. Kahl, and J.P. Schofield. 1992. The MAGIC simulation of
surface water at, and first year results from, the Bear Brook Watershed Manipulation,
Maine, USA. Environmental Pollution 77:279-286.
Paulsen, S.G., D.P. Larsen, P.R. Kaufmann, T.R. Whittier, J.R. Baker, D.V. Peck, J.McGue,
R.M. Hughes, D. McMullen, D. Stevens, J.L. Stoddard, J. Lazorchak, W. Kinney, A.R.
Selle, and R. Hjort. 1991.EMAP-Surface Waters Monitoring and Research Strategy:
Fiscal Year 1991. EPA/600/3-91/022, U.S. Environmental Protection Agency, Corvallis,
OR: 184pp.
Posch, M., J. Kamari, M. Forsius, A. Henriksen, and A. Wilander. 1997. Exceedance of critical
loads for lakes in Finland, Norway and Sweden: Reduction requirements for acidifying
nitrogen and sulfur deposition. Environmental Management 21: 291-304.
Rago, P. J., and J.G. Wiener. 1986. Does pH affect fish species richness when lake area is
considered? Transactions of the American Fisheries Society 775:438-447.
Reynolds, B., and D.A. Norris. 2001. Freshwater critical loads in Wales. Water, Air, and Soil
Pollution: Focus 7:495-505.
Rochelle, B.P., and M.R. Church. 1987. Regional patterns of sulfur retention in watersheds of the
eastern U.S. Water, Air, and Soil Pollution 3<5(l-2):61-73.
Rochelle, B.P., M.R. Church, and M.B. David. 1987. Sulfur retention at intensively studied sites
in the U.S. and Canada. Water, Air, and Soil Pollution 33:73-83.
Schindler, D.W. 1988. Effects of acid rain on freshwater ecosystems. Science 239:232-239.
Schindler, D.W., S.E.M. Kaslan, and R.H. Hesslenin. 1989. Biological improvement in lakes of
the Midwestern and Northeastern United States from acid rain. Environmental Science
and Technology 23:573-580.
Schreck, C.B. 1981. Stress and rearing of salmonids. Aquaculture 25:241-249.
Final Risk and Exposure Assessment September 2009
Appendix 4-52
-------
Aquatic Acidification Case Study
Schreck, C.B. 1982. Stress and compensation in teleostean fishes: response to social and physical
factors. Pp. 295-321 in Stress and Fish. Edited by A.D. Pickering. New York: Academic
Press.
Shannon, J.D. 1998. Calculation of Trends from 1900 through 1990 for Sulfur andNOx-N
Deposition Concentrations ofSulfate and Nitrate in Precipitation, and Atmospheric
Concentrations of SOx and NOx Species over the Southern Appalachians. Report to
Southern Appalachian Mountains Initiative, April.
Stoddard, J.L. 1990. Plan for Converting the NAPAP Aquatic Effects Long-Term Monitoring
(LTM) Project to the Temporally Integrated Monitoring of Ecosystems (TIME) Project.
International report. U.S. Environmental Protection Agency, Corvallis, OR.
Stoddard, J.L., and J.H. Kellogg. 1993. Trends and patterns in lake acidification in the state of
Vermont: evidence from the Long-Term Monitoring project. Water, Air, and Soil
Pollution 67:301-317.
Stoddard, J. L., Urquhart, N. S. Newell, A. D. Kugler, D. (1996) The Temporally Integrated
Monitoring of Ecosystems (TIME) project design. 2, Detection of regional Acidification
trends. Water Resources 32: 2529-2538
Stoddard, J.L., C.T. Driscoll, J.S. Kahl, J.H Kellogg. 1998. A regional analysis of lake
acidification trends for the northeastern U.S., 1982-1994. Environmental Monitoring and
Assessment 57:3 99-413.
Stoddard, J., J.S. Kahl, F.A. Deviney, D.R. DeWalle, C.T. Driscoll, A.T. Herlihy, J.H. Kellogg;
P.S. Murdoch, J.R. Webb, and K.E. Webster. 2003. Response of Surface Water Chemistry
to the Clean Air Act Amendments of 1990. EPA 620/R-03.001. U.S. Environmental
Protection Agency, Office of Research and Development, National Health and
Environmental Effects Research Laboratory, Research Triangle Park, NC.
Sullivan, T.J., and B.J. Cosby. 1998. Modeling the concentration of aluminum in surface waters.
Water, Air, and Soil Pollution J05:643-659.
Final Risk and Exposure Assessment September 2009
Appendix 4-53
-------
Aquatic Acidification Case Study
Sullivan, T.J., andBJ. Cosby. 2004. Aquatic Critical Load Development for the Monongahela
National Forest, West Virginia. Report Prepared for U.S. Department of Agriculture
Forest Service, Monongahela National Forest, Elkins, WV. E&S Environmental
Chemistry, Inc., Corvallis, OR.
Sullivan, T.J., C.T. Driscoll, BJ. Cosby, IJ. Fernandez, A.T. Herlihy, J. Zhai, R. Stemberger,
K.U. Snyder, J.W. Sutherland, S.A. Nierzwicki-Bauer, C.W. Boylen, T.C. McDonnell,
and N.A. Nowicki. 2006. Assessment of the Extent to which Intensively-Studied Lakes are
Representative of the Adirondack Mountain Region. Final report. New York State Energy
Research and Development Authority (NYSERDA), Albany, NY. Available at
http://nysl.nysed.gOv/uhtbin/cgisirsi/Qcwd6NzFby/NYSL/138650099/8/4298474
(accessed November 1, 2007).
Sverdrup, H., W. de Vries, and A. Henriksen. 1990. Mapping Critical Loads. Miljorapport 14.
Nordic Council of Ministers, Copenhagen, Denmark.
Urquhart, N.S., S.G. Paulsen, and D.P. Larsen. 1998. Monitoring for policy-relevant regional
trends over time. Ecological Applications S(2):246-257.
U.S. EPA (Environmental Protection Agency). 1995. Acid Deposition Standard Feasibility
Study. Report to Congress. EPA 430-R-95-001a. U.S. Environmental Protection Agency,
Office of Air and Radiation, Washington, DC.
U.S. EPA (Environmental Protection Agency). 2008. Integrated Science Assessment (ISA) for
Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report). EPA/600/R-
08/082F. U.S. Environmental Protection Agency, National Center for Environmental
Assessment-RTF Division, Office of Research and Development, Research Triangle
Park, NC. Available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201485.
Van Sickle, J., J.P. Baker, H.A. Simonin, B.P. Baldigo, W.A. Kretser, and W.E. Sharpe. 1996.
Episodic acidification of small streams in the northeastern United States: Fish mortality
in field bioassays. Ecological Applications (5:408-421.
Final Risk and Exposure Assessment September 2009
Appendix 4-54
-------
Aquatic Acidification Case Study
Webb J.R., Cosby B.J., Galloway J.N., and Hornberger G.M..1989. Acidification of native brook
trout streams in Virginia. Water Resources Research 25:1367-1377.
Webb, J.R., F.A. Deviney, J.N. Galloway, C.A. Rinehart, P.A. Thompson, and S. Wilson. 1994.
The acid-base status of native brook trout streams in the mountains of Virginia: a regional
assessment based on the Virginia Trout Stream Sensitivity Study. Prepared for the
Virginia Department of Game and Inland Fisheries, Richmond, VA.
Webb, J.R., BJ. Cosby, F.A. Deviney, J.N. Galloway, S.W. Maben, and AJ. Bulger. 2004. Are
brook trout streams in western Virginia and Shenandoah National Park recovering from
acidification? Environmental Science and Technology 35:4091-4096.
Webster, K. E., Brezonik, P. L., Holdhsen, B. J. 1993. Temporal trends in low alkalinity lakes of
the upper Midwest (1983-1989). Water Air Soil Pollution 67: 397-414
Wedemeyer, G.A., B.A. Barton, and DJ. MeLeay. 1990. Stress and acclimation. Pp. 178-198 in
Methods for Fish Biology. Edited by C.B. Schreck and P.B. Moyle. Bethesda, MD:
American Fisheries Society.
Whitehead, P.G., S. Bird, M. Hornung, BJ. Cosby, C. Neal, and P. Paricos. 1988. Stream
acidification trends in the Welsh Uplands: A modelling study of the Llyn Brianne
catchments. Journal of Hydrology 707:191-212.
Whittier, T.R., S.G. Paulsen, D.P. Larsen, S.A. Peterson, A.T. Herlihy, andP.H. Kaufmann.
2002. Indicators of ecological stress and their extent in the population of northeastern
lakes: a regional-scale assessment. BioScience 52:235-247.
Wright, R.F., BJ. Cosby, M.B. Flaten, and J.O. Reuss. 1990. Evaluation of an acidification
model with data from manipulated catchments in Norway. Nature 343:53-55.
Wright, R.F., E. Lotse, and E. Semb. 1994. Experimental acidification of alpine catchments at
Sogndal, Norway: results after 8 years. Water, Air, and Soil Pollution 72:297-315.
Final Risk and Exposure Assessment September 2009
Appendix 4-55
-------
Aquatic Acidification Case Study
Final Risk and Exposure Assessment September 2009
Appendix 4-56
-------
Aquatic Acidification Case Study
ATTACHMENT A
1. MODELING DESCRIPTIONS
1.1 MAGIC
Model of Acidification of Groundwater in Catchments (MAGIC) is a lumped-parameter
model of intermediate complexity, developed to predict the long-term effects of acidic deposition
on surface water chemistry (Cosby et al., 1985a, b). The model simulates soil solution chemistry
and surface water chemistry to predict the monthly and annual average concentrations of the
major ions in these waters. MAGIC consists of (1) a 2-10 submodel in which the concentrations
of major ions are assumed to be governed by simultaneous reactions involving sulfate (SC>42")
adsorption, cation exchange, dissolution-precipitation- speciation of aluminum (Al), and
dissolution-speciation of inorganic carbon; and (2) a mass balance submodel in which the flux of
major ions to and from the soil is assumed to be controlled by atmospheric inputs, chemical
weathering, net uptake and loss in biomass, and losses to runoff. At the heart of MAGIC is the
size of the pool of exchangeable base cations in the soil. As the fluxes to and from this pool
change over time in response to changes in atmospheric deposition, the chemical equilibria
between soil and soil solution shift, resulting in changes in surface water chemistry. Thus, the
degree and rate of change of surface water acidity depend both on flux factors and the inherent
biogeochemical characteristics of the affected soils.
Cation exchange is modeled using equilibrium (Gaines-Thomas) equations with
selectivity coefficients for each base cation and Al. SC>42" adsorption is represented by a
Langmuir isotherm. Al dissolution and precipitation are assumed to be controlled by equilibrium
with a solid phase of aluminum hydroxide (A1(OH)3). Al speciation is calculated by considering
hydrolysis reactions, as well as complexation with SC>42" and fluoride (F). The effects of carbon
dioxide (CO2) on pH and on the speciation of inorganic carbon are computed from equilibrium
equations. Organic acids are represented in the model as tri-protic analogues. Weathering and the
uptake rate of nitrogen are assumed to be constant. A set of mass balance equations for base
cations and strong acid anions are included (Cosby et al., 1985).
Given a description of the historical deposition at a site, the model equations are solved
numerically to give long-term reconstructions of surface water chemistry (for complete details of
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 1
-------
Aquatic Acidification Case Study
the model, see Cosby et al., 1985 a, b; Cosby et al., 1989). MAGIC was successfully used to
reconstruct the history of acidification and to simulate the future trends on a regional basis and in
a large number of individual catchments in both North America and Europe (e.g., Cosby et al.,
1989, 1990, 1996; Hornberger et al., 1989; Jenkins et al., 1990 a, b, c; Lepisto et al., 1988;
Norton et al., 1992; Sullivan and Cosby, 1998; Sullivan and Cosby, 2004; Whitehead et al.,
1988; Wright et al., 1990, 1994).
The input data required in this project for aquatic and soils resource modeling with the
MAGIC model (i.e., stream water, catchment, soils, deposition data) were assembled and
maintained in databases for each site modeled (electronic spreadsheets, text-based MAGIC
parameter files). Model outputs for each site were archived as text-based time-series files of
simulated variable values. The outputs were also concatenated across all sites and maintained in
electronic spreadsheets.
1.1.1 Input Data and Calibration
The calibration procedure requires that streamwater chemistry, soil chemical and physical
characteristics, and atmospheric deposition data be available for each watershed. The surface
water chemistry data needed for calibration are the concentrations of the individual base cations
(calcium [Ca2+], magnesium [Mg2+], sodium [Na+], and potassium [K+]) and acid anions
(chlorine [Cl~], SC>42", nitrate [N(V]) and the stream pH. The soil data used in the model
comprise physical properties, including soil depth and bulk density, and chemical properties,
such as soil pH, soil cation-exchange capacity, and exchangeable bases in the soil (Ca2+, Mg2+,
Na+, and K+). The deposition inputs required for calibration include the concentrations and
magnitudes of all major ions from wet, dry, and cloud deposition.
The acid-base chemistry modeling for this project was conducted using the year 2002 as
the Base Year. The effects models were calibrated to the available atmospheric deposition and
water chemistry data and then interpolated or extrapolated to yield Base Year estimates of lake
water chemistry in the year 2002, which served as the starting point for modeling of current
water chemistry (e.g., the years 2002 to 2100)
1.1.2 Lake, Stream, and Soil Data for Calibration
Several water chemistry databases were acquired for use in model calibration. Data were
derived primarily from the Environmental Monitoring and Assessment Program (EMAP) and
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 2
-------
Aquatic Acidification Case Study
Temporally Integrated Monitoring of Ecosystems (TIME) survey and monitoring efforts. The
required lake water and soil composition data for the modeling efforts included the following
measurements:
• Stream water composition— pH, acid neutralizing capacity (ANC), Ca2+, Mg2+, K+, Na+,
SO42", NCV, and Cl'
• Soil properties— thickness and total cation exchange capacity, exchangeable bases (Ca2+,
Mg2+, Na+, and K+) bulk density, porosity, and pH where available; the stream water
chemistry database also included dissolved organic and inorganic carbon, silicic acid
(H4SiO4), and inorganic monomeric Al (i.e., Ali).
1.1.3 Wet Deposition and Meteorology Data for Calibration
MAGIC requires, as atmospheric inputs for each site, estimates of the total annual
deposition (eq/ha/yr) of eight ions, and the annual precipitation volume (meters/year [m/yr]). The
eight ions are calcium (Ca2+), magnesium (Mg2+), Na+, K+, ammonium (NH4+), wet sulfate (SO4),
chlorine (Cl"), and nitrate (NOs). Total deposition of an ion at a particular site for any year can be
represented as combined wet, dry, and occult (i.e., cloud and fog) deposition:
TotDep = WetDep + DryDep + OccDep. (1)
Inputs to the MAGIC model are specified as wet deposition (the annual flux in
meq/m2/yr) and a dry, cloud and fog deposition factor (DDF, unitless), which is multiplied by the
wet deposition in order to get total deposition:
TotDep = WetDep x DDF, (2)
where DDF is the ratio of total deposition to wet deposition. It usually prescribed as equal to a
constant fraction of the wet deposition.
Given an annual wet deposition flux (WetDep), the ratio of dry deposition to wet
deposition (DryDep/WetDep), and the ratio of cloud and fog deposition to wet deposition
(OccDep/WetDep) for a given year at a site, the total deposition for that site and year is uniquely
determined.
In order to calibrate MAGIC, a time-series of total deposition is needed, beginning with a
reference calibration year and including the 140 years preceding the calibration year. The
procedure for providing a time-series of total deposition inputs to MAGIC follows.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 3
-------
Aquatic Acidification Case Study
The absolute values of wet deposition and DDF for each ion are provided for a Reference
Year at each site. For all the case study sites, a Reference Year of 2002 was used. Given the
Reference Year deposition values, deposition data for the historical and calibration periods, and
potentially any future deposition scenarios, can be estimated as a fraction of the Reference Year
value. For instance, to calculate the total deposition of a particular ion in some historical or
future year, j:
TotDepG) = [WetDep(O) x WetDepScaleG) ] x [ DDF(O) x DDF Scale(j)], (3)
where:
WetDep(O) = the Reference Year wet deposition (meq/m2/yr) of the ion
WetDepScale(j) = the scaled value of wet deposition in year j (expressed as a fraction of the
wet deposition in the Reference Year)
DDF(O) = the dry and occult deposition factor for the ion for the Reference Year
DDFScale(j) = the scaled value of the dry and occult deposition factor in year j (expressed
as a fraction of the DDF in the Reference Year).
The absolute value of wet deposition used for the Reference Year is time and space
specific—varying geographically within the region, varying locally with elevation, and varying
from year to year. It is desirable to have the estimates of wet deposition take into account the
geographic location and elevation of the site, as well as the year for which calibration data are
available. Therefore, estimates of wet deposition used for the Reference Year should be derived
from either direct measurements or a procedure (i.e., model) that has a high spatial resolution and
considers elevation effects. As described in Section 4.2.1.4, the absolute wet deposition values
used for the Reference Year in this project were derived from observed data from NADP
hybridized with high-spatially resolved estimates of rainfall.
The value of the DDF used for the Reference Year specifies the ratio between the
absolute amounts of wet and total deposition. While the wet deposition component varies
spatially and temporally, this ratio is not nearly so responsive. In large part, this is because the
varying wet deposition parameter is usually a large component of the total deposition and is
included in both the numerator and denominator of the ratio. For example, if in a given year at a
particular site, the wet deposition goes up, then the total deposition usually goes up; or, if the
elevation or aspect of a given site results in lower wet deposition, the total deposition also will
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 4
-------
Aquatic Acidification Case Study
often be lower. Therefore, estimates of the absolute values of DDF may be derived from a model
that has a relatively low spatial resolution and/or temporally smoothes the data. Estimates of the
absolute values of the DDF for the Reference Year at each site in this project were derived from
the Advanced Statistical Trajectory Regional Air Pollution (ASTRAP) model (Shannon, 1998),
as described below.
The long-term scaled sequences used to specify time-series of deposition inputs for
MAGIC simulations usually do not require detailed spatial or temporal resolution. Scaled
sequences of wet deposition or DDF (normalized to the same reference year) at neighboring sites
will be similar, even if the absolute wet deposition or DDF at the sites are different because of
factors such as local aspect or elevation. Therefore, if the scaled long-term patterns of any of
these do not vary much from place to place, estimates of the scaled sequences (as for estimates of
absolute DDF values) may be derived from a model that has a relatively low spatial resolution.
As described in the following sections, output from the ASTRAP model was used to construct
scaled sequences of both wet deposition and DDF for these case study areas.
1.1.4 Wet Deposition Data (Reference Year and Calibration Values)
The absolute values of wet deposition used for defining the Reference Year and for the
MAGIC calibrations must be highly site-specific. Estimated wet deposition data was used for
each site derived from the spatial interpolation model of Grimm and Lynch (2004), referred to
here as the Grimm model. The Grimm model is based on observed wet deposition concentrations
at NADP monitoring stations and radar-based precipitation estimates adjusted by elevation
effects, and provides a spatially-resolved estimate of wet deposition for each of the eight ions
required by MAGIC. The Grimm model makes a correction for changes in precipitation volume
(and thus wet deposition) based on the elevation at a given site. This correction arises from a
model of orographic effects on precipitation magnitudes derived from regional climatological
data.
The latitude, longitude, and elevation of the case study sites were provided as inputs to
the Grimm model. Estimates of quarterly and annual wet deposition and precipitation estimates
for each modeling site were found for the time period from 1983 through 2002. These annual
data were used to define the Reference Year and were used in conjunction with the ASTRAP
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 5
-------
Aquatic Acidification Case Study
historical deposition sequences for MAGIC calibration and simulation. The ASTRAP historical
sequences were scaled to match the Grimm estimates at each site.
1.1.5 Dry, Cloud, and Fog Deposition Data and Historical Deposition Sequences
Historical sequences of wet deposition and DDF were estimated using the ASTRAP
model. The ASTRAP model provided estimates of historical wet, dry, and occult deposition of
sulfur and oxidized nitrogen at modeled sites for the two case study areas. The ASTRAP sites
included 10 NADP deposition sites. For each of the modeled sites, ASTRAP produced wet, dry,
and occult deposition estimates of sulfur and oxidized nitrogen every 10 years, starting in 1900
and ending in 1990. The model outputs are smoothed estimates of deposition roughly equivalent
to a 10-year moving average centered on each of the output years. The wet, dry, and occult
deposition outputs of ASTRAP were used to estimate the absolute DDF for each site (using the
DryDep/WetDep and OccDep/WetDep ratios from the ASTRAP 19 output) and to set up the
scaled sequences of historical wet deposition and historical DDF for the calibration of each site
modeled in this project. Using the values and rates of change from the year 1900 ASTRAP
estimates, values for each time period going back to 1850 were estimated through linear
interpolation.
Because the ASTRAP sites are in the same region, but are not in the identical locations as
the MAGIC sites, and since deposition magnitudes are spatially- and elevation-sensitive, the
historical sequences of deposition at the ASTRAP sites were scaled to align with the deposition
estimates from the Grimm model for the MAGIC case study areas. First, the time series of wet
deposition estimates for each ASTRAP site were used to construct historical scaled sequences of
wet deposition. The absolute wet deposition outputs for the period 1850 to 1990 from each site
modeled in ASTRAP were normalized using their year 1990 values, converting them into scaled
sequences. It was then necessary to couple these historical scaled wet deposition sequences from
the year 1990 to the MAGIC Reference Year 2002. This coupling was accomplished using the
observed changes in wet deposition for the period 1983 to 2002 derived from the Grimm model.
With these site-specific deposition magnitudes and rates of change, the normalized ASTRAP
values were converted back into concentrations, though, now scaled to the deposition at the
MAGIC sites.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 6
-------
Aquatic Acidification Case Study
Because DDF is much less sensitive to location, the actual (nonscaled) estimates from
ASTRAP were used at the MAGIC sites. The value of DDF for the year 1990 was used as the
value of DDF for the Reference Year (i.e., no change was assumed for DDF for the period 1990
to 2002). The resulting time series of DDF values for the period 1900 to 2002 for each ASTRAP
site were normalized to the year 2002 values to provide historical scaled sequences of DDF at
each ASTRAP site.
1.1.6 Protocol for MA GIC Calibration and Simulation at Individual Sites
The aggregated nature of the MAGIC model requires that it be calibrated with observed
data from a system before it can be used to forecast potential system response to changes in
deposition. Calibration is achieved by specifying values of certain parameters within the model
that can be directly measured or observed in the system of interest (called fixed parameters). The
model is then run (using observed and/or assumed atmospheric and hydrologic inputs), and the
outputs (streamwater and soil chemical variables called criterion variables) are compared with
observed values of these variables. If the observed and simulated values differ, the values of
another set of parameters in the model (called optimized parameters) are adjusted to improve the
fit. After a number of iterations adjusting the optimized parameters, the simulated-minus-
observed values of the criterion variables usually converge to zero (within some specified
tolerance, or uncertainty). The model is then considered calibrated.
There are eight observed fixed parameters that are used to drive the estimate (i.e., current
soil exchangeable pool size and current output flux of each of the four base cations), and there
are eight parameters to be optimized in this procedure (i.e., the weathering and the selectivity
coefficient of each of the four base cations). If new assumptions or new values for any of the
observed fixed parameters or inputs to the model are adopted, the model must be recalibrated by
readjusting the optimized parameters until the simulated-minus-observed values of the criterion
variables again fall within the specified tolerance.
Estimates of the fixed parameters, the deposition inputs, and the target variable values to
which the model is calibrated all contain uncertainties. A "fuzzy optimization" procedure was
used for these case study sites to provide explicit estimates of the effects of these uncertainties.
The procedure consists of performing multiple calibrations at each site using random values of
the fixed parameters drawn from a range of fixed parameter values (representing uncertainty in
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 7
-------
Aquatic Acidification Case Study
knowledge of these parameters) and random values of Reference Year deposition drawn from a
range of total deposition estimates (representing uncertainty in these inputs). The final
convergence (i.e., completion) of the calibration is determined when the simulated values of the
criterion variables are within a specified "acceptable window" around the nominal observed
value. This acceptable window represents uncertainty in the target variable values being used to
calibrate the site.
Each of the multiple calibrations at a site begins with (1) a random selection of values of
fixed parameters and deposition, and (2) a random selection of the starting values of the
adjustable parameters. The adjustable parameters are then optimized using an algorithm seeking
to minimize errors between simulated and observed criterion variables. Calibration success is
judged when all criterion values simultaneously are within their specified acceptable windows,
(which may occur before the absolute possible minimum error is achieved). This procedure is
repeated 10 times for each site.
For this project, the acceptable windows for base cation concentrations in streams were
taken as +/- 2 microequivalents per liter (ueq/L) around the observed values. Acceptable
windows for soil exchangeable base cations were taken as +/- 0.2% around the observed values.
Fixed parameter uncertainty in soil depth, bulk density, cation exchange capacity, stream
discharge, and stream area were assumed to be +/- 10% of the estimated values. Uncertainty in
total deposition was +/- 10% for all ions.
The final calibrated model at the site is represented by the ensemble of parameter values
of all of the successful calibrations at the site. When performing a simulation of the site, each of
the calibrated parameter sets are run for a given historical or future scenario, generating an
ensemble of results. The results include multiple simulated values of each variable for each year,
all of which are acceptable in the sense of the calibration constraints applied in the fuzzy
optimization procedure. The median of all the simulated values within a year is taken to be "the
most likely" response for the site in that year. For this project, whenever single values for a site
are presented or used in an analysis, these values are the median of the ensemble values derived
from running each of the parameter sets for the site.
An estimate of the uncertainty (or reliability) of a simulated response to a given scenario
can also be derived from the multiple simulated values within a year resulting from the ensemble
simulations. For any year in a given scenario, the largest and smallest values of a simulated
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 8
-------
Aquatic Acidification Case Study
variable define the upper and lower 95% confidence bounds for that site's response for the
scenario under consideration. Thus, for all variables and all years of the scenario, a band of
simulated values can be produced from the ensemble simulations at a site that encompasses the
likely response (and provides an estimate of the simulation uncertainty) for any point in the
scenario. For these case study areas, whenever uncertainty estimates are presented, the estimate
is based on the range of values from the ensemble simulations for each of its sites.
In addition, uncertainty estimates for three classes of the major inputs of the model were
made through a sensitivity study, examining response of parameters and ability of the model to
attain calibration in response to variation in the following inputs:
• Soils data for calibration
• Stream water data calibration
• Deposition data calibration.
1.1.7 Combined Model Calibration and Simulation Uncertainty
The sensitivity analyses described above were designed to address specific assumptions
or decisions that had to be made to assemble the data for the 44 or 60 modeled sites in a form
that could be used for calibration of the model. In all cases, the above analyses address the
questions of what the effect would have been if alternate available choices had been taken. These
analyses were undertaken for a subset of sites for which the alternate choices were available at
the same sites. As such, the analyses above are informative, but they provide no direct
information about the uncertainty in calibration or simulation arising from the choices that were
incorporated into the final modeling protocol for all sites. That is, having made the choices about
soils assignments, high elevation deposition, and stream samples for calibration (and provided an
estimate of their inherent uncertainties), the need arises for a procedure for estimating
uncertainty at each and all of the individual sites using the final selected calibration and
simulation protocol.
These simulation uncertainty estimates were derived from the multiple calibrations at
each site provided by the "fuzzy optimization" procedure employed in this project. For each of
the modeled sites, 10 distinct calibrations were performed with the target values, parameter
values, and deposition inputs for each calibration, reflecting the uncertainty inherent in the
observed data for the individual site. The effects of the uncertainty in the assumptions made in
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 9
-------
Aquatic Acidification Case Study
calibrating the model (and the inherent uncertainties in the data available) can be assessed by
using all successful calibrations for a site when simulating the response to different scenarios of
future deposition. The model then produces an ensemble of simulated values for each site. The
median of all simulated values in a year is considered the most likely response of the site. The
simulated values in the ensemble can also be used to estimate the magnitude of the uncertainty in
the projection. Specifically, the difference in any year between the maximum and minimum
simulated values from the ensemble of calibrated parameter sets can be used to define an
"uncertainty" (or a "confidence") width for the simulation at any point in time. All 10 of the
successful model calibrations will lie within this range of values. These uncertainty widths can
be produced for any variable and any year to monitor model performance.
Direct comparison of simulated versus observed water chemistry values were compared
to determine the uncertainty and variability in the MAGIC model output. Average water
chemistry (SC>42", NCV, and ANC) simulated versus observed values during the calibration
period (i.e., reference year) were compared for all modeled sites. In addition, simulated versus
observed average yearly values for ANC for the period of 1980 to 2007 for 4 sites were
completed. The observed water chemistry data were from the, ALTM-LTM, VTSSS-LTM, and
TIME water quality measurement programs and represent annual average concentrations. The
statistic of Root Mean Squared Error (RMSE) were also calculated for predicted versus observed
values for both the calibration period and the period of 1980 to 2007. RMSE is a frequently used
measure of the differences between values predicted by a model or an estimator and the values
actually observed from the thing being modeled or estimated. The RMSE was based on an
annual average ANC over a 5-year period.
1.1.8 Results of the Uncertainty Analysis
Based on the MAGIC model simulations, the 95% confidence interval for the pre-
acidification and current average ANC concentrations of the 44 modeled lakes was 106.8 to
134.0 and 50.5 to 81.8 ueq/L, respectively, which is on average a 15 ueq/L difference in ANC
concentrations, or 10%. The 95% confidence interval for pre-acidification and current average
ANC concentrations of the 60 modeled streams was 91.9 to 110.9 and 53.4 to 62.4 ueq/L,
respectively, which is on average 8 ueq/L difference in ANC concentration, or 5%.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 10
-------
Aquatic Acidification Case Study
These direct comparisons show good agreement between simulated and observed water
quality values. Results of predicted versus observed average water chemistry during the
calibration period (i.e., reference year) are shown in Figures 1.1-1 and 1.1-2 for MAGIC
modeling. The model showed close agreement with measured values at all sites for the 1-year
comparison of modeled values. For all sites' SC>42", N(V, and ANC simulations, the RMSE for
predicted versus observed values were 0.1 |ieq/L, 0.05 |ieq/L, and 3.5 jieq/L for lakes in the
Adirondacks Case Study Area and 1.0 |ieq/L, 0.06 |ieq/L, and 1.0 |ieq/L for streams in the
Shenandoah Case Study Area. Plots of simulated and observed ANC values for the period of
1980 to 2007 are graphed in Figures 1.1-3 and 1.1-4 for two lakes in the Adirondacks Case
Study Area and for two streams Shenandoah Case Study Area. The RMSE of ANC was 11.8
|ieq/L and 4.0 jieq/L for the two lakes in the Adirondacks Case Study Area and was 7.8 jieq/L
and 5.1 jieq/L for the two streams in Shenandoah Case Study Area.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 11
-------
Aquatic Acidification Case Study
150
£•
•| 100
42", N(V
ANC, and pH during the model calibration period for each of the 44 lakes in the
Adirondacks Case Study Area. The black line is the 1:1 line.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 12
September 2009
-------
Aquatic Acidification Case Study
50 100
Observed Chemistry
150
300 -,
£• 200 -I
to
E
42", N(V.
ANC, and pH during the model calibration period for each of the 60 streams in the
Shenandoah Case Study Area. The black line is the 1:1 line.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 13
September 2009
-------
Aquatic Acidification Case Study
50
25
0)
O
-25
Observed
Simulated
Indian Lake
Dismal Pond
1975 1980 1985 1990 1995 2000 2005 2010
Figure 1.1-3. MAGIC simulated and observed values of ANC for two lakes in the
Shenandoah Case Study Area. Red points are observed data and the simulated values are
the line. The Root Mean Squared Error (RMSE) for ANC was 11.8 |ieq/L for Helton
Creek and 4.0 jieq/L for Nobusiness Creek.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 14
September 2009
-------
Aquatic Acidification Case Study
200
^150
i
"S-
0)
=: 100
0
< 50
o
en _
=5- 25
0)
z 0
oc
19
» Observed . . .. _ .
Helton Creek
Olt-w-,1 ,\nlnfi 1
•
^ 1
1
Nobusiness Creek
±~~~~ *— »—
^ ™ V ^
75 1980 1985 1990 1995 2000 2005 20
Years
10
Figure 1.1-4. MAGIC simulated and observed values of ANC for two lakes in the
Shenandoah Case Study Area. Red points are observed data and the simulated values are
the line. The Root Mean Squared Error (RMSE) for ANC was 11.8 |ieq/L for Helton
Creek and 4.0 jieq/L for Nobusiness Creek.
1.2 Critical Loads: Steady-State Water Chemistry Models
The critical load of acidity for lakes or streams was derived from present-day water
chemistry using a combination of steady-state models. Both the Steady-State Water Chemistry
(SSWC) model and First-order Acidity Balance model (FAB) is based on the principle that
excess base-cation production within a catchment area should be equal to or greater than the acid
anion input, thereby maintaining the ANC above a preselected level (Reynolds and Norris, 2001;
Posch et al. 1997). These models assume steady-state conditions and assume that all SO42 in
runoff originates from sea salt spray and anthropogenic deposition. Given a critical ANC
protection level, the critical load of acidity is simply the input flux of acid anions from
atmospheric deposition (i.e., natural and anthropogenic) subtracted from the natural (i.e.,
preindustrial) inputs of base cations in the surface water.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 15
September 2009
-------
Aquatic Acidification Case Study
Atmospheric deposition of NOX and SOX contributes to acidification in aquatic
ecosystems through the input of acid anions, such as N(V and SC>42". The acid balance of
headwater lakes and streams is controlled by the level of this acidifying deposition of NO3" and
SC>42" and a series of biogeochemical processes that produce and consume acidity in watersheds.
The biotic integrity of freshwater ecosystems is then a function of the acid-base balance, and the
resulting acidity-related stress on the biota that occupy the water.
The calculated ANC of the surface waters is a measure of the acid-base balance:
ANC = [BC]* - [AN]* (1)
where [BC]* and [AN]* are the sum of base cations and acid anions (N(V and S(V),
respectively. Equation (1) forms the basis of the linkage between deposition and surface water
acidic condition and the modeling approach used. Given some "target" ANC concentration
[ANCiimit]) that protects biological integrity, the amount of deposition of acid anions (AN) or
depositional load of acidity CL(A) is simply the input flux of acid anions from atmospheric
deposition that result in a surface water ANC concentration equal to the [ANCiimit] when
balanced by the sustainable flux of base cations input and the sinks of nitrogen and sulfur in the
lake and watershed catchment.
Critical loads for nitrogen and sulfur (CL(N) + CL(S) ) or critical load of acidity CL(A)
were calculated for each waterbody from the principle that the acid load should not exceed the
nonmarine, nonanthropogenic base cation input and sources and sinks in the catchment minus a
neutralizing to protect selected biota from being damaged:
CL(N) + CL(S) or CL(A) = BC*dep + BCW - Bcu - AN - ANCiimit (2)
where
BC deP = (BC*=Ca*+Mg*+K*+Na*), nonanthropogenic deposition flux of base cations and
BCW = the average weathering flux, producing base cations
Bcu (Bc=Ca*+Mg*+K*) = the net long-term average uptake flux of base cations in the biomass
(i.e., the annual average removal of base cations due to harvesting)
AN = the net long-term average uptake, denitrification, and immobilization of nitrogen anions
(e.g. NO3") and uptake of SO4"
= the lowest ANC-flux that protects the biological communities.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 16
-------
Aquatic Acidification Case Study
Since the average flux of base cations weathered in a catchment and reaching the lake or streams
is difficult to measure or compute from available information, the average flux of base cations
and the resulting critical load estimation were derived from water quality data (Henriksen and
Posch, 2001; Henriksen et al., 1992; Sverdrup et al., 1990). Weighted annual mean water
chemistry values were used to estimate average base cation fluxes, which were calculated from
water chemistry data collected from the Temporally Integrated Monitoring of Ecosystems
(TIME)/Long-Term Monitoring (LTM) monitoring networks, that include Adirondack Long-
term Monitoring (ALTM), Virginia Trout Stream Sensitivity Study (VTSSS), and the
Shenandoah Watershed Study (SWAS), and Environmental Monitoring and Assessment Program
(EMAP) (see Section 4. 1 .2. 1 of Chapter 4).
The preacidification nonmarine flux of base cations for each lake or stream, BC*0, is
-Bcu (3)
Thus, critical load for acidity can be rewritten as
CL(N) + CL(S) = BC*0 - AN - ANCHmit = Q ([BC*]0 - [AN] - [ANC]Hmit), (4)
where the second identity expresses the critical load for acidity in terms of catchment runoff (Q)
m/yr and concentration ([x] = X/Q).
The sink of nitrogen in the watershed is equal to the uptake (Nupt), immobilization (Nimm), and
denitrification (Nden) of nitrogen in the catchment. Thus, critical load for acidity can be rewritten
as
CL(N) + CL(S) = (fNupt + (1 - r)(Nimm + Nden)} + ( [BC]0* - [ANClimit])Q (5)
where f and r are dimensionless parameters that define the fraction of forest cover in the
catchment and the lake/catchment ratio. The in-lake retention of nitrogen and sulfur was assumed
to be negligible.
Equation 5 described the FAB model that was applied when sufficient data was available
to estimate the uptake, immobilization, and denitrification of nitrogen and the neutralization of
acid anions (e.g. N(V) in the catchment. In the case were data was not available, the contribution
of nitrogen anions to acidification was assumed to be equal to the nitrogen leaching rate (Nieach)
into the surface water. The flux of acid anions in the surface water is assumed to represent the
amount of nitrogen that is not retained by the catchment, which is determined from the sum of
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 17
-------
Aquatic Acidification Case Study
measured concentration of N(V and ammonia in the stream chemistry. This case describes the
SSWC model and the critical load for acidity is
CL(A) = Q ([BC*]0 - [ANC]limit) (6)
where the contribution of acid anions is considered as part of the exceedances calculation (see
Section 1.2.5, below).
For the assessment of current condition in both case study areas, the critical load
calculation described in Equation 6 was used for most lakes and streams. The lack of sufficient
data for quantifying nitrogen denitrification and immobilization prohibited the wide use of the
FAB model. In addition, given the uncertainty in quantifying nitrogen denitrification and
immobilization, the flux of nitrogen anions in the surface water was assumed to more accurately
reflect the contribution of NCV to acidification.
Several major assumptions are made: (1) steady-state conditions exist, (2) the effect of
nutrient cycling between plants and soil is negligible, (3) there are no significant nitrogen inputs
from sources other than atmospheric deposition, (4) ammonium leaching is negligible because
any inputs are either taken up by biota or adsorbed onto soils or nitrate compounds, and (5) long-
term sinks of sulfate in the catchment soils are negligible.
1.2.1 Preindustrial Base Cation Concentration
Present-day surface water concentrations of base cations are elevated above their steady-
state preindustrial concentrations because of base cation leaching through ion exchange in the
soil due to anthropogenic inputs of SC>42" to the watershed. For this reason, present-day surface
water base cation concentrations are higher than natural or preindustrial levels, which, if not
corrected for, would result in critical load values not in steady-state condition. To estimate the
preacidification flux of base cations, the present flux of base cations was estimated, BC*t, given
by
BC*t = BC*dep + BCw-Bcu+BCexc, (7)
where
exc = the release of base cations due to ion-exchange processes.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 18
-------
Aquatic Acidification Case Study
Assuming that deposition, weathering rate, and net uptake have not changed over time,
xc can be obtained by subtracting Equation 5 from Equation 7:
BCexc = BC t ~~ BC o (8)
This present-day excess production of base cations in the catchment was related to the
long-term changes in inputs of nonmarine acid anions (ASO 2 + ANOs) by the F-factor (see
below):
BCexc = F(ASO*2 + AN03) (9)
For the preacidification base cation flux, solving Equation 5 for BC*o and then
substituting Equation 8 for BCexc and explicitly describing the long-term changes in nonmarine
acid ion inputs:
BC*0 = BC*t - F (SO*4,t - SO*4!o +NO\t - NO*3;0) (10)
The preacidification N(V concentration, NO 3,0, was assumed to be zero.
1.2.2 F-factor
An F-factor was used to correct the concentrations and estimate preindustrial base
concentrations for lakes in the Adirondack Case Study Area. In the case of streams in the
Shenandoah Case Study Area, the preindustrial base concentrations were derived from the
MAGIC model as the base cation supply in 1860 (hindcast) because the F-factor approach is
untested in this region. An F-factor is a ratio of the change in nonmarine base cation
concentration due to changes in strong acid anion concentrations (Henriksen, 1984; Brakke et al.,
1990):
F =([BC*]t. [BC*]0)/([S04*]t - [S04*]o + [NO3*]t - [NO3*]o), (12)
where the subscripts t and 0 refer to present and preacidification conditions, respectively. If F=l,
all incoming protons are neutralized in the catchment (only soil acidification); at F=0, none of
the incoming protons are neutralized in the catchment (only water acidification). The F-factor
was estimated empirically to be in the range 0.2 to 0.4, based on the analysis of historical data
from Norway, Sweden, the United States, and Canada (Henriksen, 1984). Brakke et al. (1990)
later suggested that the F-factor should be a function of the base cation concentration:
F = sin (Ti/2 Q[BC*]t/[S]) (13)
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 19
-------
Aquatic Acidification Case Study
where
Q = the annual runoff (m/yr)
[S] = the base cation concentration at which F=l; and for [BC*]t>[S] F is set to 1. For
Norway [S] has been set to 400 milliequivalents per cubic meter (meq/m3)(circa.
8 mg Ca/L) (Brakke et al., 1990).
The preacidification SC>42" concentration in lakes, [SO4*]o, is assumed to consist of a
constant atmospheric contribution and a geologic contribution proportional to the concentration
of base cations (Brakke et al., 1989). The preacidification SC>42" concentration in lakes, [SO/Jo
was estimated from the relationship between [SO42-]o* and [BC]t* based on work completed by
Henriksen et al., 2002 as described by the following equation:
[SO42-]o* = 15 + 0.16 * [BC]t* (14)
1.2.3 ANCLimits
Four classes of ANC limits were estimated: Suitable ANC >50 ueq/L, Indeterminate
ANC 20 to 50 ueq/L, Marginal ANC 0 to 20 ueq/L, and Unsuitable ANC <0 ueq/L.
1.2.4 Sea Salt Corrections
The model applies a sea salt correction to the water chemistry concentrations. The
equations below were applied to all lakes and streams, and to all the New England states and
eastern Canadian provinces for the New England Governors and Eastern Canadian Premier
assessment. The equations correct for sea salt. An asterisk (*) indicates the value has been
corrected for sea salt. Units are in ueq/L.
Ca* = (Ca-(CLx 0.0213)) (14)
Mg* = (Mg - (CL x 0.0669)) (15)
Na* = (Na - (CL x 0.557)) (16)
K* = (K - (CL x 0.0.0206)) (17)
= (SO4-(CLx0.14)) (18)
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 20
-------
Aquatic Acidification Case Study
1. 2. 5 Critical load exceedance
It is not possible to define a maximal loading for a single total of acidity (i.e., both
nitrogen and sulfur deposition) because the acid anions sulfate and nitrate behave differently in
the way they are transported with hydrogen ions; one unit of deposition of sulfur will not have
the same net effect on surface water ANC as an equivalent unit of nitrogen deposition. However,
the individual maximum and minimum critical loads for nitrogen and sulfur are defined when
nitrogen or sulfur do not contribute to the acidity in the water. The maximum critical load for
sulfur (DLmax(S)) is the following:
CLmax(S) = [( [BC]0* - [ANQeveiDQ] (3)
when nitrogen deposition does not contribute to the acidity balance. Given the assumption that
the long-term sinks of sulfate in the catchment soils are negligible, the amount of sulfur entering
the catchment is equal to the amount loaded to the surface water. For this reason, the minimal
amount of sulfur is equal to zero:
CLmm(S) = 0 (4)
In the case of nitrogen, CLm;n(N) is the minimum amount of deposition of total nitrogen (NHX +
NOX) that catchment processes can effectively remove (e.g., Nupt + Nimm + Nden +Nret) without
contributing to the acidic balance:
CLmm(N) = fNupt + (l-r)( Nimm + Nden) (5)
The CLmax(N) is the load for total nitrogen deposition when sulfur deposition is equal to
zero:
CLmax(N) = fNupt + (l-r)( Nimm + Nden) + [( [BC]0* - [ANCieVei])Q]. (6)
In reality, neither nitrogen nor sulfur deposition will ever be zero, so the depositional load for the
deposition of one is fixed by the deposition of the other, according to the line defining in
Figure 1.2-1
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 21
-------
Aquatic Acidification Case Study
DLmin(IM)
DLmax(N)
N Deposition
Figure 1.2-1. The depositional load function defined by the model.
The thick lines indicate all possible combinations of depositional loads of nitrogen and
sulfur acidity that a catchment can receive and still maintain an ANC concentration equal to its
ANQimit. Note that in the above formulation, individual depositional loads of nitrogen and sulfur
are not specified; each combinations of depositions (Sdep and Ndep) fulfilling Equations 2 through
6.
Finally, the calculated exceedances of the critical load of acidity Ex(A), when the Nieachis
considered the contribution of acid anions acidification is the following:
Ex(A) = S*dep + Nleach - CL(A), (11)
where S*dep is the amount of sulfur deposited in the catchment (assuming that all SC>42" deposited
leaches into the waterbody) and Nieach is the amount of deposited nitrogen, Ndep, that moves into
the water.
While SO42" is assumed to be a mobile anion (Sieach = S*dep), nitrogen is to a large extent
retained in the catchment by various processes; therefore, Ndep cannot be used directly in the
exceedances calculation. Only present-day exceedances can be calculated from the leaching of
nitrogen, Nieach, which is determined from the sum of measured concentrations of N(V and
ammonia in the stream chemistry. No nitrogen deposition data are required for exceedance
calculations; however, Ex(A) quantifies only the exceedances at present rates of retention of
nitrogen in the catchment.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 22
September 2009
-------
Aquatic Acidification Case Study
1.2.6 Lake-to- Catchment-Area Ratios
Lake and catchment areas data are from the Environmental Monitoring and Assessment Program
(EMAP) (see Section 4.1.2.1 of Chapter 4).
1.2.7 Denitrification and N Immobilization in Soils
It was assumed that the denitrification fractions are related to the soil types in the
catchments. In deeply drained podzolic soils, denitrification values are generally low. However,
high values may occur in areas with peatsoils (Klemedtsson and Svensson, 1988; De Vries et al.,
1994). Therefore, the average denitrification fraction for each catchment was approximated by
the following linear relationship:
fde = 0.1+0.7fpeat (12)
where fpeat is the fraction of peatlands in the catchment area.
For the long-term immobilization of nitrogen in forest soils (Nimm), a constant value of 2
kg N/ha/yr was used. This value represents the lower end of the range suggested for European
critical load calculations (Posch et al., 1997).
1.2.8 Uncertainty and Variability
There is uncertainty associated with the parameters in the steady-state critical load model
used to estimate aquatic critical loads. The strength of the critical load estimate and the
exceedance calculation relies on the ability to estimate the catchment-average base cation supply
(i.e., input of base cations from weathering of bedrock and soils and air), runoff, and surface
water chemistry. The uncertainty associated with runoff and surface water measurements is fairly
well known. However, the ability to accurately estimate the catchment supply of base cations to
a waterbody is still poorly known. This is important because the catchment supply of base
cations from the weathering of bedrock and soils is the factor that has the most influence on the
critical load calculation and also has the largest uncertainty (Li and McNulty, 2007). Although
the approach to estimate base cation supply in the case study areas (e.g., F-factor) approach has
been widely published and analyzed in Canada and Europe, and has been applied in the United
States (e.g., Dupont et al., 2005), the uncertainty in this estimate is unclear and is likely large.
For this reason, an uncertainty analysis of the state-steady critical load model was completed to
evaluate the uncertainty in the critical load and exceedances estimations.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 23
-------
Aquatic Acidification Case Study
A probabilistic analysis using a range of parameter uncertainties was used to assess: (1)
the degree of confidence in the exceedance values and (2) the coefficient of variation (CV) of the
critical load and exceedance values. The probabilistic framework is Monte Carlo, whereby each
steady-state input parameter varies according to specified probability distributions and their
range of uncertainty (Table 1.2-1). The purpose of the Monte Carlo methods was to propagate
the uncertainty in the model parameters in the steady-state critical load model.
Table 1.2-1. Parameters Used and their Uncertainty Range. The Range of Surface Water
Parameters (e.g., CA, MG, CL, NA, NOs, 804) were Determined from Surface Water Chemistry
Data for the Period from 1992 to 2006 from the LTM-TIME Monitoring Network. Runoff(Q)
and Acidic Deposition were Set at 50% and 25%
Parameter
Q
CA
MG
CL
NA
NO3
SO4
Acidic Deposition
(NOX & SO4)
Units
(ieq/L
(ieq/L
(ieq/L
(ieq/L
(ieq/L
(ieq/L
(ieq/L
meq/L
Uncertainty range
50%
65%
64%
52%
58%
30%
57%
25%
Distribution
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Lognormal
Within the Monte Carlo analysis, model calculations were run a sufficient number of
times (i.e., 1,000 times) to capture the range of behaviors represented by all variables. The
analysis tabulated the number of lakes where the confidence interval is entirely below the critical
load, the confidence interval is entirely above the critical load, and the confidence interval
straddles zero. Similar results are given for the number of sites with all realizations above the
critical load, all realizations below the critical load, and some realizations above and some below
the critical load. An inverse cumulative distribution function for exceedances was constructed
from the 1000 model runs for each site, which describes the probability of a site to exceed its
critical load. For each site, the probability of exceeding its critical load (i.e. probability of
exceedance) is determined at the percent of the cumulative frequency distribution that lies above
zero. The probability of exceedance, where the percentage of the cumulative frequency
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 24
September 2009
-------
Aquatic Acidification Case Study
distribution lies above zero, was calculated for all sites and assigned to one of the following five
classes:
• 0-5% probability: unlikely to be exceeded;
• 5-25% probability: relatively low risk of exceedance;
• 25-75% probability: potential risk of exceedance;
• 75-95% probability: relatively high risk of exceedance;
• > 95% probability: highly likely to be exceeded.
This gives us a measure of the degree of confidence in whether the site exceeds its critical
load. The CDF for Little Hope Pond is shown in Figure 1.2-2.
The coefficient of variation (CV) was also calculated on each site for both the critical
load and exceedance calculations. The coefficient of variation represents the ratio of the standard
deviation to the mean, and it is a useful statistic for comparing the degree of variation in the data.
The coefficient of variation allows a determination of how much uncertainty (risk) comparison to
its mean.
1
_ N=1000
0.9
0.8
0.7
I 0.5 I
2 Oj4 i exceedence, where
°- | the % of the
0.3 | cumulative
( frequency chart lies
0-2 | above zero
0.1 [
0
-400 -300 -200 -100 0 100 200
Exceedance (meq/ha/yr)
Figure 1.2-2. The inverse cumulative frequency distribution for Little Hope Pond. The x-
axis shows critical load exceedance in meq/ha/yr and y-axis is the probability. The
dashed lines represent zero exceedance. In the case of Little Hope Pond, the dash line
divides mostly the probability distribution on the left hand side, indicating Little Hope
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment A - 25
-------
Aquatic Acidification Case Study
Pond has a relative low probability of being exceeded (0.3). Critical load and
exceedances values were based on a critical level of protection of ANC = 50 ueq/L.
1.2.8.1 Results of the uncertainty analysis
The means and coefficients of variation for critical load (CL(A)) and exceedances
(EX(A)) values are shown for all sites in Table 1.2-2 for four ANC limits (0, 20, 50, 100 ueq/L.
The average coefficients of variation for the various critical load values for all sites are
remarkably low except for those calculated using a critical limit of 100 ueq/L. It is noticeable
that though all the relevant input parameters have spreads of 25% to 65%, the CVs for CL(A) are
only 4%, 5%, 9%, and 100% for critical load limits of 0%, 20%, 50%, and 100% ueq/L,
respectively. In the case of the absolute value of the exceedances (EX(A)), the average CVs for
all sites are higher, but still relatively low at 18%, 17%, 25%, and 33%. The individual
coefficients of variation for each sites and an ANC limit of 50 ueq/L are shown in Figure 1.2-3.
Although the average CV is relatively small for the population of sites modeled, individual site
CV can varies from 1% to 45% for CL(A) and to 5% to over 100% for EX(A). This difference is
due to the high degree of uncertainty in site specific parameters for particular sites and a low
degree of confidence in the exceedance value itself for these sites. In addition, when the mean
value is near zero, as is the case for exceedance values, the coefficient of variation is sensitive to
small changes in the mean, which likely explains why some sites have high CV compare to
others.
Table 1.2-2. Means and Coefficients of Variation of Critical Loads and Exceedances for Surface
Water.
Parameter
CL(A)
EX(A)
Critical load Limit
(^eq/L)
0
20
50
100
0
20
50
100
Mean
(meq/L)
247.8
227.0
196.7
140.3
-178.3
-157.6
-127.2
-75.0
Coefficient of
variation (%)
4
5
9
100
18
17
25
33
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 26
September 2009
-------
Aquatic Acidification Case Study
-ion ^------------------------------------^^^^
100
on _
?
*•" KO
0
4.0
on
Q
•
_
•
•
•
•
I
+
•
1
i
CL(A) EX(A)
Figure 1.2-3. Coefficients of variation of surface water critical load for acidity CL(A)
and exceedances (EX(A)). Critical load and exceedances values were based on a critical
level of protection of ANC = 50 ueq/L.
The probability of exceedance results, where the percentage of the cumulative frequency
distribution lies above zero, are shown in Figure 1.2-4. Those areas that have less than 5%
probability of exceedance are those with a high degree of confidence that the critical loads are
not exceeded; conversely, areas with more than a 95% probability of exceedance are the most
certain to be exceeded.
For the sites in the aquatic case study areas, the probability of exceeding the critical load
at an ANC limit of 0, 20, 50, and 100 ueq/L were relatively high. The waterbodies that exceeded
their critical loads had a greater than 80% probability of doing so. The range of probability of
exceedance was from 80% to 98%, indicating a relatively high confidence that these sites
exceeded their critical load. The results suggest a relatively robust estimate of critical loads and
exceedance rates for the case study areas. It is important to note that this analysis may understate
the actual uncertainty because some of the range and distribution types of parameters are not
well known for the United States at this time.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 27
September 2009
-------
Aquatic Acidification Case Study
Critical Load Exceedances
Probability
• 0-5% - unlikely to be exceeded
5-25% - relatively low risk of exceedance
25-75% - potential risk of ecceedence
75-95% - relatively high risk of exceedence
> 95% - highly likely to be exceeded
Figure 1.2-4. Probability of exceedance of critical load for acidity for 2002.
Final Risk and Exposure Assessment
Appendix 4, Attachment A - 28
September 2009
-------
Aquatic Acidification Case Study
ATTACHMENT B
1. EMAP/TIME/LTM PROGRAMS
The EPA Environmental Monitoring and Assessment Program (EMAP) began regional
surveys of the nation's surface waters in 1991 with a survey of northeastern United States lakes.
Since then, EMAP and Regional-EMAP (REMAP) surveys have been conducted on lakes and
streams throughout the country. The objective of these EMAP surveys is to characterize
ecological condition across populations of surface waters. EMAP surveys are probability surveys
where sites are picked using a spatially balanced systematic randomized sample so that the
results can be used to make estimates of regional extent of condition (e.g., number of lakes,
length of stream). EMAP sampling typically consists of measures of aquatic biota (e.g., fish,
macroinvertebrates, zooplankton, periphyton), water chemistry, and physical habitat. Of
particular interest with respect to acidifying deposition effects were two EMAP surveys
conducted in the 1990s, the Northeastern Lake Survey and the Mid-Atlantic Highlands
Assessment of streams (MAHA). The Northeastern Lake Survey was conducted in summer from
1991 to 1994 and consisted of 345 randomly selected lakes in New York, New Jersey, Vermont,
New Hampshire, Maine, Rhode Island, Connecticut, and Massachusetts (Whittier et al., 2002).
To make more precise estimates of the effects of acidic deposition, the sampling grid was
intensified to increase the sample site density in the Adirondack Mountains and New England
Uplands areas known to be susceptible to acidic deposition. The MAHA study was conducted on
503 stream sites from 1993 to 1995 in the states of West Virginia, Virginia, Pennsylvania,
Maryland, Delaware, and the Catskill Mountain region of New York (Herlihy et al., 2000).
Sampling was done during spring baseflow. Sample sites were restricted to first through third
order streams as depicted on the USGS 1:100,000 digital maps used in site selection. To make
more precise estimates of the effects of acidic deposition, the sampling grid was intensified to
increase the sample site density in the Blue Ridge, Appalachian Plateau, and Ridge section of the
Valley and Ridge ecoregions. Results from both of these surveys were used to develop and select
the sampling sites for the Temporally Integrated Monitoring of Ecosystems (TIME) program,
which is described below.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment B - 1
-------
Aquatic Acidification Case Study
2. TEMPORALLY INTEGRATED MONITORING OF ECOSYSTEMS
AND LONG-TERM MONITORING PROGRAMS
There are two surface water chemistry monitoring programs, administered by EPA, that
are especially important to inform the assessment of aquatic ecosystem responses to changes in
atmospheric deposition. These are the TIME program (Stoddard et al., 2003) and the Long-term
Monitoring (LTM) project (Ford et al., 1993; Stoddard et al., 1998). These efforts focus on
portions of the United States most affected by the acidifying influence of sulfur and nitrogen
deposition, including lakes in the Adirondack Mountains of New York and in New England, and
streams in the Northern Appalachian Plateau and Blue Ridge in Virginia and West Virginia.
Both projects are operated cooperatively with numerous collaborators in state agencies, academic
institutions, and other federal agencies. The TIME program and LTM project have slightly
different objectives and structures, which are outlined below. Stoddard et al. (2003) conducted a
thorough trends analysis of the TIME and LTM data.
2.1 TIME Program
At the core of the TIME project is the concept of probability sampling, whereby each
sampling site is chosen statistically from a predefined target population. Collectively, the
monitoring data collected at the sites are representative of the target population of lakes or
streams in each study region. The target populations in these regions include lakes and streams
likely to be responsive to changes in acidifying deposition, defined in terms of acid neutralizing
capacity (ANC), which represents an estimate of the ability of water to buffer acid. Measurement
of Gran ANC uses the Gran technique to find the inflection point in an acid-base titration of a
water sample (Gran, 1952). In the Northeast, the TIME target population consists of lakes with a
Gran ANC <100 microequivalents per liter (ueq/L). In the mid-Atlantic, the target population is
upland streams with Gran ANC <100 ueq/L. In both regions, the sample sites selected for future
monitoring were selected from the EMAP survey sites in the region (Section AX3.2.1.1) that met
the TIME target population definition. Each lake or stream is sampled annually (in summer for
lakes; in spring for streams), and results are extrapolated with known confidence to the target
population(s) as a whole using the EMAP site population expansion factors or weights (Larsen
and Urquhart, 1993; Larsen et al., 1994; Stoddard et al., 1996; Urquhart et al., 1998). TIME sites
were selected using the methods developed by the EMAP (Herlihy et al., 2000; Paulsen et al.,
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment B - 2
-------
Aquatic Acidification Case Study
1991;). The TIME program began sampling northeastern lakes in 1991. Data from 43 lakes in the
Adirondack Mountains can be extrapolated to the target population of low ANC lakes in that
region. There are about 1,000 low-ANC Adirondack lakes, out of a total population of 1,842
lakes with surface area greater than 1 hectare (ha). Data from 30 lakes (representing about 1,500
low-ANC lakes, out of a total population of 6,800) form the basis for TIME monitoring in New
England. Probability monitoring of mid-Atlantic streams began in 1993. Stoddard et al. (2003)
analyzed data from 30 low-ANC streams in the Northern Appalachian Plateau (representing
about 24,000 kilometer (km) of low-ANC stream length out of a total stream length of 42,000
km).
The initial 1993 to 1995 EMAP-MAHA sample in the mid-Atlantic was not dense
enough to obtain enough sites in the TIME target population in the Blue Ridge and Valley and
Ridge ecoregions. In 1998, another denser random sample was conducted in these ecoregions to
identify more TIME sites. After pooling TIME target sites taken from both MAHA and the 1998
survey, there are now 21 TIME sites in the Blue Ridge and Ridge and Valley that can be used for
trend detection in this aggregate ecoregion in the mid-Atlantic in addition to the northern
Appalachian Plateau ecoregion.
2.2 LTM Program
As a complement to the statistical lake and stream sampling in TIME, the LTM Program
samples a subset of generally acid-sensitive lakes and streams that have long-term data, many
dating back to the early 1980s. These sites are sampled 3 to 15 times per year. This information
is used to characterize how some of the most sensitive aquatic systems in each region are
responding to changing deposition, as well as giving information on seasonal variation in water
chemistry. In most regions, a small number of higher-ANC (e.g., Gran ANC >100 ueq/L) sites
are also sampled, and these help to separate temporal changes due to acidifying deposition from
those attributable to other disturbances (e.g., climate, land use change). Because of the
availability of long-term records (i.e., more than two decades) at many LTM sites, their trends
can also be placed in a better historical context than those of the TIME sites, where data are only
available starting in the 1990s. Monitored water chemistry variables include pH, ANC, major
anions and cations, monomeric aluminum (Al), silicon (Si), specific conductance, dissolved
organic carbon, and dissolved inorganic carbon. The field protocols, laboratory methods, and
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment B - 3
-------
Aquatic Acidification Case Study
quality assurance procedures are specific to each team of investigators. This information is
contained in the cited publications of each research group. The EMAP and TIME protocols and
quality assurance methods are generally consistent with those of the LTM cooperators. Details of
LTM data from each region are given below.
New England lakes: The LTM project collects quarterly data from lakes in Maine
(sampled by the University of Maine) (Kahl et al., 1991; Kahl et al., 1993) and Vermont (data
collected by the Vermont Department of Environmental Conservation) (Stoddard and Kellogg,
1993; Stoddard et al., 1998). Data from 24 New England lakes were available for the trend
analysis reported by Stoddard et al. (2003) for the period 1990 to 2000. In addition to quarterly
samples, a subset of these lakes have outlet samples collected on a weekly basis during the
snowmelt season; these data are used to characterize variation in spring chemistry. The majority
of New England LTM lakes have mean Gran ANC values ranging from 20 to 100 ueq/L; two
higher ANC lakes (i.e., Gran ANC between 100 and 200 ueq/L) are also monitored.
Adirondack lakes: The trend analysis of Stoddard et al. (2003) included data from 48
Adirondack lakes, sampled monthly by the Adirondack Lake Survey Corporation (Driscoll and
Van Dreason, 1993; Driscoll et al., 1995); a subset of these lakes are sampled weekly during
spring snowmelt to help characterize spring season variability. Sixteen of the lakes have been
monitored since the early 1980s; the others were added to the program in the 1990s. The
Adirondack LTM dataset includes both seepage and drainage lakes, most with Gran ANC values
in the range -50 to 100 ueq/L; three lakes with Gran ANC between 100 ueq/L and 200 ueq/L are
also monitored.
Appalachian Plateau streams: Stream sampling in the Northern Appalachian Plateau is
conducted about 15 times per year, with the samples spread evenly between baseflow (e.g.,
summer and fall) and high flow (e.g., spring) seasons. Data from four streams in the Catskill
Mountains (collected by the U.S. Geological Survey) (Murdoch and Stoddard, 1993), and five
streams in Pennsylvania (collected by Pennsylvania State University) (DeWalle and Swistock,
1994) were analyzed by Stoddard et al. (2003). All of the northern Appalachian LTM streams
have mean Gran ANC values in the range 25 to 50 ueq/L.
Upper Midwest lakes: Forty lakes in the Upper Midwest were originally included in the
LTM project, but funding in this region was terminated in 1995. The Wisconsin Department of
Natural Resources (funded by the Wisconsin Acid Deposition Research Council, the Wisconsin
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment B - 4
-------
Aquatic Acidification Case Study
Utilities Association, the Electric Power Research Institute, and the Wisconsin Department of
Natural Resources) has continued limited sampling of a subset of these lakes, as well as carrying
out additional sampling of an independent subset of seepage lakes in the state. The data reported
by Stoddard et al. (2003) included 16 lakes (both drainage and seepage) sampled quarterly
(Webster et al., 1993), and 22 seepage lakes sampled annually in the 1990s. All of the Upper
Midwest LTM lakes exhibit mean Gran ANC values from 30 to 80 ueq/L.
Ridge/Blue Ridge streams: Data from the Ridge and Blue Ridge provinces consist of a
large number of streams sampled quarterly throughout the 1990s as part of the Virginia Trout
Stream Sensitivity Study (Webb et al., 1989), and a small number of streams sampled more
intensively (as in the Northern Appalachian Plateau). A total of 69 streams, all located in the
Ridge section of the Ridge and Valley province, or within the Blue Ridge province, and all
within the state of Virginia, had sufficient data for the trend analyses by Stoddard et al. (2003).
The data are collected cooperatively with the University of Virginia and the National Park
Service. Mean Gran ANC values for the Ridge and Blue Ridge data range from 15 to 200 ueq/L,
with 7 of the 69 sites exhibiting mean Gran ANC >100 ueq/L.
Final Risk and Exposure Assessment September 2009
Appendix 4, Attachment B - 5
-------
-------
September 2009
Appendix 5
Terrestrial Acidification Case Study
Final
EPA Contract Number EP-D-06-003
Work Assignment 3-62
Project Number 0209897.003.062
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27709
Prepared by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
INTERNATIONAL
-------
-------
Terrestrial Acidification Case Study
TABLE OF CONTENTS
Acronyms and Abbreviations xi
1.0 Background 1
1.1 Indicators, Ecological Endpoints, and Ecosystem Services 2
1.1.1 Indicators 2
1.1.2 Ecological Endpoints 7
1.1.3 Ecosystem Services 10
1.2 Case Study Areas 11
1.2.1 GIS Analysis ofNational Sensitivity 11
1.2.2 Selection of Case Study Areas 12
1.2.3 Sugar Maple 17
1.2.3.1 Kane Experimental Forest 17
1.2.3.2 Plot Selection for Kane Experimental Forest Case Study Area 18
1.2.4 Red Spruce 20
1.2.4.1 Hubbard Brook Experimental Forest 20
1.2.4.2 Plot Selection for Hubbard Brook Experimental Forest Case
Study Area 22
2.0 Approach and Methods 26
2.1 Chosen Method 30
2.1.1 Critical Load Equations and Calculations 31
2.1.1.1 Simple Mass Balance Calculations 31
2.1.1.2 Deposition Relative to Critical Load Calculations 37
2.1.1.3 Critical Load Function 38
2.1.2 Critical Load Data Requirements 39
2.1.2.1 Data Requirements and Sources 39
2.1.2.2 Select!on of Indicator Values 42
2.1.2.3 Case Study InputData 45
2.2 Critical Load Function Response Curves Associated with the Three Levels of
Protection 49
3.0 Results 50
3.1 Critical Load Estimates 50
3.1.1 Sugar Maple 50
3.1.2 Red Spruce 58
3.2 Recommended Parameter Values and Critical Loads 62
3.3 Current Conditions 63
4.0 Expansion of Critical Load Assessments for Sugar Maple and Red Spruce 75
4.1 Critical Load Assessments 75
4.2 Relationship between Atmospheric Nitrogen and Sulfur Deposition and Tree Growth 91
5.0 Uncertainty Analysis 92
5.1 Kane Experimental Forest and Hubbard Brook Experimental Forest Case Study Areas 92
5.2 Expansion of Critical Load Assessments 93
6.0 References 96
Final Risk and Exposure Assessment September 2009
Appendix 5 - i
-------
Terrestrial Acidification Case Study
Attachment A: Relationship Between Atmospheric Nitrogen and Sulfur Deposition and
Sugar Maple and Red Spruce Tree Growth 1
1. Introduction 1
2. Source of Data for Analyses 2
3. Regression Analyses Methodology and Results 7
4. Additional Sources of Variability Influencing the Relationships between Tree Volume
Growth and Nitrogen Deposition and Critical Load Exceedance 18
4.1 State-Specific Variables 18
4.2 Dead Trees 19
4.3 Other Factors 19
5. Conclusions 21
LIST OF FIGURES
Figure 1.1-1. Conceptual impacts of acidifying deposition on soil Ca2+ depletion, tree
physiology, and forest ecosystem health and sustainability (recreated from
DeHayes et al., 1999)
Figure 1.1-2. Areal coverages of red spruce and sugar maple tree species within the
continental United States (USFS, 2006) 10
Figure 1.2-1. Map of areas of potential sensitivity of red spruce and sugar maple to
acidification in the United States (see Table 1.2-1 for listing of data
sources to produce this map) 12
Figure 1.2-2. Location of the Kane Experimental Forest (Horsley etal., 2000) 17
Figure 1.2-3. The seven plots used to evaluate critical loads of acidity in the Kane
Experimental Forest 19
Figure 1.2-4. Location of theHubbard Brook Experimental Forest 21
Figure 1.2-5. Vegetation cover (NLCD, 2001) and location of Watershed 6 of Hubbard
Brook Experimental Forest 24
Figure 1.2-6. Grid units within Watershed 6 of Hubbard Brook Experimental Forest. The
red outline delineates the spruce-fir forest type. The dotted grid cell areas
indicate the grid units with high proportions of red spruce and represents
the composite plot area for the Hubbard Brook Experimental Forest Case
Study Area 25
Figure 1.2-7. Location of case study plots within Watershed 6 of Hubbard Brook
Experimental Forest 26
Figure 2.1-1. The critical load function created from the calculated maximum and
minimum levels of total nitrogen and sulfur deposition (eq/ha/yr). The
grey areas show deposition levels less than the established critical loads.
The red line is the maximum critical level of sulfur deposition (valid only
when nitrogen deposition is less than the minimum critical level of
nitrogen deposition [blue dotted line]). The flat line portion of the curves
Final Risk and Exposure Assessment September 2009
Appendix 5 - ii
-------
Terrestrial Acidification Case Study
indicates nitrogen deposition corresponding to the CLm;n(N) (i.e., nitrogen
absorbed by nitrogen sinks within the system) 39
Figure 2.1-2. The relationship between the Bc/Al ratio in soil solution and the percentage
of tree species (found growing in North America) exhibiting a 20%
reduction in growth relative to controls (after Sverdrup and Warfvinge,
1993b) 44
Figure 2.1-3. The relationship between soil solution Bc/Al ratio and stem or root growth
in sugar maple (from Sverdrup and Warfvinge, 1993b) 44
Figure 2.1-4. The relationship between soil solution Bc/Al ratio and biomass or root
growth in red spruce (from Sverdrup and Warfvinge, 1993b) 45
Figure 2.2-1. An example of the critical load function response curves associated with the
three (Be/Al)crit ratios and the associated levels of protection of tree health.
The flat line portion of the curves indicates total nitrogen deposition
corresponding to the CLm;n (nitrogen absorbed by nitrogen sinks within the
system) 50
Figure 3.1-1. The critical load function response curves detailing the lowest critical load
estimates for Plot 1 of the Kane Experimental Forest (refer to Table 3.1-1
for the parameters corresponding to each of the curves). The flat line
portion of the curves indicates total nitrogen deposition corresponding to
the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system) 56
Figure 3.1-2. The critical load function response curves detailing the highest critical load
estimates for Plot 1 of the Kane Experimental Forest (refer to Table 3.1-1
for the parameters corresponding to each of the curves). The flat line
portion of the curves indicates total nitrogen deposition corresponding to
the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system 57
Figure 3.1-3. The critical load function response curves detailing the lowest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area
(refer to Table 3.1-10 for the parameters corresponding to each of the
curves). The flat line portion of the curves indicates total nitrogen
deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen
sinks within the system) 60
Figure 3.1-4. The critical load function response curves detailing the highest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area
(refer to Table 3.1-10 for the parameters corresponding to each of the
curves). The flat line portion of the curves indicates total nitrogen
deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen
sinks within the system) 60
Figure 3.3-1. Plot 1 Kane Experimental Forest critical load function response curves,
detailing the lowest critical load estimates for Kane Experimental Forest
Case Study Area (refer to Table 3.1-1 for the parameters corresponding to
each of the curves). The 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels ((N+S)COmb) were greater than the critical loads of
nitrogen and sulfur at all levels of protection ((Bc/Al)crit= 0.6, 1.2, and
Final Risk and Exposure Assessment September 2009
Appendix 5 - iii
-------
Terrestrial Acidification Case Study
10.0). The flat line portion of the curves indicates total nitrogen deposition
corresponding to the CLm;n (N) (nitrogen absorbed by nitrogen sinks
within the system) 68
Figure 3.3-2. Plot 1 critical load function response curves, detailing the highest maximum
deposition load estimates for Kane Experimental Forest Case Study Area
(refer to Table 3.1-1 for the parameters corresponding to each of the
curves). The 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels ((N+S)comb) were greater than the critical load of total nitrogen and
sulfur deposition calculated with the highest level of protection (Bc/Al)crit
= 10.0). The flat line portion of the curves indicates total nitrogen
deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen
sinks within the system) 69
Figure 3.3-3. Critical load function response curves, detailing the lowest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area
(refer to Table 3.1-10 for the parameters corresponding to each of the
curves). The 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels ((N+S)COmb) were greater than the critical load of total nitrogen and
sulfur calculated with the highest and the intermediate levels of protection
((Bc/Al)crit= 1.2 and 10.0). The flat line portion of the curves indicates
total nitrogen deposition corresponding to the CLmin(N) (nitrogen absorbed
by nitrogen sinks within the system) 70
Figure 3.3-4. Critical load function response curves, detailing the highest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area
(refer to Table 3.1-10 for the parameters corresponding to each of the
curves). The 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels ((N+S)comb) were less than the critical loads associated with all three
(Be/Al)crit ratios. The flat line portion of the curves indicates total nitrogen
deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen
sinks within the system) 71
Figure 3.3-5. Critical load function response curves for the three selected critical loads
conditions (corresponding to the three levels of protection) for the Kane
Experimental Forest Case Study Area. The 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels ((N+S)comb) were greater than the
highest and intermediate level of protection critical loads. The flat line
portion of the curves indicates total nitrogen deposition corresponding to
the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system) 72
Figure 3.3-6. Critical load function response curves for the three selected critical loads
conditions (corresponding to the three levels of protection) for the
Hubbard Brook Experimental Forest Case Study Area. The 2002
CMAQ/NADP total nitrogen and sulfur deposition levels ((N+S)comb) were
greater than the highest level of protection critical loads. The flat line
portion of the curves indicates total nitrogen deposition corresponding to
the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system) 73
Final Risk and Exposure Assessment September 2009
Appendix 5 - iv
-------
Terrestrial Acidification Case Study
Figure 3.3-7. The influence of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels (NHX-N) on the critical load function response curve, and
in turn, the maximum critical loads of sulfur (CLmax(S)) and oxidized
nitrogen (NOX-N) for the selected highest protection critical load for the
Kane Experimental Forest Case Study Area. The critical load of oxidized
nitrogen (NOX-N) is 661 eq/ha/yr (910 - 249 eq/ha/yr). The CLmin(N)
(nitrogen absorbed by nitrogen sinks within the system) corresponds to the
value depicted in Figure 3.3-5 74
Figure 3.3-8. The influence of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels (NHX-N) on the critical load function response curve and,
in turn, the maximum critical loads of sulfur (CLmax(S)) and oxidized
nitrogen (NOX-N) for the selected highest protection critical load for the
Hubbard Brook Experimental Forest Case Study Area. The critical load of
oxidized nitrogen (NOX-N) is 328 eq/ha/yr (487 - 159 eq/ha/yr). The
CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system)
corresponds to the value depicted in Figure 3.3-6 75
Figure 4.1-1. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the lowest protection
critical load (Bc/Al(crit) = 0.6) in the following: none of the sugar maple
plots, <50% of the sugar maple plots, and >50% of the sugar maple plots 86
Figure 4.1-2. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the intermediate protection
critical load (Bc/Al(crit) = 1.2) in the following: none of the sugar maple
plots, <50% of the sugar maple plots, and >50% of the sugar maple plots 87
Figure 4.1-3. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the highest protection
critical load (Bc/Al(crit) = 10.0) in the following: none of the sugar maple
plots, <50% of the sugar maple plots, and >50% of the sugar maple plots 88
Figure 4.1-4. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the lowest protection
critical load (Bc/Al(crit) = 0.6) in the following: none of the red spruce
plots, <50% of the red spruce plots, and >50% of the red spruce plots 89
Figure 4.1-5. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the intermediate protection
critical load (Bc/Al(crit) = 1.2) in the following: none of the red spruce
plots, <50% of the red spruce plots, and >50% of the red spruce plots 90
Figure 4.1-6. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the highest protection
critical load (Bc/Al(crit) = 10.0) in the following: none of the red spruce
plots, <50% of the red spruce plots, and >50% of the red spruce plots 91
Final Risk and Exposure Assessment September 2009
Appendix 5 - v
-------
Terrestrial Acidification Case Study
ATTACHMENT FIGURES
Figure 1-1. Hypothesized relationships between tree growth and nitrogen deposition and
critical load exceedance. When deposition does not exceed the critical
load, growth is stimulated by nitrogen deposition. When deposition
exceeds the critical load (deposition is greater than the critical load),
growth is reduced. (This figure is a modification of a curve describing
forest productivity as a function of long-term chronic nitrogen additions
outlined in Aber et al., 1995) A-2
Figure 3-1. Areas of the continental United States that were covered during the last
glacial event (~ 20,000 ybp) (Reed and Bush, 2005) A-13
LIST OF TABLES
Table 1.1-1. Literature Support for Selected Indicators of Acidification 2
Table 1.1-2. Key Indicators of Acidification Due to Nitrogen and Sulfur 3
Table 1.1-3. Summary of Linkages between Acidifying Deposition, Biogeochemical
Processes That Affect Ca2+, Physiological Processes That Are Influenced
by Ca2+, and Effect on Forest Function 8
Table 1.2-1. Summary of Mapping Layers, Selected Indicator, and Selected Ecological
Endpoint for the Terrestrial Acidification Case Study 12
Table 1.2-2. Compilation of Potential Areas for the Terrestrial Acidification Case Study
(i.e., for Studying Red Spruce) as Identified in the Literature 14
Table 1.2-3. Major Studies at the Kane Experimental Forest 18
Table 1.2-4. Characteristics of the Case Study Plots in the Kane Experimental Forest 20
Table 1.2-5. Major Studies at the Hubbard Brook Experimental Forest 22
Table 2.1-1. Soil Texture Classes as a Function of Clay and Sand Content 34
Table 2.1-2. Parent Material Classes for Common FAO Soil Types 34
Table 2.1-3. Weathering Rate Classes as a Function of Texture and Parent Material
Classes 35
Table 2.1-4. Data Requirements and Sources for Calculating Critical Loads for Total
Nitrogen and Sulfur Deposition in Hubbard Brook Experimental Forest
and Kane Experimental Forest 40
Table 2.1-5. The Three Indicator (Bc/Al)crit Soil Solution Ratios Used in This Case Study
and the Corresponding Levels of Protection to Tree Health and Critical
Loads 43
Table 2.1-6. Input Values for the Calculation of Critical Load in Hubbard Brook
Experimental Forest and Kane Experimental Forest 47
Final Risk and Exposure Assessment September 2009
Appendix 5 - vi
-------
Terrestrial Acidification Case Study
Table 2.1-7. Soil Characteristics in the Seven Plots of the Kane Experimental Forest Case
Study Area for the Calculation of the Base Cation Weathering Rate
Parameters 48
Table 2.1-8. Annual Volume Growth by Tree Species in Each of the Seven Plots of the
Kane Experimental Forest Case Study Area for the Calculation of Nutrient
Uptake (Bcu and Nu) 48
Table 2.1-9. Specific Gravity and Nutrient Concentrations by Biomass Component (Bark
and Bole Wood) and by Tree Species for the Calculation of Nutrient
Uptake (Bcuand Nu) in the Kane Experimental Forest Case Study Area 49
Table 3.1-1. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 1 of the Kane Experimental Forest
Case Study Area 51
Table 3.1-2. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (K^b), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 2 of the Kane Experimental Forest
Case Study Area 52
Table 3.1-3. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 3 of the Kane Experimental Forest
Case Study Area 52
Table 3.1-4. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 4 of the Kane Experimental Forest
Case Study Area 53
Table 3.1-5. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb) and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 5 of the Kane Experimental Forest
Case Study Area 53
Table 3.1-6. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 6 of the Kane Experimental Forest
Case Study Area 54
Table 3.1-7. Critical Loads Calculated with the Different Base Cation Weathering,
Gibbsite Equilibrium Constant (^gibb), and Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values in Plot 7 of the Kane Experimental Forest
Case Study Area 54
Table 3.1-8. Ranges of Critical Load Values (eq/ha/yr) (with and without the Influence of
Nutrient Uptake and Removal with Tree Harvest) for the Seven Plots of
the Kane Experimental Forest Case Study Area (Both Kgm values and
methods to estimate BCW were used in these calculations to present the
Final Risk and Exposure Assessment September 2009
Appendix 5 - vii
-------
Terrestrial Acidification Case Study
range of critical loads estimated using all combinations of the parameter
values.) 55
Table 3.1-9. Comparison of the Critical Load Values Determined in This Case Study and
the Critical Load Values Determined by McNulty et al. (2007) for the
Seven Plots in the Kane Experimental Forest Case Study Area 58
Table 3.1-10. Critical Load Calculated with the Different Base Cation Weathering and
Gibbsite Equilibrium Constant (K^b) Parameter Values in the Hubbard
Brook Experimental Forest Case Study Area 59
Table 3.1-11. Summary of the Critical Load Values Determined by Other Studies
Conducted in the Hubbard Brook Experimental Forest (negative values are
equal to 0 eq/ha/yr) 61
Table 3.2-1. Critical Loads Selected to Represent the Three Levels of Protection in the
Kane Experimental Forest and Hubbard Brook Experimental Forest Case
Study Areas 62
Table 3.3-1. Ranges of Differences between the 2002 CMAQ/NADP Total Nitrogen and
Sulfur Deposition Levels ((N+S)comb) and the Estimated Critical Load
Values (with and without the Influence of Nutrient Uptake and Removal
[Nu and Bcu]) for the Seven Plots of the Kane Experimental Forest Case
Study Area 64
Table 3.3-2. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 1 of the Kane Experimental Forest Case Study
Area 64
Table 3.3-3. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)COmb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 2 of the Kane Experimental Forest Case Study
Area 65
Table 3.3-4. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 3 of the Kane Experimental Forest Case Study
Area 65
Table 3.3-5. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)COmb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 4 of the Kane Experimental Forest Case Study
Area 66
Final Risk and Exposure Assessment September 2009
Appendix 5 - viii
-------
Terrestrial Acidification Case Study
Table 3.3-6. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 5 of the Kane Experimental Forest Case Study
Area 66
Table 3.3-7. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 6 of the Kane Experimental Forest Case Study
Area 67
Table 3.3-8. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods, Two Gibbsite Equilibrium
Constants [^gibb], and Two Base Cation (Bcu) and Nitrogen (Nu) Uptake
Parameter Values) in Plot 7 of the Kane Experimental Forest Case Study
Area 67
Table 3.3-9. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)comb) and the Critical Load Values (with Two
Base Cation Weathering Estimation Methods and Two Gibbsite
Equilibrium Constants [^gibb]) in the Hubbard Brook Experimental Forest
Case Study Area 68
Table 4.1-1. Number and Location of USFS FIA Permanent Sampling Plots (each plot is
0.07 ha) Used in the Analysis of Critical Loads For the Full Geographic
Ranges of Sugar Maple and Red Spruce 76
Table 4.1-2. Gibbsite Equilibrium (Kgibb) Determined by Percentage of Soil Organic
Matter 79
Table 4.1-3. Parent Material Acidity Classifications for Base Cation (BCW) Estimations 79
Table 4.1-4. Parent Material and Descriptive Modifier Characteristics (within the
SSURGO Soils [USDA-NRCS, 2008c] andUSGS Geology [USGS,
2009b] Databases) Used to Classify Parent Material Acidity 80
Table 4.1-5. Ranges of Critical Load Values by Level of Protection (Bc/Al^t) = 0.6, 1.2,
and 10.0) and by State for the Full Distribution Ranges of Sugar Maple
and Red Spruce 83
Table 4.1-6. Percentages of Plots, by Protection Level (Bc/Al(crit) = 0.6, 1.2, and 10.0) and
by State, where 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels Were Greater Than the Critical Loads for Sugar Maple and Red
Spruce 85
Table 5.2-1. Differences and Percentage Differences in Plot-Level Critical Load
Estimates Associated with the Misclassification of Parent Material Acidity
for the Full Range Assessment of Sugar Maple 94
Final Risk and Exposure Assessment September 2009
Appendix 5 - ix
-------
Terrestrial Acidification Case Study
Table 5.2-2. Differences and Percentage Differences in Plot-Level Critical Load
Estimates Associated with the Misclassification of Parent Material Acidity
for the Full Range Assessment of Red Spruce 95
ATTACHMENT TABLES
Table 2-1. Summary of Plot-Level Data Used in the Regression Analyses for Sugar
Maple Volume and Growth, Nitrogen Deposition, and Critical Load
Exceedances (based on critical loads calculated with Bc/Al = 10.0) A-4
Table 2-2. Summary of Plot-Level Data Used in the Regression Analyses for Red Spruce
Volume and Growth, Nitrogen Deposition, and Critical Load Exceedances
(based on critical loads calculated with Bc/Al = 10.0) A-6
Table 3-la. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Sugar Maple Tree Growth and Nitrogen Deposition (for plots
where deposition did not exceed critical loads calculated with Bc/Al =
10.0) A-8
Table 3-lb. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Red Spruce Tree Growth and Nitrogen Deposition (for plots
where deposition did not exceed critical loads calculated with Bc/Al =
10.0) A-9
Table 3-2a. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Sugar Maple Tree Growth and Critical Load Exceedance (for
plots where nitrogen and sulfur deposition was greater than critical loads
calculated with Bc/Al = 10.0) A-9
Table 3-2b. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Red Spruce Tree Growth and Critical Load Exceedance (for
plots where nitrogen and sulfur deposition was greater than critical loads
calculated with Bc/Al = 10.0) A-10
Table 3-3. Summary of Plot-Level Data for Sugar Maple Volume and Growth and
Critical Load Exceedances North of the Glaciation Line (for plots where
nitrogen and sulfur deposition exceeded critical loads calculated with
Bc/Al = 10.0) A-14
Table 3-4. Summary of Plot-Level Data for Red Spruce Volume and Growth and Critical
Load Exceedances North of the Glaciation Line (for plots where nitrogen
and sulfur deposition exceeded critical loads calculated with Bc/Al = 10.0)... A-16
Table 3-5. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Sugar Maple Tree Growth and Critical Load Exceedance
North of the Glaciation Line (for plots where deposition exceeded critical
loads calculated with Bc/Al = 10.0) A-17
Table 3-6. Results from the Multivariate Ordinary Least Squares Linear Regression
Analyses of Red Spruce Tree Growth and Critical Load Exceedance,
North of the Glaciation Line (for plots where deposition exceeded critical
loads calculated with Bc/Al = 10.0) A-18
Final Risk and Exposure Assessment September 2009
Appendix 5 - x
-------
Terrestrial Acidification Case Study
ACRONYMS AND ABBREVIATIONS
Al
A13+
ANC
ANCle,cnt
Be
(Bc/Al)crit
BC
BCdep
Bcu
BCW
n 2+
Ca
cr
Cldep
CLmax(N)
CLmm(N)
CLF
CLRTAP
cm
CMAQ
eq/ha/yr
FAO
Fe3+
FGROWCFAL
FIA
ft3
GIS
ha
HBEF
HNO3
ICP
ISA
K+
Kgibb
KEF
kg
km
m
Mg
mm
Mn
2+
2+
base cation (Ca ^ + 1C
base cation (Ca2+ + K4
base cation (Ca2+ + K4
base cation (Ca2+ + K4
aluminum2+'3+
trivalent aluminum
acid neutralizing capacity
forest soil acid neutralizing capacity of critical load leaching (calculated
value)
+ Mg2+)
+ Mg2+) to aluminum ratio (selected indicator value)
+ Mg2+ + Na+)
+ Mg2+ + Na+) deposition
uptake of base cations (Ca2+ + K+ + Mg2+) by trees
base cation (Ca2+ + K+ + Mg2+ + Na+) weathering
calcium
chloride
chloride deposition
maximum critical load of nitrogen
maximum critical load of sulfur
minimum critical load
minimum critical load of nitrogen
critical load function
Convention on Long-Range Transboundary Air Pollution
centimeter
Community Multiscale Air Quality
equivalents per hectare per year
Food and Agriculture Organization
trivalent (ferrous) iron
Net annual sound cubic-foot growth of a live tree on forest land
Forest Inventory and Analysis National Program
cubic feet
geographic information systems
hectare
Hubbard Brook Experimental Forest
nitric acid
International Cooperative Programme
Integrated Science Assessment
potassium
the gibbsite equilibrium constant
Kane Experimental Forest
kilogram
kilometer
meter
magnesium
millimeter
manganese
2+,4+
Final Risk and Exposure Assessment
Appendix 5 - xi
September 2009
-------
Terrestrial Acidification Case Study
mol
N
Nde
N,
Nu
(N+S)comb
Na+
NADP
NEG/ECP
NH4+
NHX
NCV
NOX
NOy
NRCS
OLS
S
8MB
S02
SO42"
SOX
SSURGO
STATSGO
UNECE
USDA
USFS
USGS
VOLCFNET
ybp
mole
nitrogen
denitrification
nitrogen immobilization
uptake of nitrogen by trees
combined nitrogen and sulfur deposition
sodium
National Atmospheric Deposition Program
Conference of New England Governors and Eastern Canadian Premiers
ammonium
total reduced nitrogen
nitrate
nitrogen oxides
oxidized nitrogen
Natural Resources Conservation Service
ordinary least squares
sulfur
Simple Mass Balance
sulfur dioxide
sulfate
sulfur oxides
Soil Survey Geographic Database
State Soil Geographic Database
United Nations Economic Commission for Europe
U.S. Department of Agriculture
U.S. Forest Service
U.S. Geological Survey
total net volume per tree
years before present
Final Risk and Exposure Assessment
Appendix 5 - xii
September 2009
-------
Terrestrial Acidification Case Study
1.0 BACKGROUND
The selection and performance of case studies represent Steps 3 and 4, respectively, of
the 7-step approach to planning and implementing a Risk and Exposure Assessment of total
nitrogen, nitrogen oxides (NOX) (as a component of total nitrogen), and sulfur oxides (SOX)
deposition on ecosystems, as presented in the April 2008 Scope and Methods Plan for Risk
Exposure Assessment (U.S. EPA, 2008a). Step 4 entails evaluating the current Community
Multiscale Air Quality Model (CMAQ) modeling results for 2002 and the 2002 National
Atmospheric Deposition Program (NADP) monitoring data for total nitrogen and sulfur
deposition loads on, and effects to, a chosen case study assessment area, including ecosystem
services. This case study evaluates the current wet and dry atmospheric nitrogen and sulfur
deposition load to terrestrial ecosystems and the role atmospheric deposition can play in the
acidification of a terrestrial ecosystem.
Deposition of NOX and SOX can result in acidification of certain terrestrial ecosystems.
Because ecosystems and species may respond differently, case studies have been used to
illustrate potential effects of acidification on sensitive species. This report presents the
quantitative approach used to analyze the acidification effects of total nitrogen, NOX (as a
component of total nitrogen), and SOX deposition on red spruce and sugar maple.
Acidification
Acidification is the process of increasing the acidity of a system (e.g., lake, stream, forest
soil). Within soils, acidification occurs through increases in hydrogen ions or protons. Terrestrial
acidification occurs as a result of both natural biogeochemical processes and acidifying
deposition where strong acids are deposited into the soil. Acidifying deposition increases
concentrations of nitrogen and sulfur in the soil, which accelerates the leaching of sulfate (SC>42")
and nitrate (N(V) from the soil to drainage water. Under natural conditions (i.e., low
atmospheric deposition of nitrogen and sulfur), the limited mobility of anions in the soil controls
the rate of base cation leaching. However, acidifying deposition of nitrogen and sulfur species
can significantly increase the concentration of anions in the soil, leading to an accelerated rate of
base cation leaching, particularly the leaching of calcium (Ca2+) and magnesium (Mg2+) cations.
If soil base saturation (i.e., the concentration of exchangeable base cations as a percentage of the
total cation exchange capacity. Cation exchange capacity, the sum total of exchangeable cations
Final Risk and Exposure Assessment September 2009
Appendix 5-1
-------
Terrestrial Acidification Case Study
that a soil can absorb, is 20% to 25%, or lower, inorganic aluminum2+'3+ (Al) can become
mobilized, leading to the leaching of Al into soil waters and surface waters (Reuss and Johnson,
1985). This is an important effect of acidifying deposition because inorganic Al is toxic to tree
roots, fish, algae, and aquatic invertebrates (U.S. EPA, 2008c, Sections 3.2.1.5, 3.2.2.1, and
3.2.3).
1.1 INDICATORS, ECOLOGICAL ENDPOINTS, AND ECOSYSTEM
SERVICES
1.1.1 Indicators
The Integrated Science Assessment (ISA) for Oxides of Nitrogen and Sulfur-Ecological
Criteria (Final Report) (ISA) (U.S. EPA, 2008c) identified a variety of indicators supported by
the literature that can be used to measure the effects of acidification in soils (Table 1.1-1). Table
1.1-2 provides a general summary of these indicators by indicator groups.
Table 1.1-1. Literature Support for Selected Indicators of Acidification
Citation
Main Finding
Soil Base Saturation
Reuss, 1983
Cronan and Grigal, 1995
Lawrence et al., 2005
Bailey et al., 2004
Johnson et al., 1991; Joslin and Wolfe, 1992
When base saturation less than 15% to 20%, then
exchange ion chemistry is dominated by inorganic Al.
When base saturation below about 15% in the soil, B-
horizon could lead to impacts from Al stress.
When base saturation declines from 30% to 20% in the
upper soil, B-horizon showed decreases in diameter
growth of Norway spruce.
Sugar maple mortality found at Ca2+ saturation less
than 2% and Mg2+ saturation less than 0.5% in the
upper soil B-horizon.
When base saturation below about 20%, base cation
reserves are so low that Al exchange dominates.
Al Concentrations
Johnson et al., 1991; Joslin and Wolfe, 1992
Cronan and Grigal, 1995; Eagar et al., 1996
When base saturation below about 20%, base cation
reserves are so low that Al exchange dominates.
There is a 50% chance of negative effects on tree
growth if the molar ratio of Ca2+/Al in soil solution is
as low as 1.0. There is a 100% chance for negative
effects on growth at molar ratio value below 0.2.
Final Risk and Exposure Assessment
Appendix 5-2
September 2009
-------
Terrestrial Acidification Case Study
Citation
Johnson et al., 1994a, b
DeWittetal., 2001
Main Finding
Ca2+/Al ratios above 1.0 were found in a forestland
experiencing high mortality over the course of 4 years.
Ca2+/Al ratios of Norway spruce stand below 0.5
showed reduced Mg2+ concentrations in needles in the
third year.
Carbon/Nitrogen Ratio
Aber et al., 2003; Ross et al., 2004
Increased effects of nitrification occur only in soil with
carbon/nitrogen ratio below about 20 to 25.
Source: U.S. EPA 2008c, Section 3.2.2.1.
Table 1.1-2. Key Indicators of Acidification Due to Nitrogen and Sulfur
Key Indicator Group
Acid anions
Base cations
Acidity
Carbon
Metals
Biological
Examples of Indicators
SO42 , NO3
Ca2+, Mg2+, BC (sum of
Ca2+,Mg2+,K+andNa+),
Be (sum of Ca2+, Mg2+, and
K+)
pH, acid neutralizing
capacity
Carbon/nitrogen ratio
A13+, Fe3+
Tree health, community
structure, species
composition, taxonomic
richness, Index of Biotic
Integrity
Description
Trends in these concentrations reflect recent
trends in atmospheric deposition (especially
SO42") and in ecosystem responses to long-term
deposition (notably NO3" and desorbed SO42").
These cations are mobilized by weathering
reactions and cation exchange. They respond
indirectly to decreases in SO42" and NO3"
because a reduced input of acids will lead to a
reduction of neutralizing processes in the soil,
thereby reducing the release of base cations to
soil water and runoff water. (Base saturation is
included within this category.)
These indicators reflect the outcomes of
interactions between the changing
concentration of acid anions and base cations.
The carbon/nitrogen ratio of soil indicates
alterations to the nitrogen biogeochemical
cycle.
These metals are mobilized as a response to the
deposition of SO42" and NO3".
Ecological effects occur at four levels:
individual, population, community, and
ecosystem. Metrics have been developed for
each level to assess the negative effects of
acidification.
Note: BC = base cation (Caz+ + K+ + Mg/+ + Na+); Be = base cation (Ca2+ + K+ + Mg2+); K+ = potassium;
Na+ = sodium; A13+ = trivalent aluminum; Fe3+ = trivalent (ferrous) iron
2+
Much of the literature discussing terrestrial acidification focuses on Ca and Al as the
primary indicators of detrimental effects for trees and other terrestrial vegetation. As such, this
Final Risk and Exposure Assessment
Appendix 5-3
September 2009
-------
Terrestrial Acidification Case Study
discussion of indicators of terrestrial acidification focuses on these two parameters and the
interaction between them. The use of these indicators—in combination and through the
evaluation framework that will be described within this case study—ultimately combines all
indicator categories described in Table 1.1-1 except the carbon category. Ca2+ and Al are the
focus of the analysis because both of these indicators are strongly influenced by soil acidification
and both have been shown to have quantitative links to tree health, including aluminum's
interference with Ca2+uptake and Al toxicity to roots.
A detailed description of the influences of Al on Ca2+ is provided by Schaberg et al.
(2001)1:
Decreases in concentrations of exchangeable calcium are generally attributed to
displacement by hydrogen ions, which can originate from either acidifying
deposition or uptake of cations by roots (Johnson et al., 1994a; Richter et al.,
1994). A regional survey of soils in northeastern red spruce forests in 1992-93
(fig. 2)2 has revealed that decreases in exchangeable calcium concentrations in the
Oa horizon (a layer within the forest floor, where uptake of nutrients is greatest)
can also result from increased concentrations of exchangeable aluminum, which
originated in the underlying mineral soil (Lawrence et al., 1995). By lowering the
pH of the aluminum-rich mineral soil, acid deposition can increase aluminum
concentrations in soil water through dissolution and ion-exchange processes.
Once in solution, the aluminum (although not a nutrient) is taken up by roots and
transported through the trees to be eventually deposited on the forest floor in
leaves and branches.
A continued buildup of aluminum in the Oa horizon can (1) decrease the
availability of calcium for roots (Lawrence et al., 1995), (2) lower the efficiency
of calcium uptake because aluminum is more readily taken up than calcium when
the ratio of calcium to aluminum in soil water is less than 1 (Cronan and Grigal,
1995), and (3) be toxic to roots at high concentrations (Lawrence et al., 1995).
The relationship between Ca2+ and Al and tree health is summarized in the ISA (U.S.
EPA, 2008c, Section 3.2.2.1), as excerpted below3:
Al may be toxic to tree roots. Plants [exposed to] high Al concentration in soil
solution often have reduced root growth, which restricts the ability of the plant to
take up water and nutrients, especially Ca (Parker et al., 1989) (Figure 3-5 [of
U.S. EPA, 2008c]). Ca is well known as an ameliorant for Al toxicity to roots in
soil solution, as well as to fish in a stream. However, because inorganic Al tends
1 References contained within this quotation are not included in the References section of this case study report.
2 Figure 2 is not included in this case study report.
3 References contained within this quotation are not included in the References section of this case study report.
Final Risk and Exposure Assessment September 2009
Appendix 5-4
-------
Terrestrial Acidification Case Study
to be increasingly mobilized as soil Ca is depleted, elevated concentrations of
inorganic Al tend to occur with low levels of Ca in surface waters. Mg, and to a
lesser extent Na and K, have also been associated with reduced Al toxicity.
Dissolved Al concentrations in soil solution at spruce-fir study sites in the
southern Appalachian Mountains frequently exceed 50 uM and sometimes exceed
100 uM (Eagar et al., 1996; Johnson et al., 1991; Joslin and Wolfe, 1992a). All
studies reviewed by Eagar [et al.] (1996) showed a strong correlation between Al
concentrations and NO3 concentrations in soil solution. They surmised that the
occurrence of periodic large pulses of NOs" in solution were important in
determining Al chemistry in the soils of southern Appalachian Mountain spruce-
fir forests.
The negative effect of Al mobilization on Ca uptake by tree roots was proposed
by Shortle and Smith (1988). Substantial evidence of this relationship has
accumulated over the past two decades through field studies (Kobe et al., 2002;
McLaughlin and Tjoelker, 1992; Minocha et al., 1997; Schlegel et al., 1992;
Shortle et al., 1997) and laboratory studies (see review by Cronan and Grigal,
1995; Sverdrup and Warfvinge, 1993). Based on these studies, it is clear that high
inorganic Al concentration in soil water can be toxic to plant roots. The toxic
response is often related to the concentration of inorganic Al relative to the
concentration of Ca, expressed as the molar ratio of Ca to inorganic Al in soil
solution. As a result, considerable effort has been focused on determining a
threshold value for the ratio of Ca to Al that could be used to identify soil
conditions that put trees under physiological stress.
Building on the explanation of the relationship between Ca2+, Al, and tree health, a figure
developed by DeHayes et al. (1999), clearly shows the connections between nitrogen and sulfur
acidifying deposition and Ca2+ within an ecosystem (Figure 1.1-1). The authors used solid lines
to denote known connections and dotted lines to present potential impacts. While the authors did
not specify that increases in Al within the soils would occur with reductions in biologically
available Ca2+ pools, this impact is expected as detailed in the previous paragraphs. The final
process represented in Figure 1.1-1 completes the linkage from the indicator of Ca2+ (and
therefore Al) to the effects on the ecosystem services for the terrestrial area.
Final Risk and Exposure Assessment September 2009
Appendix 5-5
-------
Terrestrial Acidification Case Study
Acid Ra
*
Reductions in Biologically
Important Calcium Pools
/\ -•
Membrane
Disruption
Disruptions in Stress
Response Systems "^
\ /
Predisposition to Stress-induced Injury
t
^^^^^^ Soil Calcium
^^^^^^ uepienon
*****
Potential Secondary
( Environmental Stresses
* % Air pollutants
* Temperature perturbati
(high, low, variable)
Pathogens
Drought
- Heavy metals
*
*
9
»
on
1
•
/
Disruptions in Forest Ecosystem Health/Stability
2+
Figure 1.1-1. Conceptual impacts of acidifying deposition on soil Ca depletion,
tree physiology, and forest ecosystem health and sustainability (recreated from
DeHayes et al., 1999).
In summary, based on the ISA (U.S. EPA, 2008c) and supporting literature, soil Ca2+ and
Al are suitable chemical indicators to represent the impacts of acidification in soils and to
provide a linkage between soil acidification and tree health. Therefore, the Ca/Al ratio in soil
solution was selected as the basis for the indicator in this case study to evaluate critical loads of
acidity in terrestrial systems. Within the calculations of critical loads, the base cation (Be) to Al
ratio (Bc/Al), consisting of molar equivalents of Ca2+, Mg2+, and K+, was used to represent the
Ca/Al indicator. This Bc/Al ratio was selected because it is the most commonly used indicator or
critical ratio (Bc/Al(crit)) in the Simple Mass Balance (8MB) model used to estimate critical acid
loads in the European Union, Canada, and the United States (McNulty et al., 2007; Ouimet et al.,
2006; UNECE, 2004), and the 8MB model was applied to this case study (see Section 2.1 for
Final Risk and Exposure Assessment
Appendix 5-6
September 2009
-------
Terrestrial Acidification Case Study
description of model). In addition, tree species show similar sensitivities to Ca/Al and Bc/Al soil
solution ratios. Therefore, the Bc/Al ratio represents a good indicator of the negative impacts of
soil acidification on terrestrial vegetation. Sverdrup and Warfvinge (1993b), in a metadata
analysis of laboratory and field studies, reported that the critical Bc/Al ratios for a large variety
of tree species ranged from 0.2 to 0.8. This range is similar to that described by Cronan and
Grigal (1995) for Ca/Al. In their metadata assessment of studies examining sensitivities to the
Ca/Al ratio, plant toxicity or nutrient antagonism was reported to occur at Ca/Al ratios ranging
from 0.2 to 2.5.
1.1.2 Ecological Endpoints
The tree species most commonly associated with the impacts of acidification due to
atmospheric nitrogen and sulfur deposition include red spruce (Picea rubens), a coniferous tree
species, and sugar maple (Acer sacchamm), a deciduous tree species. Both species are found in
the eastern United States, and soil acidification is widespread throughout this area (Warby et al.,
2009).
Red spruce is found scattered throughout high-elevation sites in the Appalachian
Mountains, including the southern peaks (Figure 1.1-2). Noticeable levels of the canopy red
spruce died within the Adirondack, Green, and White mountains in the 1970s and 1980s.
Although a variety of conditions, such as changes in climate and exposure to ozone, may impact
the growth of red spruce (Fincher et al., 1989; Johnson et al., 1988), acidifying deposition has
been implicated as one of the main factors causing this decline. Based on the research conducted
to date, acidic deposition can cause a depletion of base cations in upper soil horizons, Al toxicity
to tree roots, and accelerated leaching of base cations from foliage (U.S. EPA, 2008c, Section
3.2.2.3). Such nutrient imbalances and deficiencies can reduce the ability of trees to respond to
stresses, such as insect defoliation, drought, and cold weather damage (DeHayes et al., 1999;
Driscoll et al., 2001), thereby decreasing tree health and increasing mortality. Additional
linkages between acidifying deposition and red spruce physiological responses are indicated in
Table 1.1-3. Within the southeastern United States, periods of red spruce decline slowed after
the 1980s, when a corresponding decrease in SC>2 emissions, and therefore acidic deposition, was
recorded (Webster et al., 2004).
Final Risk and Exposure Assessment September 2009
Appendix 5-7
-------
Terrestrial Acidification Case Study
Sugar maple is found throughout the northeastern United States and the central
Appalachian Mountain region (Figure 1.1-2). This species has been declining in the eastern
United States since the 1950s. Studies on sugar maple have found that one source of this decline
in growth is related to both acidifying deposition and base-poor soils on geologies dominated by
sandstone or other base-poor substrates (Bailey et al., 2004; Horsley et al., 2000). These site
conditions are representative of the conditions expected to be most susceptible to impacts of
acidifying deposition because of probable low initial base cation pools and high base cation
leaching losses (U.S. EPA, 2008c, Section 3.2.2.3). The probability of a decrease in crown vigor
or an increase in tree mortality has been noted to increase at sites with low Ca2+ and Mg2+ as a
result of leaching caused by acidifying deposition (Drohan and Sharpe, 1997). Low levels of
Ca2+ in leaves and soils were related to lower rates of photosynthesis and higher antioxidant
enzyme activity in sugar maple stands in Pennsylvania (St. Clair et al., 2005). Additionally, plots
of sugar maples in decline were found to have Ca2+/Al ratios less than 1, as well as lower base
cation concentrations and pH values compared to plots of healthy sugar maples (Drohan et al.,
2002). Sugar maple regeneration has also been noted to be restricted under conditions of low soil
Ca2+ levels (Juice et al., 2006). These indicators have been shown to be related to the deposition
of atmospheric nitrogen and sulfur. Additional linkages between acidifying deposition and sugar
maple physiological responses are indicated in Table 1.1-3.
Table 1.1-3. Summary of Linkages between Acidifying Deposition, Biogeochemical Processes
That Affect Ca2+, Physiological Processes That Are Influenced by Ca2+, and Effect on Forest
Function
Biogeochemical Response to
Acidifying Deposition
Leach Ca2+ from leaf membrane
Reduce the ratio of Ca2+/Al in
soil and soil solutions
Reduce the ratio of Ca2+/Al in
soil and soil solutions
Reduce the availability of
nutrient cations in marginal soils
Physiological Response
Decrease the cold tolerance of
needles in red spruce
Dysfunction in fine roots of red
spruce blocks uptake of Ca2+
More energy is used to acquire
Ca2+ in soils with low Ca2+/Al
ratios
Sugar maples on drought-prone
or nutrient-poor soils are less
able to withstand stresses
Effect on Forest Function
Loss of current-year needles in
red spruce
Decreased growth and increased
susceptibility to stress in red
spruce
Decreased growth and increased
photosynthetic allocation to red
spruce roots
Episodic dieback and growth
impairment in sugar maple
Source: Fenn et al., 2006a.
Final Risk and Exposure Assessment
Appendix 5-8
September 2009
-------
Terrestrial Acidification Case Study
Although the main focus of the Terrestrial Acidification Case Study is an evaluation of
the negative impacts of nitrogen and sulfur deposition on soil acidification and tree health, it
should be recognized that, under certain conditions, nitrogen and sulfur deposition can have a
positive impact on tree health. Nitrogen limits the growth of many forests (Chapin et al., 1993;
Killam, 1994; Miller, 1988); therefore, in such forests, nitrogen deposition may act as a fertilizer
and stimulate growth. Forests where critical acid loads are not exceeded by nitrogen and sulfur
deposition could potentially be included within this group of forests that respond positively to
deposition. These potential positive growth impacts of nitrogen and sulfur deposition are
discussed further, and the results of analyses are presented in Attachment A at the end of this
Appendix.
In summary, among the potential influencing
factors, including elevated ozone levels and changes in
climate, the acidification of soils is one of the factors that
estimate critical deposition loads of
acidity in this case study.
can negatively impact the health of red spruce and sugar
End Point: The health of sugar
maple. Mortality and susceptibility to disease and injury can be increased and growth decreased
with acidifying deposition. Therefore, the health of sugar maple and red spruce was used as the
endpoints (ecological responses) to evaluate acidification in terrestrial systems. "Health," in the
context of this case study, is defined as the physiological condition of a tree, which impacts
growth and/or mortality.
Final Risk and Exposure Assessment September 2009
Appendix 5-9
-------
Terrestrial Acidification Case Study
~^\ States
Red Spruce Range
J Sugar Maple Range
SOurce:USFS Forest Inventory and Analysis
Program (2006)
Figure 1.1-2. Areal coverages of red spruce and sugar maple tree species
within the continental United States (USFS, 2006).
1.1.3 Ecosystem Services
Ecosystem services are generally defined as the benefits individuals and organizations
obtain from ecosystems. In the Millennium Ecosystem Assessment (MEA, 2005), ecosystem
services are classified into four main categories:
• Provisioning—includes products obtained from ecosystems
• Regulating—includes benefits obtained from the regulation of ecosystem processes
• Cultural—includes the nonmaterial benefits people obtain from ecosystems through
spiritual enrichment, cognitive development, reflection, recreation, and aesthetic
experiences
• Supporting—includes those services necessary for the production of all other ecosystem
services (MEA, 2005).
Final Risk and Exposure Assessment
Appendix 5-10
September 2009
-------
Terrestrial Acidification Case Study
A number of impacts on the ecological endpoints of forest health, water quality, and
habitat exist, including the following:
• Decline in forest aesthetics—cultural
• Decline in forest productivity—provisioning
• Increases in forest soil erosion and reductions in water retention—cultural and regulating.
Recognizing that many ecosystem services have not been adequately studied, the
ecosystem services highlighted in this case study will include economic values associated with
red spruce and sugar maple wood volume production.
1.2 CASE STUDY AREAS
The selection of case study areas to evaluate terrestrial acidification was based on
geographic information systems (GIS) mapping (locations recommended by the ISA [U.S. EPA,
2008c; Sections 3.2, 4.2, and Annex B]), and the availability of data for the selected indicators
and ecological endpoints, as presented in relevant literature and databases.
1.2.1 GIS Analysis of National Sensitivity
A GIS analysis was performed on datasets and datalayers of physical, chemical, and
biological properties to map areas of potential sensitivity to acidification in the United States
(Table 1.2-1). Ranges of sugar maple and red spruce were mapped by extracting counties with
plots that contained either sugar maple or red spruce from the U.S. Forest Service (USFS) Forest
Inventory and Analysis (FIA) database (http://fia.fs.fed.us/tools-data/). To characterize soil
acidity, soil pH was mapped with State Soil Geographic Database (STATSGO) soils
(http://www.soilinfo.psu.edu/index.cgi?soil_data&conus&data_cov) and USFS Forest Soils
datalayers. Soil thickness was also extracted from the STATSGO soils data. Areas with bedrock
with high acid neutralizing capacity (ANC) were determined by using the karst topography
dataset from the National Atlas of the United States (Tobin and Weary, 2005). Karst topography
is a landscape formed by the dissolution of soluble rock (e.g., limestone, dolomite); caves,
springs, and sinkholes are common features of this type of landscape (USGS, 2009a). Locations
with sugar maple or red spruce, soil pH <5.0, soils <51 centimeters (cm) in depth, and low ANC
bedrock (not dominated by carbonate rocks) were selected to represent areas with potential
sensitivity to acidification (Figure 1.2-1).
Final Risk and Exposure Assessment September 2009
Appendix 5-11
-------
Terrestrial Acidification Case Study
Table 1.2-1. Summary of Mapping Layers, Selected Indicator, and Selected Ecological Endpoint
for the Terrestrial Acidification Case Study
Targeted
Ecosystem
Effect
Terrestrial
acidification
due to
nitrogen and
sulfur
Selected
Indicator
Ca2+/Al
ratio
Selected
Biological
Endpoint
Reduced
health of
red spruce
and sugar
maple
Mapping Layers
Sugar maple and red spruce coverages (USFS, 2006)
Soil pH (USDA, 1994; USFS, 2008, Dr. Charles Perry,
personal communication)
Soil depth (USDA, 1994)
Karst topography (Tobin and Weary, 2005)
The Bc/Al (Be = Ca , Mg , and K ) is used to represent the Ca /Al ratio indicator in the acid load
calculations (described further at the end of Section 1.1.1).
I Slates
Potentially Sensilive to Terrestrial Acidification
0 250 500 750 1,000
km
Figure 1.2-1. Map of areas of potential sensitivity of red spruce and sugar maple to
acidification in the United States (see Table 1.2-1 for listing of data sources to produce
this map).
1.2.2 Selection of Case Study Areas
Following the identification of regions of potential sensitivity to acidification, Risk and
Exposure Assessment sites recommended by the Science Advisory Board—Ecological Effects
Final Risk and Exposure Assessment
Appendix 5-12
September 2009
-------
Terrestrial Acidification Case Study
Subcommittee (U.S. EPA, 2005) and found in the ISA (U.S. EPA, 2008c, Appendix A) and in
the body of published and unpublished literature were reviewed to determine the most suitable
locations for the Hubbard Brook Experimental Forest (HBEF) and Kane Experimental Forest
(KEF) case study areas.
Selection of a location for studying the sugar maple focused on the Allegheny Plateau
region in Pennsylvania, where a large proportion of published and unpublished research has been
focused. A significant amount of the research work in the Allegheny Plateau region has been
sponsored by the USFS and has produced extensive datasets of soil and tree characteristics
(Horsley et al., 2000; Bailey et al., 2004; Hallett et al., 2006). The USFS-designated KEF was
selected as the area for studying the sugar maple as part of the Terrestrial Acidification Case
Study. The KEF has been the focus of several long-term studies since the 1930s.
Selection of a case study area for studying the red spruce involved the review of a variety
of regions. Four studies that examined the relationship between the Ca2+/Al soil solution ratio
and tree health were identified, and relevant soil and tree information for each of the study
regions was compiled (Table 1.2-2). A review of this information led to the selection of the
HBEF in New Hampshire's White Mountains as the area for the study of red spruce in the
Terrestrial Acidification Case Study. The HBEF was also recommended by the ISA (U.S. EPA,
2008c, Appendix A) as a good location for the Risk and Exposure Assessment. This forest has
experienced high atmospheric nitrogen and sulfur deposition levels and low Ca2+/Al soil solution
ratios. It has been the subject of extensive nutrient investigations and has provided a large dataset
from which to work on the case study.
Final Risk and Exposure Assessment September 2009
Appendix 5-13
-------
Terrestrial Acidification Case Study
Table 1.2-2. Compilation of Potential Areas for the Terrestrial Acidification Case Study (i.e., for Studying Red Spruce) as Identified
in the Literature
Site Name
Balsam High
Top, NC
Clingman's
Dome, TN
Double Spring
Gap, TN
Mount LeConte,
TN
Mount Sterling,
TN
Richland Balsam
Mountain, NC
Spruce Mountain,
NC
Sleepers River,
VT
Groton, VT
Elevation
(m)
1,641
2,020
1,678
2,010
1,772
1,941
1,695
NA
520
Size of Tree
Population
Spruce-fir forest6
Spruce-fir forest6
Spruce-fir forest6
Spruce-fir forest6
Spruce-fir forest6
Spruce-fir forest6
Spruce-fir forest6
Red spruce
dominated with low
exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Availability of
Field Data
University study
University study
University study
University study
University study
University study
University study
Not selected in
studies
USFS study
location
Ecological
Importance
Great Smoky
Mountains
National Park
Great Smoky
Mountains
National Park
Great Smoky
Mountains
National Park
Great Smoky
Mountains
National Park
Great Smoky
Mountains
National Park
Blue Ridge
Parkway
Great Smoky
Mountains
National Park
Reported Impacts
Nearly 100% chance of
negative forest health
effects
Nearly 100% chance of
negative forest health
effects
Nearly 100% chance of
negative forest health
effects
75% chance of a negative
forest health effects
Nearly 100% chance of
negative forest health
effects
Nearly 100% chance of
negative forest health
effects
Nearly 100% chance of
negative forest health
effects
Site did not contain
sufficient number of
healthy, mature red
spruce for study
No specific references at
this time, but
disturbances are known
to have occurred
Ca2+:Al and
Al/Ca2+
Ratios
0.094a
0.084a
0.053a
0.567a
0.07a
0.07a
0.128a
NA
0.3b
Deposition
Load
(kg/ha/yr)
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
5.3C
Source(s)
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Bintz and
Butcher, 2007
Shortle et al,
1997
Shortle et al.,
1997;Wargoet
al.,2003
Final Risk and Exposure Assessment
Appendix 5-14
September 2009
-------
Terrestrial Acidification Case Study
Site Name
Rowland, ME
Bartlett, NH
Kossuth, ME
Hubbard Brook,
NH
Whiteface
Mountain, NY
Crawford Notch,
NH
Elevation
(m)
60
525
100
755
950
670
Size of Tree
Population
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Availability of
Field Data
USFS study
location
USFS
Experimental
Forest (1,052 ha);
red spruce covers
highest slopes
USFS study
location
USFS
Experimental
Forest (3, 138 ha);
red spruce
abundant at
higher elevations
and on rock
outcrops
USFS study
location
USFS study
location
Ecological
Importance
Within the White
Mountains
Within the White
Mountains
Within the White
Mountains
Reported Impacts
No specific references at
this time, but
disturbances are known
to have occurred
No specific references at
this time, but
disturbances are known
to have occurred
No specific references at
this time, but
disturbances are known
to have occurred
Acid-extractable Al in the
forest floor increased
over the past two decades
at the HBEF, and ratios
of Al to Ca2+ in mineral
soil solutions (but not
forest floor solutions)
were strongly correlated
with exchangeable Al
content in the forest floor.
Contained neither
evidence of unusual
mortality or current tree
decline; winter injury
events reported (Lazarus
etal.,2004)
50% chance of negative
forest health effects;
mortality of red spruce
was significant, but most
of the remaining trees
were in good to fair
health
Ca2+:Al and
Al/Ca2+
Ratios
0.4b
0.8b
0.8b
0.8b
0.8b
l.lb
Deposition
Load
(kg/ha/yr)
3.1C
4.9C
2.8C
6.0C
7.9C
5.5C
Source(s)
Shortle et al,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al. 2003
Shortle et al.,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al.,2003
Final Risk and Exposure Assessment
Appendix 5-15
September 2009
-------
Terrestrial Acidification Case Study
Site Name
Big Moose Lake,
NY
Bear Brook, ME
Cone Pond, NH
Mt. Abraham,
VT
Mt. Ascutney,
VT
Elevation
(m)
550
400
610
NA
762
Size of Tree
Population
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
gradient of forest
floor exchangeable
Al/Ca2+ ratios
Red spruce
dominated with a
high exchangeable
Al/Ca2+ ratios
Series of high
elevation spruce-fir
forest nitrogen
addition plots f
Availability of
Field Data
USFS study
location
USFS study
location
USFS study
location
Not selected in
studies
USFS study
location
Ecological
Importance
Within the White
Mountains
Within the Green
Mountains
Nitrogen additions
to system
Reported Impacts
50% chance of negative
forest health effects
75% chance of negative
forest health effects
Nearly 100% chance of
negative forest health
effects
Site did not contain
sufficient number of
healthy, mature red
spruce for study; forest
floor solution Al/ Ca2+
ratio above the 50%
chance level
Reduction in live basal
area on the high nitrogen
addition plots versus
control plots
Ca2+:Al and
Al/Ca2+
Ratios
1.2b
1.9b
5.2b
7.1b
NA
Deposition
Load
(kg/ha/yr)
6.4C
3.8C
5.4C
NA
Additions'1
Source(s)
Shortle et al,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al.,2003
Shortle et al.,
1997; Wargoet
al.,2003
Shortle et al.,
1997
McNulty etal.,
2005
a Molar Ca27Al ratio (Bintz and Butcher, 2007).
b Oa horizon Al/Ca2+ ratios (Wargo et al., 2003).
0 Estimated wet nitrogen deposition (Lilleskov et al., 2008).
d In addition to ambient total nitrogen deposition, paired plots each received 15.7 kilograms (kg) N/ha/yr (low nitrogen addition), 31.4 kg N/ha/yr (high nitrogen
addition) or no nitrogen addition (control) from 1988 to 2002.
e High elevation sites in the Southern Appalachians—The sites are located in the Great Smoky Mountains National Park and Richland Balsam Mountain on the
Blue Ridge Parkway. Sites were selected because of the presence of a spruce-fir forest with a northwest slope aspect within 10 km of a trailhead at elevations
between 1,650 and 2,025 meters (m).
f Red spruce grew in large patches (>1 ha) at elevations above 725 m. Red spruce comprised > 80% of the total basal area in all plots; the remainder of the other
tree species were divided among balsam fir, red maple, mountain maple, and birch.
NA= Not available
Final Risk and Exposure Assessment
Appendix 5-16
September 2009
-------
Terrestrial Acidification Case Study
1.2.3 Sugar Maple
1.2.3.1 Kane Experimental Forest
The KEF (USFS, 1999, 2008b) is located on the eastern edge of the Allegheny National
Forest, 5.6 kilometers (km) south of Kane, PA (Figure 1.2-2). It is comprised of 703 hectare (ha)
of forestland and ranges in elevation from about 550 to 640 meters (m) above sea level, primarily
on flat to gently sloping land. The climate of the KEF is humid temperate; the average annual
temperature is 6°C. The forest receives approximately 110 cm of precipitation per year, mostly
as rain, including 10 cm/month during the growing season.
'EHNSYLVAklA '*
Figure 1.2-2. Location of the Kane Experimental Forest (Horsley et al., 2000).
The forest soils on the Allegheny Plateau are derived from shales and sandstones. In
general, these soils are very stony and exist as extremely stony loams and sandy loams. They are
strongly acidic. The major soil series are the well-drained Hazelton series, the moderately well-
drained to somewhat poorly drained Cookport series, and the somewhat poorly drained Cavode
series.
The forest stands on the KEF are typical of the Allegheny Plateau. They resulted from a
series of cuttings made in the original hemlock-beech-maple stands starting as early as the mid-
1880s. Currently, the KEF contains second-growth stands ranging from 60 to about 100 years of
age, several third-rotation stands 20 or 40 years old, and one tract with remnant old growth. Most
stands are even-aged, with black cherry, maples, and beech being the main overstory species.
Final Risk and Exposure Assessment
Appendix 5-17
September 2009
-------
Terrestrial Acidification Case Study
The KEF was formally established in 1932, although research there began as early as
1927 or 1928. The forest's primary mission has been forest management research, and the
current research focus is centered on three topics: regeneration and forest renewal stand
dynamics, silviculture, and sugar maple decline. Table 1.2-3 summarizes major studies at the
KEF related to the sugar maple and chemical criterion that can be used in calculating critical
loads of atmospheric nitrogen and sulfur deposition.
Table 1.2-3. Major Studies at the Kane Experimental Forest
Authors
Year
Title
Key Finding
Horsley.et al.
2000
Factors Associated
with the Decline-
Disease of Sugar
Maple on the
Allegheny Plateau
The most important factors determining sugar maple
health were foliar levels of Mg2+ and Mn and
defoliation history. The decline disease of sugar
maple appears to be the result of an interaction
between Mg2+ (and perhaps Mn) nutrition and stress
caused by defoliation.
Bailey et al.
2005
Thirty Years of
Change in Forest
Soils of the
Allegheny Plateau,
Pennsylvania
Between 1967 and 1997, there were significant
decreases in exchangeable Ca2+ and Mg2+
concentrations and pH at all soil depths.
Exchangeable Al concentrations increased at all
depths at all sites; however, increases were only
significant in upper soil horizons. At most of the
sites, losses of Ca2+ and Mg2+ on a pool basis were
much larger than could be accounted for in biomass
accumulation, suggesting the leaching of nutrients
; OO O O
off site.
Note: Mn = manganese.
1.2.3.2 Plot Selection for Kane Experimental Forest Case Study Area
Plots were selected for the KEF Case Study Area based on the location of permanent
sampling plots and the presence of sugar maple trees. Permanent sampling plots are 0.1 acres
(0.04 ha) in size and are evenly distributed and spaced (200 m east-west between plots)
throughout the forest. Only plots that were not located within an active research study and
contained a basal area of at least 20% sugar maple were selected (Table 1.2-4). A total of seven
plots (0.28 ha total) were identified based on these criteria and were used for the KEF Case
Study Area (Figure 1.2-3). Site, stand, and soil characteristics of these seven plots are presented
in Table 1.2-4
Final Risk and Exposure Assessment
Appendix 5-18
September 2009
-------
Terrestrial Acidification Case Study
Kane Experimental Forest
3 4
Legend
• Case Study Plots (0.1 acre)
NLCD Classifications
, Developed, Open Space
^B Deciduous Forest
^H Evergreen Forest
I Mixed Forest
I | Scrub/Shrub
^] Grassland/Herbaceous
^) Pasture/Hay
New York
Ohio
Pennsylvania
West Virginia
West Virginia
Maryland
Legend
I I Kane Forest Location
Figure 1.2-3. The seven plots used to evaluate critical loads of acidity in the Kane
Experimental Forest.
Final Risk and Exposure Assessment
Appendix 5-19
September 2009
-------
Terrestrial Acidification Case Study
Table 1.2-4. Characteristics of the Case Study Plots in the Kane Experimental Forest
Plot
Number
1
2
3
4
5
6
7
KEF Plot
ID
2,920
2,150
950
850
760
650
560
Location
78°47'35"W
41°35'41"N
78°46'33"W
41°36'5"N
78°44'50"W
41°35'50"N
78°44'40"W
41°35'51"N
78°44'33"W
41°36'01"N
78°44'24"W
41°35'56"N
78°44'16"W
41°35'59"N
Elevation
(m)
583.9
616.5
530.2
548.9
589.0
530.6
527.7
Stand
Basal Area
(m2/ha)
62.8
44.8
28.8
22.4
24.0
30.7
23.63
Representation
of Sugar Maple
in Stand (% of
Basal Area)
32%
23%
36%
44%
28%
35%
33%
Soil Type
(Soil Series)
CpB
(Cockport)
HaC
(Harleton)
BxD
(Buchanan)
HaF
(Harleton)
CpD
(Cockport)
HaF
(Harleton)
HaF
(Harleton)
Note: W = west; N = north.
1.2.4 Red Spruce
1.2.4.1 Hubbard Brook Experimental Forest
The FfBEF is located in the southern part of the White Mountain National Forest in
Grafton County, central New Hampshire (Figure 1.2-4). The experimental forest consists of an
oblong basin approximately about 8-km long by 5-km wide, and covers 3,138 ha. Hubbard
Brook is the single major stream draining the basin. Elevations within the HBEF range from 222
to 10,015 m. The climate of HBEF is predominantly continental, with a January temperature
average of -9°C and an average July temperature of 18°C. Annual precipitation at the HBEF
averages about 1,400 millimeters (mm), with one-third to one-quarter as snow.
Final Risk and Exposure Assessment
Appendix 5-20
September 2009
-------
Terrestrial Acidification Case Study
O
County
State
| White Mountain National Fore
Hubbard Brook Experimental
Forest Location
Figure 1.2-4. Location of the Hubbard Brook Experimental Forest.
Soils at the HBEF are predominantly well-drained Spodosols (Typic Haplorthods)
derived from glacial till, with sandy loam textures. Principal soil series are the sandy loams of
the Berkshire series, along with the Skerry, Becket, and Lyman series. These soils are acidic (i.e.,
pH about 4.5 or less) and relatively infertile (i.e., base saturation of mineral soil ~ 10%).
Although highly variable, soil depths, including unweathered till, average about 2.0 m from
surface to bedrock.
The HBEF is entirely forested, mainly with deciduous northern hardwoods. Red spruce is
abundant at higher elevations and on rock outcrops. Logging in the area began in the late 1880s
and ended around 1917. The present second-growth forest is even-aged and composed of about
80% to 90% hardwoods and 10% to 20% conifers.
The HBEF was established in 1955 as a major center for hydrologic research in New
England, and in 1963, the Hubbard Brook Ecosystem Study was founded. In 1988, the HBEF
Final Risk and Exposure Assessment
Appendix 5-21
September 2009
-------
Terrestrial Acidification Case Study
was designated as a Long-Term Ecological Research site. Research at the HBEF has been in
progress for more than 50 years and has focused on hydrometeorological monitoring,
biogeochemical nutrient cycling, and stand dynamics. Table 1.2-5 summarizes major studies that
were related to red spruce and calculated critical loads of nitrogen and sulfur at the HBEF
(HBES, 2008b; Pardo and Driscoll, 1996; USFS, 2008a).
Table 1.2-5. Major Studies at the Hubbard Brook Experimental Forest
Authors
Driscoll et al.
Pardo and
Driscoll
Palmer et al.
Siccama et al.
Year
1989
1996
2004
2007
Title
Changes in the chemistry of
surface waters: 25 -year results at
the HBEF
Critical loads for nitrogen
deposition: case studies at two
northern hardwood forests
Long-term trends in soil solution
and stream water chemistry at
the HBEF; relationship with
landscape position
Population and biomass
dynamics of trees in a northern
hardwood forest at HBEF
Key Finding
A decline in the sum of base cations in
surface water paralleled the sulfate
decline in atmospheric deposition,
preventing any long-term decrease in
stream acidity. There have been no
significant long-term trends in
precipitation inputs or stream outflow
ofNO3".
Critical loads for nitrogen deposition
with respect to acidity ranged from 0 to
630 eq/ha/yr; critical loads with respect
to effects of elevated nitrogen
(eutrophication and nutrient
imbalances) ranged from 0 to 1,450
eq/ha/yr.
Significant declines in strong acid
anion concentrations were accompanied
by declines in base cation
concentrations in soil solutions draining
the Oa and Bs soil horizons at all
elevations. Persistently low Ca2+/Al;
ratios (<1) in Bs-horizon soil solutions
at these sites may be evidence of
continuing Al stress to trees.
Tree data from 1991 to 2001, including
total aboveground biomass, in-growth
of >10 cm DBH trees, mortality,
biomass by type, aboveground net
primary productivity, and net
ecosystem productivity.
Note: eq/ha/yr = equivalents per hectare per year; DBH = diameter at breast height.
1.2.4.2 Plot Selection for Hubbard Brook Experimental Forest Case Study Area
Selection of plots for the HBEF Case Study Area was restricted to Watershed 6 (Figure
1.2-5). This watershed is 13.2 ha and is maintained as the biogeochemical control watershed for
Final Risk and Exposure Assessment
Appendix 5-22
September 2009
-------
Terrestrial Acidification Case Study
research studies. It consists of typical northern hardwood species (e.g., sugar maple, beech,
yellow birch) on the lower 90% of its area and by a montane boreal transition forest of red
spruce, balsam fir, and white birch (e.g., spruce-fir forest type) on the highest 10% of its area.
The watershed is divided into 208 25x25-m2 grid cells. This grid system and the 2002 Forest
Inventory for the watershed were used to identify the nine grid units (units 9, 14, 15, 21 to 24,
32, and 33) within the northeast portion of the watershed that contain large portions of red spruce
trees (Figure 1.2-6). These nine grid cells were combined into a 0.56-ha plot for the HBEF Case
Study Area (Figure 1.2-7). This case study plot is located at 43°57'N, 71°44'W and is 762.0
to769.3 m in elevation. Soils within the plot are from the Tunbridge-Lyman soil association and
consist of Tunbridge and Lyman soil series with smaller inclusions of Marlowe and Peru soils.
Red spruce accounts for 18.8% of the total basal area (131.3 m2/ha) in the plot area.
Final Risk and Exposure Assessment September 2009
Appendix 5-23
-------
Terrestrial Acidification Case Study
Hubbard Brook Experimental Forest
Legend
I | Watershed Boundaries
NLCD Classifications
I I Developed, Open Space
03^] Deciduous Forest
^^| Evergreen Forest
] Mixed Forest
| | Scrub/Shrub
I I Pasture/Hay
^H Cultivated Crops
| | Woody Wetlands
New
York
New
Hampshire
Maine
Massachusetts
Legend
—I Hubbard Brook
' ' Forest Location
Figure 1.2-5. Vegetation cover (NLCD, 2001) and location of Watershed 6 of Hubbard
Brook Experimental Forest.
Final Risk and Exposure Assessment
Appendix 5-24
September 2009
-------
Terrestrial Acidification Case Study
- ---!!
T
„ J * * » | M -
-*—(!i - W -fc—it—1 t—-»
fl i 1 « i « I « 1* *f \
- !J_ V _ 3i _i_±_£-±
• •- -
ikliMl:!
a
i - - r s - s i u i.
I « - i> m I in I IP IM I w [
b.«i -a—a - w_»—*•—a
, * !1 •-,''! '".'!-,_
(i
-
i
—
Ltlr-
i •
• -
..
•e I w I n>
"
U| I U
•
Figure 1.2-6. Grid units within Watershed 6 of Hubbard Brook Experimental Forest. The
red outline delineates the spruce-fir forest type. The dotted grid cell areas indicate the
grid units with high proportions of red spruce and represents the composite plot area for
the Hubbard Brook Experimental Forest Case Study Area.
Final Risk and Exposure Assessment
Appendix 5-25
September 2009
-------
Terrestrial Acidification Case Study
Legend
l~l Watershed Boundaries
11 ' Case Study Location
Land Cover Classifications
j I Developed. Open Space
iH Deciduous Foresl
^H Evergreen Forest
I I Mixed Foresl
I 1 Scrub/Shrub
1 I Paslure/Hay
^H Cultivated Crops
~1 Woody Wetlands
Figure 1.2-7. Location of case study plots within Watershed 6 of Hubbard Brook
Experimental Forest.
2.0 APPROACH AND METHODS
The ISA (U.S. EPA, 2008c, Section 3.1.1) identified critical load assessments as a
suitable approach to evaluate the potential impacts of anthropogenic pollution on biological end
points and ecosystem impairment. A critical load is "a quantitative estimate of ecosystem
exposure to one or more pollutants below which significant harmful effects on specified sensitive
elements of the environment do not occur, according to present knowledge" (McNulty et al.,
2007). Critical loads of acidity from atmospheric nitrogen and sulfur deposition for an ecosystem
have been specifically defined as "the highest deposition of acidifying compounds that will not
cause chemical changes leading to long-term harmful effects on ecosystem structure and
function" (Nilsson and Grennfelt, 1988). "The basic idea of the critical load concept is to balance
the depositions that an ecosystem is exposed to with the capacity of this ecosystem to buffer the
Final Risk and Exposure Assessment
Appendix 5-26
September 2009
-------
Terrestrial Acidification Case Study
input (e.g., the acidity input buffered by the weathering rate), or to remove it from the system
(e.g., nitrogen by harvest) without harmful effects within or outside the system" (UNECE, 2004).
European countries have been using critical load assessments for many years to
determine the impacts of atmospheric nitrogen and sulfur deposition in forest ecosystems. These
studies have served as the platform for informing policy related to the control and reduction of
emissions of acidifying pollutants. The International Cooperative Programme (ICP) on
Modelling and Mapping Critical Loads and Levels and Air Pollution Effects, Risks and Trends
has published a series of manuals (the most recent in 2004) to provide guidance on calculating
and mapping critical loads. These manuals helped parties to the United Nations Economic
Commission for Europe (UNECE) Convention on Long-Range Transboundary Air Pollution
(CLRTAP) meet their obligations and conduct effects and risk assessments (UNECE, 2004).
Canada has also completed critical load evaluations in support of efforts to design emission-
reduction programs (Jeffries and Lam, 1993; RMCC, 1990). Critical load modeling was included
in the 7997 Canadian Acid Rain Assessment (Jeffries, 1997) for several regions in eastern
Canada.
The establishment and analysis of critical loads within the United States is relatively new.
The Conference of New England Governors and Eastern Canadian Premiers (NEG/ECP) funded
studies that used critical load-based methods to estimate sustainable acidifying deposition rates
and exceedances for upland forests representative of the New England states and the eastern
Canadian Provinces in 2000 to 2001 (NEG/ECP Forest Mapping Group, 2001). More recently,
McNulty et al. (2007) completed a national critical load assessment for U.S. forest soils at a 1-
km2 scale.
Within the ISA (U.S. EPA, 2008c, Section D.2.2), EPA detailed an 8-step protocol to
define the basic critical load question in any analysis. Those steps are repeated here:
1. Identify the ecosystem disturbance that is occurring (e.g., acidification, eutrophication).
Not all disturbances will occur in all regions or at all sites, and the degree of disturbance
may vary across landscape areas within a given region or site.
2. Identify the landscape receptors that are subjected to the disturbance (e.g., forests,
surface waters, crops). Receptor sensitivity may vary locally and/or regionally, and the
hierarchy of those receptors that are most sensitive to a particular kind of disturbance
may vary as well.
Final Risk and Exposure Assessment September 2009
Appendix 5-27
-------
Terrestrial Acidification Case Study
3. Identify the biological indicators within each receptor that are affected by atmospheric
deposition (i.e., individual organism, species, population, or community characteristics).
Indicators will vary geographically and perhaps locally within a given receptor type.
4. Establish the critical biological responses that define "significant harm" to the biological
indicators (e.g., presence/absence, loss of condition, reduced productivity, species shifts).
Significant harm may be defined differently for biological indicators that are already at
risk from other stressors or for indicators that are perceived as "more valued."
5. Identify the chemical indicators or variables that produce or are otherwise associated
with the harmful responses of the biological indicators (e.g., streamwater pH, lake Al
concentration, soil base saturation). In some cases, the use of relatively easily measured
chemical indicators (e.g., surface water pH or acid neutralizing capacity [ANC]) may be
used as a surrogate for chemical indicators that are more difficult to measure (e.g., Al
concentration).
6. Determine the critical chemical limits for the chemical indicators at which the harmful
responses to the biological indicators occur (e.g., pH < 5, base saturation < 5%, inorganic
Al concentration greater than 2 umol). Critical limits may be thresholds for indicator
responses, such as presence/absence, or may take on a continuous range of values for
continuous indicator responses, such as productivity or species richness. Critical limits
may vary regionally or locally depending on factors such as temperature, existence of
refugia, or compensatory factors (e.g., high Ca2+ concentration mitigates the toxicity of
Al to fish and plant roots).
7. Identify the atmospheric pollutants that control (affect) the pertinent chemical indicators
(e.g., deposition of SC>42", N(V, ammonium [NH4+], nitric acid [HNOs]). Multiple
pollutants can affect the same chemical variable. The relative importance of each
pollutant in producing a given chemical response can vary spatially and temporally.
8. Determine the critical pollutant loads (e.g., kg/ha/yr total deposition of sulfur or
nitrogen) at which the chemical indicators reach their critical limits. Critical pollutant
loads usually include both wet and dry forms of pollutant deposition. The critical
pollutant load may vary regionally within a receptor or locally within a site (e.g., as
factors such as elevation or soil depth vary) and may vary temporally at the same location
(e.g., as accumulated deposition alters chemical responses).
Final Risk and Exposure Assessment September 2009
Appendix 5-28
-------
Terrestrial Acidification Case Study
As shown in the eight steps above, a variety of indicators and responses can be
incorporated into the estimation of a critical load, the point at which ecological impacts occur.
Varying any one of these will result in a different critical load estimate. As a result, there is no
single definitive critical load for an ecosystem. In this case study, terrestrial acidification was
evaluated using the chemical indicator of the Bc/Al ratio in the soil solution (as a surrogate for
Ca2+/Al—discussed earlier at the end of Section 1.1.1) and biological indicators (ecological
endpoints) of red spruce and sugar maple tree health and growth. The critical chemical limits
discussed above allow for the calculation of multiple critical loads, depending on the level of
protection of interest. Three base cation to aluminum ratio - critical load (Bc/Al)crit ratio values
were applied in this case study to provide a range of protection (i.e., low, intermediate, high) to
tree health and growth, and these values (Bc/Al)crit ratios) are detailed in Section 2.1.2.2.
Several methodological approaches can be taken to estimate critical loads in terrestrial
ecosystems. Three of the most commonly used methods are empirically derived estimations,
steady-state mass-balance model estimations, and dynamic model estimations (Bull et al., 2001;
Bobbink et al., 2003; Jenkins et al., 2003; McNulty et al., 2007; UNECE, 2004).
The UNECE CLRTAP has used the empirically-derived estimation approach within their
mapping framework. Empirically derived critical load estimates of atmospheric nitrogen
deposition for specific receptor groups within natural and seminatural terrestrial ecosystems and
wetland ecosystems were first presented in a background document for the 1992 workshop on
critical loads held under the UNECE CLRTAP Convention at Lokeberg (Sweden) (Bobbink et
al., 1992). Updates to the empirically derived loads were completed for a 2007 update to the
2004 Manual on Methodologies and Criteria for Modeling and Mapping Critical Loads and
Levels and Air Pollution Effects, Risks, and Trends (henceforth referred to as the TCP Mapping
and Modeling Manual) (UNECE, 2004). Empirically derived critical loads can provide good
estimates of the impacts of acidifying deposition on terrestrial systems. However, they require
data from studies that establish the impacts of varying loads (e.g., amount and duration) on
ecosystem processes and attributes and have a limited ability to extrapolate to other systems with
different characteristics.
The mass-balance model estimation method to calculate critical loads consists of simple
models that relate chemical indicators (e.g., related to or indicative of biological impact of
acidifying deposition) to the deposition levels observed in an ecosystem. The chemical indicator
Final Risk and Exposure Assessment September 2009
Appendix 5-29
-------
Terrestrial Acidification Case Study
used in the mass balance calculations must have a proven relationship to the biological indicator.
With the mass-balance approach, critical loads are calculated by relating the flow of acidifying
agents (i.e., base cations and other ions) and nutrients into, out of, and within an ecosystem.
These mass-balance models are steady state and offer estimates of critical loads for time frames
based on the data used to evaluate the mass balance (UNECE, 2004). To accurately characterize
the steady-state ecosystem condition and impacts of acidifying deposition, it is important to use
long-term averages of input fluxes in the mass-balance calculations. Benefits of the simple mass-
balance approach are its ease of use, moderate data requirements, and applicability over a large
area (Pardo and Driscoll, 1996). Disadvantages, however, include an inability to incorporate
changes or ecosystem responses into the modeled critical load estimates.
Dynamic-model estimation methods simulate the processes of pollutant fate and transport
into, out of, and within a system on a temporally varying basis. They are more data intensive
than mass-balance models and require the modeling of temporal rates and processes in addition
to the mass balance of acidifying agents, base cations, and nutrients. Some dynamic models
involve the integration of hydrologic, geochemical, and biological processes, but such models
are still of limited use in determining critical loads (Pardo and Driscoll, 1996). An advantage of
dynamic models is that they allow for an estimation or prediction of ecosystem response over
time and under different acidifying deposition scenarios (Pardo and Duarte, 2007).
2.1 CHOSEN METHOD
The Simple Mass Balance (8MB) model, outlined in the TCP Mapping and Modeling
Manual (UNECE, 2004) to determine terrestrial critical loads, was used to estimate the critical
loads of acidifying nitrogen and sulfur deposition in the KEF and HBEF (i.e., for sugar maple
and red spruce, respectively) case study areas. This model is currently the most commonly used
approach to estimate critical loads and has been widely applied in Europe (Sverdrup and de
Vries, 1994), the United States (McNulty et al., 2007; Pardo and Duarte, 2007), and Canada
(Watmough et al., 2006; Ouimet et al., 2006). Although a limitation of the 8MB model is that it
is a steady-state model, as stated by the UNECE (2004), "Since critical loads are steady-state
quantities, the use of dynamic models for the sole purpose of deriving critical loads is somewhat
inadequate." In addition, if dynamic models are "used to simulate the transition to a steady state
for the comparison with critical loads, care has to be taken that the steady-state version of the
Final Risk and Exposure Assessment September 2009
Appendix 5-30
-------
Terrestrial Acidification Case Study
dynamic model is compatible with the critical load model" (UNECE, 2004). Therefore, the
selection of the 8MB model is the most suitable approach for this case study examining critical
loads for sugar maple and red spruce.
The 8MB model examines a long-term, steady-state balance of base cation, chloride, and
nutrient inputs, "sinks," and outputs within an ecosystem. With this model, base cation
equilibrium is assumed to equal the system's critical load. It is a single-layer model, where
assumptions stipulate that the soil layer is a homogeneous unit at least as deep as the rooting
zone, so that the nutrient cycle can be ignored. This allows the model to focus directly on growth
and uptake processes. There are several additional assumptions that are included with application
of the 8MB model (UNECE, 2004):
• All evapotranspiration occurs on the top of the soil profile
• Percolation is constant through the soil profile and occurs only vertically
• Physico-chemical constants are assumed to be uniform throughout the whole soil profile
• Internal fluxes (e.g., weathering rates, nitrogen immobilization) are independent of soil
chemical conditions (e.g., pH).
The 8MB model relates atmospheric nitrogen and sulfur deposition to a critical load by
incorporating mass balances for nitrogen and sulfur within the soils with the charge balance of
ions in the soil leaching flux. This model accounts for the processes that add and remove
nitrogen and sulfur, as well as base cations and other charged elemental species, from the soil.
Although this model analyzes both total nitrogen and sulfur deposition loads, it does not
allow for the analysis of the specific effects of the different total reactive nitrogen species.
However, as stated in Chapter 5 of the ICP Mapping and Modeling Manual, "the possible
differential effects of the deposited nitrogen species (oxidized nitrogen [NOy] or reduced
nitrogen [NHX]) are insufficiently known to make a differentiation between these nitrogen
species for critical load establishment" (UNECE, 2004). Therefore, attempting an analysis of the
impacts of different nitrogen species was not seen as necessary.
2.1.1 Critical Load Equations and Calculations
2.1.1.1 Simple Mass Balance Calculations
The 8MB model used to estimate critical loads of acidity in this case study is presented in
Equation 1. The full derivation of this equation is detailed in the ICP Mapping and Modeling
Final Risk and Exposure Assessment September 2009
Appendix 5-31
-------
Terrestrial Acidification Case Study
Manual (UNECE, 2004). Unless otherwise stated, all variables are expressed in units of eq/ha/yr.
Equivalent, or "eq," is a unit that removes the influence of molecular weight and is equivalent to
"mole." For example, 1 g of Ca is equal to 0.25 eq (1 g/the molecular weight of Ca = 40.08).
CL(S + N) = BCdep - Cldep + BCW -Bcu + N, + Nu + Nde - ANCle,cnt (1)
where
CL(S+N) = forest soil critical load for combined nitrogen and sulfur acidifying
deposition ((N+S)comb)
BCdep = base cation (Ca2+ + K+ + Mg2+ + Na+) deposition4
Cldep = chloride deposition
BCW = base cation (Ca2+ + K+ + Mg2+ + Na+) weathering
Bcu = uptake of base cations (Ca2+ + K+ + Mg2+) by trees
N; = nitrogen immobilization
Nu = uptake of nitrogen by trees
Nde = denitrification
ANCie,crit = forest soil acid neutralizing capacity of critical load leaching
NOTE: There is a distinction between the base cation variables base cation (BC) and Be. BC
includes all four base cations (Ca2+ + K+ + Mg2+ and K+), whereas Be only includes three
cations—those that are taken up by vegetation (Ca2+ + K+ + Mg2+) (UNECE, 2004). Terms in the
8MB equations that are directly related to or impact vegetation use the Be variable.
Some of these parameters had defined or selected input values (BCdep, Cldep, N;, Nu and
Nde), while four of these parameters, including BCW, Bcu, Nu and ANCie,crit, required calculation.
Two methods were used to calculate BCW in this case study; the clay-substrate method
and the soil type-texture approximation method. The clay-substrate method has been used by
many researchers in North America (Ouimet et al., 2006; Watmough et al., 2006; McNulty et al.,
2007; Pardo and Duarte 2007), and the soil type-texture approximation is one of the methods
outlined in the TCP Mapping and Modeling Manual (UNECE, 2004). Base cation weathering is
4 The ICP Mapping and Modeling Manual (UNECE, 2004) recommends that wet deposition be corrected for sea salt
on sites within 70 km of the coast. Both the HBEF and KEF case study areas are greater than 70 km from the
coast, so this correction was not used.
Final Risk and Exposure Assessment September 2009
Appendix 5-32
-------
Terrestrial Acidification Case Study
the most influential and most difficult-to-estimate parameter within the 8MB model (Whitfield et
al., 2006; Li and McNulty, 2007). Therefore, these two methods were chosen to provide a range
of BCW estimates within which the correct value probably lies (discussed further in Section 5).
Base cation weathering was calculated with the clay-substrate method using equations
outlined by McNulty et al. (2007) (Equations 1 to 3). This method relies on a combination of
parent material and clay percentage to determine the soil weathering rate. Parent material acidity
was determined by silica content (see Table 3 in McNulty et al., 2007).
Acid Substrate: BCe = (56.7 x %clay)- (p.32 x (%clay)2) (2)
Intermediate Substrate: BCe = 500 + (53.6 x %clay)- (o. 18 x (%clay)2) (3)
Basic Substrate: BCe = 500 + (59.2 x %clay) (4)
where
BCe = empirical soil base cation (Ca2+ + K+ + Mg2+ + Na+) weathering rate
(eq/ha/yr)
% clay = the percentage of clay within the soil.
The empirical base cation weathering rate was corrected for soil temperature and depth of
the rooting zone soil (Sverdrup and de Vries, 1994; Hodson and Langan, 1999; van der Salm and
de Vries, 2001; UNECE, 2004; Watmough et al., 2006; Whitfield et al., 2006; Pardo and Duarte
2007; NEG/ECP Forest Mapping Group, 2001) to determine the final BCW as outlined in
Equations 5 and 6.
T~»/^ T~»/^ V V2-6+273 j i z,/j-r i y i /r-\
BCC =BCe xev JJ (5)
BCW = BCC x depth (6)
where
BCC = base cation (Ca2+ + K+ + Mg2+ + Na+) weathering rate corrected for
temperature (eq/ha/yr/m)
A = Arrhenius constant (3,600 kelvin [K])
Final Risk and Exposure Assessment September 2009
Appendix 5-33
-------
Terrestrial Acidification Case Study
Tm = mean annual soil (or air) temperature (°C)
Depth = the depth of rooting zone mineral soil (m).
Base cation weathering was calculated with the soil type-texture approximation method
using Tables 2.1-1, 2.1-2, and 2.1-3 and Equation 7 (UNECE, 2004). Similar to the clay-
substrate method, the soil type-texture approximation method also requires a combination of soil
texture and parent material acidity to calculate base cation weathering. The soil texture class was
determined by percentages of clay and sand (Table 2.1-1), and the parent material acidity was
classified according to Food and Agriculture Organization (FAO) soil types (Table 2.1-2). Soil
texture class and parent material acidity was then combined to determine the weathering rate
class (Table 2.1-3)
Table 2.1-1. Soil Texture Classes as a Function of Clay and Sand Content
Texture Class
1
2
3
4
5
Name
Coarse
Medium
Medium fine
Fine
Very fine
Definition
clay <18% and sand >65%
clay <35% and sand >15%, but
clay>18%ifsand>65%
clay <35% and sand <15%
35%60%
Source: UNECE, 2004.
Table 2.1-2. Parent Material Classes for Common FAO Soil Types
Parent Material
Acidic
Intermediate
Basic
Organic
FAO Soil Type
Ah, Ao, Ap, B, Ba, Bd, Be, Bf, Bh, Bm, Bx, D, Dd, De, Dg, Gx, I, Id, le, Jd, P,
Pf, Pg, Ph, PI, Po, Pp, Q, Qa, Qc, Qh, Ql, Rd, Rx, U, Ud, Wd
A, Af, Ag, Bv, C, Cg, Ch, Cl, G, Gd, Ge, Gf, Gh, Gi, Gl, Gm, Gs, Gt, H, Hg, Hh,
HI, J, Je, Jm, Jt, L, La, Ld, Lf, Lg, Lh, Lo, Lp, Mo, R, Re, V, Vg, Vp, W, We
F, T, Th, Tm, To, TV
O, Od, Oe, Ox
Source: UNECE, 2004.
Final Risk and Exposure Assessment
Appendix 5-34
September 2009
-------
Terrestrial Acidification Case Study
Table 2.1-3. Weathering Rate Classes as a Function of Texture and Parent Material Classes
Parent
Material
Acidic
Intermediate
Basic
Organic
Texture Class
1
1
2
2
2
3
4
5
3
3
4
5
4
6
6
6
5
6
6
6
Class 6 for Oe and class 1 for other organic soils
Source: UNECE, 2004.
273+J
(7)
where
z = rooting zone soil depth (m)
WRc = weathering rate class
T = average annual soil temperature (°C)
A = Arrhenius constant (3,600 K)
Base cation (Bcu) and nitrogen (Nu) uptake were calculated for this case study using the
equation outlined by McNulty et al. (2007) (Equation 8). These terms represent nutrients that are
taken up from the soil and used to support tree growth and maintenance but are eventually
returned to the system through litter senescence and decay. In a forest stand that does not
experience biomass removal, these nutrients are internally cycled and not lost from the system.
Under this scenario, both Bcu and Nu would be given values of 0 equivalents per hectare per year
(eq/ha/yr) in the 8MB calculations. However, in a managed stand that is harvested, base cations
and nitrogen taken up by the trees are removed from the forest system with tree harvesting and
are, therefore, considered a loss or output from the system within the 8MB calculations.
Watershed 6 in HBEF is a reference watershed and is not harvested. Therefore, for the
FffiEF Case Study Area, the Bcu and Nu variables were assumed to have a value of 0 eq/ha/yr
because biomass and nutrients are not removed from these plots. In contrast, most of the stands
in the KEF are harvested, and therefore, as discussed further in Section 3.1.1 Bcu and Nu were
estimated for the seven plots in the KEF Case Study Area. Equation 8 was modified, as
Final Risk and Exposure Assessment
Appendix 5-35
September 2009
-------
Terrestrial Acidification Case Study
necessary, to estimate uptake in the bark and bole for nitrogen, Ca2+, Mg2+, and K+. These
calculations were conducted for each species on each plot.
Uptake (eq/ha/yr) = AVI x NC x SG x % bark x 0.65 (8)
where
AVI = average forest volume increment (m3/ha/yr)
NC = base cation ((Ca2+ + K+ + Mg2+ ) or nitrogen nutrient concentration in bark and
bole (%)
SG = specific gravity of bark and bole wood (g/cm3)
% bark = percentage of volume growth that is allotted to bark
65% = average aboveground tree volume that is removed from the site (Birdsey,
1992; Hall et al., 1998; Martin et al., 1998).
Acid neutralizing capacity (ANC(ie,Crit)) represents the buffering or acid neutralizing
capacity of the soil, and the selection of the chemical indicator for the effects on the biological
receptor or ecological endpoint occurs within the calculation of ANC(ie,Crit). Several formulations
for ANC(ie,crit) exist, depending on which indicator is being used to examine the critical load for
the biological receptor (endpoint), sensitivity to pH conditions, or sensitivity to the toxic effects
of Al. A large proportion of the research indicates Al toxicity in relation to Ca2+ depletion as the
main indicator of red spruce and sugar maple mortality and decline. Therefore, for the estimates
of critical loads for these two species at HBEF or KEF, Ca2+ and Al concentrations applied
through the base cation to aluminum (Bc/Al)crit indicator ratio were used in the ANC(ie,Crit)
calculations according to Equation 9. As outlined in the end of Section 1.1.1, the Bc/Al ratio is a
good surrogate for the Ca2+/Al indicator and is the most commonly used indicator (Bc/Al(crit)) in
estimations of acid load (McNulty et al., 2007; Ouimet et al., 2006; UNECE, 2004).
xl/3
1.5:
Bcdep+Bcw-BCu
KM x —
Bcdep+Bc-Bc
-1.5x^-— (9)
where
Final Risk and Exposure Assessment September 2009
Appendix 5-36
-------
Terrestrial Acidification Case Study
Q = annual runoff in m3/ha/yr
Bcdep = base cation (Ca2+ + K+ + Mg2+) deposition5
Bcw = soil base cation (Ca2++ K+ + Mg2+) weathering6
Bcu = base cation (Ca2++ K+ + Mg2+) uptake by trees
.Kgibb = the gibbsite equilibrium constant (a function of forest soil organic matter
content that affects Al solubility) (UNECE, 2004)
(Bc/Al)crit = the base cation to aluminum ratio (indicator)
Base cation weathering (Bcw) in the ANC(ie,Crit) parameter was calculated using the two
methods described earlier: the clay-substrate and soil type-texture approximation methods.
However, sodium (Na+) typically accounts for 10% to 30% of base cation weathering (BCW)
(Sverdrup and deVries 1994), and therefore, the Bcw, which only consists of Ca2+, K+, and Mg2+,
was determined by multiplying BCW by 0.80, the mid-range of the Na+ proportional content.
2.1.1.2 Deposition Relative to Critical Load Calculations
If total nitrogen and sulfur deposition (combined) ((N+S)COmb) is greater than the
calculated critical load for acidity, the soil is no longer able to neutralize the acidifying
deposition, and there is increased likelihood of environmental harm (McNulty et al., 2007).
Deposition of (N+S)comb (expressed as eq/ha/yr) that is greater than the critical load was
calculated in this case study by comparing the 8MB estimated critical load to the CMAQ/NADP
total nitrogen and sulfur deposition levels as outlined in Equation 3.
Ex(S + N)dep = Sdep + Ndep - CL(S + N) (10)
where
Ex = exceedance of the forest soil critical nitrogen and sulfur loads
(S+N)deP = the deposition of sulfur and nitrogen.
5 Bcdep is not the same as BCdep used in Equation 1. BCdep includes Ca2+, K+, Mg2+, and Na+, whereas Bcdep includes
base cations that are taken up by vegetation (i.e., only includes Ca2+, K+, and Mg2+),
6 Bcw is not the same as BCW used in Equation 1. BCW includes Ca2+, K+, Mg2+, and Na+, whereas Bcw includes base
cations that are taken up by vegetation (i.e., only includes Ca2+, K+, and Mg2+).
Final Risk and Exposure Assessment September 2009
Appendix 5-37
-------
Terrestrial Acidification Case Study
2.1.1.3 Critical Load Function
The critical load function (CLF) expresses the relationship between all combinations of
total nitrogen and sulfur deposition ((N+S)comb) and the critical load of an ecosystem. To define
the CLF, minimum and maximum critical load levels for both total nitrogen and sulfur
deposition must be determined (UNECE, 2004). These maximum and minimum levels were
calculated in this case study using Equations 1 1 through 13 (UNECE, 2004).
CLmax(s)=BCdep -Cldep +BCW -BCU - ANCle,cnt (11)
CLmm(N) = NI+Nu+Nde (12)
CL
CLmax(N) = CLmm(N)+ - (13)
1
e
where
fde = denitrification fraction (0
-------
Terrestrial Acidification Case Study
CLmax(S)
o
in
O
a
4*
Q
CLmm(N)
CLmax(N)
N Deposition
Figure 2.1-1. The critical load function created from the calculated maximum and
minimum levels of total nitrogen and sulfur deposition (eq/ha/yr). The grey areas show
deposition levels less than the established critical loads. The red line is the maximum
critical level of sulfur deposition (valid only when nitrogen deposition is less than the
minimum critical level of nitrogen deposition [blue dotted line]). The flat line portion of
the curves indicates nitrogen deposition corresponding to the CLm;n(N) (i.e., nitrogen
absorbed by nitrogen sinks within the system).
2.1.2 Critical Load Data Requirements
2.1.2.1 Data Requirements and Sources
Atmospheric, hydrologic, soil, bedrock geology, and tree measurement data are necessary
to evaluate critical loads associated with total nitrogen and sulfur deposition. The specific data
requirements to satisfy Equations 1 through 13 and calculate critical loads and CLF for this case
study are presented in Table 2.1-4. This table also outlines the sources of these data specific to
the two case study areas. Cloud deposition of nitrogen was not included in the critical load
calculations because of the lack of available data. However, it should be noted that cloud
deposition coupled with wet and dry deposition can result in 6 to 20 times greater total nitrogen
deposition at high elevation relative to low elevation sites (Baumgardner et al., 2003). Therefore,
total nitrogen deposition and the degree to which total nitrogen deposition exceeds the critical
load at the HBEF Case Study Area may be underestimated.
Final Risk and Exposure Assessment
Appendix 5-39
September 2009
-------
Terrestrial Acidification Case Study
Table 2.1-4. Data Requirements and Sources for Calculating Critical Loads for Total Nitrogen
and Sulfur Deposition in Hubbard Brook Experimental Forest and Kane Experimental Forest
DATA
Total nitrogen
and sulfur
deposition — wet
and dry
Base cation
(Ca2+, Mg2+,
Na+, K+)
deposition — wet
Base cation
(Ca2+, Mg2+,
Na+, K+)
deposition — dry
Chlorine (Cl~)
deposition — wet
Chlorine (CF)
deposition — dry
Runoff
Mean annual
soil temperature
Soil horizon
depth
Percentage of
clay by soil
horizon
Percentage of
sand by soil
horizon
Percentage of
organic matter
by soil horizon
DATA NAME AND TYPE
Name
CMAQ/
NADP
NADP
CASTNET
NADP
CASTNET
Annual run-
off (1:
7,500,000
scale)
Soil
temperature
data (HBEF/
KEF)
SSURGO
SSURGO
SSURGO
SSURGO
Type
GIS
datalayers
GIS datalayer
GIS datalayer
GIS datalayer
GIS datalayer
GIS datalayer
Database
(HBEF)/
Peer-
reviewed
journal
articles
(KEF)
GIS datalayer
GIS datalayer
GIS datalayer
GIS datalayer
DATA SOURCE
Hubbard Brook
Experimental Forest
Provided by U.S.
Environmental Protection
Agency (EPA)/NADP,
2003a, e, h
NADP, 2003b, d, f g
U.S. EPA, 2008b
NADP, 2003c
U.S. EPA, 2008b
Gebertetal., 1987
HBES, 2008c
USDA-NRCS, 2008b
USDA-NRCS, 2008b
USDA-NRCS, 2008b
USDA-NRCS, 2008b
Kane Experimental
Forest
Provided by EPA/
NADP, 2003a, e, h
NADP, 2003b, d, f, g
U.S. EPA, 2008b
NADP, 2003c
U.S. EPA, 2008b
Gebertetal., 1987
Carter and Ciolkosz,
1980
USDA-NRCS, 2008a
USDA-NRCS, 2008a
USDA-NRCS, 2008a
USDA-NRCS, 2008a
Final Risk and Exposure Assessment
Appendix 5-40
September 2009
-------
Terrestrial Acidification Case Study
DATA
Gibbsite
equilibrium
constant (Kslbb)
Parent material/
bedrock
Food and
Agriculture
Organization
(FAO) soil type
Nitrogen
immobilization
(NO
Denitrification
(Nde)
Stand
composition
Annual volume
increment (AVI)
by species
Percentage
allocation of
growth to bark
by species
Specific gravity
(bark and bole
wood) by
species
DATA NAME AND TYPE
Name
Selected Kglbb
values
Map of
bedrock
geology
Map of
dominant soil
types
Selected N,
value
Selected Nde
value
Forest
inventory
(HBEF)/
SILVAH
(KEF)
Forest
inventory
(HBEF) /
SILVAH
(KEF)
Selected %
allocation
values
Selected
specific
gravity
values
Type
Peer-
reviewed
journal
articles and
literature
GIS datalayer
GIS datalayer
Peer-
reviewed
journal
article and
literature
Peer-
reviewed
journal
articles
Database
(HBEF)/
mensuration
model (KEF)
Database
(HBEF)/
mensuration
model (KEF)
Peer-
reviewed
journal
article
Peer-
reviewed
journal
article
DATA SOURCE
Hubbard Brook
Experimental Forest
Ouimet et al, 2006;
Watmough et al., 2006;
UNECE, 2004
USGS, 2000
FAO, 2007
McNulty et al., 2007;
UNECE, 2004
McNulty et al., 2007;
Ouimet et al., 2006;
Watmough et al., 2006
HBES, 2008a
HBES, 2008a
McNulty et al., 2007
Jenkins et al., 2001
Kane Experimental
Forest
Ouimet et al., 2006;
Watmough et al., 2006;
UNECE, 2004
PADCNR, 2001
FAO, 2007
McNulty et al., 2007;
UNECE, 2004
McNulty et al., 2007;
Ouimet et al., 2006;
Watmough et al., 2006
Thomasma et al., 2008
Thomasma et al., 2008
McNulty et al., 2007
Jenkins etal., 2001
Final Risk and Exposure Assessment
Appendix 5-41
September 2009
-------
Terrestrial Acidification Case Study
DATA
Nutrient
concentration
(bark and bole
wood) by
species —
nitrogen, Ca2+,
Mg2+, K+
Percentage
biomass (bark
and bole)
removal during
harvest
DATA NAME AND TYPE
Name
Nitrogen,
Ca2+, K+,
Mg2+
concentra-
tions in bark
and bole
wood by
species
Selected
value
Type
Forest
Service
Report
Peer
reviewed
journal
article
DATA SOURCE
Hubbard Brook
Experimental Forest
Pardo et al, 2004
McNulty et al., 2007
Kane Experimental
Forest
Pardo et al., 2004
McNulty et al., 2007
Note: CMAQ = Community Multiscale Air Quality Model; NADP = National Atmospheric Deposition
Program; CASTNET = Clean Air Status and Trends Network; GIS = Geographic Information System;
SSURGO = Soil Survey Geographic Database; SILVAH = Silviculture of Allegheny Hardwoods
2.1.2.2 Selection of Indicator Values
As described at the end of Section 1.1.1, the Bc/Al ratio ((Bc/Al)crit) in the soil solution
was selected as the indicator for the calculation of critical loads in this case study. The (Bc/Al)crit
connects the acid-influenced chemical status of the soil with the tree response: as the ratio
decreases, tree health and growth can be impaired because of reduced uptake of base cations and
increased Al toxicity. Most studies that calculate critical loads of acidity set the (Bc/Al)crit ratio
to 1.0 or 10.0 (McNulty et al., 2007; NEG/ECP Forest Mapping Group, 2001; Pardo and Duarte,
2007; UNECE, 2004). The (Bc/Al)crit ratio of 1.0 is a common default value in European forests
(UNECE, 2004) and has been applied to coniferous forests in the United States (McNulty et al.,
2007). A (Bc/Al)crit ratio of 10.0 is a more conservative ratio and has been applied to hardwood
forests in the United States (McNulty et al., 2007), in Canadian forests (Ouimet et al., 2006;
Watmough et al., 2006), and in systems where maintained tree health is required (NEG/ECP
Forest Mapping Group, 2001). Soil solution Bc/Al ratios of 10.0 are less likely to reduce soil
base saturation and are not known to impair tree vigor or growth.
Cronan and Grigal (1995) conducted a meta-analysis of research investigating the
relationship between soil solution Ca2+/Al ratio and growth of 18 tree species. They found a 50%
chance of negative impacts on tree growth or nutrition when the soil solution Ca2+/Al ratio was
Final Risk and Exposure Assessment
Appendix 5-42
September 2009
-------
Terrestrial Acidification Case Study
as low as 1.0, a 75% chance when the soil solution ratio was as low as 0.5, and nearly a 100%
chance of impaired tree growth or nutrition when the soil solution Ca2+/Al molar ratio was as low
as 0.2. In a similar meta-analysis of studies that explored the relationship between Bc/Al and tree
growth, Sverdrup and Warfvinge (1993b) reported the Bc/Al ratio at which growth was reduced
by 20% relative to control trees. Figure 2.1-2 presents the findings of Sverdrup and Warfvinge
(1993b) based on 46 of the tree species that grow in North America. This summary indicates that
there is a 50% chance of negative tree response (i.e., greater than 20% reduced growth) at a soil
solution Bc/Al ratio of 1.2. Sverdrup and Warfvinge (1993b) also presented the results of studies
conducted on individual tree species. Figures 2.1-3 and 2.1-4 show growth in sugar maple and
red spruce, respectively. According to these figures, sugar maple growth was reduced by 20%
and red spruce growth was reduced by 35% (relative to controls) at a Bc/Al ratio of 0.6.
Three Bc/Al ratio ((Bc/Al)crit) values were used in this case study to evaluate different
levels of protection associated with total nitrogen and sulfur deposition: 0.6, 1.2, and 10 (Table
2.1-5). The (Bc/Al)crit ratio of 0.6 represents the highest level of impact (lowest level of
protection) to tree health and growth; as much as 75% of 46 tree species found in North America
experience reduced growth at this ratio (Sverdrup and Warfvinge, 1993b). Both red spruce and
sugar maple show at least a 20% reduction in growth at the 0.6 (Bc/Al)crit ratio. The (Bc/Al)crit
ratio of 1.2 is considered to represent a moderate level of impact; the growth of 50% of tree
species (found growing in North America) were negatively impacted at this soil solution ratio
(Figure 2.1-4). The (Bc/Al)crit ratio of 10.0 was selected to represent the lowest level of impact
(greatest level of protection) to tree growth; it is the most conservative value used in studies that
have calculated critical loads in the United States and Canada (NEG/ECP Forest Mapping
Group, 2001; McNulty et al., 2007; Watmough et al., 2004).
Table 2.1-5. The Three Indicator (Bc/Al)cr;t Soil Solution Ratios Used in This Case Study and the
Corresponding Levels of Protection to Tree Health and Critical Loads
Indicator (Bc/Al)crit Soil
Solution Ratio
0.6
1.2
10.0
Level of Protection to Tree
Health
Low
Intermediate
High
Critical Load
High
Intermediate
Low
Final Risk and Exposure Assessment September 2009
Appendix 5-43
-------
Terrestrial Acidification Case Study
10
fl
o
"o
2 1 -
0
C/3
a
0
cS
-------
Terrestrial Acidification Case Study
G
O
(J
I
s
o
»*»»*
ff)
120
100
80
60
40
20
0
-o
A McQuattie and Schier (1990)
* Schier (1984)
A Ohooetal.<1988}
• Hutchinson et al, (1985)
O Thornton et al. (1987)
* Joslin and Wolfe (L 989)
0,01 0,1 1 10 100 1,000
Soil solution (Ca+Mg4-K)/Al molar ratio
Figure 2.1-4. The relationship between soil solution Bc/Al ratio and biomass or root
growth in red spruce (from Sverdrup and Warfvinge, 1993b).
2.1.2.3 Case Study Input Data
The data used to calculate critical loads for sugar maple and red spruce in the KEF and
HBEF of this case study are presented in Tables 2.1-5, 2.1-6, 2.1-7, and 2.1-8. The majority of
the data was specific to the case study areas and was compiled from published research studies
and models, site-specific databases, or spatially-explicit GIS datalayers. However, several of the
parameters, including denitrification (Nde), nitrogen immobilization (N;), the gibbsite equilibrium
constant (Kgibb), and rooting zone soil depth required the use of default values or values used in
published critical load assessments. Denitrification loss of nitrogen (Nde) was assumed to be 0.0
eq/ha/yr because both the KEF and FffiEF study plots are upland forests and denitrification is
considered negligible in such forests (McNulty et al., 2007; Ouimet et al., 2006; Watmough et
al., 2006). The TCP Mapping and Modeling Manual (UNECE, 2004) reported values of N;in the
soil, ranging from 14.3 to 35.7 eq/ha/yr in colder climates and up to 71.4 eq/ha/yr in warmer
climates. Nitrogen immobilization (N;) was set to 42.86 eq/ha/yr (the average of the colder and
warmer climate immobilization rates) for both forests in this case study. This approach and value
was also used by McNulty et al. (2007) for forests in the United States. Two values of the ^g;bb,
Final Risk and Exposure Assessment
Appendix 5-45
September 2009
-------
Terrestrial Acidification Case Study
300 and 3,000 m6/eq2, were used in the calculations of critical loads because the 300 m6/eq2
value is a widely used default value (UNECE, 2004; McNulty et al., 2007), and the 3,000 m6/eq2
value has been used to map critical loads in Canada (Ouimet et al., 2006; Watmough et al.,
2006). The 3,000 m6/eq2 constant is also the highest K^\^ value associated with soils with low
organic matter contents (UNECE, 2004). Fifty cm (0.5 m) was selected to represent the depth of
the rooting zone layer in this case study. Fine roots, which are responsible for the vast majority
of nutrient uptake, are typically concentrated in the upper 10 to 20 cm of soil (van der Salm and
de Vries, 2001). These roots are most susceptible to the impacts of acidification. Therefore, a 0.5
m depth has been suggested as a suitable rooting zone depth in the calculation of critical loads
for forest soils (Sverdrup and de Vries, 1994; Hodson and Langan, 1999).
As detailed in the preceding section, the three (Bc/Al)crit ratio values, associated with
three levels of forest protection, were used in the critical load calculations for this case study.
The 0.6, 1.2, and 10.0 (Bc/Al)crit ratios were applied to both the KEF and FffiEF case study areas.
As outlined earlier, base cation weathering rates were calculated using two methods; the
clay-substrate method and the soil type-texture association method. The data presented in
Tables 2.1-6 and 2.1-7 were used for these calculations. Similarly, base cation (Bcu) and
nitrogen (Nu) uptake values were calculated in two different ways for the two case study areas. In
HBEF, Bcu and Nu were assumed to be 0 eq/ha/yr because Watershed 6 is a reference watershed
and does not have a history or future of harvesting. Biomass (and the nutrients contained therein)
would, therefore, not have been removed from site. In KEF, two sets of values were used to
model two scenarios and estimate Bcu and Nu in the 8MB model calculations. In the first
scenario, it was assumed that the tree biomass was not harvested. Therefore, Nu and Bcw, in this
scenario, were set to 0 eq/ha/yr. In the second scenario, the case study plots were assumed to be
managed and harvested on a regular basis. Values of Bcu and Nu for this scenario were therefore
calculated using tree data (Tables 2.1-6, 2.1-8, and 2.1-9) and Equation 8 in Section 2.1.1.1. The
calculation of critical loads with the different Bcu and Nu values allowed for a comparison of the
influence of forest harvesting on the estimates of critical loads. The removal of nitrogen and base
cations with harvesting can significantly reduce the critical load of total nitrogen and sulfur
acidifying deposition in an ecosystem; the uptake and removal of base cations reduces the
capacity of the system to neutralize acidifying deposition.
Final Risk and Exposure Assessment September 2009
Appendix 5-46
-------
Terrestrial Acidification Case Study
Table 2.1-6. Input Values for the Calculation of Critical Load in Hubbard Brook Experimental
Forest and Kane Experimental Forest
DATA
2002 CMAQ/NADP total nitrogen and
sulfur deposition levels — wet and dry
(eq/ha)
Average annual (2000 to 2007) base
cation (Ca2+, Mg2+, Na+, K+)
deposition — wet (eq/ha)
Average annual (2000 to 2007) base
cation (Ca2+, Mg2+, Na+, K+)
deposition — dry (eq/ha)
Average annual (2000 to 2007)
chloride (Cl~) deposition — wet (eq/ha)
Average annual chloride (Cl~)
deposition — dry (eq/ha)
Runoff (m3/ha/yr)
Average annual soil temperature (to
0.5 m) (°C)
Rooting Zone Soil depth (m)
Percentage clay (in top 0.5 m of soil) a
Percentage sand (in top 0.5 m of soil) a
Percentage organic matter (in top 0.5 m
of soil)3
Gibbsite equilibrium constant (Kslbb)
(m6/eq2)
Parent material/bedrock13
Food and Agriculture Organization
(FAO) soil type
Nitrogen immobilization (N;)
(eq/ha/yr)
Denitrification (Nde) (eq/ha/yr)
Stand composition (in common name
alphabetical order)
CASE STUDY AREA
Hubbard Brook
Experimental Forest
Nitrogen = 60 1.07,
Sulfur = 233. 08 11
Ca2+ =15.87, Mg2+ = 5.69, K+ =
4.27, Na+ = 35.78
Ca2+ = 0.29, Mg2+ = 0.14,
K+ = 0.18,Na+ = 0.83
45.48
(2004 to 2007) 0.37
7,620
(1989 to 1998)7.29
0.5
6.4
57.4
3.6
300 and 3,000
Quartz, mica, schist, and
quartzite
Orthic Podzol (Po)
42.86
0
American Beech (Fagus
grandifolia), Balsam Fir (Abies
balsamea), Birch spp. (Betula
spp.), Mountain Ash (Sorbus
americana), Red Maple (Acer
rubrum), Red Spruce, Striped
Maple (Acer pensylvanicum)
and Sugar Maple
Kane Experimental Forest
Nitrogen = 967.54,
Sulfur = 646.37
Ca2+ = 33.03, Mg2+ = 7.96, K+ =
6.33, Na+= 18.31
Ca2+= 1.08,Mg2+ = 0.38,
K+ = 0.34, Na+ = 0.82
35.36
(2003 to 2007) 0.17
6,350
(1976 to 1979)7.90
0.5
See Table 2.1-7
See Table 2.1-7
See Table 2.1-7
300 and 3,000
Sandstone, conglomerate, and
shale
Dystric Cambisol (Bd)
42.86
0
American Beech (Fagus
grandifolia), Birch spp. (Betula
spp. ), Black cherry (Prunus
serotina), Cucumber Tree
(Magnolia acuminata), Eastern
Hemlock (Tsuga canadensis),
Red Maple (Acer rubrum), and
Sugar Maple
Final Risk and Exposure Assessment
Appendix 5-47
September 2009
-------
Terrestrial Acidification Case Study
DATA
Annual volume increment (AVI) by
species (m3/ha/yr)
Percentage allocation of growth to bark
by species
Specific gravity (bark and bole wood)
by species (g/cm3)
Nutrient concentration (% nitrogen,
Ca2+, Mg2+, K+) (bark and bole wood)
by tree species
Percentage biomass (bark and bole)
removal during harvest
CASE STUDY AREA
Hubbard Brook
Experimental Forest
—
—
—
—
Kane Experimental Forest
See Table 2.1-8
11% — coniferous species /
15% — deciduous species
See Table 2.1-9
See Table 2.1-9
65
' Determined by weighted average by horizon depth and soil series coverage
b Based on dominant mineralogy
Table 2.1-7. Soil Characteristics in the Seven Plots of the Kane Experimental Forest Case Study
Area for the Calculation of the Base Cation Weathering Rate Parameters
Soil Attribute (in
top 0.5 m of soil)
Percentage Clay
Percentage Sand
Percentage Organic
Matter
PLOT
1
23.3
39.9
1.4
2
20.1
27.0
0.7
3
21.3
26.2
1.1
4
20.1
27.0
0.7
5
23.3
39.9
1.4
6
20.1
27.0
0.7
7
20.1
27.0
0.7
Table 2.1-8. Annual Volume Growth by Tree Species in Each of the Seven Plots of the Kane
Experimental Forest Case Study Area for the Calculation of Nutrient Uptake (Bcu and Nu)
PLOT
1
2
3
4
5
6
7
TREE SPECIES ANNUAL VOLUME GROWTH (m3/ha/yr)
American
Beech
NA
NA
0.8
0.9
0.9
NA
0.6
Birch
spp.
NA
NA
0.2
NA
3.0
NA
0.1
Black
Cherry
3.3
3.0
1.3
NA
NA
1.8
0.8
Cucumber
Tree
NA
NA
NA
NA
0
NA
NA
Eastern
Hemlock
NA
NA
0.1
NA
NA
NA
0.1
Red
Maple
0.8
NA
0.3
1.1
NA
0.1
NA
Sugar
Maple
2.8
1.4
1.3
0.4
0.6
1.3
0.7
NA = Not applicable.
Final Risk and Exposure Assessment
Appendix 5-48
September 2009
-------
Terrestrial Acidification Case Study
Table 2.1-9. Specific Gravity and Nutrient Concentrations by Biomass Component (Bark and
Bole Wood) and by Tree Species for the Calculation of Nutrient Uptake (Bcu and Nu) in the Kane
Experimental Forest Case Study Area
SPECIES
American
Beech
Birch spp
Black
Cherry
Cucumber
Tree
Eastern
Hemlock
Red Maple
Sugar
Maple
BARK
Specific
Gravity
(g/cm3)
0.50
0.61
0.50
0.50
0.34
0.55
0.54
%N
0.75
0.40
0.00
0.54
0.27
0.43
0.51
«*
0.22
0.12
0.00
0.21
0.15
0.20
0.31
Mg°2+
0.05
0.04
0.00
0.06
0.03
0.05
0.06
o/ pa2+
/o ca
2.81
0.78
2.69
2.15
0.74
1.30
2.23
BOLE
Specific
Gravity
(g/cm3)
0.56
0.61
0.47
0.52
0.38
0.49
0.56
%N
0.11
0.09
0.00
0.10
0.08
0.09
0.10
«*
0.07
0.05
0.00
0.09
0.09
0.08
0.07
Mg°2+
0.02
0.02
0.00
0.02
0.01
0.02
0.02
Ca°2+
0.07
0.08
0.08
0.11
0.07
0.11
0.13
Note: N = nitrogen.
2.2 CRITICAL LOAD FUNCTION RESPONSE CURVES ASSOCIATED
WITH THE THREE LEVELS OF PROTECTION
The three (Bc/Al)crit ratio values (0.6, 1.2, and 10.0) used to evaluate level of protection
to tree health and growth correspond to three different critical load values of total nitrogen and
sulfur deposition ((N+S)COmb). A critical load based on 0.6 (Bc/Al)crit ratio is the least stringent
load and offers the least protection to forests of the three critical loads. The 1.2 (Bc/Al)crit ratio
results in an intermediate critical load of total nitrogen and sulfur deposition with moderate
protection to tree health. A critical load based on a 10.0 (Bc/Al)crit ratio would be the most
stringent load and offers the greatest protection to the health of trees.
As outlined in Section 2.1.1.3, critical loads of acidity can be translated into CLF
relationships; the CLF defines the combinations of (N+S)comb that are equal to the calculated
combined critical load of an ecosystem. Therefore, to provide an indication of all combinations
of total nitrogen and sulfur deposition associated with the three different levels of protection,
CLF response curves were produced with the three (Bc/Al)crit ratios. From the 0.6, 1.2, and 10.0
Final Risk and Exposure Assessment
Appendix 5-49
September 2009
-------
Terrestrial Acidification Case Study
(Bc/Al)crit ratios, the critical loads of (N+S)COmb and corresponding CLmax(N), CLm;n(N) and
CLmax(S) values were calculated to generate CLF response curves for each case study area.
Figure 2.2-1 provides an example of the CLF response curves associated with the three levels of
protection.
o
o
a
CLF based on lowest
protection level
(Bc/Al(crit)) = 0.6
CLF based on
intermediate protection
level (Bc/Al(crit)) = 1.2
CLF based on highest
protection level
(Bc/Al(crit)) = 10.0
N Deposition
Figure 2.2-1. An example of the critical load function response curves associated with
the three (Be/Al)crit ratios and the associated levels of protection of tree health. The flat
line portion of the curves indicates total nitrogen deposition corresponding to the CLm;n
(nitrogen absorbed by nitrogen sinks within the system).
3.0 RESULTS
3.1 CRITICAL LOAD ESTIMATES
3.1.1 Sugar Maple
The estimates of critical loads for sugar maple in the seven plots in KEF are presented in
Tables 3.1-1 through 3.1-7. The critical load estimates for all seven plots are summarized in
Table 3.1-8 and ranged from 728 eq/ha/yr to 2,998 eq/ha/yr of combined total nitrogen and
sulfur deposition ((N+S)COmb)- The ranges of critical loads associated with the three (Bc/Al)crit
ratios differed by level of impact to tree health, but there was some overlap between the ranges
of values. The lowest level of protection, (Be/Al)crit ratio = 0.6, had the highest critical loads that
ranged from 1,132 eq/ha/yr to 2,998 eq/ha/yr of (N+S)COmb. The intermediate level of protection,
(Be/Al)crit ratio = 1.2, had critical loads ranging from 1,033 eq/ha/yr to 2,079 eq/ha/yr. The
(Be/Al)crit ratio of 10.0, corresponding to the most protective level for tree health, had critical
Final Risk and Exposure Assessment
Appendix 5-50
September 2009
-------
Terrestrial Acidification Case Study
load values ranging from 728 eq/ha/yr to 1,139 eq/ha/yr of (N+S)COmb. Plot location, the two
different methods to estimate base cation weathering, the two K&bb values, and the inclusion of
the influence of nutrient uptake and removal (Bcu and Nu greater than 0 eq/ha/yr) in the critical
load calculations all influenced the critical loads for the three (Be/Al)crit ratio values. In general,
Plot 1 had both the lowest and highest critical load values. This was largely due to the method
used to calculate BCW (clay-substrate method) and the relative high percentages of clay and
organic matter in the soil, which accounted for the high critical load values. The low critical load
values in Plot 1 were caused by the comparatively high amount of base cation and nitrogen
uptake by the trees. The K&bb constant also influenced the critical load values, with 300 m6/eq2
causing higher critical loads than 3000 m6/eq2. Similarly, the clay-substrate method to estimate
BCW produced higher critical load values than did the soil type-texture approximation method.
The inclusion of the influence of nutrient uptake and removal by trees (Bcu and Nu greater than 0
eq/ha/yr) in the calculations of critical loads resulted in a large decrease in the critical values,
especially for the (Be/Al)crit ratio of 0.6, which offers the lowest level of protection to tree health.
Table 3.1-1. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant (Kgibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 1 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
2,998
2,079
1,139
1,940
1,473
960
tfgibb = 3000
m6/eq2
2,678
1,825
1,014
1,692
1,276
862
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
2,633
1,832
1,002
1,553
1,208
814
tfgibb = 3000
m6/eq2
2,328
1,591
883
1,332
1,033
728
Final Risk and Exposure Assessment
Appendix 5-51
September 2009
-------
Terrestrial Acidification Case Study
Table 3.1-2. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant C^gibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 2 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu and Nu
NOT
Included
Bcu and Nu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
2,711
1,885
1,032
2,009
1,481
910
tfgibb = 3000
m6/eq2
2,403
1,641
911
1,749
1,275
808
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
2,633
1,832
1,002
1,928
1,426
879
tfgibb = 3000
m6/eq2
2,328
1,591
883
1,673
1,223
779
Table 3.1-3. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant (Kgibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 3 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu andNu
NOT
Included
Bcu andNu
Included
Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr)
Clay-Substrate Method
Ksibb = 300
m6/eq2
2,817
1,957
1,071
2,228
1,626
983
tfgibb = 3,000
m6/eq2
2,504
1,709
949
1,954
1,408
875
Soil Type-Texture
Approximation Method
Ksibb = 300
m6/eq2
2,633
1,832
1,002
2,039
1,497
912
Agibb ~ 3,000
m6/eq2
2,328
1,591
883
1,776
1,288
809
Final Risk and Exposure Assessment
Appendix 5-52
September 2009
-------
Terrestrial Acidification Case Study
Table 3.1-4. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant C^gibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 4 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
2,711
1,885
1,032
2,388
1,707
990
tfgibb = 3,000
m6/eq2
2,403
1,641
911
2,101
1,479
878
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
2,633
1,832
1,002
2,308
1,653
960
tfgibb = 3,000
m6/eq2
2,328
1,591
883
2,025
1,429
850
Table 3.1-5. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant (Kgibb) and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 5 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu and Nu
NOT
Included
Bcu and Nu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr )
Clay-Substrate Method
Ksibb = 300
m6/eq2
2,998
2,079
1,139
2,545
1,849
1,118
tfgibb = 3,000
m6/eq2
2,678
1,825
1,014
2,256
1,620
1,004
Soil Type-Texture
Approximation Method
Agibb = 300
m6/eq2
2,633
1,832
1,002
2,173
1,596
978
Agibb ~ 3,000
m6/eq2
2,328
1,591
883
1,903
1,382
873
Final Risk and Exposure Assessment
Appendix 5-53
September 2009
-------
Terrestrial Acidification Case Study
Table 3.1-6. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant C^gibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 6 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr )
Clay-Substrate Method
tfgibb = 300
m6/eq2
2,711
1,885
1,032
2,212
1,599
946
tfgibb = 3,000
m6/eq2
2,403
1,641
911
1,936
1,380
838
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
2,633
1,832
1,002
2,132
1,545
916
tfgibb = 3,000
m6/eq2
2,328
1,591
883
1,861
1,329
810
Table 3.1-7. Critical Loads Calculated with the Different Base Cation Weathering, Gibbsite
Equilibrium Constant (Kgibb), and Base Cation (Bcu) and Nitrogen (Nu) Uptake Parameter Values
in Plot 7 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Load (eq/ha/yr )
Clay-Substrate Method
Ksibb = 300
m6/eq2
2,711
1,885
1,032
2,368
1,693
981
tfgibb = 3,000
m6/eq2
2,403
1,641
911
2,082
1,466
869
Soil Type-Texture
Approximation Method
Ksibb = 300
m6/eq2
2,633
1,832
1,002
2,289
1,639
951
Agibb ~ 3,000
m6/eq2
2,328
1,591
883
2,007
1,415
840
Final Risk and Exposure Assessment
Appendix 5-54
September 2009
-------
Terrestrial Acidification Case Study
Table 3.1-8. Ranges of Critical Load Values (eq/ha/yr) (with and without the Influence of
Nutrient Uptake and Removal with Tree Harvest) for the Seven Plots of the Kane Experimental
Forest Case Study Area (Both Kgibb values and methods to estimate BCW were used in these
calculations to present the range of critical loads estimated using all combinations of the
parameter values.)
Nutrient
Uptake in
Critical
Load
Bcu and
NUNOT
Included
Bcu and
Nu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Critical Loads (eq/ha/yr )
Plotl
2,328 to
2,998
1,591 to
2,079
883 to
1,139
1,332 to
1,940
1,033 to
1,473
728 to
960
Plot 2
2,328 to
2,711
1,591 to
1,885
883 to
1,032
1,673 to
2,009
1,223 to
1,481
779 to
910
Plot3
2,328 to
2,817
1,591 to
1,957
883 to
1,071
1,776 to
2,228
1,288 to
1,626
809 to
983
Plot 4
2,328 to
2,711
1,591 to
1,885
883 to
1,032
2,025 to
2,388
1,429 to
1,707
850 to
990
PlotS
2,328 to
2,998
1,591 to
2,079
883 to
1,139
1,903 to
2,545
1,3 82 to
1,849
873 to
1,118
Plot 6
2,328 to
2,711
1,591 to
1,885
883 to
1,032
1,861 to
2,212
1,329 to
1,599
8 10 to
946
Plot?
2,328 to
2,711
1,591 to
1,885
883 to
1,032
2,007 to
2,368
1,415 to
1,693
840 to
981
Two series of CLF response curves corresponding to the three (Be/Al)crit ratio values (i.e.,
0.6, 1.2, and 10.0) for Plot 1 of the KEF Case Study Area are shown in Figures 3.1-1 and 3.1-2.
Plot 1 had the highest and lowest critical load estimates. Therefore, the extreme values of this
plot were graphed to capture the range of critical loads associated with the three levels of
protection in the KEF Case Study Area. Figure 3.1-1 presents the CLF response curves
associated with the lowest critical load estimates. This scenario corresponded to the BCW
calculated using the soil type-texture approximation method, the 3000 m6/eq2 Xgibb constant, and
the inclusion of the influence of nutrient uptake and removal (Bcu and Nu greater than 0 eq/ha/yr)
in the calculation of critical load. Nitrogen uptake (Nu) substantially increased the minimum
critical load of nitrogen (CLm;n(N)) in this CLF relationship. Figure 3.1-2 shows the CLF
response curves associated with the highest critical load estimates in Plot 1. These estimates
occurred with BCW calculated using the clay-substrate method and the 300 m6/eq2 Ksibb constant.
Nutrient uptake (i.e., Bcu and Nu) was set to 0 eq/ha/yr in this calculation of critical load.
Final Risk and Exposure Assessment
Appendix 5-55
September 2009
-------
Terrestrial Acidification Case Study
3500 -i
a
It
o s;
CO
1026
727
422
n
c
*xxO^
, 306 728
- Low Protection (Bc/Al = 0.6)
- Intermediate Protection (Bc/Al =1.2)
- - - High Protection (Bc/Al = 10.0)
CLmin(JN)
V
1332 3
N Deposition
(eq/ha/yr)
5
00
Figure 3.1-1. The critical load function response curves detailing the lowest critical load
estimates for Plot 1 of the Kane Experimental Forest (refer to Table 3.1-1 for the
parameters corresponding to each of the curves). The flat line portion of the curves
indicates total nitrogen deposition corresponding to the CLm;n(N) (nitrogen absorbed by
nitrogen sinks within the system).
Final Risk and Exposure Assessment
Appendix 5-56
September 2009
-------
Terrestrial Acidification Case Study
3500
2955
2037
Q
t/3
1096
Low Protection (Bc/Al = 0.6)
~ Intermediate Protection (Bc/Al =1.2)
- - - High Protection (Bc/Al = 10.0)
--CLmin(N)
43
1139
2079
2998
3500
N Deposition
(eq/ha/yr)
Figure 3.1-2. The critical load function response curves detailing the highest critical load
estimates for Plot 1 of the Kane Experimental Forest (refer to Table 3.1-1 for the
parameters corresponding to each of the curves). The flat line portion of the curves
indicates total nitrogen deposition corresponding to the CLm;n(N) (nitrogen absorbed by
nitrogen sinks within the system.
The critical loads calculated for the KEF Case Study Area are consistent with critical
loads determined by other studies conducted on forests in the Allegheny Plateau. McNulty et al.
(2007), in their evaluation of critical loads across the United States, calculated loads of 1,061 to
1,146 eq/ha/yr , corresponding to the location of the seven case study plots in KEF (Table 3.1-9).
These values are very similar to the ranges (910 to 1,139 eq/ha/yr) determined in this case study,
using similar parameter values. McNulty et al. (2007) used the 8MB model to calculate critical
load, the clay-substrate method to estimate BCW and the indicator value of 10.0 for (Bc/Al)crit for
hardwood tree species. It is not known which K^ constant was used or if nutrient uptake and
removal (Bcu and Nu greater than 0 eq/ha/yr) was included in their calculations.
Final Risk and Exposure Assessment
Appendix 5-57
September 2009
-------
Terrestrial Acidification Case Study
Table 3.1-9. Comparison of the Critical Load Values Determined in This Case Study and the
Critical Load Values Determined by McNulty et al. (2007) for the Seven Plots in the Kane
Experimental Forest Case Study Area
Case Study
Plot
1
2
3
4
5
6
7
Critical Load (eq/ha/yr)
Case Study*
Nutrient Uptake (Bcu
and Nu) NOT Included
1,139
1,032
1,071
1,032
1,139
1,032
1,032
Nutrient Uptake (Bcu
and Nu) Included
960
910
983
990
1,118
946
981
McNulty et al. (2007)
1,146
1,144
1,061
1,061
1,061
1,061
1,064
The case study values in this table are those calculated with Kslbb = 300 m6/eq2 and (Bc/Al)cnt = 10.0,
and the clay-substrate method to estimate base cation weathering.
3.1.2 Red Spruce
The estimates of critical loads of acidity for red spruce in the HBEF Case Study Area are
presented in Table 3.1-10. The critical load estimates for this case study area were lower than
those for KEF, and ranged from 391 eq/ha/yr to 2,568 eq/ha/yr of combined total nitrogen and
sulfur deposition ((N+S)comb). Similar to the KEF Case Study Area, the ranges of critical loads
associated with the three (Be/Al)crit ratios differed by level of protection to tree health. The least
stringent, least protective level, (Be/Al)crit ratio = 0.6, had the highest critical loads that ranged
from 991 eq/ha/yr to 2,568 eq/ha/yr of (N+S)COmb The intermediate level of protection, (Bc/Al)crit
ratio = 1.2, had critical loads ranging from 697 eq/ha/yr to 1,801 eq/ha/yr. The (Be/Al)crit ratio of
10.0, corresponding to the most stringent, most protective level for tree protection, had critical
load values ranging from 391 eq/ha/yr to 987 eq/ha/yr of (N+S)comb. The Kg{b\, and method to
calculate BCW also influenced the critical load estimates for the HBEF Case Study Area, with the
Xgibb value of 300 m6/eq2 resulting in higher critical load values than 3000 m6/eq2. In contrast,
the soil type-texture approximation method to estimate BCW caused higher critical load values in
the FffiEF Case Study Area. These trends in the results were largely due to the relatively low
clay and organic matter concentrations in the soils in FffiEF Case Study Area compared to the
Final Risk and Exposure Assessment
Appendix 5-58
September 2009
-------
Terrestrial Acidification Case Study
KEF Case Study Area; this lower clay content resulted in much lower BCW rates using the clay-
substrate compared to soil type-texture method.
Table 3.1-10. Critical Load Calculated with the Different Base Cation Weathering and Gibbsite
Equilibrium Constant (Kgibb) Parameter Values in the Hubbard Brook Experimental Forest Case
Study Area
(Bc/Al)crit Ratio
0.6
1.2
10.0
Critical Load (eq/ha/yr)
Clay-Substrate Method
^gibb = 300 m
eq2
1,237
892
487
tfgibb = 3,000
m6/eq2
991
697
391
Soil Type-Texture Approximation
Method
tfgibb = 300
m6/eq2
2,568
1,801
987
tfgibb = 3,000
m6/eq2
2,232
1,534
856
Two series of CLF response curves that indicate the combined total nitrogen and sulfur
deposition ((N+S)comb) levels for the three (Be/Al)crit ratios (i.e., 0.6, 1.2, and 10.0) for the HBEF
Case Study Area are shown in Figures 3.1-3 and 3.1-4. These two sets of critical load estimates
were selected to provide an indication of the range of critical loads associated with the three
levels of protection for red spruce health in the HBEF Case Study Area.
Figure 3.1-3 shows the CLF response curves corresponding to the lowest critical loads.
This scenario occurred with the BCW calculated using the clay-substrate method and the 300
m6/eq2 Ksibb constant. Figure 3.1-4 shows the CLF response curves associated with the highest
critical load estimates in the HBEF Case Study Area. These estimates were calculated with BCW
estimated using the soil type-texture approximation method and the 3,000 m6/eq2 ^g;bb constant.
Final Risk and Exposure Assessment
Appendix 5-59
September 2009
-------
Terrestrial Acidification Case Study
onnn
£H
.2 ^
If
a =S
U a-
Q
00
948
654
348
(
Low Protection (Bc/Al = 0.6)
Intermediate Protection (Bc/Al = 1.2)
- - - High Protection (Bc/Al = 10.0)
CLrmn(JN)
*s ^^^^
o>\
, 43 391 697 991 y
N Deposition
(eq/ha/yr)
00
Figure 3.1-3. The critical load function response curves detailing the lowest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area (refer to Table
3.1-10 for the parameters corresponding to each of the curves). The flat line portion of
the curves indicates total nitrogen deposition corresponding to the CLm;n(N) (nitrogen
absorbed by nitrogen sinks within the system).
3000
2525
3 £
11
JTT ^^.
944
Low Protection (Bc/Al = 0.6)
Intermediate Protection (Bc/Al =1.2)
High Protection (Bc/Al =10.0)
CLmin(N)
043
987
1801
2568
3000
N Deposition
(eq/ha/yr)
Figure 3.1-4. The critical load function response curves detailing the highest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area (refer to Table
3.1-10 for the parameters corresponding to each of the curves). The flat line portion of
the curves indicates total nitrogen deposition corresponding to the CLm;n(N) (nitrogen
absorbed by nitrogen sinks within the system).
Final Risk and Exposure Assessment
Appendix 5-60
September 2009
-------
Terrestrial Acidification Case Study
The HBEF Case Study Area has been the site of extensive research, and several studies
have estimated critical loads of acidity for the experimental forest (Table 3.1-11). Using the
8MB model and parameter values similar or equivalent to those used in this case study, McNulty
et al. (2007) calculated a critical load of 516 eq/ha/yr for the location of the case study plot. The
NEG/ECP Forest Mapping Group (2005) that conducted a detailed assessment of critical loads in
several of the northeastern states and the eastern provinces of Canada determined critical load
values around 1,500 eq/ha/yr for the area of the case study plot in HBEF (NEG/ECP Forest
Mapping Group, 2005). This group used a modified version of the 8MB model to estimate
critical loads. Pardo and Driscoll (1996) also calculated critical loads for the HBEF. They
evaluated four charge- and mass-balance models (steady-state water chemistry, nitrogen mass
balance, base cation mass balance, and steady-state mass balance models) over three different
time periods (between 1965 and 1988) and found values ranging from -433 to 1,452 eq/ha/yr
(negative values are equal to 0 eq/ha/yr). Therefore, the critical load values of 391 to 2,181
eq/ha/yr calculated in this case study are consistent with those from earlier studies that used
similar and different methods or models to estimate critical loads. The higher estimates (i.e.,
2,568 eq/ha/yr) in this case study can be attributed to the soil type-texture approximation method
of calculating BCW and the lowest (Bc/Al)crit (0.6) value.
Table 3.1-11. Summary of the Critical Load Values Determined by Other Studies Conducted in
the Hubbard Brook Experimental Forest (negative values are equal to 0 eq/ha/yr)
Critical Load Model
Used for Estimation
Simple Mass Balance
Steady- State Water
Chemistry
Nitrogen Mass Balance
Basic Cation Mass
Balance
Modified Basic Cation
Mass Balance
Steady-State Balance
Critical Load Values for Each Study (eq/ha/yr)
Pardo and
Driscoll (1996)
-
-4 to 1 1
133 to 1,452
62 to 133
Oto 1,405
-433 to 630
McNulty et
al. (2007)
516
-
-
-
-
-
NEG/ECP Forest
Mapping Group
(2005)
-1,500
-
-
-
-
-
Case Study
391 to 2,568
-
-
-
-
-
Source: Pardo and Driscoll, 1996.
Final Risk and Exposure Assessment
Appendix 5-61
September 2009
-------
Terrestrial Acidification Case Study
3.2 RECOMMENDED PARAMETER VALUES AND CRITICAL LOADS
Within the ranges of critical loads estimated for the KEF and HBEF case study areas,
three critical loads were selected to represent the conditions associated with the three levels of
protection (Bc/Al(crit) = 0.6, 1.2, and 10.0) for sugar maple in KEF and for red spruce in FffiEF
(Table 3.2-1). For the KEF Case Study Area, these critical load values, in order of lowest to
highest level of protection were 2,009, 1,481 and 910 eq/ha/yr (for Bc/Al(crit)= 0.6, 1.2, and 10.0,
respectively). For the HBEF Case Study Area, these values, in order of lowest to highest level of
protection, were 1,237, 892, and 487 eq/ha/yr (for Bc/Al^t) = 0.6, 1.2, and 10.0, respectively).
These critical load estimates were derived using the clay-substrate method to estimate
BCW and a K^,b of 300 m6/eq2. For the KEF Case Study Area, nutrient uptake and removal with
tree harvest (Bcu and Nu) was also included in the critical load estimates, and within the
constraints of the selected parameters, the plot with the most conservative (i.e., lowest critical
load) was selected to represent the full KEF Case Study Area. The parameter values were set to 0
eq/ha/yr in the HBEF Case Study Area because the study plots in this experimental forest are not
actively managed or harvested. The selection of these parameters and methods was based on the
best available recommendations of scientists and research efforts, to date. When field
assessments and measurements are not possible, the clay-substrate method is one of the most
commonly used methods to estimate base cation weathering in North America (Ouimet et al.,
2006; Watmough et al., 2006; McNulty et al., 2007; Pardo and Duarte 2007), and the 300 m6/eq2
value of the K^,b is a recommended default value (UNECE, 2004). For the KEF Case Study
Area, the influence of nutrient uptake and removal (i.e., Bcu and Nu greater than 0 eq/ha/yr) was
included because the forest has been and will likely continue to be actively harvested (USFS,
1999).
Table 3.2-1. Critical Loads Selected to Represent the Three Levels of Protection in the Kane
Experimental Forest and Hubbard Brook Experimental Forest Case Study Areas
Protection Level (Bc/Al(crit) ratio)
Low (Bc/Al(cnt) = 0.6)
Medium (Bc/Al(cnt) = 1.2)
High (Bc/Al(cnt)= 10.0)
Critical Load Values for Each Case Study Area (eq/ha/yr)
KEF
2,009
1,481
910
HBEF
1,237
892
487
Final Risk and Exposure Assessment
Appendix 5-62
September 2009
-------
Terrestrial Acidification Case Study
3.3 CURRENT CONDITIONS
This section discusses the impact of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels relative to the critical loads estimated for the KEF and HBEF case study areas.
The atmospheric deposition of total nitrogen and sulfur ((N+S)comb) in both the HBEF and KEF
case study areas was elevated. According to 2002 CMAQ output, the KEF Case Study Area
received 13.6 kilograms (kg) N/ha (967.5 eq/ha) and 20.7 kg S/ha (646.4 eq/ha), and the HBEF
Case Study Area experienced 8.4 kg N/ha (601.1 eq/ha) and 7.5 kg S/ha (233.1 eq/ha). When
these deposition levels were compared to the critical loads calculated using the three (Bc/Al)crit
ratio values, the CMAQ-modeled (N+S)COmb deposition loads were found to be both greater than
and less than the three critical loads for the two case study areas (Tables 3.3-1 to 3.3-9, Figures
3.3-1 to 3.3-4) . In all plots of the KEF Case Study Area, the 2002 CMAQ/NADP total nitrogen
and sulfur deposition levels ((N+S)COmb) were greater than the range of total nitrogen and sulfur
allowable for the most stringent critical load (where (Bc/Al)crit= 10.0). Similarly, in the HBEF
Case Study Area, the modeled (N+S)comb deposition was greater than the critical loads estimated
using the (Bc/Al)crit = 10.0 ratio and the clay-substrate method to estimate BCW,. However,
combined 2002 CMAQ/NADP total nitrogen and sulfur deposition levels were less than the
critical loads estimated with (Bc/Al)crit= 10.0 and BCW determined by the soil type-texture
approximation. The 2002 CMAQ/NADP total nitrogen and sulfur deposition levels ((N+S)COmb)
were less than the ranges of total nitrogen and sulfur allowable for the least stringent critical load
(where (Bc/Al)crit= 0.6) for both the HBEF Case Study Area and all plots of the KEF Case Study
Area. The only exception to this trend was in Plot 1 of the KEF Case Study Area, where the
removal of base cations and nitrogen with harvesting (Bcu and Nu greater than 0 eq/ha/yr) were
included in the load calculations. The variability in the results comparing 2002 CMAQ/NADP
total nitrogen and sulfur deposition levels to calculated acid loads shows the strong influence of
the Bc/Al(crit) indicator ratio (reflecting the level of tree protection) in critical load estimates.
Final Risk and Exposure Assessment September 2009
Appendix 5-63
-------
Terrestrial Acidification Case Study
Table 3.3-1. Ranges of Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur
Deposition Levels ((N+S)COmb) and the Estimated Critical Load Values (with and without the
Influence of Nutrient Uptake and Removal [Nu and Bcu]) for the Seven Plots of the Kane
Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and Estimated
Critical Loads (eq/ha/yr)
Plotl
-1,384
to -7 14
-465 to
23*
475* to
731*
-327 to
282*
141* to
581*
654* to
886*
Plot 2
-1,098
to -7 14
-271 to
23*
582* to
731*
-395 to
-59
132* to
391*
704* to
835*
Plot3
-1,203
to -7 14
-343 to
23*
543* to
731*
-6 14 to
-162
-12 to
326*
631* to
805*
Plot 4
-1,098
to -7 14
-271 to
23*
582* to
731*
-774 to
-411
-93 to
185*
624* to
764*
PlotS
-1,384
to -7 14
-465 to
23*
475* to
731*
-931 to
-289
-235 to
231*
496* to
741*
Plot 6
-1,098
to -7 14
-271 to
23*
582* to
731*
-598 to
-247
15* to
285*
668* to
804*
Plot?
1,098 to
-714
-271 to
23*
582* to
731*
-755 to
-393
-79 to
199*
633* to
774*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)comb deposition.
Table 3.3-2. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 1 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Deposition Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
-1,384
-465
475*
-327
141*
654*
tfgibb = 3,000
-1,064
-211
600*
-78
338*
751*
Soil Type-Texture
Approximation Method
tfgibb = 300
-1,019
-218
612*
61*
406*
800*
tfgibb = 3,000
-714
23*
731*
282*
581*
886*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)comb deposition.
Final Risk and Exposure Assessment
Appendix 5-64
September 2009
-------
Terrestrial Acidification Case Study
Table 3.3-3. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 2 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads ( eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
-1,098
-271
582*
-395
132*
704*
tfgibb = 3,000
-789
-27
703*
-135
339*
806*
Soil Type-Texture
Approximation Method
tfgibb = 300
-1,019
-218
612*
-314
188*
735*
tfgibb = 3,000
-714
23*
731*
-59
391*
835*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)comb deposition.
Table 3.3-4. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 3 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
-1,203
-343
543*
-614
-12
631*
tfgibb = 3,000
m6/eq2
-890
-95
665*
-340
206*
738*
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
-1,019
-218
612*
-425
117*
702*
tfgibb = 3,000
m6/eq2
-714
23*
731*
-162
326*
805*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)COmb deposition.
Final Risk and Exposure Assessment
Appendix 5-65
September 2009
-------
Terrestrial Acidification Case Study
Table 3.3-5. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 4 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
-1,098
-271
582*
-774
-93
624*
tfgibb = 3,000
m6/eq2
-789
-27
703*
-487
134*
736*
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
-1,019
-218
612*
-694
-39
654*
^gibb ~ 3,000
m6/eq2
-714
23*
731*
-411
185*
764*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)COmb deposition.
Table 3.3-6. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 5 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
-1,384
-465
475*
-931
-235
496*
tfgibb = 3,000
m6/eq2
-1,064
-211
600*
-642
-6
609*
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
-1,019
-218
612*
-559
18*
636*
tfgibb = 3,000
m6/eq2
-714
23*
731*
-289
231*
741*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)COmb deposition.
Final Risk and Exposure Assessment
Appendix 5-66
September 2009
-------
Terrestrial Acidification Case Study
Table 3.3-7. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 6 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
-1,098
-271
582*
-598
15*
668*
tfgibb = 3,000
m6/eq2
-789
-27
703*
-322
234*
776*
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
-1,019
-218
612*
-518
69*
698*
tfgibb = 3,000
m6/eq2
-714
23*
731*
-247
285*
804*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)COmb deposition.
Table 3.3-8. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods, Two Gibbsite Equilibrium Constants [£gibb], and Two Base Cation (Bcu) and Nitrogen
(Nu) Uptake Parameter Values) in Plot 7 of the Kane Experimental Forest Case Study Area
Nutrient
Uptake in
Critical
Load
Estimate
Bcu andNu
NOT
Included
Bcu andNu
Included
(Bc/Al)crit
Ratio
0.6
1.2
10.0
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and
Estimated Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300
m6/eq2
-1,098
-271
582*
-755
-79
633*
tfgibb = 3,000
m6/eq2
-789
-27
703*
-469
148*
745*
Soil Type-Texture
Approximation Method
tfgibb = 300
m6/eq2
-1,019
-218
612*
-675
-25
663*
^gibb ~ 3,000
m6/eq2
-714
23*
731*
-393
199*
774*
* Indicates a positive value or deposition greater than the critical load based on CMAQ-modeled 2002
(N+S)COmb deposition.
Final Risk and Exposure Assessment
Appendix 5-67
September 2009
-------
Terrestrial Acidification Case Study
Table 3.3-9. Differences between the 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition
Levels ((N+S)comb) and the Critical Load Values (with Two Base Cation Weathering Estimation
Methods and Two Gibbsite Equilibrium Constants [^gibb]) in the Hubbard Brook Experimental
Forest Case Study Area
(Bc/Al)crit Ratio
0.6
1.2
10.0
Difference between CMAQ/NADP (N+S)comb Deposition and Estimated
Critical Loads (eq/ha/yr)
Clay-Substrate Method
tfgibb = 300 m6
eq2
-403
-58
347*
tfgibb = 3,000
m6/eq2
-157
137*
443*
Soil Type-Texture Approximation
Method
tfgibb = 300
m6/eq2
-1,734
-967
-153
tfgibb = 3,000
m6/eq2
-1,398
-700
-21
3500 -i
|1
Q
-------
Terrestrial Acidification Case Study
3500
2955
2037
a!
1096
Low Protection (Bc/Al = 0.6)
— Intermediate Protection (Bc/Al =1.2)
- - High Protection (Bc/Al = 10.0)
--CLmin(N)
• 2002 CMAQ N and S Deposition
43
1139
2079
2998
3500
N Deposition
(eq/ha/yr)
Figure 3.3-2. Plot 1 critical load function response curves, detailing the highest
maximum deposition load estimates for Kane Experimental Forest Case Study Area
(refer to Table 3.1-1 for the parameters corresponding to each of the curves). The 2002
CMAQ/NADP total nitrogen and sulfur deposition levels ((N+S)COmb) were greater than
the critical load of total nitrogen and sulfur deposition calculated with the highest level of
protection (Bc/Al)crit= 10.0). The flat line portion of the curves indicates total nitrogen
deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the
system).
Final Risk and Exposure Assessment
Appendix 5-69
September 2009
-------
Terrestrial Acidification Case Study
3000 -|
£H
I?
M ;s
0 jf
a =5
-------
Terrestrial Acidification Case Study
u
Q
3000
2525
1758
944
Low Protection (Bc/Al = 0.6)
Intermediate Protection (Bc/Al = 1.2)
- - - High Protection (Bc/Al = 10.0)
--CLmin(N)
• 2002 CMAQ N and S Deposition
043
987
1801
2568
3000
N Deposition
(eq/ha/yr)
Figure 3.3-4. Critical load function response curves, detailing the highest critical load
estimates for the Hubbard Brook Experimental Forest Case Study Area (refer to Table
3.1-10 for the parameters corresponding to each of the curves). The 2002 CMAQ/NADP
total nitrogen and sulfur deposition levels ((N+S)COmb) were less than the critical loads
associated with all three (Be/Al)crit ratios. The flat line portion of the curves indicates total
nitrogen deposition corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen sinks
within the system).
As outlined in Section 3.2, critical loads of 2,009, 1,481 and 910 eq/ha/yr were selected
to represent the three levels of increasing protection for the KEF Case Study Area, and 1,237,
892 and 487 eq/ha/yr were the critical loads selected for the HBEF Case Study Area. These
estimates are based on the critical load parameters suggested and most frequently used by
scientists and previous research. When compared to the 2002 CMAQ-modeled deposition levels,
it was evident that the deposition levels were greater than the most protective critical load
(Bc/Al(crit) = 10.0) for both case study areas and also greater than the intermediate protection
critical load (Bc/Al(CIit) = 1.2) for KEF (Figures 3.3-5 and 3.3-6)). In these comparisons, 2002
CMAQ/NADP total nitrogen and sulfur deposition levels exceeded the KEF Case Study Area
critical load by 132 to 704 eq/ha/yr and exceeded the HBEF Case Study Area's critical load by
347 eq/ha/yr. Similar results have been reported in other studies which have assessed the two
case study areas. McNulty et al. (2007) and Pardo and Driscoll (1996) found that deposition
levels were greater than the estimated critical loads in the FffiEF area. McNulty et al. (2007) also
Final Risk and Exposure Assessment
Appendix 5-71
September 2009
-------
Terrestrial Acidification Case Study
reported that 2002 CMAQ/NADP total nitrogen and sulfur deposition levels in the KEF
exceeded the calculated critical loads for the KEF Case Study Area. These results suggest that
the health of red spruce at FffiEF and sugar maple at KEF may have been compromised by the
acidifying nitrogen and sulfur deposition received in 2002.
3500
o
CX
X
Q
1796
1268
697
0
Low Protection (Bc/Al = 0.6)
Intermediate Protection (Bc/Al =1.2)
- - - - High Protection (Bc/Al = 10.0)
CLmin(N)
• 2002 CMAQ N and S Deposition
0 213
910
1481
2009
3500
N Deposition
(eq/ha/yr)
Figure 3.3-5. Critical load function response curves for the three selected critical loads
conditions (corresponding to the three levels of protection) for the Kane Experimental
Forest Case Study Area. The 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels ((N+S)comb) were greater than the highest and intermediate level of protection
critical loads. The flat line portion of the curves indicates total nitrogen deposition
corresponding to the CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system).
Final Risk and Exposure Assessment
Appendix 5-72
September 2009
-------
Terrestrial Acidification Case Study
3000 -i
S3
.9 ^
.-a >^
° 2
ex =5
-------
Terrestrial Acidification Case Study
curves for both areas. For both case studies, the maximum sulfur critical load (CLmax(S)) and the
maximum nitrogen critical load (CLmax(N)), as NOX, were lowered.. In the calculations for the
KEF Case Study Area, the CLmax(S) was reduced by 5% to 661 eq/ha/yr, and in the HBEF Case
Study Area calculations, the CLmax(S) was reduced by 26% to 328 eq/ha/yr. Similarly, the
CLmax(N) (as NOX) for KEF was reduced by 27% to 661 eq/ha/yr, and the CLmax(N) (as NOX) for
HBEF was reduced by 33% to 328 eq/ha/yr.
1°00
Is
£ ^ CLmax(S) 697
u cr
Q £,
m
"^•i.
O NHX-N deposition (fixed amount)
H NOX-N deposition
• 2002 CMAQ N and S deposition
• • CLmm(N)
~l i i
0
CLmin(N)
N Deposition
(eq/ha/yr)
•
_U
• _ 1200
Figure 3.3-7. The influence of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels (NHX-N) on the critical load function response curve, and in turn, the
maximum critical loads of sulfur (CLmax(S)) and oxidized nitrogen (NOX-N) for the
selected highest protection critical load for the Kane Experimental Forest Case Study
Area. The critical load of oxidized nitrogen (NOX-N) is 661 eq/ha/yr (910 - 249
eq/ha/yr). The CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system)
corresponds to the value depicted in Figure 3.3-5.
Final Risk and Exposure Assessment
Appendix 5-74
September 2009
-------
Terrestrial Acidification Case Study
y^n
.2 ^ CLmax(S) 444 -
'« ^S
O 2
CX ^
Q ^
G NHX-N deposition (fixed amount)
H NOX-N deposition
• 2002 CMAQ N and S deposition
' ' CLmm(N)
~\
0
CLmin(N)
'
N Deposition
(eq/ha/yr)
•
f""v
I
487 750
Figure 3.3-8. The influence of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels (NHX-N) on the critical load function response curve and, in turn, the
maximum critical loads of sulfur (CLmax(S)) and oxidized nitrogen (NOX-N) for the
selected highest protection critical load for the Hubbard Brook Experimental Forest Case
Study Area. The critical load of oxidized nitrogen (NOX-N) is 328 eq/ha/yr (487 - 159
eq/ha/yr). The CLm;n(N) (nitrogen absorbed by nitrogen sinks within the system)
corresponds to the value depicted in Figure 3.3-6.
4.0 EXPANSION OF CRITICAL LOAD ASSESSMENTS FOR
SUGAR MAPLE AND RED SPRUCE
4.1 CRITICAL LOAD ASSESSMENTS
Although the KEF and HBEF case studies estimated critical loads for red spruce and
sugar maple in two locations and established that the 2002 CMAQ/NADP total nitrogen and
sulfur deposition levels were greater than the calculated loads, these results cannot be
extrapolated directly to represent the critical load condition for the full distribution ranges of the
two tree species. Critical loads are largely determined by soil characteristics, and these
characteristics vary by location. Therefore, to gain an understanding of the range of critical load
Final Risk and Exposure Assessment
Appendix 5-75
September 2009
-------
Terrestrial Acidification Case Study
values experienced by sugar maple and red spruce, it is necessary to calculate critical loads in
multiple locations throughout the ranges of the two species.
Critical load calculations were applied to multiple locations within 24 states for sugar
maple and in 8 states for red spruce. Individual site locations within each State were determined
by the USFS FIA database permanent sampling plots' locations on forestland7 (timberland8 for
New York, Arkansas, Kentucky and North Carolina), each covering 0.07 ha. Only database
information for nonunique9, permanent sampling plots that supported the growth of sugar maple
or red spruce and had the necessary soil, parent material, atmospheric deposition, and runoff data
were included in the analyses. With these restrictions, 4,992 of the 14,669 sugar maple plots and
763 of the 2,875 red spruce plots were included in the calculations of the plot-specific critical
loads (Table 4.1-1). Although only a subset of the total sugar maple and red spruce plots were
included in the analyses, the results are thought to capture accurately the range and trends of
critical loads of the two species. Due to the randomness of the plot restrictions, it is unlikely that
a bias was incorporated into the analyses.
Table 4.1-1. Number and Location of USFS FIA Permanent Sampling Plots (each plot is 0.07
ha) Used in the Analysis of Critical Loads For the Full Geographic Ranges of Sugar Maple and
Red Spruce
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kansas
Sugar Maple
13
10
35
29
306
13
NA
Red Spruce
-
-
-
-
-
-
-
7 Forestland is defined as, "land at least 10 percent stocked by forest trees of any size, or formerly having such tree
cover, and not currently developed for non-forest uses, with a minimum area classification of 1 acre." (USFS,
2002a).
8 Timberland is defined as, "forest land capable of producing in excess of 20 cubic feet per acre per year and not
legally withdrawn from timber production, with a minimum area classification of 1 acre." (USFS, 2002b).
9 Nonunique permanent sampling plot locations are those that have critical load attribute values (e.g., soils, runoff,
and atmospheric deposition) that are not distinct and are repeated within a 250-acre area of the plot location. This
"confidentiality" filter is a requirement of the USFS to prevent the disclosure of data that can be directly linked to
a location on private land. To comply with the necessary "confidentiality," full coverages of the data required for
the critical load deposition calculations were given to the USFS, and the USFS matched and provided the data to
each nonunique, permanent sampling plot.
Final Risk and Exposure Assessment
Appendix 5-76
September 2009
-------
Terrestrial Acidification Case Study
State
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Rhode Island
South Carolina
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
TOTAL
Sugar Maple
14
271
4
33
633
289
147
82
6
485
17
374
285
NA
NA
319
114
175
378
960
4,992
Red Spruce
-
560
-
3
-
-
-
55
-
52
1
-
NA
-
-
1
11
NA
7
-
763
Note: NA = data not available for State; "-" = tree species not present on
forestland in State.
The parameter values selected for the 8MB calculations of critical loads for all plots
within the ranges of sugar maple and red spruce included wet deposition of base cations (Na+,
Ca+2, Mg+2, and K+) and chlorine, the clay-substrate method to estimate BCW, three levels of
protection ((Bc/Al)crit ratio = 0.6, 1.2 and 10.0), a 0.5 m rooting zone soil depth, and the N;
(42.86 eq/ha/yr) and Nde (0 eq/ha/yr) default values used for the HBEF and KEF case study
areas. TheKgibb constant ranged from 100 to 950 m6/eq2, and was determined by average organic
matter content, as outlined by McNulty et al. (2007) (Table 4.1-2). Nutrient (Nu) and base cation
(BCU) uptake were not included in the 8MB calculations because it was not possible to determine
the harvesting status of the individual sampling plots. Corrections for sea salt influence were not
applied to the wet deposition because such corrections were found to over-correct deposition
estimates (McNulty et al., 2007).
Final Risk and Exposure Assessment
Appendix 5-77
September 2009
-------
Terrestrial Acidification Case Study
Similar to the KEF and HBEF case studies, the U.S. Department of Agriculture- Natural
Resources Conservation Service (USDA-NRCS) SSURGO soils database (USDA-NRCS,
2008c) was the main source used to estimate BCW in the calculations of critical loads for the full
distribution ranges of sugar maple and red spruce. The U.S. Geological Survey (USGS) state-
level integrated map database for the United States (USGS, 2009b) was used as a secondary
source of information, when necessary. Parent material acidity was inferred from the parent
material attribute in the SSURGO soils database. The estimated total silica and ferromagnesium
content, relative to the mineral assemblage typical of the rock or sediment type, were used to
classify parent material as acidic, intermediate, or basic, according to the classification table
(Table 4.1-3) outlined by McNulty et al. (2007) from Gray and Murphy (1999). When possible,
classification of the parent material silica content was determined by the range of rock types
provided as examples in Table 4.1-3. When rock types were not clearly indicated in the parent
material attribute, parent material acidity was classified using a systematic protocol involving the
consideration of descriptive modifiers that suggest a probable range of silica or ferromagnesium
content (Table 4.1-4). In cases where the parent material attribute in the SSURGO soils database
was not populated or was too nondescriptive to classify, acidity rating of parent material was
inferred from the USGS state-level spatial geology databases. The criteria applied to the soils
data were also used in interpretation of these USGS spatial datasets, along with general
observation related to spatial patterns of local and regional geologic settings that suggested
characteristics of igneous and metamorphic petrogenesis and implied sedimentary deposit!onal
mechanisms and environments. If parent material acidity was classified as "organic" or "other"
or could not be determined by either the SSURGO or USGS geology databases, the critical load
was not estimated for the location. The BCW values in the critical load assessments were not
corrected for temperature because the soil temperature attribute in the SSURGO soils database
was missing data for most of the plot locations. Average air temperature was not used as a
substitute because McNulty et al. (2007) determined that corrections for air temperatures were
more suitable for northern climates, presumably where the temperature corrections were derived
(i.e., Scandinavia).
Final Risk and Exposure Assessment September 2009
Appendix 5-78
-------
Terrestrial Acidification Case Study
Table 4.1-2. Gibbsite Equilibrium (Kgibb) Determined by Percentage of Soil Organic Matter
Soil Type Layer
Mineral soils: C layer
Soils with low organic matter: B/C layers
Soils with some organic material: A/E layers
Peaty and organic soils: organic layers
Organic
Matter %
<5
5tol5
15 to 70
>70
Kgibb (m6/eq2)
950
300
100
9.5
Source: Modification of table from McNulty et al. (2007).
Table 4.1-3. Parent Material Acidity Classifications for Base Cation (BCW) Estimations
Parent
Material
Classification
Acidic
Intermediate
Basic
Parent
Material
Category
Extremely
siliceous
Highly
siliceous
Transitional
siliceous/
Intermediate
Intermediate
Mafic
Ultramafic
Calcareous
Silica
Content
>90%
72% to
90%
62% to
72%
52% to
62%
45% to
52%
<45%
Low
Calcium-
Ferromagnesium
Content
Extremely low
(generally <3%)
Low
(generally 3% to
7%)
Moderately low
(generally 7% to
14%)
Moderate
(generally 14% to
20%)
(generally 20% to
30%)
Very high
(generally >30%)
CaCO3 dominate
other bases variable
Examples
Quartz sands (i.e., beach,
alluvial, or Aeolian), chert,
quartzite, quartz reefs, and
silicified rocks
Granite, rhyolite, adamellite,
dellenite, quartz sandstone,
and siliceous tuff
Granodiorite, dacite,
trachyte, syenite, most
greywacke, most lithic
sandstone, most argillaceous
rocks, and
siliceous/intermediate tuff
Monzonite, trachy-andesite,
diorite, andesite,
intermediate tuff, as well as
some greywacke, lithic
sandstone, and argillaceous
rock
Gabbro, dolerite, basalt, and
mafic tuff (uncommon)
Serpentinite, dunite,
peridotite, amphibolite, and
tremolite-chlorite-talc schists
Limestone, dolomite,
calcareous shale, and
calcareous sands
Final Risk and Exposure Assessment
Appendix 5-79
September 2009
-------
Terrestrial Acidification Case Study
Parent
Material
Classification
Organic
Other
Parent
Material
Category
Organic
Alluvial
Sesquioxide
Silica
Content
Low
Variable*
Variable*
Calcium-
Ferromagnesium
Content
Organic matter
dominates bases
variable
Variable
Variable,
dominated by
sesquioxides
Examples
Peat, coal, and humified
vegetative matter
Variable
Laterite, bauxite, ferruginous
sandstone, and ironstone
Category not defined by silica content
Source: Modified from McNulty et al. (2007).
Table 4.1-4. Parent Material and Descriptive Modifier Characteristics (within the SSURGO
Soils [USDA-NRCS, 2008c] and USGS Geology [USGS, 2009b] Databases) Used to Classify
Parent Material Acidity
Parent Material or
Modifier Characteristic
Glacial deposits, till,
outwash, or drift (without
modifiers)
Glacial lacustrine (without
modifiers)
Glaciomarine (without
modifiers)
Marine deposits (without
modifiers)
Loess and eolian (without
modifiers)
Alluvium (without
modifiers)
Residuum (without
modifier)
Colluvium (without
modifiers)
Deposits, till, or outwash
(without modifiers)
Saprolite (without
modifiers)
Acidity Classification
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Acidic
Not able to classify
Intermediate
Not able to classify
Not able to classify
Rational
Probable mixture of different mineralogies
Probable mixture of different mineralogies
Probable mixture of different mineralogies
Probable mixture of different mineralogies
Probable mixture of different mineralogies
Probable composition, predominantly
quartz
Mineralogy unknown
Probable mixture of different mineralogies
located close to source area
Mineralogy unknown
Mineralogy unknown
Final Risk and Exposure Assessment
Appendix 5-80
September 2009
-------
Terrestrial Acidification Case Study
Parent Material or
Modifier Characteristic
Sandy modifier
Loamy modifier
Skeletal modifier
Red, brown, ferric, iron
modifier
Opposing mineralogies
Multiple layers described
Acidity Classification
Acidic
Intermediate
Classification based on
top two layer
descriptions
Basic
Intermediate
Classification based on
top layer description
Rational
Probable high silica content
Equal contributions of chemical properties
(intermediate ion exchange capacity) from
the three soil textures (i.e., sand, silt, clay)
Skeletal layer is very thin and spotty, so top
two layers considered main source of
parent material
Probable high iron content (associated with
potentially higher) cation ion-exchange
capacity
Considered average of two mineralogies
Top layer considered main source of soil
parent material
The calculated critical loads for the three levels of protection (Bc/Al(Crit) = 0.6, 1.2 and
10.0) were compared to 2002 CMAQ/NADP total nitrogen and sulfur deposition levels to
determine which plots with sugar maple and/or red spruce experienced deposition levels greater
than the critical load values.
Based on the 8MB calculations for the three levels of protection (Bc/Al(CIit) = 0.6, 1.2 and
10.0), critical loads of acidifying deposition for sugar maple in the 24 states were found to range
from 107 to 6,008 eq/ha/yr (Table 4.1-5). Critical loads for red spruce in the 8 states ranged from
180 to 4,278 eq/ha/yr. In a comparison of the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels and calculated critical loads, 3% to 75% of all sugar maple plots and 3% to
36% of all red spruce plots were found to have 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels greater than the critical loads; the highest protection critical loads (Bc/Al(Crit) =
10.0) had the highest frequency of exceedance (Table 4.1-6). Aggregated by State, a large
proportion of the sugar maple and red spruce plots showed high levels of critical load
exceedance for the highest protection level (Bc/Al(Crit) = 10.0) and comparatively lower
exceedance frequency at the lowest protection level ((Bc/Al(Crit) = 0.6)) (Table 4.1-6, Figures
4.1-1 to 4.1-6). In general, New Hampshire displayed the greatest degree of critical load
exceedance at all protection levels for both species.
Final Risk and Exposure Assessment
Appendix 5-81
September 2009
-------
Terrestrial Acidification Case Study
Collectively, given the limitations and uncertainties associated with the 8MB model to
estimate critical acid loads (see Section 4.3.9 in Chapter 4 and Section 5 in Appendix 5 for
further description), these results suggest that the health of at least a portion of the sugar maple
and red spruce growing in the United States may have been compromised with the 2002
CMAQ/NADP total nitrogen and sulfur deposition levels; even with the lowest level of
protection, half the states contained sugar maple and red spruce stands that were negatively
impacted by acidifying deposition. At the highest level of protection (Bc/Al(crit) = 10.0), the
apparent impact of the 2002 CMAQ/NADP total nitrogen and sulfur deposition levels was much
greater. A large proportion of sugar maple (>80% of plots in 13 of 24 states) and the majority of
red spruce (100% of plots in 5 of 8 states) experienced deposition levels that exceeded the
critical loads. If this high protection critical load accurately represents the conditions of the two
species, a large proportion of both sugar maple and red spruce, throughout their ranges, were
most likely negatively impacted under 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels.
Final Risk and Exposure Assessment September 2009
Appendix 5-82
-------
Terrestrial Acidification Case Study
Table 4.1-5. Ranges of Critical Load Values by Level of Protection (Bc/Al(crit) = 0.6, 1.2, and 10.0) and by State for the Full
Distribution Ranges of Sugar Maple and Red Spruce
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kansas
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Ranges of Critical Load Values (eq/ha/yr)
Sugar Maple
Bc/Al = 0.6
1,592 to 5,337
2,239 to 4,290
1,5 19 to 2,468
2,543 to 3,671
1,478 to 5,859
2,260 to 3,791
NA
2,044 to 3,994
746 to 4,284
2,066 to 3,090
791 to 2,4 14
400 to 6,008
220 to 4,9 16
978 to 4,891
5 80 to 1,994
1,452 to 2,651
503 to 4,467
1,415 to 3,444
1,226 to 4,986
1,026 to 4,047
Bc/Al = 1.2
1,1 14 to 3,638
1,536 to 2,913
1,05 8 to 1,702
1,730 to 2,485
1,020 to 3,971
1,533 to 2,560
NA
1,390 to 2,707
535 to 2,983
1,417 to 2,122
566 to 1,661
294 to 4,070
166 to 3,3 18
681 to 3,304
419 to 1,439
1,029 to 1,824
370 to 3,039
1,0 10 to 2,426
855 to 3,366
723 to 2,752
Bc/Al = 10.0
617 to 2,015
857 to 1,623
581 to 941
965 to 1,390
573 to 2,2 14
854 to 1,424
NA
749 to 1,497
295 to 1,620
929 to 1,178
319to919
169 to 2,269
107 to 1,861
377 to 1,843
236 to 780
566 to 1,012
209 to 1,686
558 to 1,319
469 to 1,877
402 to 1,530
Red Spruce
Bc/Al = 0.6
-
-
-
-
-
-
-
-
599 to 4,278
-
1,706 to 1,736
-
-
-
418 to 1,994
-
526 to 3, 146
1256
-
NA
Bc/Al = 1.2
-
-
-
-
-
-
-
-
439 to 2,979
-
1,191 to 1,213
-
-
-
324 to 1,439
-
386 to 2,156
926
-
NA
Bc/Al = 10.0
-
-
-
-
-
-
-
-
249 to 1,623
-
656 to 669
-
-
-
180 to 780
-
217 to 1,195
501
-
NA
Final Risk and Exposure Assessment
Appendix 5-83
September 2009
-------
Terrestrial Acidification Case Study
State
Rhode Island
South Carolina
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Combined (all plots)
Ranges of Critical Load Values (eq/ha/yr)
Sugar Maple
Bc/Al = 0.6
NA
NA
921 to 5,755
479 to 5,660
1,036 to 5, 852
369 to 4,134
400 to 5,031
220 to 6,008
Bc/Al = 1.2
NA
NA
653 to 3,901
351 to 3,846
726 to 3,968
270 to 2,8 19
290 to 3,393
166 to 4,070
Bc/Al = 10.0
NA
NA
351 to 2, 175
201 to 2, 142
4 10 to 2,208
152 to 1,560
166 to 1,898
107 to 2,269
Red Spruce
Bc/Al = 0.6
-
-
2,065
846 to 2,305
NA
2,300 to 3,634
-
418 to 4,278
Bc/Al = 1.2
-
-
1,433
6,016 to 1,648
NA
1,610 to 2,533
-
324 to 2,979
Bc/Al = 10.0
-
-
788
336 to 888
NA
884 to 1,382
-
180 to 1,623
Note: NA = data not available for state; "-" = tree species not present on forestland in state.
Final Risk and Exposure Assessment
Appendix 5-84
September 2009
-------
Terrestrial Acidification Case Study
Table 4.1-6. Percentages of Plots, by Protection Level (Bc/Al(crit) = 0.6, 1.2, and 10.0) and by
State, where 2002 CMAQ/NADP Total Nitrogen and Sulfur Deposition Levels Were Greater
Than the Critical Loads for Sugar Maple and Red Spruce
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kansas
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Rhode Island
South Carolina
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Combined
(all plots)
Percentage of Plots Where Critical Load is Exceeded (%)
Sugar Maple
Bc/Al =
0.6
0
0
0
0
0.3
0
NA
0
0
0
6
6
2
0.7
29
0
6
0
1
7
NA
NA
0.3
2
2
2
2
3
Bc/Al = 1.2
23
0
23
0
12
0
NA
0
0.7
25
33
14
7
2
38
67
20
6
16
22
NA
NA
o
J
7
9
8
10
12
Bc/Al =
10.0
31
10
100
66
87
23
NA
86
20
100
100
70
30
46
84
100
95
71
95
98
NA
NA
50
99
59
95
82
75
Red Spruce
Bc/Al = 0.6
-
-
-
-
-
-
-
-
0.2
-
0
-
-
-
27
-
14
0
-
NA
-
-
0
2
NA
0
-
3
Bc/Al = 1.2
-
-
-
-
-
-
-
-
0.5
-
100
-
-
-
38
-
15
0
-
NA
-
-
0
6
NA
0
-
5
Bc/Al = 10.0
-
-
-
-
-
-
-
-
16
-
100
-
-
-
78
-
79
100
-
NA
-
-
100
100
NA
100
-
36
Note: NA = data not available for state; "-" = tree species not present on forestland in state
Final Risk and Exposure Assessment
Appendix 5-85
September 2009
-------
Terrestrial Acidification Case Study
Sugar Maple (Bc/AI=0.6)
Legend
no exceedances
<50% exceedance
>50% exceedance
Sugar Maple state
wild data nol available
Figure 4.1-1. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the lowest protection critical load
(Bc/Al(crit) = 0.6) in the following: none of the sugar maple plots, <50% of the sugar
maple plots, and >50% of the sugar maple plots.
Final Risk and Exposure Assessment
Appendix 5-86
September 2009
-------
Terrestrial Acidification Case Study
Sugar Maple (Bc/AI=1.2}
Legend
no exceedances
<50% exceedance
>50% exceedance
Sugar Maple state
with data nol available
Figure 4.1-2. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the intermediate protection critical load
(Bc/Al(crit) = 1.2) in the following: none of the sugar maple plots, <50% of the sugar
maple plots, and >50% of the sugar maple plots.
Final Risk and Exposure Assessment
Appendix 5-87
September 2009
-------
Terrestrial Acidification Case Study
Sugar Maple (Bc/AI=10.0)
Legend
no exceedances
<50% exceedance
>50% exceedance
Sugar Maple state
wilh data nol available
Figure 4.1-3. States where sugar maple is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the highest protection critical load
(Bc/Al(crit) = 10.0) in the following: none of the sugar maple plots, <50% of the sugar
maple plots, and >50% of the sugar maple plots.
Final Risk and Exposure Assessment
Appendix 5-88
September 2009
-------
Terrestrial Acidification Case Study
Red Spruce (Bc/AI=0.6)
Legend
no exceedances
<50% exceedance
^50% exceedance
Red Spruce state
with data not available
Figure 4.1-4. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the lowest protection critical load
(Bc/Al(crit) = 0.6) in the following: none of the red spruce plots, <50% of the red spruce
plots, and >50% of the red spruce plots.
Final Risk and Exposure Assessment
Appendix 5-89
September 2009
-------
Terrestrial Acidification Case Study
Red Spruce (Bc/AI=1.2)
Legend
no exceedances
<50% exceedance
>50% exceedance
Red Spruce slate
with data not available
Figure 4.1-5. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the intermediate protection critical load
(Bc/Al(crit) = 1.2) in the following: none of the red spruce plots, <50% of the red spruce
plots, and >50% of the red spruce plots.
Final Risk and Exposure Assessment
Appendix 5-90
September 2009
-------
Terrestrial Acidification Case Study
Red Spruce (Bc/AI=10.0)
Legend
no exceedances
<50% exceedance
£50% exceedance
Red Spruce state
with data not available
4.2
Figure 4.1-6. States where red spruce is found and where 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels exceeded the highest protection critical load
(Bc/Al(crit) = 10.0) in the following: none of the red spruce plots, <50% of the red spruce
plots, and >50% of the red spruce plots.
RELATIONSHIP BETWEEN ATMOSPHERIC NITROGEN AND
SULFUR DEPOSITION AND TREE GROWTH
The impacts of the 2002 CMAQ/NADP total nitrogen and sulfur deposition and critical
load exceedances on sugar maple and red spruce growth throughout the full ranges of the two
species is presented and discussed in Attachment A.
Final Risk and Exposure Assessment
Appendix 5-91
September 2009
-------
Terrestrial Acidification Case Study
5.0 UNCERTAINTY ANALYSIS
5.1 KANE EXPERIMENTAL FOREST AND HUBBARD BROOK
EXPERIMENTAL FOREST CASE STUDY AREAS
Despite the extensive use of the 8MB model to estimate critical loads, there is uncertainty
regarding the output from the model and calculations. To a large degree, this uncertainty comes
from the dependence of the 8MB calculations on assumptions made by the researcher and the
use of default values. Parameters including base cation weathering (BCW and Bcw), ANCie,Crit,
-Kgibb, Nu, N;, Nde, and Bcu are rarely measured at each location and must be selected based on the
literature or on other calculations and models. In an analysis conducted by Li and McNulty
(2007), it was determined that BCW and ANCie,Crit were the main sources of uncertainty, with
each respectively contributing 49% and 46% to the total variability in critical load estimates. It
has, therefore, been suggested that the calculation of critical loads using a relevant range of
parameter values can provide the foundation for an uncertainty analysis (Li and McNulty, 2007;
Hall et al., 2001; Hodson and Langan 1999); it is likely that the correct critical load of a system
will be contained within the range of load estimates from such an approach. If all or a large
majority of estimates indicate that the critical load of a system is exceeded with 2002
CMAQ/NADP total nitrogen and sulfur deposition levels, it is likely that deposition is greater
than the critical load and that the trees and vegetation in that system are being negatively
impacted by acidification. Conversely, if deposition is not greater than the majority of critical
load estimates, there can be greater confidence that acidifying deposition is not impacting the
system. Under a scenario of a near equal number of estimates indicating exceedance and
nonexceedance, it is not possible to determine the actual acidification status of a system with
confidence. Nonetheless, such results do suggest that the system is near the critical load level and
should be monitored or assessed more thoroughly.
In this case study, multiple values were used for several parameters in the 8MB
calculations for KEF and HBEF; BCW was calculated with two methods, two values of Kg{\,b
constant were used, and three indicator values of (Bc/Al)crit were evaluated. Therefore, it was
possible to use the range of output values from the calculations to access the certainty of the
acidification status of the HBEF and KEF case study areas. For both sugar maple and red spruce,
a similar number of estimates indicated deposition levels greater than the critical loads; the
Final Risk and Exposure Assessment September 2009
Appendix 5-92
-------
Terrestrial Acidification Case Study
critical loads associated with the most stringent, most protective Bc/Alcrit ratio indicator (Bc/Alcrit
= 10.0) were frequently lower than the 2002 CMAQ/NADP total nitrogen and sulfur deposition
levels. Conversely, the critical loads calculated with the (Bc/Al)crit ratio indicative of a high risk
to tree health (Bc/Alcrit = 0.6) were higher than the 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels. The intermediate indicator ratio ((Bc/Al)crit) had critical load estimates that
were either exceeded or not exceeded by 2002 CMAQ/NADP total nitrogen and sulfur
deposition levels . The patterning of the results suggests that the 2002 CMAQ/NADP total
nitrogen and sulfur deposition levels were very close to, if not greater than, the critical loads of
the two case study areas, and both ecosystems are likely to be sensitive to any future changes in
the levels of nitrogen and sulfur acidifying deposition.
A more thorough, quantified uncertainty analysis of the parameters that are selected for
the 8MB method calculations of critical acid loads is recommended for future analyses.
5.2 EXPANSION OF CRITICAL LOAD ASSESSMENTS
Critical load estimates for individual plots within the distribution ranges of sugar maple
and red spruce were calculated using the clay-substrate method to estimate BCW. As discussed
earlier, the BCW term within the 8MB model is one of the most influential terms in the
calculation of a critical load, and the determination of this BCW value is strongly influenced by
the classified acidity of the soil parent material. In large-scale analyses, descriptions of the
mineralogy of parent material underlying the soil may be missing, nondescriptive, only
suggestive of mineralogy, or may only represent the dominant mineralogy in a large area (and
therefore not accurately capture the smaller-scale variation in mineralogy). Therefore, it is
possible to misclassify the parent material acidity in the BCwterm.
In the analyses of critical loads for the full distribution ranges of sugar maple and red
spruce in this report, two fine-scale databases (i.e., SSURGO soils [USDA-NRCS, 2008c] and
USGS state-level geology [USGS, 2009b] databases) were used as the sources for parent
material mineralogy to allow for location-specific mineralogy descriptions. In addition, a
systematic protocol similar to that used in Europe (UNECE, 2004) and Australia (Gray and
Murphy, 1999), and based on known and probable silica and ferromagnesium content, spatial
patterns of local and geologic settings, and implied deposit!onal mechanisms and environments
was used to determine the parent material acidity classifications. Therefore, steps were taken to
Final Risk and Exposure Assessment September 2009
Appendix 5-93
-------
Terrestrial Acidification Case Study
determine accurate, location-specific acidity classifications. Nonetheless, parent material in some
of the plots may have been misclassified.
To evaluate the degree to which critical load estimates could change with a
misclassification of parent material acidity, a simple analysis of absolute (eq/ha/yr) and
percentage change associated with misclassifications of parent materials was conducted, using
the critical loads associated with the three levels of protection ((Bc/Al)crit = 0.6. 1.2 and 10. 0) for
sugar maple and red spruce. The differences between all combinations of critical loads calculated
with basic, intermediate, and acidic parent materials were determined, and these difference
values were expressed as a percentage of the original critical load estimates (Tables 5.2-1 and
5.2-2). For example, the percentage difference associated with the misclassification of an
intermediate parent material as acidic would be calculated as the absolute value of (CLintermediate ~
^J-^acidJ/ ^^intermediate •
Table 5.2-1. Differences and Percentage Differences in Plot-Level Critical Load Estimates
Associated with the Misclassification of Parent Material Acidity for the Full Range Assessment
of Sugar Maple
Bc/Al(crit)
Ratio
0.6
1.2
Misclassification of
Parent Material
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Difference between Critical
Loads (eq/ha/yr)
Range of
Values
737tol,559
784 to 3, 631
0 to 2,072
737 to 1,559
784 to 3, 631
0 to 2,072
493 to 1,045
527 to 2,433
0 to 1,388
493 to 1,045
527 to 2,433
0 to 1,388
Average
797
1,018
222
797
1,018
222
537
686
149
537
686
149
Median
784
936
156
784
936
156
529
634
105
529
634
105
% Difference between
Critical Loads (%)
Range of
Values
25 to 492
42 to 492
Oto36
20 to 83
30 to 83
Oto26
24 to 428
40 to 428
Oto36
19 to 81
29 to 81
Oto26
Average
58
70
8
35
40
7
56
67
8
34
39
7
Median
51
61
7
34
38
7
50
60
7
33
37
6
Final Risk and Exposure Assessment
Appendix 5-94
September 2009
-------
Terrestrial Acidification Case Study
Bc/Al(crit)
Ratio
10.0
Bc/Al(ci.it) Ratio
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Misclassification of Parent
Material
Range of
Values
275 to 583
293 to 1,357
0 to 774
275 to 583
293 to 1,357
0 to 774
Range
of
Values
298
381
83
298
381
83
Averag
e
294
351
58
294
351
58
Difference between Critical
Loads (eq/ha/yr)
Median
24 to 376
41 to 376
Oto36
20 to 79
29 to 79
Oto26
Range
of
Values
56
67
8
34
39
7
Averag
e
50
60
7
33
37
6
Table 5.2-2. Differences and Percentage Differences in Plot-Level Critical Load Estimates
Associated with the Misclassification of Parent Material Acidity for the Full Range Assessment
of Red Spruce
Bc/Al(Crit)
Ratio
0.6
1.2
10.0
Misclassification of
Parent Material
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Acidic as Intermediate
Acidic as Basic
Intermediate as Basic
Intermediate as Acidic
Basic as Acidic
Basic as Intermediate
Difference between Critical
Loads (eq/ha/yr)
Range of
Values
772 to 1,340
853 to 1,584
0 to 655
772 to 1,340
853 to 1,584
0 to 655
521 to 969
578 to 1,075
0 to 444
521 to 969
578 to 1,075
0 to 444
289 to 511
320 to 594
0 to 246
289 to 511
320 to 594
0 to 246
Average
831
965
134
831
965
134
567
658
91
567
658
91
312
362
50
312
362
50
Median
829
932
99
829
932
99
566
637
67
566
637
67
311
350
37
311
350
37
% Difference between
Critical Loads (%)
Range of
Values
27 to 453
44 to 453
Oto 16
22 to 82
31 to 82
Oto 14
27 to 392
43 to 392
Oto 16
21 to 80
30 to 80
Oto 14
27 to 345
45 to 345
Oto 16
21 to 78
31 to 78
Oto 14
Average
69
78
6
40
43
5
66
74
5
38
42
5
66
75
5
39
42
5
Median
67
74
5
40
43
5
64
71
5
39
42
4
65
73
5
39
42
5
Final Risk and Exposure Assessment
Appendix 5-95
September 2009
-------
Terrestrial Acidification Case Study
The comparisons of critical loads revealed that changes in critical load values could range
from 0 to 3,631 eq/ha/yr for sugar maple and 0 to 1,584 eq/ha/yr for red spruce with the
misclassification of parent material acidity. These ranges correspond to percentage differences
ranging from 0% to 492% and 0% to 453% for sugar maple and red spruce, respectively. The
results also indicate that the biggest impacts of a misclassification on critical load estimates
would occur with an acidic parent material being misclassified as basic; the average percentage
changes in the estimated critical loads, in such a scenario, were 67% to 70% for sugar maple and
74% to 78% for red spruce, and the median percentage changes were 60% to 61% and 71% to
74% for the two species, respectively. In contrast, the smallest impacts on critical load estimates
would occur when a basic parent material was incorrectly classified as intermediate and vice
versa. In this scenario, the average and median percentage changes in critical load estimates were
only 7% to 8% and 6% to 7% for sugar maple and 5% to 6% and 4% to 5% for red spruce. Given
the potential significant impacts of a misclassification of parent material acidity on critical load
estimates, this potential source of error should be considered in the accuracy and application of
the critical load estimates.
6.0 REFERENCES
Aber, J.D., C.L. Goodale, S.V. Ollinger, M.-L. Smith, A.H. Magill, M.E. Martin, R.A. Hall, and
J.L. Stoddard. 2003. Is nitrogen deposition altering the nitrogen status of northeastern
forests? Bioscience 53:375-389.
Aber, J.D., A. Magill, S.G. McNulty, R.D. Boone, KJ. Nadelhoffer, M. Downs, and R. Hallett.
1995. Forest Biogeochemistry and primary production altered by nitrogen saturation.
Water, Air, and Soil Pollution 55:1665-1670.
Allen, H.W. 2001. Silvicultural treatments to enhance productivity. Pp. 129-139 in The Forests
Handbook. Vol. 2. Edited by J.Evans. Oxford, UK: Blackwell Science Ltd. Available at:
http://books.google.com/books?id=pH47NTIPLVIC&pg=PA129&lpg=PA129&dq=silvic
ultural+treatments+to+enhance+productivity+the+forests+handbook&source=bl&ots=Be
oXcPLqzZ&sig=KPS_wT8FX-YOi9UKhSRop-OK9qs&hl=en#v=onepage&q=&f=false
(accessed September 9, 2009).
Final Risk and Exposure Assessment September 2009
Appendix 5-96
-------
Terrestrial Acidification Case Study
Bailey S.W., S.B. Horsley, R.P. Long, and R.A. Hallett. 2004. Influence of edaphic factors on
sugar maple nutrition and health on the Allegheny plateau. Soil Science Society of
America Journal 68:243-252.
Bailey, S.W., S.B. Horsley, and R.P. Long. 2005. Thirty years of change in forest soils of the
Allegheny Plateau, Pennsylvania. Soil Science Society of America Journal (59:681-690.
Baumgardner, R.E., S.S. Isil, T.F. Lavery, C.M. Rogers, and V.A. Mohnen. 2003. Estimates of
cloud water deposition at mountain Acid Deposition Program Sites in the Appalachian
Mountains. Journal of Air and Waste Management Association 53:291-308.
Bintz, W.W., and DJ. Butcher. 2007. Characterization of the health of southern Appalachian red
spruce (Piceae rubens) through determination of calcium, magnesium, and aluminum
concentrations of foliage and soil. Microchemical Journal 57:170-174.
Birdsey, R.A. 1992. Carbon Storage and Accumulation in United States Ecosystems. U.S. Forest
Service General Technical Report WO-59. U.S. Department of Agriculture, Forest
Service, Northeast Forest Experiment Station, Radnor, PA.
Bobbink R., D. Boxman, E. Fremstad, G. Heil, A. Houdijk, and J. Roelofs. 1992. Critical loads
for nitrogen eutrophication of terrestrial and wetland ecosystems based upon changes in
vegetation and fauna. Pp. 111-159 in Critical Loads for Nitrogen, Edited by Grennfelt
and Thornelof. Miljorapport 1992:41. Nordic Council of Ministers, Copenhagen,
Denmark.
Bobbink R., M. Ashmore, S. Braun, W. Fliickiger, and IJJ. Van den Wyngaert. 2003. Empirical
nitrogen critical loads for natural and semi-natural ecosystems: 2002 update. Pp. 43-170
in Empirical Critical Loads for Nitrogen. Edited by B. Achermann and R. Bobbink.
Environmental Documentation 164, Swiss Agency for the Environment, Forests and
Landscape (SAEFL), Berne, Switzerland.
Briggs, R.D., Hornbeck, J.W., Smith, C.T., Lemin, R.C., and McCormack, M.L. 2000. Long-
term effects of forest management on nutrient cycling in spruce- fir forests. Forest
Ecology and Management 138: 285-299.
Final Risk and Exposure Assessment September 2009
Appendix 5-97
-------
Terrestrial Acidification Case Study
Bull, K., B. Achermann, V. Bashkin, R. Chrast, G. Fenech, M. Forsius, H.-D. Gregor, R.
Guardans, T. Haussmann, V. Hayes, B. Kvaeven, M. Lorenz, L. Lundin, G. Mills, M.
Posch, B.L. Skjelkvale, and MJ. Ulstein. 2001. Coordinated effects monitoring and
modeling for developing and supporting international air pollution control agreements.
Water, Air, and Soil Pollution 730:119-130.
Carter, B.J., and EJ. Ciolkosz. 1980. Soil temperature regimes of the central Appalachians. Soil
Science Society of America Journal 44:1052-1058.
Chapin, F.S., III, L. Moilanen, and K. Kielland. 1993. Preferential use of organic nitrogen for
growth by a non-mycorrhizal arctic sedge. Nature 361:150-153.
Cronan, C.S., and D.F. Grigal. 1995. Use of calcium/aluminum ratios as indicators of stress in
forest ecosystems. Journal of Environmental Quality 24:209-226.
De Witt, H.A., J. Mulder, P.H. Nygaard, and D. Aamlid. 2001. Testing the aluminum toxicity
hypothesis: A field manipulation experiment in mature spruce forest in Norway. Water,
Air, and Soil Pollution 730:995-1000.
DeHayes, D.H., P.G. Schaberg, GJ. Hawley, and G.R. Strimbeck. 1999. Acid rain impacts on
calcium nutrition and forest health. Bioscience ¥9(10):789-800.
Driscoll, C.T., G.E. Likens, L.O. Hedin, J.S. Eaton, and F.H. Bormann. 1989. Changes in the
chemistry of surface waters: 25-year results at the Hubbard Brook Experimental Forest,
NH. Environmental Science and Technology 23(2): 137-143.
Driscoll, C.T., G.B. Lawrence, AJ. Bulger, TJ. Butler. C.S. Cronan, C. Eager, K.F. Lambert,
G.E. Likens, J.L. Stoddard, and K.C. Weathers. 2001. Acidic deposition in the
northeastern United States: Sources and inputs, ecosystem effects, and management
strategies. Bioscience 57(3): 180-198.
Drohan, J. R., and W.E. Sharpe. 1997. Long-term changes in forest soil acidity in Pennsylvania,
U.S.A. Water, Air, and Soil Pollution 95:299-311.
Final Risk and Exposure Assessment September 2009
Appendix 5-98
-------
Terrestrial Acidification Case Study
Drohan, P.J., S.L. Stout, and G.W. Petersen. 2002. Sugar maple (Acer saccharum Marsh.)
decline during 1979-1989 in northern Pennsylvania. Forest Ecology and Management
170:1-11.
Eagar, C., H. Van Miegroet, S.B. McLaughline, and N.S. Nicholas. 1996. Evaluation of effects of
acidic deposition to terrestrial ecosystems in class I areas of the Southern Appalachians.
Technical Report. Southern Appalachian Mountains Initiative, Asheville, NC.
FAO (Food and Agriculture Organization). 2007. Digital Soil Map of the World. GIS datalayer.
United Nations, Food and Agriculture Organization, GeoNetwork, Rome, Italy. Available
at: http://www.fao.org/geonetwork/srv/en/main.home (accessed September 9, 2009).
Fenn, M.E., V.M. Perea-Estrada, L.I. de Bauer, M. Perez-Suarez, D.R. Parker, and V.M. Cetina-
Alcala. 2006a. Nutrient status and plant growth effects of forest soils in the Basin of
Mexico. Environmental Pollution 140(2): 187-199.
Fenn, M.E., Huntington, T.G., McLaughlin, S.B., Eager, C., Gomez, A., and Cook, R.B. 2006b.
Status of soil acidification in North America. Journal of Forest Science 52(special issue):
3-13.
Fincher, J., J.R. Gumming, R. G. Alscher, G. Rubin, and L Weinstein. 1989. Long-term ozone
exposure affects winter hardiness of red spruce (Picea rubens Sarg.) seedlings. New
Phytologist 773:85-96.
Gebert, W.A., DJ. Graczyk, and W.R. Krug, 1987. Average Annual Runoff in the United States,
1951-80: U.S. Geological Survey Hydrologic Investigations Atlas HA-710, Scale
1:7,500,000. GIS datalayer. U.S. Department of Interior, U.S. Geological Survey,
Madison, WI. Available at: http://water.usgs.gov/GIS/dsdl/runoff.eOO.gz (accessed
September 9, 2009).
Gray, J.M., and B.W. Murphy. 1999. Parent Material and Soils - A Guide to the Influence of
Parent Material on Soil Distribution in Eastern Australia. Technical Report Number 45.
NSW Department of Land and Water Conservation, Sydney, Australia. Available at:
Final Risk and Exposure Assessment September 2009
Appendix 5-99
-------
Terrestrial Acidification Case Study
http://www.dnr.nsw.gov.au/care/soil/soil_pubs/parent_material.html (accessed September
9, 2009).
Hall, I, K. Bull, I. Bradley, C. Curtis, P. Freer-Smith, M. Hornung, D. Howard, S. Langan, P.
Loveland, B. Reynolds, J. Ullyett, and T. Warr. 1998. Status of UK Critical Loads and
Exceedences. Part 1 - Critical Loads and Critical Loads Maps. Report to the Department
of Environment Transport and the Regions. Biological Records Centre, Institute of
Terrestrial Ecology, Centre for Ecology and Hydrology, Monks Wood Experimental
Station, Abbots Ripton, Huntingdonshire, England.
Hall, J., B. Reynolds, S. Langan, M. Hornung, F. Kennedy, and J. Aherne. 2001. Investigating
the uncertainties in the simple mass balance equation for acidity critical loads for
terrestrial ecosystems in the United Kingdom. Water, Air and Soil Pollution: Focus 1:
43-56.
Hallett, R.A., S.W. Bailey, S.B. Horsley, and R.P. Long. 2006. Influence of nutrition and stress
on sugar maple at a regional scale. Canadian Journal of Forest Research 36:2235-2246.
HBES (Hubbard Brook Ecosystem Study). 2008a. Forest Inventory of a Northern Hardwood
Forest: Watershed 6 (the biogeochemical reference watershed) 2002. Online database.
Hubbard Brook Ecosystem Study, North Woodstock, NH. Available at:
http://www.hubbardbrook.org/data/dataset.php?id=35 (accessed December 18, 2008).
HBES (Hubbard Brook Ecosystem Study). 2008b. Hubbard Brook Ecosystem Study. Online
information. U.S. Department of Agriculture, Forest Service, Water and Ecosystems
Project, Northern Research Station, Hubbard Brook Experimental Forest, Durham, NH.
Available http://www.hubbardbrook.org (accessed June 20, 2008).
HBES (Hubbard Brook Ecosystem Study). 2008c. Soil Temperature Data. Online database.
Hubbard Brook Ecosystem Study, North Woodstock, NH. Available at
http://www.hubbardbrook.org/data/dataset.php?id=65 (accessed December 18, 2008).
Final Risk and Exposure Assessment September 2009
Appendix 5-100
-------
Terrestrial Acidification Case Study
Hodson, M.E., and S.J. Langan. 1999. Considerations of uncertainty in setting critical loads of
acidity of soils: the role of weathering rate determination. Environmental Pollution 106:
73-81.
Horsley, S.B., R.P. Long, S.W. Bailey, R.A. Hallett, and T.J. Hall. 2000. Factors associated with
the decline disease of sugar maple on the Allegheny Plateau. Canadian Journal of Forest
Research 30:1365-1378.
Jeffries, D.S. (ed.). 1997. Canadian Acid Rain Assessment. Volume 3, The Effects of Canada's
Lakes, Rivers and Wetlands. Environment Canada, Gatineau, Quebec.
Jeffries, D.S., and D.C.L. Lam. 1993. Assessment of the effect of acid deposition on Canadian
lakes: determination of critical loads for sulphate deposition. Water Science Technology
25:183-187.
Jenkins, A., B.J. Cosby, R.C. Ferrier, T. Larssen, and M. Posch. 2003. Assessing emission
reduction targets with dynamic models: deriving target load functions for use in
integrated assessment. Hydrology and Earth System Services 7(4):609-617.
Jenkins, J.C., R.A. Birdsey, and Y. Pan. 2001. Biomass and NPP estimation for the mid-Atlantic
region (USA) using plot-level forest inventory data. Ecological Applications 77:1174-
1193.
Johnson, A.H., S.B. Andersen, and T.G. Siccama. 1994a. Acid rain and soils of the Adirondacks.
I. Changes in pH and available calcium, 1930-1984. Canadian Journal of Forest
Research 24:39-45.
Johnson, A.H., A.J. Friedland, E.K. Miller, and T.G. Siccama. 1994b. Acid rain and soils of the
Adirondacks. III. Rates of soil acidification in a montane spruce-fir forest at Whiteface
Mountain, New York. Canadian Journal of Forest Research 24:663-669.
Johnson, C.E., A.H. Johnson, and T.G. Siccama. 1991. Whole-tree clear-cutting effects on
exchangeable cations and soil acidity. Soil Science Society of America Journal
55(2):502-508.
Final Risk and Exposure Assessment September 2009
Appendix 5-101
-------
Terrestrial Acidification Case Study
Johnson, A.H., E.R. Cook, and T.G. Siccama. 1988. Climate and red spruce growth and decline
in the northern Appalachians. Proceedings of the National Academy of Sciences 85:
5369-5373.
Johnson, D.W. 1984. Sulfur cycling in forests. Biogeochemistry 7:29-43.
Joslin, J.D., and M.H. Wolfe. 1992. Red spruce soil chemistry and root distribution across a
cloud water deposition gradient. Canadian Journal of Forest Research 22:893-904.
Juice, S.M., T.J. Fahey, T.G. Siccama, C.T. Driscoll, E.G. Denny, C. Eagar, N.L. Cleavitt, R.
Minocha, and A.D. Richardson. 2006. Response of sugar maple to calcium addition to
northern hardwood forest. Ecology 57:1267-1280.
Killam, K. 1994. Soil Ecology. Cambridge, UK: Cambridge University Press.
Lawrence, G.B., A.G. Lapenis, D. Berggren, B.F. Aparin, K.T. Smith, W.C. Shortle, S.W.
Bailey, D.L. VarlyGuin, and B. Babikov. 2005. Climate dependency of tree growth
suppressed by acid deposition effects on soils in northwest Russia. Environmental
Science and Technology 39:2004-2010.
Lazarus, B.E., P.G. Schaberg, D.H. DeHayes, and G.J. Hawley. 2004. Severe red spruce winter
injury in 2003 creates unusual ecological event in the northeastern United States.
Canadian Journal of Forest Research 34:1784-1788.
Li, H., and S.G. McNulty. 2007. Uncertainty analysis on simple mass balance model to calculate
critical loads for soil acidity. Environmental Pollution 149:315-326.
Lilleskov, E.A., P.M. Wargo, K.A. Vogt, and D.J. Vogt. 2008. Mycorrhizal fungal community
relationship to root nitrogen concentration over a regional atmospheric nitrogen
deposition gradient in the northeastern USA. Canadian Journal of Forest Research 38:
1260-1266.
Martin, J.G, B.D. Kloeppel, T.L. Schaefer, D.L. Kimbler, and S.G. McNulty. 1998.
Aboveground biomass and nitrogen allocation often deciduous southern Appalachian
tree species. Canadian Journal of Forest Research 25:1648-1659.
Final Risk and Exposure Assessment September 2009
Appendix 5-102
-------
Terrestrial Acidification Case Study
McNulty, S.G., and J. Boggs. In press. A Connectional Framework for Redefining Forest Soil
Critical Acid Loads under a Changing Climate. Environmental Pollution.
McNulty, S.G., J. Boggs, J.D. Aber, L. Rustad, and A. Magill. 2005. Red spruce ecosystem level
changes following 14 years of chronic nitrogen fertilization. Forest Ecology and
Management 279:279-291.
McNulty, S.G., E.G. Cohen, H. Li, and J.A. Moore-Myers. 2007. Estimates of critical acid loads
and exceedences for forest soils across the conterminous United States. Environmental
Pollution 149:2% 1-292.
MEA (Millennium Ecosystem Assessment Board). 2005. Ecosystems and Human Well-being:
Current State and Trends, Volume 1. Edited by R. Hassan, R. Scholes, and N. Ash.
Washington: Island Press. Available at
http://www.millenniumassessment.org/documents/document.766.aspx.pdf
Miller, H.G. 1988. Long-term effects of application of nitrogen fertilizers on forest sites. Pp. 97-
106 in Forest Site Evaluation and Long-Term Productivity. Edited by D.W. Cole and S.P.
Gessel. Seattle, WA: University of Washington Press.
NADP (National Atmospheric Deposition Program). 2003 a. Annual Ammonium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003b. Annual Calcium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003c. Annual Chloride Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003d. Annual Magnesium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
Final Risk and Exposure Assessment September 2009
Appendix 5-103
-------
Terrestrial Acidification Case Study
NADP (National Atmospheric Deposition Program). 2003e. Annual Nitrate Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003f. Annual Potassium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003g. Annual Sodium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003h. Annual Sulfate Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NEG/ECP Forest Mapping Group (Conference of New England Governors and Eastern
Canadian Premiers Forest Mapping Group). 2001. Protocol for Assessment and Mapping
of Forest Sensitivity to Atmospheric S and N Deposition: Acid Rain Action Plan-Action
Item 4: Forest Mapping Research Project. Prepared by NEG/ECP Forest Mapping
Group. Available at
http://www.nrs.fs.fed.us/clean_air_water/clean_water/cri tical_loads/local-
resources/docs/NEGECP_F orest_Sensitivity_Protocol_5_21_04.pdf.
NEG/ECP Forest Mapping Group (Conference of New England Governors and Eastern
Canadian Premiers Forest Mapping Group). 2005. Assessment of Forest Sensitivity to
Nitrogen and Sulfur Deposition in New Hampshire and Vermont. Report submitted to the
New Hampshire Department of Environmental Services, Concord, NH.
NLCD (National Land Cover Dataset). 2001. National Land Cover Database 2001. GIS
datalayer. Multi-Resolution Land Characteristics Consortium, U.S. Geological Survey,
Sioux Falls, SD. Available at http://www.mrlc.gov/mrlc2k_nlcd.asp.
Final Risk and Exposure Assessment September 2009
Appendix 5-104
-------
Terrestrial Acidification Case Study
Nilsson, J., and P. Grennfelt (eds.). 1988. Critical Loads for Sulphur and Nitrogen: A Report
from a Workshop at Skokloster, Sweden. Milj0rapport 1988:15. Nordic Council of
Ministers, Copenhagen, Denmark.
Ouimet, R., J.-D. Moore, and L. Duchesne. 2008. Effects of experimental acidification and
alkalinization on soil and growth and health of Acer saccharum Marsh. Journal of Plant
Nutrition and Soil Science 171:868-871.
Ouimet, R., P.A. Arp, S.A. Watmough, J. Aherne, and I. DeMerchant. 2006. Determination and
mapping critical loads of acidity and exceedences for upland forest soils in Eastern
Canada. Water, Air and Soil Pollution 772:57-66.
PA DCNR (Pennsylvania Department of Conservation and Natural Resources). 2001. Digital
Bedrock Geology of Pennsylvania. GIS datalayer. Pennsylvania Bureau of Topographic
and Geologic Survey, Department of Conservation and Natural Resources, Harrisburg,
PA. Available at http://www.dcnr.state.pa.us/topogeo/mapl/bedmap.aspx.
Palmer, S.M., C.T. Driscoll, and C.E. Johnson. 2004. Long-term trends in soil solution and
stream water chemistry at the Hubbard Brook Experiment Forest: relationship with
landscape position. Biogeochemistry 68:51-70.
Pardo, L.H., and C.T. Driscoll. 1996. Critical loads for nitrogen deposition: case studies at two
northern hardwood forests. Water, Air, and Soil Pollution 59:105-128.
Pardo, L.H., and N. Duarte. 2007. Assessment of Effects of Acidic Deposition on Forested
Ecosystems in Great Smoky Mountains National Park using Critical Loads for Sulfur and
Nitrogen. U.S. Department of Agriculture, Forest Service. Prepared for Tennessee Valley
Authority, S. Burlington, VT.
Pardo, L.H., M. Robin-Abbott, N. Duarte, and E.K. Miller. 2004. Tree Chemistry Database
(Version 1.0). General Technical Report NE-324. U.S. Department of Agriculture, Forest
Service, Northeastern Research Station, Newton Square, PA.
Final Risk and Exposure Assessment September 2009
Appendix 5-105
-------
Terrestrial Acidification Case Study
Reed, J.C. and Bush, C.A. 2005. Generalized Geologic Map of the Conterminous United States
Edition 1.2. 1:7,500,000 scale. U.S. Geological Survey. Available at
http://pubs.usgs.gov/atlas/geologic (accessed on: November 1, 2008).
Reuss, J.O. 1983. Implications of the calcium-aluminum exchange system for the effect of acid
precipitation on soils. Journal of Environmental Quality 72:591-595.
Reuss, J.O., and D.W. Johnson. 1985. Effect of soil processes on the acidification of water by
acid deposition. Journal of Environmental Quality 14:26-31.
RMCC (Federal/Provincial Research and Monitoring Coordinating Committee). 1990. The 1990
Canadian Long-Range Transport of Air Pollutants and Acid Deposition Assessment
Report. Federal/Provincial Research and Monitoring Coordinating Committee,
Environment Canada, Ottawa, Canada.
Ross, D.S., G.B. Lawrence, and G. Fredriksen. 2004. Mineralization and nitrification patterns at
eight northeastern USA forested research sites. Forest Ecology Management 188:317-
335.
Schaberg, P.G., D.H. DeHayes, and GJ. Hawley. 2001. Anthropogenic calcium depletion: a
unique threat to forest ecosystem health? Ecosystem Health 7:214-228.
Shortle, W.C., K.T. Smith, R. Minocha, G.B. Lawrence, andM.B. David. 1997. Acidic
deposition, cation mobilization, and biochemical indicators of stress in healthy red
spruce. Journal of Environmental Quality 2(5:871-876.
Siccama, T.G., TJ. Fahey, C.E. Johnson, T.W. Sherry, E.G. Denny, E.B. Girdler. G.E. Likens,
and P. A. Schwarz. 2007. Population and biomass dynamics of trees in a northern
hardwood forest at Hubbard Brook. Canadian Journal of Forest Research 3 7(4): 737-
749.
St. Clair, S.B., J.E. Carlson, and J.P. Lynch. 2005. Evidence for oxidative stress in sugar maple
stands growing on acidic, nutrient imbalanced forest soils. Oecologia 7 ¥5:258-369.
Final Risk and Exposure Assessment September 2009
Appendix 5-106
-------
Terrestrial Acidification Case Study
Stone, D.M. 1986. Effect of thinning and nitrogen fertilization on diameter growth of pole-size
sugar maple. Canadian Journal of Forest Research 16: 1245-1249.
Sverdrup, H. Personal communication. 2009. Communication between Harald Sverdrup (Lund
University, Sweden) and Jennifer Phelan (RTI International, USA) during the Canadian
Council of Ministers of the Environment Acid Deposition Critical Loads: Status of
Methods and Indicators Workshop, March 18 and 19, 2009. Ottawa, Ontario, Canada.
Sverdrup, H., and P. Warvfmge. 1993a. Calculating field weathering rates using a mechanistic
geochemical model PROFILE. Applied Geochemistry 5:273-283.
Sverdrup, H., and P. Warfvinge. 1993b. The effect of soil acidification on the growth of trees,
grass and herbs as expressed by the (Ca+ Mg+ K)/Al ratio. Reports in Ecology and
Environmental Engineering 2. Lund University, Department of Chemical Engineering,
Lund, Sweden.
Sverdrup, H., and W. de Vries. 1994. Calculating critical loads for acidity with the simple mass
balance equation. Water, Air, and Soil Pollution 72:143-162.
Thomasma, S., P. Knopp, and S. Stout. 2008. SILVAH Version 5.6. Online computer tool. U.S.
Department of Agriculture, Forest Service, North Eastern Forest Experiment Station,
Irvine, PA. Available at: http://www.nrs.fs.fed.us/tools/silvah/ (accessed September 9,
2009).
Tobin, B.D., and D.J. Weary. 2005. Engineering Aspects ofKarst. GIS datalayer. National Atlas
of the United States, Reston, VA. Available at:
http://www.nationalatlas.gov/mld/karstOm.html (accessed September 9, 2009).
UNECE (United Nations Economic Commission for Europe). 2004. Manual on Methodologies
and Criteria for Modeling and Mapping Critical Loads and Levels and Air Pollution
Effects, Risks, and Trends. Convention on Long-Range Transboundary Air Pollution,
Geneva Switzerland. Available at http://www.icpmapping.org (accessed August 16,
2006).
Final Risk and Exposure Assessment September 2009
Appendix 5-107
-------
Terrestrial Acidification Case Study
U.S. EPA (Environmental Protection Agency). 2005. Advisory on Plans for Ecological Effects
Analysis in the Analytical Plan for EPA 's Second Prospective Analysis - Benefits and
Costs of the Clean Air Act, 1990-2020. EPA-COUNCIL-ADV-05-001. U.S.
Environmental Protection Agency, Office of the Administrator, Science Advisory Board,
Washington, DC. June 23.
U.S. EPA (Environmental Protection Agency). 2008a. April 2008 Scope and Methods Plan for
Risk Exposure Assessment. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 2008b. Clean Air Status and Trends Network
(CASTNET). GIS datalayer. Clean Air Mapping and Analysis Program, U.S.
Environmental Protection Agency, Office of Atmospheric Programs, Clean Air Markets
Division, Washington, DC. Available at http://epamap4.epa.gov/cmap/viewer.htm.
U.S. EPA (Environmental Protection Agency). 2008c. Integrated Science Assessment for Oxides
of Nitrogen and Sulfur Environmental Criteria. Final. EPA/600/R-08/082F. U.S.
Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment, Research Triangle Park, NC. December.
USDA (United States Department of Agriculture). 1994. Soil Information for Environmental
Modeling and Ecosystem Management: State Soil Geographic (STATSGO) Database.
Soil datasets. U.S. Department of Agriculture, Natural Resources Conservation Service,
National Cartography and Geospatial Center, Washington, DC. Available at
http://www.soilinfo.psu.edu/index.cgi?soil_data&conus&data_cov.
USDA-NRCS (United States Department of Agriculture-Natural Resources Conservation
Service). 2008a. Soil Survey Geographic (SSURGO) Database for Cameron and Elk
Counties, PA. GIS datalayer. U.S. Department of Agriculture, Natural Resources
Conservation Service, Washington, DC. Available at http://datagateway.nrcs.usda.gov.
USDA-NRCS (United States Department of Agriculture-Natural Resources Conservation
Service). 2008b. Soil Survey Geographic (SSURGO) Database for Grafton County Area,
Final Risk and Exposure Assessment September 2009
Appendix 5-108
-------
Terrestrial Acidification Case Study
New Hampshire. GIS datalayer. U.S. Department of Agriculture, Natural Resources
Conservation Service, Washington, DC. Available at: http://datagateway.nrcs.usda.gov.
USDA-NRCS (United States Department of Agriculture-Natural Resources Conservation
Service). 2008c. Soil Survey Geographic (SSURGO) Database. U.S. Department of
Agriculture, Natural Resources Conservation Service, Fort Worth, TX. Available at:
http://datagateway.nrcs.usda.gov.
USFS (U.S. Forest Service). 1999. Kane Experimental Forest. Online information. U.S.
Department of Agriculture, Forest Service, Kane Experimental Forest, Forestry Sciences
Laboratory, Irvine, PA. Available at http://www.fs.fed.us/ne/warren/kane.html (accessed
June 17, 2008).
USFS (U.S. Forest Service). 2002a. Forest Inventory and Analysis National Program: Area of
Forest Land, 2002. U.S. Department of Agriculture, Forest Service, Forest Inventory and
Analysis National Program, Arlington, VA. Available at http://fia.fs.fed.us/tools-
data/maps/descr/for_land.asp (accessed March 3, 2009).
USFS (U.S. Forest Service). 2002b. Forest Inventory and Analysis National Program: Area of
Timberland, 2002. U.S. Department of Agriculture, Forest Service, Forest Inventory and
Analysis National Program, Arlington, VA. Available at http://fia.fs.fed.us/tools-
data/maps/descr/tim_land.asp (accessed MarchS, 2009).
USFS (U.S. Forest Service). 2006. Forest Inventory Analysis (FIA) Database. U.S. Department
of Agriculture, Forest Service, Forest Inventory and Analysis National Program,
Arlington, VA. Available at http://fia.fs.fed.us/tools-data.
USFS (U.S. Forest Service). 2008a. HubbardBrook Experimental Forest. Online information.
U.S. Department of Agriculture, Forest Service, Northeastern Research Station, Newtown
Square, PA. Available at
http://www.fs.fed.us/ne/newtown_square/research/experimental-forest/hubbard-
brook.shtml (accessed June 17, 2008).
Final Risk and Exposure Assessment September 2009
Appendix 5-109
-------
Terrestrial Acidification Case Study
USFS (U.S. Forest Service). 2008b. Kane Experimental Forest. Online information. U.S.
Department of Agriculture, Forest Service, Northeastern Research Station, Newtown
Square, PA. Available at
http://www.fs.fed.us/ne/newtown_square/research/experimental-forest/kane.shtml
(accessed June 17, 2008).
USFS (U.S. Forest Service) 2008c. The Forest Inventory and Analysis Database: Database
Description and Users Manual Version 3.0 for Phase 2. Forest Inventory and Analysis
Program, U.S. Department of Agriculture, Forest Service. 243 pp.
USFS (U.S. Forest Service). 2008. Personal communication from Dr. Charles (Hobie) Perry,
USD A Forest Service, Northern Research Station, 1992 Folwell Avenue, St. Paul, MN
55108. August 20.
USGS (U.S. Geological Survey). 2000. Bedrock Geologic Map of the Hubbard Brook
Experimental Forest, Grafton County, New Hampshire. GIS datalayer. U.S. Geological
Survey, Reston, VA. Available at: http://pubs.usgs.gov/openfile/ofOO-045/.
USGS (U.S. Geological Survey). 2009a. Karst. Online information. U.S. Department of the
Interior, U.S. Geological Survey, Office of Ground Water, Reston, VA. Available at
http://water.usgs.gov/ogw/karst/ (accessed February 2009).
USGS (U.S. Geological Survey). 2009b. State Geological Map Compilation. U.S. Department of
the Interior, U.S. Geological Survey, Reston, VA. Available at:
http://tin.er.usgs.gov/geology/state (accessed January 28, 2009).
Van der Salm, C., and W. de Vries. 2001. A review of the calculation procedure for critical acid
loads for terrestrial ecosystems. The Science of the Total Environment 277:11-25
Warby, R.A.F., C.E. Johnson, and C.T. Driscoll. 2009. Continuing acidification of organic soils
across the northeastern USA: 1984-2001. Soil Science Society of America Journal
73:274-284.
Wargo, P.M., K. Vogt, D. Vogt, Q. Holifield, J. Tilley, G. Lawrence, and M. David. 2003.
Vitality and chemistry of roots of red spruce in forest floors of stands with a gradient of
Final Risk and Exposure Assessment September 2009
Appendix 5-110
-------
Terrestrial Acidification Case Study
soil Al/Ca ratios in the northeastern United States. Canadian Journal of Forest Research
33:635-652.
Watmough, S.A., J. Aherne, and PJ. Billion. 2004. Critical Loads Ontario: Relating Exceedance
of the Critical Load with Biological Effects at Ontario Forests. Report 2. Environmental
and Resource Studies, Trent University, ON, Canada.
Watmough, S., J. Aherne, P. Arp, I. DeMerchant, and R. Ouimet. 2006. Canadian experiences in
development of critical loads for sulphur and nitrogen. Pp. 33-38 in Monitoring Science
and Technology Symposium: Unifying Knowledge for Sustainability in the Western
Hemisphere Proceedings RMRS-P-42CD. Edited by C. Aguirre-Bravo, PJ. Pellicane,
D.P. Burns, and S. Draggan. U.S. Department of Agriculture, Forest Service, Rocky
Mountain Research Station, Fort Collins, CO.
Webster, K.L., IF. Creed, N.S. Nicholas, and H.V. Miegroet. 2004. Exploring interactions
between pollutant emissions and climatic variability in growth of red spruce in the Great
Smoky Mountains National Park. Water, Air, and Soil Pollution 759:225-248.
Whitfield, C.J., S.A. Watmough, J. Aherne, and PJ. Dillon. 2006. A comparison of weathering
rates for acid-sensitive catchments in Nova Scotia, Canada and their impact on critical
load calculations. Geoderma 136: 899-911.
Final Risk and Exposure Assessment September 2009
Appendix 5-111
-------
Terrestrial Acidification Case Study
Final Risk and Exposure Assessment September 2009
Appendix 5-112
-------
Terrestrial Acidification Case Study
ATTACHMENT A:
RELATIONSHIP BETWEEN ATMOSPHERIC
NITROGEN AND SULFUR DEPOSITION AND SUGAR
MAPLE AND RED SPRUCE TREE GROWTH
1. INTRODUCTION
Nitrogen and sulfur deposition in forest systems can have either positive or negative
impacts on tree growth. The growth of many forests in North America is limited by nitrogen
availability (Chapin et al., 1993; Killam, 1994; Miller, 1988), and nitrogen fertilization is often a
key component of forest management (Allen, 2001). Therefore, in such nitrogen-limited systems,
nitrogen deposition may stimulate tree growth. In contrast, nitrogen additions in some systems
can sometimes be greater than what trees require and can negatively impact tree health and
growth (Aber et al., 1995; McNulty et al., 2005). Forests where atmospheric deposition of
nitrogen and sulfur is greater than the critical load may be examples of such a condition. When
critical loads are exceeded, tree health and growth may be compromised both directly and
indirectly because of soil nutrient deficiencies and imbalances caused by acidic deposition and
the leaching of base cations from the soil. Tree growth may be reduced and/or trees may have an
increased susceptibility to drought and pest damage, aluminum (Al) toxicity in roots, reduced
tolerance to cold, and a greater propensity to frost injury (Driscoll et al., 2001; DeHayes et al.,
1999; Fenn et al., 2006b; McNulty et al., 2005; Ouimet et al., 2008). In the context of acidifying
deposition of nitrogen and sulfur, the positive versus negative impact of deposition on tree
growth may depend largely upon whether the critical load is exceeded by the deposition level
and may follow a modified version of the inverted U-shape relationship hypothesized by Aber et
al. (1995) for forest systems that receive chronic, long-term nitrogen additions (Figure 1-1). If
nitrogen and sulfur deposition is less than the critical load, tree growth may be stimulated
because of a fertilizer effect of nitrogen deposition. Conversely, when the deposition is greater
than the critical load, tree vigor and growth may be reduced because of the negative impacts of
soil acidification. The transition point between growth stimulation and impairment would occur
when deposition is equal to the critical load.
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 1
-------
Terrestrial Acidification Case Study
Nitrogen Deposition
J3
O
Critical Load Exceedance
Figure 1-1. Hypothesized relationships between tree growth and nitrogen deposition and
critical load exceedance. When deposition does not exceed the critical load, growth is
stimulated by nitrogen deposition. When deposition exceeds the critical load (deposition
is greater than the critical load), growth is reduced. (This figure is a modification of a
curve describing forest productivity as a function of long-term chronic nitrogen additions
outlined in Aber et al., 1995).
To assess the impacts of nitrogen and sulfur deposition on sugar maple and red spruce in
the Terrestrial Acidification Case Study, the relationships between net annual volume growth
and (1) nitrogen deposition and (2) critical load exceedance were examined empirically across
the geographical ranges of the two species.
2. SOURCE OF DATA FOR ANALYSES
Data from plots in the U.S. Forest Service (USFS) Forest Inventory and Analysis (FIA)
database used in Section 4 of the Terrestrial Acidification Case Study (Appendix 5) were applied
to this assessment. Highest protection level critical loads (Bc/Al = 10.0), 2002 CMAQ/NADP
nitrogen and sulfur deposition, and USFS FIA database net annual individual tree volume
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 2
September 2009
-------
Terrestrial Acidification Case Study
growth10 and tree volume11 for all live sugar maple and red spruce trees in each plot were used in
the analyses. Critical load exceedances and nitrogen deposition were determined using the 2002
CMAQ/NADP modeled nitrogen and sulfur deposition estimates. The tree volumes and annual
growth values were from the most recent FIA measurement period, and the interval between
measurements (to determine growth rates) for the plots ranged from 1 to 11 years. Trees with
negative growth values (but that were not dead) were included in the analyses to account for the
potential indirect impacts of nitrogen and sulfur deposition. All trees that had "0" volume values
were excluded from the analyses. Given these data restrictions, a total of 1,059 sugar maple and
419 red spruce plots were included in the tree volume growth-nitrogen deposition regression
analyses, and a total of 2,988 sugar maple and 194 red spruce plots were included in the tree
growth-critical load exceedance analyses. Volumes and volume growth for the sugar maple and
red spruce trees in each plot were averaged to produce single values of each parameter for each
species. Tables 2-1 and 2-2 summarize the plot-level FIA sugar maple and red spruce data used
to model the relationships between tree growth and nitrogen deposition and critical load
exceedance.
10 FGROWCFAL or GROWCFAL (i.e., the net annual sound cubic-foot growth of a live tree (of trees >12.7 cm
diameter). Growth values may be negative because of loss of volume due to death or damage, rot, broken top, or
other natural causes. Source: FIA database (USFS, 2008c).
11 VOLCFNET (i.e., the net cubic foot volume of wood in the central stem of a tree, beginning at 12.7 cm in
diameter or larger up the stem to a minimum 10.16 cm top outer-bark diameter. This measurement does not
include rotten, missing, and form cull. Source: FIA database (USFS, 2008c).
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 3
-------
Terrestrial Acidification Case Study
Table 2-1. Summary of Plot-Level Data Used in the Regression Analyses for Sugar Maple Volume and Growth, Nitrogen Deposition,
and Critical Load Exceedances (based on critical loads calculated with Bc/Al = 10.0)
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Total
Number
of Plots
12
8
33
25
266
8
14
242
4
27
596
257
122
72
6
280
13
55
NITROGEN DEPOSITION REGRESSION
ANALYSIS
Number of
Plots with
N+S
Deposition
Lower Than
CL
9
7
—
8
31
6
2
191
—
—
178
178
64
12
—
16
4
1
Average
Nitrogen
Deposition
(eq/ha/yr)
821.9
676.3
-
796.1
899.0
881.3
815.2
393.9
-
-
563.2
618.8
717.9
457.1
-
677.1
663.7
846.0
Average
Tree
Volume
Growth
(ni3/yr)
0.0111
0.0098
-
0.0093
0.0177
0.0114
0.0129
0.0079
-
-
0.0055
0.0074
0.0072
0.0053
-
0.0097
-0.0038
0.0033
Average
Tree
Volume
(m3)
0.5254
0.3469
-
0.2967
0.5649
0.2329
0.4126
0.2696
-
-
0.2995
0.2482
0.2422
0.1963
-
0.5243
0.6127
0.0892
CRITICAL LOAD EXCEEDANCE REGRESSION
ANALYSIS
Number of
Plots with
N+S
Deposition
Greater
Than CL
3
1
33
17
235
2
12
51
4
27
418
79
58
60
6
264
9
54
Average
Critical
Load
Exceedance
(eq/ha/yr)
434.8
105.7
487.8
145.1
439.8
48.1
412.3
130.4
599.9
473.3
242.2
156.1
171.5
378.4
601.4
437.9
299.3
554.6
Average
Tree
Volume
Growth
(ni3/yr)
0.0094
0.0098
0.0093
0.0070
0.0169
0.0053
0.0154
0.0105
0.0102
0.0029
0.0109
0.0101
0.0051
0.0088
0.0125
0.0101
0.0158
0.0115
Average
Tree
Volume
(m3)
0.1171
0.3539
0.2787
0.2009
0.3856
0.1234
0.3563
0.3229
0.3734
0.3657
0.3069
0.2555
0.2462
0.3043
0.3569
0.3443
0.3777
0.3854
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 4
September 2009
-------
Terrestrial Acidification Case Study
State
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Total Observations
(used in
calculations)
Total
Number
of Plots
270
264
162
104
337
870
4,047
NITROGEN DEPOSITION REGRESSION
ANALYSIS
Number of
Plots with
N+S
Deposition
Lower
Than CL
7
132
2
41
19
151
1,059
Average
Nitrogen
Depositio
n
(eq/ha/yr)
707.7
770.4
743.4
712.2
768.2
733.0
1,059
Average
Tree
Volume
Growth
(m3/yr)
0.0185
0.0111
0.0092
0.0130
0.0148
0.0107
1,059
Average
Tree
Volume
(m3)
0.3778
0.3179
0.2935
0.2827
0.2824
0.3078
1,059
CRITICAL LOAD EXCEED ANCE
REGRESSION ANALYSIS
Number of
Plots with
N+S
Deposition
Greater
Than CL
263
132
160
63
318
719
2,988
Average
Critical
Load
Exceedanc
e
(eq/ha/yr)
557.9
161.7
301.7
291.6
352.1
185.2
2,988
Average
Tree
Volume
Growth
(m3/yr)
0.0115
0.0128
0.0077
0.0126
0.0104
0.0087
2,988
Average
Tree
Volume
(m3)
0.3576
0.3486
0.4112
0.3012
0.2825
0.3040
2,988
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 5
September 2009
-------
Terrestrial Acidification Case Study
Table 2-2. Summary of Plot-Level Data Used in the Regression Analyses for Red Spruce Volume and Growth, Nitrogen Deposition,
and Critical Load Exceedances (based on critical loads calculated with Bc/Al = 10.0)
State
Maine
Massachusetts
New Hampshire
New York
Tennessee
Vermont
West Virginia
Total Observations
(used in calculations)
Total
Number
of Plots
483
3
42
18
1
60
6
613
NITROGEN DEPOSITION REGRESSION
ANALYSIS
Number of
Plots with
N+S
Deposition
Lower
Than CL
405
_
10
4
-
-
-
419
Average
Nitrogen
Deposition
(eq/ha/yr)
390.3
-
450.1
615.0
-
-
-
419
Average
Tree
Volume
Growth
(m3/yr)
0.0075
-
0.0060
0.0066
-
-
-
419
Average
Tree
Volume
(m3)
0.2487
-
0.1826
0.6146
-
-
-
419
CRITICAL LOAD EXCEEDANCE REGRESSION
ANALYSIS
Number of
Plots with
N+S
Deposition
Greater
Than CL
78
3
32
14
1
60
6
194
Average
Critical
Load
Exceedance
(eq/ha/yr)
133.1
628.5
369.0
282.5
420.2
292.4
257.6
194
Average
Tree
Volume
Growth
(m3/yr)
0.0074
0.0040
0.0063
0.0044
0.0196
0.0073
0.0109
194
Average
Tree
Volume
(m3)
0.2446
0.2028
0.2451
0.2210
0.4634
0.3281
0.3289
194
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 6
September 2009
-------
Terrestrial Acidification Case Study
3. REGRESSION ANALYSES METHODOLOGY AND RESULTS
The impacts of nitrogen deposition and critical load exceedance on individual tree
volume growth of sugar maple and red spruce were examined empirically using multivariate
ordinary least squares (OLS) linear regression analyses, testing the positive and negative linear
relationships represented in Figure 1-1. Plots where 2002 nitrogen and sulfur deposition did not
exceed the critical load (based on Bc/Al = 10.0 critical load estimates) were used in the tree
volume growth-nitrogen deposition analyses, and plots where deposition exceeded the critical
load were used in the regression analyses comparing critical load exceedance and volume growth
(Tables 2-1 and 2-2). In both sets of analyses for both species, the explanatory variables
included the linear term of nitrogen deposition or critical load exceedance (both expressed as
equivalents per hectare per year [eq/ha/year]) for each plot, linear and squared terms of average
tree volumes in cubic meters (m3), and a categorical (dummy) variable for each State (with
Arkansas and Connecticut arbitrarily selected as the reference categories for sugar maple for the
nitrogen deposition and critical load exceedance regressions, respectively. For red spruce, New
York was selected for the nitrogen deposition regression, and Vermont was selected for the
critical load exceedance regression). Tree volume was included as an explanatory variable of tree
growth because tree age, the preferred explanatory variable, was not available for this dataset.
Tree age and tree volume are highly correlated, and volume growth is influenced by tree size, so
tree volume was seen as an appropriate surrogate explanatory variable. The State variables were
included in the analyses to control for unobserved sources of variation in tree growth related to a
plot's general geographic location. Examples of potential unobserved factors include differences
in data collection methods and measurements across reporting State, climatic factors, and
geological characteristics.
The results of the linear regression analyses to test the influences of nitrogen deposition
and critical load exceedance on sugar maple and red spruce volume growth are presented in
Tables 3-la-b and 3-2a-b, respectively
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 7
-------
Terrestrial Acidification Case Study
Table 3-la. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses
of Sugar Maple Tree Growth and Nitrogen Deposition (for plots where deposition did not exceed
critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Nitrogen Deposition
Average Tree Volume
Square of Average Tree Volume
Alabama
Illinois
Indiana
Iowa
Kentucky
Maine
Michigan
Minnesota
Missouri
New Hampshire
New York
North Carolina
Ohio
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth (m3/yr)
Coefficient
-0.00278
1.22xlO'5
0.01296
-8.32 xlO'4
-0.00242
-0.00131
0.00338
5.19xlO'4
6.37 xlO'4
0.00249
-0.00238
-4.20 xlO'4
-0.00167
-3.34 xlO'6
-0.00190
-0.01658
-0.00535
0.00790
5.46xlO'4
-7.88 x 10'4
0.00360
0.00468
6.76 xlO'4
t-statistic
-0.53
3.22
6.44
-1.39
-0.39
-0.21
0.65
0.08
0.07
0.52
-0.51
-0.09
-0.34
0.00
-0.35
-2.17
-0.41
1.22
0.12
-0.08
0.72
0.87
0.14
p-value
0.5965
0.0013
O.0001
0.1645
0.6941
0.8350
0.5139
0.9392
0.9479
0.6034
0.6134
0.9286
0.7302
0.9995
0.7301
0.0299
0.6812
0.2241
0.9080
0.9356
0.4693
0.3847
0.8858
1,059
0.1175
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 8
September 2009
-------
Terrestrial Acidification Case Study
Table 3-lb. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses
of Red Spruce Tree Growth and Nitrogen Deposition (for plots where deposition did not exceed
critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Nitrogen Deposition
Tree Volume
Square of Tree Volume
Maine
New Hampshire
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth (m3/yr)
Coefficient
0.00704
-6.69 xlO'6
0.00955
-0.00528
0.00121
4.5 x 10'4
t-statistic
1.57
-1.12
3.72
-2.31
0.44
0.15
p-value
0.1166
0.2679
0.0002
0.0215
0.6626
0.8825
419
0.0351
Table 3-2a. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses
of Sugar Maple Tree Growth and Critical Load Exceedance (for plots where nitrogen and sulfur
deposition was greater than critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Critical Load Exceedance
Average Tree Volume
Square of Average Tree Volume
Alabama
Illinois
Indiana
Iowa
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
Dependent Variable: Average Tree Growth
(m3/yr)
Coefficient
0.00430
-1.43 x 10'6
0.02001
2.58 x 10'4
0.00341
-0.00112
0.00544
-0.00142
0.00449
-1.12x 10'4
-7.89x 10'4
-0.00814
7.77 x 10'4
8.48 x 10'4
-0.00389
t-statistic
1.35
-0.92
10.81
0.34
0.32
-0.21
1.65
-0.11
0.74
-0.03
-0.08
-1.76
0.24
0.23
-1.00
p-value
0.1787
0.3571
O.OOOl
0.7365
0.7523
0.8347
0.0991
0.9135
0.4564
0.9776
0.9337
0.0786
0.8095
0.8194
0.3192
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 9
September 2009
-------
Terrestrial Acidification Case Study
Explanatory Variables
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth
(m3/yr)
Coefficient
-0.00120
0.00190
-5.42x 10'4
0.00431
1.85 x 10'4
8.33x 10'4
0.00172
-0.00448
0.00261
9.14x 10'4
-0.00151
t-statistic
-0.31
0.24
-0.17
0.64
0.05
0.25
0.49
-1.32
0.68
0.28
-0.48
p-value
0.7556
0.8108
0.8685
0.5226
0.9626
0.7994
0.6225
0.1879
0.4957
0.7783
0.6347
2,988
0.1329
Table 3-2b. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses
of Red Spruce Tree Growth and Critical Load Exceedance (for plots where nitrogen and sulfur
deposition was greater than critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Critical Load Exceedance
Tree Volume
Square of Tree Volume
Maine
Massachusetts
New Hampshire
New York
West Virginia
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth
(m3/yr)
Coefficient
0.00628
-5.62xlO'6
0.00611
0.00463
1.26xlO'5
-1.73 xlO'4
2.68 x 10'4
-0.00200
0.00331
t-statistic
5.15
-2.30
1.38
1.11
0.01
-0.06
0.25
-1.43
1.64
p-value
O.OOOl
0.0223
0.1685
0.2705
0.9890
0.9524
0.7993
0.1549
0.1020
194
0.2009
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 10
September 2009
-------
Terrestrial Acidification Case Study
The results of the linear regressions analyzing the relationships between nitrogen
deposition and tree volume growth differed by species. For sugar maple, nitrogen deposition had
a positive and significant (at the/><0.05 level, /?=0.00013) impact on the aboveground volume
growth of individual trees (Table 3-la). These results suggest that the 2002 nitrogen additions,
that ranged from 332 to 1,146 eq/ha/yr (4.7 to 16.1 kg N/ha/yr), acted as a fertilizer and
stimulated growth on sites where nitrogen and sulfur deposition did not exceed the critical load.
On these sites, nitrogen most likely posed a limitation to sugar maple growth. There are few
studies that have demonstrated a positive growth response of sugar maple to nitrogen additions.
Stone (1986) found that sugar maple diameter growth was not impacted by nitrogen 100 to 300
kg N/ha fertilizers applied three years before or after thinning, and similar results have been
determined in other studies (see references in Stone, 1986). In contrast to sugar maple, the red
spruce in the tree growth-nitrogen deposition analyses displayed a negative, nonsignificant (at
the/XO.05 level, p=0.2679) relationship between deposition and growth (Table 3-lb). These
results suggest that red spruce growth may not have been limited by nitrogen on sites where the
critical load was not exceeded by nitrogen and sulfur deposition; nitrogen deposition in the range
of 255 to 667 eq/ha/yr (3.6 to 9.3 kg N/ha/yr) did not stimulate the growth of red spruce. Briggs
et al. (2000) reported that diameter growth of red spruce was unchanged by the application of
200 kg N/ha on pre-commercially thinned spruce-fir stands in Maine. McNulty et al. (2005)
found that although ambient nitrogen deposition of 5.4 kg N/ha/yr over the course of 14 years
was associated with a 20% increase in basal area of a red spruce-dominated spruce-fir forest,
supplemental additions of 15.4 kg N/ha/yr resulted in a 20% decrease in live basal area.
Therefore, the documented response of red spruce to nitrogen additions is mixed.
The results of the analyses determining the relationships between critical load exceedance
and tree growth showed similar trends for both sugar maple and red spruce; growth declined with
increasing levels of critical load exceedance. For red spruce, the relationship was significant at
the 5% level (p=0.0223) and negative, indicating that red spruce growth may be negatively
impacted by deposition levels that exceed the critical load (Table 3-2b). Decreased basal area of
red spruce in response to chronic additions of nitrogen have been reported in several other
studies (Aber et al., 1995; McNulty et al., 2005), and authors of these studies speculated that
these negative growth responses were due to sites reaching a level of nitrogen saturation.
Although the tree growth-critical load exceedance relationship was also negative for sugar
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 11
-------
Terrestrial Acidification Case Study
maple, it was not statistically significant at the 5% level (p=0.3571), suggesting that sugar maple
may not be negatively impacted by nitrogen and sulfur deposition levels greater than critical
loads (Table 3-2a). However, shortcomings of at least one of the variables used within the 8MB
calculations of critical loads may, in part, have contributed to the lack of a significant
relationship between tree growth and critical load exceedance. As discussed in Appendix 5, the
base cation weathering (BCW) term that accounts for the contributions of base cations from the
weathering of soil minerals and parent material is one of the most influential terms in the 8MB
model. Li and McNulty (2007) determined that 49% of the variability in critical load estimates
was due to this term. Within the United States and Canada, BCW for critical load assessments is
commonly estimated using the clay-substrate model (Ouimet et al., 2006; Watmough et al.,
2006; McNulty et al., 2007; Pardo and Duarte, 2007), and this model was also used in the
Terrestrial Acidification Case Study. Critical load experts from both the United States and
Canada have commented that the clay-substrate model method performs well in young soils that
have formed since the last glaciation (20,000 years before present (ybp)), but may not be suitable
for older, more weathered soils south of the most recent glacial advancement. Therefore, to
account for the potential influence of a poorer estimation of BCW in plots south of the glaciation
line (Figure 3-1), a second set of multivariate, OLS linear regression analyses were conducted,
using the same specifications (i.e., plots where nitrogen and sulfur deposition exceeded the
critical load calculated with Bc/Al=10.0), as described above. However, these analyses were
restricted to data from plots that had been covered by the last glaciation (i.e., north of the
glaciation line). Limiting the analysis to plots north of the glaciation line led to analyzing 4%
fewer red spruce plots and 26% fewer sugar maple plots (Tables 3-3 and 3-4).
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 12
-------
Terrestrial Acidification Case Study
Approximate area affected by most recent glaciation
Figure 3-1. Areas of the continental United States that were covered during the last
glacial event (~ 20,000 ybp) (Reed and Bush, 2005).
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 13
September 2009
-------
Terrestrial Acidification Case Study
Table 3-3. Summary of Plot-Level Data for Sugar Maple Volume and Growth and Critical Load
Exceedances North of the Glaciation Line (for plots where nitrogen and sulfur deposition
exceeded critical loads calculated with Bc/Al = 10.0)
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
Total Observations
(used in calculations)
Total
Number of
Plots with
N+S
Deposition
Greater
Than CL
3
1
33
17
235
2
12
51
4
27
418
79
58
60
6
264
9
54
263
132
160
63
318
719
2,988
Number of
Plots North
of the
Glaciation
Line with
N+S
Deposition
Greater
Than CL
0
0
33
12
204
2
0
51
0
27
418
79
18
60
6
264
0
26
126
0
160
0
0
719
2,205
Average CL
Exceedance
(eq/ha/yr)
NA
NA
487.76
117.17
390.90
48.07
NA
130.42
NA
473.31
242.17
156.06
84.02
378.40
601.39
437.94
NA
452.60
387.35
NA
301.67
NA
NA
185.16
2,205
Average
Tree Volume
Growth
(m3/yr)
NA
NA
0.009
0.007
0.018
0.005
NA
0.011
NA
0.003
0.011
0.010
0.012
0.009
0.013
0.010
NA
0.013
0.011
NA
0.008
NA
NA
0.009
2,205
Average
Tree Volume
(m3)
NA
NA
0.279
0.227
0.397
0.123
NA
0.323
NA
0.366
0.307
0.256
0.246
0.304
0.357
0.344
NA
0.545
0.366
NA
0.411
NA
NA
0.304
2,205
NA = not applicable. Average Critical Load Exceedance, Average Tree Volume Growth, and Average
Tree Volume values could not be determined because there were no sugar maple plots north of the
glaciation line.
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 14
September 2009
-------
Terrestrial Acidification Case Study
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kentucky
Maine
Maryland
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
TOTAL Observations
(used in calculations)
Total
Number of
Plots with
N+S
Deposition
Greater
than CL
3
1
33
17
235
2
12
51
4
27
418
79
58
60
6
264
9
54
263
132
160
63
318
719
2988
Number of
Plots with
N+S
Deposition
Greater than
CL North of
Glaciation
Line
0
0
33
12
204
2
0
51
0
27
418
79
18
60
6
264
0
26
126
0
160
0
0
719
2205
Average CL
Exceedance
(eq/ha/yr)
NA
NA
487.76
117.17
390.90
48.07
NA
130.42
NA
473.31
242.17
156.06
84.02
378.40
601.39
437.94
NA
452.60
387.35
NA
301.67
NA
NA
185.16
2205
Average
Tree
Volume
Growth
(m3/yr)
NA
NA
0.009
0.007
0.018
0.005
NA
0.011
NA
0.003
0.011
0.010
0.012
0.009
0.013
0.010
NA
0.013
0.011
NA
0.008
NA
NA
0.009
2205
Average
Tree
Volume (m3)
NA
NA
0.279
0.227
0.397
0.123
NA
0.323
NA
0.366
0.307
0.256
0.246
0.304
0.357
0.344
NA
0.545
0.366
NA
0.411
NA
NA
0.304
2205
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 15
September 2009
-------
Terrestrial Acidification Case Study
Table 3-4. Summary of Plot-Level Data for Red Spruce Volume and Growth and Critical Load
Exceedances North of the Glaciation Line (for plots where nitrogen and sulfur deposition
exceeded critical loads calculated with Bc/Al = 10.0)
State
Maine
Massachusetts
New Hampshire
New York
Tennessee
Vermont
West Virginia
Total Observations
(used in
calculations)
Total
Number of
Plots with
N+S
Deposition
Greater
Than CL
78
3
32
14
1
60
6
194
Number of
Plots with
N+S
Deposition
Greater
Than CL
North of
Glaciation
Line
78
3
32
14
0
60
0
187
Average CL
Exceedance
(eq/ha/yr)
133.1048
628.5439
368.9527
282.5433
NA
292.4329
NA
187
Average
Tree
Volume
Growth
(m3/yr)
0.007
0.004
0.006
0.004
NA
0.007
NA
187
Average
Tree
Volume
(m3)
0.245
0.203
0.245
0.221
NA
0.328
NA
187
NA = not applicable. Average Critical Load Exceedance, Average Tree Volume Growth, and Average
Tree Volume values could not be determined because there were no red spruce plots north of the
glaciation line.
The results from the linear regression analyses for sugar maple and red spruce, north of
the glaciation line, are presented in Tables 3-5 and 3-6, respectively. Similar to the first linear
regression analyses that included all critical load exceedance plots, the relationship between tree
growth and critical load exceedance was negative for both species and was statistically
significant at the 5% level (p-value of 0.0354) for red spruce. However, in contrast to the first
linear regression analyses, the relationship for the sugar maple plots north of the glaciation line
was significant at the 10% level (p-value of 0.1008), and occurred despite the 26% reduction of
plots used in the analysis. These results suggest that a larger portion of the variation in sugar
maple growth could be accounted for by critical load exceedance when the analysis was
restricted to plots north of the glaciation line and lend support to the hypothesis that the clay-
substrate model to estimate BCW may not be suitable for older soils. Future assessment of critical
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 16
September 2009
-------
Terrestrial Acidification Case Study
loads may, therefore, want to consider exploring other BCW models to estimate critical loads of
atmospheric nitrogen and sulfur deposition. Models such as PROFILE (Sverdrup and Warfvinge,
1993 a), which are based on soil mineralogy, may provide better estimates of the contribution of
base cations from soil weathering.
Table 3-5. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses of
Sugar Maple Tree Growth and Critical Load Exceedance North of the Glaciation Line (for plots
where deposition exceeded critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Critical Load Exceedance
Average Tree Volume
Square of Average Tree Volume
Illinois
Indiana
Iowa
Maine
Massachusetts
Michigan
Minnesota
Missouri
New Hampshire
New lersey
New York
Ohio
Pennsylvania
Vermont
Wisconsin
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth (m3/yr)
Coefficient
0.004875
-3.344 xlO'6
0.021150
8.944 x 10'4
-0.001884
0.005452
-0.002052
-0.000895
-0.008403
0.000222
0.000210
0.001850
-0.001647
0.001956
-0.000817
-0.002104
-0.000803
-0.005168
-0.002195
t-statistic
1.48
-1.64
10.12
1.1
-0.31
1.63
-0.16
-0.22
-1.82
0.07
0.06
0.35
-0.43
0.25
-0.25
-0.45
-0.23
-1.51
-0.68
p-value
0.1385
0.1008
<0001
0.27
0.755
0.1029
0.8743
0.8245
0.0685
0.9456
0.9553
0.7255
0.6696
0.8042
0.8035
0.6522
0.8177
0.131
0.4958
2,205
0.1722
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 17
September 2009
-------
Terrestrial Acidification Case Study
Table 3-6. Results from the Multivariate Ordinary Least Squares Linear Regression Analyses of
Red Spruce Tree Growth and Critical Load Exceedance, North of the Glaciation Line (for plots
where deposition exceeded critical loads calculated with Bc/Al = 10.0)
Explanatory Variables
Intercept
Critical Load Exceedance
Tree Volume
Square of Tree Volume
Maine
Massachusetts
New Hampshire
New York
Number of Observations
Adjusted R2
Dependent Variable: Average Tree Growth (m3/yr)
Coefficient
0.006034
-5.162X10'6
0.005590
S.lOOxlO'3
0.000285
-0.000132
0.000435
-0.001805
t-statistic
4.96
-2.12
1.26
1.23
0.32
-0.05
0.42
-1.32
p-value
<.0001
0.0354
0.2093
0.2218
0.7489
0.9629
0.6736
0.1897
187
0.1963
4. ADDITIONAL SOURCES OF VARIABILITY INFLUENCING THE
RELATIONSHIPS BETWEEN TREE VOLUME GROWTH AND
NITROGEN DEPOSITION AND CRITICAL LOAD EXCEEDANCE
In addition to potential inaccuracy in the BCW variable in the estimation of critical loads,
there are additional sources of variation that may have influenced the relationships between
critical load exceedance and the growth of red spruce and sugar maple. These sources of
variability may have also influenced the relationships between tree growth and nitrogen
deposition.
4.1 State-Specific Variables
Although State dummy variables and current tree volume were included as covariates in
the regression analyses to account for the influences of location and tree size on tree growth,
additional factors could be included in future regression analyses. For example, incorporating
latitude/longitude and elevation in analyses could remove the influence of location on tree
growth. Similarly, site index could remove the influence of site quality on the growth of red
spruce and sugar maple. Total basal area of the stand would remove the influence of stand
Final Risk and Exposure Assessment
Appendix 5, Attachment A - 18
September 2009
-------
Terrestrial Acidification Case Study
density and site occupancy. Measurement year and time between measurements could help
remove the influence of year-to-year variation in conditions and measurement methodology on
the growth data. Climatic variation (e.g., rainfall, temperature) could be included to account for
the influence of drought and frost conditions (McNulty and Boggs, in press) on tree growth. It is
recommended that future analyses comparing tree growth and nitrogen deposition or critical load
exceedance take additional sources of variability into account. Much of this data is included in
the FIA database.
4.2 Dead Trees
The regression relationship between tree growth and nitrogen deposition or critical load
exceedance may also be improved with the inclusion of dead trees in the analyses. In the FIA
database, when a tree is recorded as dead for the first time, the total volume of that tree is
considered negative volume growth over the most recent measurement period.12 As described
earlier, atmospheric deposition of nitrogen and sulfur can indirectly result in tree mortality.
Therefore, it may be appropriate to include tree mortality in an evaluation of the relationship
between tree growth and nitrogen deposition or critical load exceedance. However, the validity
of including dead tree negative volume growth measurements (as calculated by the FIA
database) in the regression analyses was uncertain, and, therefore, the inclusion of dead trees was
not pursued in the analyses reported in this Attachment.
4.3 Other Factors
The FIA sugar maple and red spruce tree data used in the analyses may have also
introduced variability and a source of error in the analyses. As discussed in Appendix 5, due to
restriction factors, not all sugar maple and red spruce plots were included in the analyses. It is
uncertain to what degree, if any, these restrictions may have biased the results. The influence of
12 Dead tree volume growth is calculated as a difference in volumes (v2-vl) divided by the time between sequential
measurement period (t2-tl). When a tree is recorded as dead, it is assigned a v2 value of "0." Therefore, the
associated volume growth is equal to the entire tree volume divided by the difference in the number of years
between the current and last measurement cycle.
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 19
-------
Terrestrial Acidification Case Study
tree ingress13 may also not have been completely accounted for in the analyses and may have
introduced another source of error. According to USFS FIA database methodology, trees must be
at least 12.7 centimeters (cm) in diameter to be included in the VOLCFNET (total net volume
per tree) tree volume table. When they reach this size, the full volume of the stem is incorporated
into the volume growth measurements, and, in many cases, these measurements would be larger
than the actual annual growth rates. Trees with 0 m3 VOLCFNET volume values were excluded
from the analyses to at least partially account for this influence. With the data that were made
available for the analyses, it was not possible to determine other possible ingress trees. The use
of VOLCFNET tree volume as a covariate of FGROWCFAL (net annual sound cubic-foot
growth of a live tree on forest land) tree growth may have also introduced a small source of
error. VOLCFNET is based on merchantable volume (e.g., pulp and sawlog), whereas
FGROWCFAL is based on growth of sound wood. The difference between the two
measurements is cull wood that is sound but not merchantable due to circumstances such as the
location on the tree or tree branchiness. The removal of trees with 0 m3 VOLCFNET volume
from the analyses removed at least a portion of the trees that would increase the influence of cull
wood on the covariate relationship between the two variables. Measurement error may also have
introduced another source of error to the analyses. Tree volumes and volume growth are based
on the measurements conducted on the main stem of the tree. Slight differences in measurements
conducted by different crews and in different years could have introduced some error to the
volume and growth estimates. Based on the FIA data provided by the USFS, it was not possible
to determine the degree to which each of these various source of error influenced the data, nor
was it possible to determine if reduction or elimination of these sources of error would change
and/or improve the regression analyses. Attempts to minimize these potential sources of error are
recommended in future analyses of the relationships between tree growth and nitrogen deposition
and critical load exceedance.
13 "Ingress" refers to trees that appear that are included in the dataset for the first time. Usually ingress is a result of
trees reaching a certain size, but occasionally, ingress can also occur when new trees are included in the
measurements. Steps are taken to identify these new trees and estimate previous size and growth, but sometimes
they may be missed.
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 20
-------
Terrestrial Acidification Case Study
5. CONCLUSIONS
In conclusion, the results from these analyses do suggest that nitrogen and sulfur
deposition can result in either a positive or negative impact of tree volume growth. When
deposition levels do not exceed the critical load of a site, tree growth may be stimulated by the
nitrogen additions, as was seen in the sugar maple analyses. Conversely, when critical loads are
exceeded by nitrogen and sulfur deposition, tree growth may be negatively impacted. This
negative growth trend was seen in both the red spruce and sugar maple (restricted to plots North
of the glaciation line). The trends suggested by these results indicate that large-scale analyses
using the USFS FIA database may provide a useful tool with which to examine both the positive
and negative impacts of nitrogen and sulfur deposition on tree species in the United States.
Final Risk and Exposure Assessment September 2009
Appendix 5, Attachment A - 21
-------
-------
September 2009
Appendix 6
Aquatic Nutrient Enrichment Case Study
Final
EPA Contract Number EP-D-06-003
Work Assignment 3-62
Project Number 0209897.003.062
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27709
Prepared by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
INTERNATIONAL
-------
-------
Aquatic Nutrient Enrichment Case Study
TABLE OF CONTENTS
Acronyms and Abbreviations v
1.0 Background 1
1.1 Indicators, Ecological endpoints, and Ecosystem Services 4
1.2 Case Studies 9
1.2.1 National Overview of Sensitive Areas 9
1.2.2 Use of ISA Information and Rationale for Site Selection 12
1.2.3 Potomac River and Potomac Estuary 17
1.2.4 Neuse River andNeuse River Estuary 20
2.0 Approach and Methods 26
2.1 Modeling 27
2.2 Chosen Method 30
2.2.1 SPARROW 31
2.2.1.1 Background and Description 31
2.2.1.2 Key Definitions for Understanding SPARROW Modeling 36
2.2.1.3 Concepts of Importance to Case Study—SPARROW
Application 38
2.2.2 ASSETS Eutrophication Index 39
2.2.2.1 Background and Description 39
2.2.2.2 Applications and Updates 46
2.2.3 Assessments Using Linked SPARROW and ASSETS El 46
2.2.3.1 Back Calculation Method 52
2.2.3.2 Uncertainty Bounds on TNS, OEC Min/Max Values 53
3.0 Re suits 59
3.1 Current Conditions 59
3.1.1 Summary of Results for the Potomac River/Potomac Estuary Case Study
Area 59
3.1.1.1 SPARROW Assessment 60
3.1.1.2 ASSETS El Assessment 66
3.1.2 Summary of Results for the Neuse River/Neuse River Estuary Case
Study Area 70
3.1.2.1 SPARROW Assessment 70
3.1.2.2 ASSETS El Assessment 77
3.2 Alternative Effects Levels 80
3.2.1 Potomac River Watershed 80
3.2.2 Neuse River Watershed 89
4.0 Implications for Other Systems 98
5.0 Uncertainty 101
6.0 Conclusions 105
7.0 References 105
Final Risk and Exposure Assessment September 2009
Appendix 6 - i
-------
Aquatic Nutrient Enrichment Case Study
LIST OF FIGURES
Figure 1.1-1. Descriptions of the five eutrophication indicators used in NOAA' s NEEA
(Bricker et al., 2007a) 6
Figure 1.1-2. A simplified schematic of eutrophication effects on an aquatic ecosystem 7
Figure 1.1-3. An illustrated representation of eutrophication measures through the use of
indicators and influencing factors from NOAA's NEEA (Bricker et al.,
2007a) 7
Figure 1.2-1. The relationship between mean dissolved inorganic nitrogen concentration
and mean wet inorganic nitrogen in unproductive lakes in different regions
in North America and Europe (Bergstrom and Jansson, 2006) 11
Figure 1.2-2. Areas potentially sensitive to aquatic nutrient enrichment 12
Figure 1.2-3. The Potomac River Watershed and Potomac Estuary 18
Figure 1.2-4. TheNeuse River Watershed andNeuse River Estuary 23
Figure 2.2-1. Modeling methodology for case study 31
Figure 2.2-2. Mass balance description applied to the SPARROW model formulation 32
Figure 2.2-3. Conceptual illustration of a reach network 33
Figure 2.2-4. SPARROW model components (Schwarz et al., 2006) 35
Figure 2.2-5. Influencing factors/Overall Human Influence index description and decision
matrix (Bricker et al., 2007a) 40
Figure 2.2-6. Overall Eutrophic Condition index description and decision matrix (Bricker
etal., 2007a) 43
Figure 2.2-7. Detailed descriptions of primary and secondary indicators of eutrophication
(Bricker et al., 2007a) 44
Figure 2.2-8. Determined Future Outlook index description and decision matrix (Bricker
etal., 2007a) 45
Figure 2.2-9. Example response curve of instream total nitrogen concentrations to
atmospheric deposition loads 48
Figure 2.2-10. Example of response for case study analysis (Bricker et al., 2007b) 49
Figure 2.2-11. ASSETS El response curve 51
Figure 2.2-12a. Back calculation analysis scenario A: no uncertainty 55
Figure 2.2-12b. Back calculation analysis scenario B: uncertainty in ASSETS El
assessment 56
Figure 2.2-12c. Back calculation analysis scenario C: uncertainty in both ASSETS El
assessment and nitrogen loading assessment 57
Figure 2.2-13. Example of an improvement by one ASSETS El score category in a back
calculation assessment 58
Figure 2.2-14. Example for resulting change in atmospheric nitrogen loads due to
improvement in ASSETS El score in back calculation assessment 59
Figure 3.1-1 a. Atmospheric deposition yields of oxidized nitrogen over the Potomac
River and Potomac Estuary watershed 62
Figure 3.1-lb. Atmospheric deposition yields of reduced nitrogen over the Potomac River
and Potomac Estuary watershed 63
Figure 3.1-lc. Atmospheric deposition yields of total nitrogen over the Potomac River
and Potomac Estuary watershed 64
Final Risk and Exposure Assessment September 2009
Appendix 6 - ii
-------
Aquatic Nutrient Enrichment Case Study
Figure 3.1-2. Total nitrogen yields from all sources as predicted using the Version 3 of
the Chesapeake Bay SPARROW application with updated 2002
atmospheric deposition inputs 65
Figure 3.1-3. Source contributions to Potomac Estuary nitrogen load 66
Figure 3.1-4. The ASSETS El scores for the Potomac Estuary (Bricker et al., 2006) 68
Figure 3.1-5a. Atmospheric deposition yields of oxidized nitrogen over the Neuse River
andNeuse River Estuary watershed 72
Figure 3. l-5b. Atmospheric deposition yields of reduced nitrogen over the Neuse River
and Neuse River Estuary watershed 73
Figure 3. l-5c. Atmospheric deposition yields of total nitrogen over the Neuse River and
Neuse River Estuary watershed 74
Figure 3.1-6. Total nitrogen yields from all sources in the Neuse River watershed as
predicted by a SPARROW modeling application for the Neuse, Tar-
Pamlico, and Cape Fear rivers' watersheds with 2002 data inputs 76
Figure 3.1-7. Source contributions to Neuse River Estuary total nitrogen load 77
Figure 3.2-1. Response curve relating instream total nitrogen concentration to total
nitrogen atmospheric deposition load for the Potomac River watershed 81
Figure 3.2-2. Fitted Overall Eutrophic Condition curve for target ASSETS EI=2, median
TNatm*i (i = run 280) 87
Figure 3.2-3. Response curve relating instream total nitrogen concentration to total
nitrogen atmospheric deposition load for the Neuse River/Neuse River
Estuary Case Study Area 90
Figure 3.2-4. Fitted Overall Eutrophic Condition curve for target ASSETS EI=2, median
TNatm'i (i = run 287) 95
Figure 3.2-5. Theoretical SPARROW response curves demonstrating relative influence of
sources on nitrogen loads to an estuary 97
Figure 4-1. Preliminary classifications of estuary typology across the nation (modified
from Bricker et al., 2007a) 100
Final Risk and Exposure Assessment September 2009
Appendix 6 - iii
-------
Aquatic Nutrient Enrichment Case Study
LIST OF TABLES
Table 1.1-1. Key Indicators of Nutrient Enrichment Due toNr, Including NOX 5
Table 1.1-2. Assessment Ecological Endpoints for Nutrient Enrichment Due to
Deposition of Total Reactive Nitrogen, Including NOX 7
Table 1.1-3. Ecosystem Services for Aquatic Systems Affected by Nutrient Enrichment 8
Table 1.2-1. Summary of Indicators, Mapping Layers, and Models for Targeted
Ecosystems 10
Table 1.2-2. Nitrogen Deposition Level vs. EPA Total Nitrogen Criteria for Lakes and
Reservoirs 11
Table 1.2-3. Science Advisory Board/Ecological Effects Subcommittee Listing of
Potential Assessment Areas for Evaluation of Benefits of Decreases in
Atmospheric Deposition with Respect to Aquatic Nutrient Enrichment 14
Table 1.2-4. Potential Assessment Areas for Aquatic Nutrient Enrichment Identified in
the ISA 15
Table 1.2-5. Physical Characteristics of the Potomac Estuary 18
Table 1.2-6. Hydrological Characteristics of the Potomac Estuary 19
Table 1.2-7. Physical Characteristics of the Neuse River Estuary 23
Table 1.2-8. Neuse River Watershed Land Use and Population 24
Table 1.2-9. Hydrological Characteristics of the Neuse River Estuary 25
Table 2.1-1. Examples of SPARROW Applications 36
Table 3.1-1. Potomac Estuary Current Condition Overall Human Influence Index Score 69
Table 3.1-2. Model Parameters for 2002 Current Condition SPARROW Application for
the Neuse River Watershed 74
Table 3.1-3. Model Evaluation Statistics for 2002 Current Condition SPARROW
Application for the Neuse River Watershed 75
Table 3.1-4. Current Condition Overall Eutrophic Condition Index Score for the Neuse
River/Neuse River Estuary Case Study Area 79
Table 3.2-1. Potomac River Watershed Alternative Effects Levels 81
Table 3.2-2. Historical Potomac River Total Nitrogen Loads and Concentrations 82
Table 3.2-3. Additional Potomac Estuary Overall Eutrophic Condition Index Scores for
Alternative Effects Levels 83
Table 3.2-4. Summary Statistics for Target ASSETS El Scenarios for the Potomac
Estuary 87
Table 3.2-5. Neuse River/Neuse River Estuary Case Study Area Alternative Effects
Levels 90
Table 3.2-6. Annual Average Instream Total Nitrogen Concentrations in the Neuse River 91
Table 3.2-7. Additional Neuse River Estuary Overall Eutrophic Condition Index Scores
for Alternative Effects Levels 92
Table 3.2-8. Summary Statistics for Target ASSETS El Scenarios for the Neuse
River/Neuse River Estuary Case Study Area 95
Table 4-1. Typology Group Categorizations 99
Final Risk and Exposure Assessment September 2009
Appendix 6 - iv
-------
Aquatic Nutrient Enrichment Case Study
ACRONYMS AND ABBREVIATIONS
ASSETS El
AQUATOX
CASTNet
CBP
CE-QUAL
CMAQ
C02
DFO
DIN
EDA
EES
EMAP
GIS
GT/MEL
HAB
HNO3
HSPF
INCA
ISA
km
km2
LANDSCAN
m
m2
m3
MAGIC
MEA
mg/L
mi2
MODMON
N2O
NADP
NCCR
NCDWQ
NEEA
NH3
NH4+
NOAA
NO
NO2
N03
NOX
Nr
Assessment of Estuarine Trophic Status eutrophication index
a simulation model for aquatic systems
Clean Air Status and Trends Network
Chesapeake Bay Program
a mathematical model of water quality
Community Multiscale Air Quality model.
carbon dioxide
determined future outlook
dissolved inorganic nitrogen
Estuarine Drainage Areas
Ecological Effects Subcommittee
Environmental Monitoring and Assessment Program
geographic information systems
Georgia Tech Hydrologic Model/Multiple Element Limitation
harmful algal bloom
nitric acid
Hydrologic Simulation Program-FORTRAN
Integrated Nitrogen in Catchments
Integrated Science Assessment
kilometer
square kilometer
a worldwide population database
meter
square meter
cubic meter
Model of Acidification of Groundwater in Catchments
Millennium Ecosystem Assessment
milligrams per liter
square mile
Modeling and Monitoring Project
nitrous oxide
National Atmospheric Deposition Program
National Coastal Condition Report
North Carolina Department of Water Quality
National Estuarine Eutrophi cation Assessment
ammonia gas
ammonium
National Oceanic and Atmospheric Administration
nitric oxide
nitrogen dioxide
nitrate
nitrogen oxides
reactive nitrogen
Final Risk and Exposure Assessment
Appendix 6 - v
September 2009
-------
Aquatic Nutrient Enrichment Case Study
OEC
OKI
psu
PnET-BCG
QA
QC
QUAL2K
RCA/ECOMSED
RF1
RHESSys
RTI
SAGT
SAV
sox
SPARROW
STORE!
TN
TNatm
TNS
ug/L
USGS
VIF
VIMS
WASP
WATERSN
Overall Eutrophic Condition
Overall Human Influence
practical salinity unit
biogeochemical model
quality assurance
quality control
Enhanced Stream Water Quality Model
Row Column AESOP/Estuary and Coastal Ocean Model with Sediment
Transport
Reach File version 1
Regional Hydro-Economic Simulation System
RTI International
Atlantic and the eastern Gulf of Mexico, as well as the Tennessee River
basin
submerged aquatic vegetation
sulfur oxides
Spatially Referenced Regression on Watershed
STOrage and RETrieval
total nitrogen
total nitrogen atmospheric deposition load
instream total nitrogen concentration
microgram per liter
U.S. Geological Survey
Variance Inflation Factor
Virginia Institute of Marine Science
Water Quality Analysis Simulation Program
Watershed Assessment Tool for Evaluating Reduction Strategies for
Nitrogen
Final Risk and Exposure Assessment
Appendix 6 - vi
September 2009
-------
Aquatic Nutrient Enrichment Case Study
1.0 BACKGROUND
One classification of effects targeted for this Risk and Exposure Assessment is nitrogen
and sulfur enrichment of ecosystems in response to deposition of nitrogen oxides (NOX) and
sulfur oxides (SOX). Nutrient enrichment effects are caused by nitrogen or sulfur deposition, but
are dominated by nitrogen deposition, which is the focus of this case study. Nutrient enrichment
can result in eutrophi cation in aquatic systems (see Section 4.3 of'the Integrated Science
Assessment (ISA) for Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report) (ISA)
(U.S. EPA, 2008a).
Because ecosystems may respond differently to nutrient enrichment, it is necessary to
first perform Risk and Exposure Assessment case studies unique to the effect and ecosystem
type. The feasibility of consolidating the effects and/or ecosystems in the Risk and Exposure
Assessment was assessed, and where feasible, a broader characterization was performed.
However, some ecosystems and their effects may be too unique to consolidate into a broad
characterization.
Upon completion of all risk and exposure assessment case studies, the results of the
assessments performed for unique combinations of effects and ecosystem types are presented
together to facilitate decision making on the total effects of nitrogen and sulfur deposition.
Ecosystem services that relate to the effects are identified and valued, if possible. Ecosystem
services provide an additional way to compare effects across various ecosystems.
The selection and performance of case studies represent Steps 3 and 4, respectively, of
the seven-step approach to planning and implementing a risk and exposure assessment, as
presented in the April 2008 Draft Scope and Methods Plan for Risk/Exposure Assessment:
Secondary NAAQS Review for Oxides of Nitrogen and Oxides of Sulfur (U.S. EPA, 2008b). Step
4 entails evaluating the current nitrogen and sulfur loads and effects to a chosen case study
assessment area, including ecosystems services. This case study evaluates the current nitrogen
deposition load to aquatic ecosystems; in particular, estuarine systems and the role atmospheric
deposition can play in the eutrophi cation of an aquatic ecosystem. Note that volatilization of
nitrogen from these systems is not considered within the case study because of the lack of
quantitative data on denitrification and volatilization rates for estuarine systems across the
United States (U.S. EPA, 2008a).
Final Risk and Exposure Assessment September 2009
Appendix 6-1
-------
Aquatic Nutrient Enrichment Case Study
Eutrophication
Eutrophication is the process whereby a body of water becomes over-enriched in
nutrients, resulting in increased productivity. As productivity increases with concomitant
increases in organic matter production, dissolved oxygen levels in the waterbody may decrease
and result in hypoxia (i.e., low dissolved oxygen levels). Total reactive nitrogen (Nr) can
promote eutrophication in inland freshwater ecosystems, as well as in estuarine and coastal
marine ecosystems, ultimately reducing biodiversity because of the lack of available oxygen
needed for the survival of many species of aquatic plants and animals. Total Nr includes all
biologically, chemically, and radiatively active nitrogen compounds in the atmosphere and
biosphere, such as ammonia gas (NHa), ammonium (NH4+), nitric oxide (NO), nitrogen dioxide
(NC>2), nitric acid (HNOs), nitrous oxide (IS^O), nitrate (NOs ), and organic compounds (e.g.,
urea, amines, nucleic acids) (U.S. EPA, 2008b).
Freshwater Aquatic Ecosystems
A freshwater lake or stream must be nitrogen-limited to be sensitive to nitrogen-mediated
eutrophi cation. Although conventional wisdom holds that most lakes and streams in the United
States are limited by phosphorus, recent evidence illustrates examples of lakes and streams that
are limited by nitrogen and show symptoms of eutrophi cation in response to nitrogen addition.
For example, surveys of lake nitrogen concentrations and trophic status along gradients of
nitrogen deposition show increased inorganic nitrogen concentrations and productivity to be
correlated with atmospheric nitrogen deposition (Bergstrom and Jansson, 2006). Additional
information supporting the connection between nitrogen loading and eutrophication in freshwater
systems is provided in EPA's ISA (U.S. EPA, 2008a, Sections 3.3.2.3 and 3.3.3.2).
Estuarine and Coastal Marine Ecosystems
Estuarine and coastal marine ecosystems are highly important to human and ecological
welfare through the ecosystem services they provide (e.g., fisheries, recreation). "Because the
productivity of estuarine and nearshore marine ecosystems is generally limited by the availability
of Nr, an excessive contribution of Nr from sources of water and air pollution can contribute to
eutrophication" (U.S. EPA, 2008a, Section 4.3.4.1). The National Oceanic and Atmospheric
Administration's (NOAA's) National Estuarine Eutrophication Assessment (NEEA) examined
more than 140 estuaries along the coasts of the conterminous United States. The assessment
Final Risk and Exposure Assessment September 2009
Appendix 6-2
-------
Aquatic Nutrient Enrichment Case Study
examined a range of symptoms of eutrophication, including algal blooms, hypoxia, and
vegetation growth. Findings from the study concluded that 65% of the assessed systems had
moderate to high overall eutrophic conditions (OECs) (Bricker et al., 2007a). Increasingly,
individual estuarine ecosystems have become the center of intensive studies on nutrient
enrichment/eutrophication causes and effects. Within the Chesapeake Bay, studies of the
frequency of phytoplankton blooms and the extent and severity of hypoxia revealed overall
increases in these detrimental effects (Officer et al., 1984). Within the Pamlico Estuary in North
Carolina, similar trends have been observed and studied by Paerl et al. (1998). Sources identified
within these assessments range from atmospheric deposition to fertilizer applications and other
land use-based applications.
Estuarine and coastal marine ecosystems experience a range of ecological problems
associated with nutrient enrichment. Because the productivity of estuarine and nearshore marine
ecosystems is generally limited by the availability of Nr, an excessive contribution of Nr from
sources of water and atmospheric pollution can contribute to eutrophication. Some of the most
important environmental effects include increased algal blooms, the occurrence of bottom-water
hypoxia, and decreases in fishery populations and the abundance of seagrass habitats (Boynton et
al., 1995; Valiela and Costa, 1988; Howarth et al., 1996; Paerl, 1995, 1997; Valiela et al., 1990).
There is broad scientific consensus that nitrogen-driven eutrophication in shallow U.S.
estuaries has increased over the past several decades and that environmental degradation of
coastal ecosystems is now a widespread occurrence (Paerl et al., 2001). For example, the
frequency of phytoplankton blooms and the extent and severity of hypoxia have increased in the
Chesapeake Bay (Officer et al., 1984), the Pamlico Estuary in North Carolina (Paerl et al., 1998),
and along the continental shelf adjacent to the Mississippi and Atchafalaya river discharges to
the Gulf of Mexico (Eadie et al., 1994). A recent national assessment of eutrophic conditions in
estuaries found that 65% of the assessed systems had moderate to high OECs (Bricker et al.,
2007a). Estuaries with high OECs were generally those that received the greatest nitrogen loads
from all sources, including atmospheric and land-based sources (Bricker et al., 2007a).
Final Risk and Exposure Assessment September 2009
Appendix 6-3
-------
Aquatic Nutrient Enrichment Case Study
1.1 INDICATORS, ECOLOGICAL ENDPOINTS, AND ECOSYSTEM
SERVICES
Major indicators for nutrient enrichment to aquatic systems from atmospheric deposition
of total Nr require measurements based on available monitoring stations for wet deposition
(National Atmospheric Deposition Program [NADPJ/National Trends Network) and limited
networks for dry deposition (Clean Air Status and Trends Network [CASTNet]). Wet deposition
monitoring stations can provide more information on an extensive range of nitrogen species than
is possible for dry deposition monitoring stations. This creates complications in developing
estimates for total nitrogen (TN) deposition levels because dry deposition data sources will likely
be underestimated because of the use of fixed deposition velocities that do not reflect local
conditions at the time of measurement, under-representation of monitoring sites in certain
landscapes, and omission of some Nr species in the measurements (U.S. EPA, 2008a, Section
2.5).
For aquatic ecosystems, the indicators for "nutrient enrichment" effects reflect a
combination of inputs from all media (e.g., air, discharges to water, diffuse runoff, groundwater
inputs). Major aquatic system indicators include nutrient loadings (Heinz Center for Science,
2007), excess algal standing crops, or in larger waterbodies, anoxia (i.e., absence of dissolved
oxygen) and/or hypoxia in bottom waters (see Table 1.1-1). For nitrogen, loadings or
concentration values related to total Nr (a combination of nitrates, nitrites, organic nitrogen, and
total ammonia) are encouraged for inclusion in numeric criteria as part of EPA-approved state
water quality standards (U.S. EPA, 2000). Given the nature of the major indicators for
atmospheric deposition and indicators for aquatic and terrestrial ecological systems, a data-fusion
approach that combines monitoring indicators with modeling inputs and outputs is often used
(Howarth, 2007).
Final Risk and Exposure Assessment September 2009
Appendix 6-4
-------
Aquatic Nutrient Enrichment Case Study
Table 1.1-1. Key Indicators of Nutrient Enrichment Due to Nr, Including NOX
Key Indicator
Group
Examples of Indicators
Description
Nitrogen deposition
Nitrate or ammonia
From wet or dry deposition monitoring stations
and networks
Nitrogen throughfall
deposition
Nitrate, ammonia, organic
nitrogen
Special measurements in terrestrial ecosystem
with corrections for nitrogen intercepted by
plant canopies
Nitrogen loadings
and fluxes to
receiving waters
Total nitrogen or
constituent species
combined with flow data
from gauged stations
Reflects a combination of inputs from all media
(e.g., air, discharges to water, diffuse runoff,
and groundwater inputs); relative role of air
deposition should ideally be compared with air
deposition data and also with available
(preferably multimedia) models
Other indicators of
aquatic system
nutrient enrichment
(eutrophication)
Algal standing crop
(plankton and periphyton);
anoxia/hypoxia for
estuaries and large rivers
Reflects a combination of inputs from all media
(e.g., air, discharges to water, diffuse runoff,
and groundwater inputs); relative role of air
deposition should ideally be compared with air
deposition data and also with available
(preferably multimedia) models
Nitrogen is an essential nutrient for estuarine and marine ecosystem fertility and is often
the algal growth-limiting nutrient (U.S. EPA, 2008a; Section 3.3.5.3). Excessive nitrogen
contributions can cause habitat degradation, algal blooms, toxicity, hypoxia, anoxia, fish kills,
and decreases in biodiversity (Paerl et al., 2002). To evaluate these impacts, five ecological
indicators were used in NOAA's recent NEEA of estuary trophic condition: chlorophyll a,
macroalgae, dissolved oxygen, nuisance/toxic algal blooms, and submerged aquatic vegetation
(SAV) (Bricker et al., 2007a).
Figure 1.1-1, excerpted from the NOAA's NEEA Update, provides a brief description of
each of the indicators. Further interactions between the indicators are described in the following
text. For greater detail on each of the indicators, including previous findings and study areas,
refer to the ISA (U.S. EPA, 2008a, Sections 3.3.2, 3.3.3, 3.3.5, 3.3.8, 4.3.4, and C.5) and the
NEEA Update (Bricker et al., 2007a).
Figure 1.1-2 provides a simplified progression of the indicators as the estuarine waters
become more eutrophic. In the NEEA Update (Bricker et al., 2007a), an illustrated relationship
between the OEC, water quality and ecological indicators, and influencing factors (e.g., nitrogen
loads) is presented (Figure 1.1-3).
Final Risk and Exposure Assessment
Appendix 6-5
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Indicators of eutrophication do not provide a direct link to the ecological benefits of the
ecosystem. Because of this, the eutrophication-impact ecological endpoints and the ecosystem
services affected must be identified and related to the quantifiable indicators. Table 1.1-2
provides some examples of the ecological endpoints associated with the indicators of
eutrophi cation. As described in the introduction, the ecological endpoints are ecological entities
and their impacts. For instance, an indicator may be low dissolved oxygen, but the ecological
endpoint or impact of having low dissolved oxygen is a decrease in the populations offish that
are highly sensitive to dissolved oxygen conditions.
Primary symptoms
Description
Chlorophyll a
(Phy to plankton)
Mac real gal blooms
Secondary symptoms
A measure used to indicate the amount of microscopic algae
(phytoplankton) growing in a water body. High concentrations can lead
to low dissolved oxygen levels as a result of decomposition.
Large algae commonly referred to as "seaweed." Bfooms can cause
losses of submerged aquatic vegetation by blocking sunlight.
Additionally, blooms may smother immobile shellfish, corals, or other
habitat. The unsightly nature of some blooms may impact tourism due
to the declining value of swimming, fishing, and boating.
Description
Dissolved
oxygen
Submerged
aquatic vegetation
Low dissolved oxygen is a eutrophic symptom because it occurs as a
result of decomposing organic matter (from dense algal blooms), which
sinks to the bottom and uses oxygen during decay. Low dissolved
oxygen can cause fish kilts, habitat loss, and degraded aesthetic values,
resulting in the loss of tourism and recreational water use.
Loss of submerged aquatic vegetation (SAV) occurs when dense alga)
blooms caused by excess nutrient additions (and absence of grazers)
decrease water clarity and light penetration. Turbidity caused by other
factors (e,g., wave energy, color) similarly affects SAV. The loss of SAV can
have negative effects on an estuary's functionality and may impact
some Fisheries due to loss of a critical nursery habitat.
Thought to be caused by a change in the natural mixture of nutrients
chat occurs when nutrient inputs increase over a long period of time.
These blooms may release toxins that kill hsh and shellfish. Human
health problems may also occur due to the consumption of
contaminated shellfish or from inhalation of airborne toxins. Many
nuisance/toxic blooms occur naturally, some are advected into
estuaries from the ocean; the role of nutrient enrichment is unclear.
Figure 1.1-1. Descriptions of the five eutrophication indicators used in NOAA's NEEA
(Bricker et al., 2007a).
Final Risk and Exposure Assessment
Appendix 6-6
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Nutrient inputs
Excessive growth of phytoplankton and macroalgae (grazers cannot
control growth)
Decreased water clarity/decreased light penetration/decreased
dissolved oxygen
SAV inhibition
Nuisance/toxic algal blooms
Low dissolved oxygen/hypoxia
Invertebrates and fish kills
Figure 1.1-2. A simplified schematic of eutrophication effects on an aquatic ecosystem.
No Problem/low Moderate low Moderate Moderate high High
i
o
w
a.
~-
i.
o
fc
=
Few symptoms occur Symptoms occur Symptoms occur Symptoms occur Symptoms occur
at more than episodically and/or less regularly less regularlyand/or periodically or
minimal levels. owera small to and tor over a over a medium to persistently and/or
medium area. medium area extensive area. over an extensive area-
Inflaen cing factors
(loads and suscptibilitti)
Key to symbols:
Submerged aquatic
vegetation
Chlorophyll a
Nuisance/toxic
blooms
Macroalgae
fj Disso Ived oxyge n
Figure 1.1-3. An illustrated representation of eutrophi cation measures through the use of
indicators and influencing factors from NOAA's NEEA (Bricker et al., 2007a).
Table 1.1-2. Assessment Ecological Endpoints for Nutrient Enrichment
Due to Deposition of Total Reactive Nitrogen, Including NOX
Assessment Ecological Endpoint
Fish abundance/population
Water quality, color, clarity
Species richness/community structure
Final Risk and Exposure Assessment
Appendix 6-7
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Assessment Ecological Endpoint
Habitat quality, including benthos and shoreline
Surface scum, odors
Continuing to link the indicators and ecological endpoints to the ecological processes of
value to society brings us to the ecosystem services related to eutrophication. Examples are
provided in Table 1.1-3. The example of dissolved oxygen and the resulting decrease in fish
population was used to identify the ecosystem services offish catch rate and fish kills, which
support both food and materials and recreational uses of the ecosystem.
Table 1.1-3. Ecosystem Services for Aquatic
Systems Affected by Nutrient Enrichment
Ecosystem Service
Fisheries
• Fish catch rate
• Fishable area
• Size/extent of fish kills
Recreation
• Boating
• Swimming
• Beach conditions
Tourism
• Aesthetics
Risk of illness
• Drinking water quality
• Contaminated fish
The methods of connecting the ecological endpoints and ecosystem services related to
eutrophi cation are beyond the scope of this case study, but they have been examined in another
study (RTI, 2008). Rather, the remaining discussion focuses on determining and detailing the
indicator measures as a function of the changing atmospheric deposition inputs of Nr, including
NOX.
Ecosystem services are generally defined as the benefits individuals and organizations
obtain from ecosystems. In the Millennium Ecosystem Assessment (MEA), ecosystem services
are classified into four main categories:
• Provisioning. Includes products obtained from ecosystems.
• Regulating. Includes benefits obtained from the regulation of ecosystem processes.
Final Risk and Exposure Assessment
Appendix 6-8
September 2009
-------
Aquatic Nutrient Enrichment Case Study
• Cultural. Includes the nonmaterial benefits people obtain from ecosystems through spiritual
enrichment, cognitive development, reflection, recreation, and aesthetic experiences.
• Supporting. Includes those services necessary for the production of all other ecosystem services
(MEA, 2005).
A number of impacts on the ecological endpoints offish population, water quality, and
habitat quality and the related ecosystem services exist, including the following:
• Fish kills - provisioning and cultural
• Surface scum - cultural
• Fish/water contamination - provisioning and cultural
• Decline in fish population - provisioning and cultural
• Decline in shoreline quality (e.g., erosion) - cultural and regulating
• Poor water clarity and color - cultural
• Unpleasant odors - cultural.
The goal of the Aquatic Nutrient Enrichment Case Study was to focus on fisheries,
recreation, and tourism. Attempts have been made to link fisheries (e.g., closings, decreased
species richness) quantitatively to eutrophication symptoms through monitoring data, and
recreation activities qualitatively through user surveys. The symptoms of eutrophication defined
by Bricker et al., (2007a) were pursued as the ecosystem ecological endpoints to link to these
ecosystem services.
1.2 CASE STUDIES
1.2.1 National Overview of Sensitive Areas
The selection of case study areas specific to eutrophication began with national
geographic information systems (GIS) mapping. Spatial datasets were reviewed that included
physical, chemical, and biological properties indicative of eutrophication potential in order to
identify sensitive areas of the United States (Table 1.2-1). The analysis then led to combining
the eutrophic estuaries from NOAA's Coastal Assessment Framework, along with areas that
exceed the nutrient criteria for lakes/reservoirs (U.S. EPA, 2002), as compared with wet nitrogen
deposition, to define areas of national aquatic nutrient enrichment sensitivity.
Final Risk and Exposure Assessment September 2009
Appendix 6-9
-------
Aquatic Nutrient Enrichment Case Study
Table 1.2-1. Summary of Indicators, Mapping Layers, and Models for Targeted Ecosystems
Targeted
Ecosystem
Effect
Indicator(s)
Mapping Layers
Model(s)
Aquatic
nutrient
enrichment and
eutrophication
Nitrate and ammonia, total
nitrogen (major reactive
nitrogen species)
Al toxicity data
Chlorophyll a (e.g., algal
standing crop)
Anoxia/hypoxia (e.g.,
primarily estuaries and tidal
rivers)
Nitrogen loadings for sub-
watersheds or larger basins
and ED As
EPA NCCR Water Quality
Index and NOAA Estuarine
Coastal Eutrophication
Index
Diatom data for nitrogen-
limited systems
STORET retrievals
USGS National Water
Quality Assessment Program
information
USGS SPARROW attributes,
information
Water quality standards
nutrient criteria for rivers and
lakes
EPA, NCCR, and NOAA
estuarine eutrophication
indicators
NOAA EDAs
EPA/NOAA airsheds for
major Atlantic and Gulf
estuaries CMAQ (e.g.,
nitrogen) by hydrological unit
code
USGS
SPARROW
PnET-BCG
Note: EDAs = Estuarine Drainage Areas; NCCR = National Coastal Condition Report; STORET =
STOrage and RETrieval; USGS = U.S. Geological Survey; SPARROW = Spatially Referenced
Regression on Watershed; CMAQ = Community Multiscale Air Quality; PnET-BCG = a
biogeochemical model
Bergstrom and Jansson (2006) compiled dissolved inorganic nitrogen (DIN) data from
4,296 lakes (i.e., 195 lakes from United States/Canada and 4,101 lakes from Europe). They
found that the mean lake DIN concentrations were strongly correlated to the mean wet DIN
deposition over large areas of Europe and North America (Figure 1.2-1). The equation for this
correlation is:
Iog7= 1.34xlogX-1.55 (r =0.70; PO.001)
(1)
where
Y is lake water DIN (microgram per liter [ug/L]), and
X is wet deposition (kilograms [kg] N/km2/yr).
EPA recommended TN criteria for lakes and reservoirs for 12 aggregated ecoregions in
2002 (U.S. EPA, 2002).
Final Risk and Exposure Assessment
Appendix 6-10
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Based on equation (1), nitrogen deposition level (X: kg N/ha/yr) associated with EPA TN
criteria (Y: |ig/L) for each aggregated ecoregion can be calculated by equation (2), and the
results are listed in Table 1.2-2.
_ 1fl(logr+1.55)/1.34
7100
(2)
800 n
0 200 400 (500 800 1000 1200 1400 1600
Wet DIN-depositkm
-------
Aquatic Nutrient Enrichment Case Study
The resulting map reveals areas of highest potential sensitivity to nitrogen deposition as
shown in Figure 1.2-2.These areas are identified in blue as nutrient sensitive estuaries contained
in NOAA's Coastal Assessment Framework, and in red in areas where deposition exceeds the
nutrient criteria. Yellow areas indicate those areas that are below the nutrient criteria, but are
within 5 kg/ha/yr of exceeding the criterion. White areas do not have an EPA nutrient criterion
for lakes/reservoirs.
Total Wet N Dep Exceedance L J -9 99 - -5 NOAA CAP
^•-26 41--15 D-4.99-0 I I Nutrient Criteria Btfy
I 1-14.99--10 ^•0.01-1.92 Lakes
Figure 1.2-2. Areas potentially sensitive to aquatic nutrient enrichment.
1.2.2 Use of ISA Information and Rationale for Site Selection
The potential case study areas identified by the Ecological Effects Subcommittee (EES)
of the Advisory Council on Clean Air Compliance Analysis were considered for examining the
ecological benefits of decreasing atmospheric deposition. Nutrient enrichment-relevant case
study areas suggested by the EES (U.S. EPA, 2005) are reproduced in Table 1.2-3. The ISA
(U.S. EPA, 2008b) also recommends case study areas as candidates for risk and exposure
Final Risk and Exposure Assessment
Appendix 6-12
September 2009
-------
Aquatic Nutrient Enrichment Case Study
assessments; Table 1.2-4 contains potential assessment areas for aquatic nutrient enrichment.
Additionally, Howarth and Marino (2006) provide a comprehensive summary of the literature
and scientific findings on eutrophication over the past 3 decades. This summary has led to the
general consensus that freshwater lakes and estuaries differ in terms of nutrient limitation as the
cause of eutrophi cation, and that nitrogen is the limiting element to primary production in coastal
marine ecosystems in the temperate zone.
For purposes of the Risk and Exposure Assessment, two regions were selected for case
study analysis to which a common methodology could be applied—Chesapeake Bay and the
Pamlico Sound. For aquatic nutrient enrichment, special emphasis was given to the Chesapeake
Bay region because it has been the focus of many previous studies and modeling efforts, and it is
currently one of the few systems within the United States in which economic-related ecosystem
services studies have been conducted. The Pamlico Sound, an economically important estuary
because of its fisheries, has been studied and modeled greatly by the universities and has also
been known to exhibit symptoms of extreme eutrophi cation. The following factors were
considered in choosing these case study areas:
• Availability of atmospheric deposition data
• Availability of existing water quality modeling that accounts for the role of atmospheric deposition
• A large, mainstem river that feeds a system with adequate hydrologic unit code delineation and
point- and nonpoint-source input data
• Scientific stature of the case study area
• Scalability and generalization opportunities for risk analysis results from the case studies.
These estuarine ecosystems have been the subjects of extensive research that provides the
data needed for a first phase of quantitative analysis of the role of nitrogen deposition in
eutrophi cation. Other candidate estuarine systems could be evaluated for potential future
analyses, whereas freshwater ecosystems in the western United States would most likely require
a separate analysis.
Because the Chesapeake Bay and Pamlico Sound are fed by multiple river systems, the
case study was scaled to one main stem river and associated estuary for each system: the
Potomac River and Potomac Estuary for the Chesapeake Bay and the Neuse River and Neuse
River Estuary for the Pamlico Sound. Details on each estuarine system are provided below.
Final Risk and Exposure Assessment September 2009
Appendix 6-13
-------
Aquatic Nutrient Enrichment Case Study
Table 1.2-3. Science Advisory Board/Ecological Effects Subcommittee Listing of Potential
Assessment Areas for Evaluation of Benefits of Decreases in Atmospheric Deposition with
Respect to Aquatic Nutrient Enrichment
Ecosystem/
Region
Main CAA
Pollutant(s)
Percentage(s)
Attributable to
Atmospheric
Deposition
Quantitative
Ecological and
Economic
Information
EES Comments
Coastal
Waquoit Bay
Chesapeake
Bay
Long Island
Sound
Barnegat Bay
Tampa Bay
Gulf of
Maine
Casco Bay
Rocky
Mountains
Nitrogen
Nitrogen
Nitrogen;
mercury
Nitrogen
Nitrogen;
mercury
Nitrogen
Nitrogen;
mercury
Nitrogen
30%
20% to 30%
Nitrogen = 23%
to 35%; Mercury
= ?
50% total;
direct deposition
30% to 39%
Nitrogen = 25%
to 30%
Low
Nitrogen = 30%
to 40%
Mercury = 84% to
92%
Nearly 100%
Yes
Yes
Yes
Yes
Yes
?
Yes
Yes
High priority. Higher loading
from nondepositional sources
may confound analysis.
High priority. Loading from
diverse sources, particularly
agricultural, may confound
analysis.
High priority. High nitrogen
loading from wastewater
treatment plants may
confound analysis.
High priority. Direct linkage
of ecological effects with
atmospheric deposition;
quantitative economic data
exist.
Medium priority. Examined in
previous EPA efforts.
Variability in loading data
may confound analysis.
Low priority. Linkage of
nitrogen loadings and
ecological impacts is not well
established. Major source of
nitrogen is open-ocean influx.
Medium priority. Good data
on ecological and economic
impacts are available.
Medium priority. Levels of
nitrogen loading much lower
than for northeastern
locations. Economic data may
be lacking.
Final Risk and Exposure Assessment
Appendix 6-14
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 1.2-4. Potential Assessment Areas for Aquatic Nutrient Enrichment Identified in the ISA
Area
Adirondack
Mountains
Chesapeake
Bay
Alpine and
subalpine
communities of
the eastern
slope of the
Rocky
Mountains, CO
Indicator
Aquatic nutrient
enrichment;
terrestrial nutrient
enrichment;
mercury
methylation
Aquatic nutrient
enrichment;
aquatic nitrogen-
limited
eutrophication
Aquatic nutrient
enrichment;
terrestrial nutrient
enrichment
Detailed
Indicator
Foliar N
concentration;
NO3" leaching;
C:N ratio; N
mineralization;
nitrification;
denitrification
Watershed N
sources;
chlorophyll a;
dissolved oxygen;
submerged
aquatic
vegetation
Biomass
production; NO3"
leaching; species
richness
Area Studies
PIRLA I and II;
Adirondack Lakes
Survey; Episodic
Response Project;
EMAP
NA
NA
Models
MAGIC;
PnET-BGC
NA
NA
References in U.S. EPA, 2008a
Baker and Laflen, 1983; Baker et al., 1990b;
Baker et al., 1990c; Baker et al., 1996; Benoit et
al., 2003; Chen and Driscoll, 2004; Confer et al.,
1983; Cumming et al., 1992; Driscoll et al.,
1987a; Driscoll et al., 1991; Driscoll et al., 1998;
Driscoll et al., 2001a; Driscoll et al., 200 Ib;
Driscoll et al., 2003b; Driscoll et al., 2003c;
Driscoll et al., 2007a; Driscoll et al., 2007b; Evers
et al., 2007; GAO, 2000; Havens et al., 1993; Ito
et al., 2002; Johnson et al., 1994b; Landers et al.,
1988; Lawrence et al., 2007; NAPAP, 1998;
Siegfried et al., 1989; U.S. EPA, 2003; Sullivan
et al., 1990; Sullivan et al., 2006a; Sullivan et al.,
2006b; U.S. EPA, 1995b; Van Sickle et al., 1996;
Whittier et al., 2002; Wigington et al., 1996; Zhai
etal.,2007
Bricker et al., 1999; Bricker et al., 2007; Boesch
et al., 2001; Boyer et al., 2002; Boyer and
Howarth, 2002; Cooper and Brush, 1991; Fisher
and Oppenheimer, 1991; Harding and Perry,
1997; Howarth, 2007; Kemp et al., 1983; Malone,
1991, 1992; Officer et al., 1984; Orth and Moore,
1984;Twilleyetal., 1985
Baron et al., 1994; Baron et al., 2000; Baron,
2006; Bowman, 2000; Bowman and Steltzer,
1998; Bowman et al., 1993; Bowman et al., 1995;
Bowman et al., 2006; Burns, 2004; Fenn et al.,
2003a; Fisk et al., 1998; Korb and Ranker, 2001;
Rueth et al., 2003; Seastedt and Vaccaro, 2001;
Sherrod and Seastedt, 2001; Steltzer and
Bowman, 1998; Suding et al., 2006; Williams and
Tonnessen, 2000; Williams et al.,1996a; Wolfe et
al., 2001
Source
ISA,
Section
3.2.4,
3.4.1,
4.2.2,
Annex B
ISA,
Section
3.3.2,
3.3.8,
4.3.4,
Annex C
ISA,
Sections
3.2.1,
3.2.2,
3.3.2,
3.3.3,
3.3.5,
3.3.8,4.5,
Annex C
andD
Final Risk and Exposure Assessment
Appendix 6-15
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Area
Beartooth
Mountain, WY
Pamlico
Estuary, NC
Rocky
Mountain
National Park,
CO
Lake Tahoe,
CA
Indicator
Aquatic nutrient
enrichment
Aquatic nitrogen
limited
eutrophication
Aquatic nutrient
enrichment
Aquatic nutrient
enrichment
Detailed
Indicator
Algae
composition
switch
Hypoxia;
phytoplankton
bloom
Diatom shifts
Primary
productivity;
chlorophyll a
Area Studies
NA
NA
NA
NA
Models
NA
NA
NA
NA
References in U.S. EPA, 2008a
Sarosetal.,2003
Paerletal., 1998
Interlandi and Kilham, 1998
Goldman, 1988; Jassby etal., 1994
Source
ISA,
Sections
3.3.5,
4.4.3,4.5,
Annex B
andC
ISA,
Sections
3.3.2,
3.3.3
ISA,
Sections
3.3.3,
3.3.5,
3.3.8,
4.3.3,4.5,
AnnexC
ISA,
Sections
3.3.3,
3.3.5,
3.3.8,
Annex C
Source: U.S.EPA, 2008a
Note: CAA = Clean Air Act; PIRLA = paleoecological investigation of recent lake acidification; EMAP = Environmental Monitoring and Assessment
Program; MAGIC = Model of Acidification of Groundwater in Catchments; PnET-BGC = a biogeochemical model; NCV = nitrate; NA = not applicable.
Final Risk and Exposure Assessment
Appendix 6-16
September 2009
-------
Aquatic Nutrient Enrichment Case Study
1.2.3 Potomac River and Potomac Estuary
The Chesapeake Bay is the largest of 130 estuaries in the United States. This commercial
and recreational resource serves more than 15 million people who live in and near its watershed
(i.e., drainage basin). The bay produces approximately 500 million pounds of oysters, crabs, and
other seafood per year. The richness of its species can be seen in the value of the bay's annual
fish harvest, which is estimated at more than $100 million. The Chesapeake Bay estuary receives
approximately 50% of its water from the Atlantic Ocean in the form of saltwater. The other half
of the water (i.e., fresh water) drains into the bay from a large 165,800-square-kilometer (km2)
(64,000-square-mile [mi2]) drainage watershed. Among the 150 major rivers and streams in the
Chesapeake Bay drainage basin are the James, Potomac, York, Rappahannock, Patuxent, and
Susquehanna rivers. The Potomac River watershed comprises about 22% of the land area and
30% of the population of the total Chesapeake Bay watershed. As a result, pollution loads from
the Potomac River have a significant impact on the health of the bay. The Chesapeake Bay
contains on average more than 68 trillion liters (18 trillion gallons) of water (Atkins and
Anderson, 2009).
The Potomac River is approximately 413 miles (665 km) long, with a drainage area of
approximately 14,670 mi2 (38,000 km2) and a population of approximately 5,350,000 people. It
begins at Fairfax Stone, WV, and runs to Point Lookout, MD. In terms of area, this makes the
Potomac River the fourth largest river along the Atlantic Coast of the United States and the
twenty-first largest in the United States as a whole (Fact-index.com, 2009). As shown in Figure
1.2-3, as well as in Table 1.2.5 and Table 1.2-6, the Potomac River contains diverse watersheds
in terms of topography, elevation (e.g., extending into the Shenandoah Mountains), and nutrient
point and nonpoint sources (e.g., forestland, farmland, and the Washington, DC, metropolitan
area). The Potomac River watershed lies in five geological
provinces: the Appalachian Plateau, Ridge and Valley, Blue
Atmospheric deposition accounts
for between 5% and 15% to 20%
of the Potomac River watershed's
total nitrogen load according to
published research (U.S. EPA,
2000; Boyeretal., 2002,
respectively). Additional expert
•1 c- <-«/ ^n/ ~™/ ^1 11, rr~T estimates put the contribution by
to contribute from 5% to 15%-20% of the watershed s TN atmospheric deposition at 30% to
load (U.S. EPA, 2000; Boyer et al., 2002, respectively). 40% of the total load'
Ridge, Piedmont Plateau, and Coastal Plain. The watershed
is approximately 12% urbanized, 36% agricultural use, and
52% forested. Atmospheric deposition has also been reported
Final Risk and Exposure Assessment September 2009
Appendix 6-17
-------
Aquatic Nutrient Enrichment Case Study
Potomac River Watershed Study Area
Watershed boundary layers and mapping data were provided by US^g^
The site URL is: http://md.water.usgs.gov/gis/chesbay/sparrow3/doc/rew3vfttfffl^,cti(|n1
Legend
I I Potomac River
Watershed
— Potomac River Basin
Stream Network
0 TIME/LTM Sites
© Calibration Sites
HUC8 Watersheds
| | 2070001
| | 2070002
| | 2070003
| | 2070004
I | 2070005
| | 2070006
| | 2070007
| | 2070008
| | 2070009
| | 2070010
| | 2070011
—i Segmented
-1 SPARROW
Watersheds
0 15 30 GO
Figure 1.2-3. The Potomac River Watershed and Potomac Estuary.
Table 1.2-5. Physical Characteristics of the Potomac Estuary
Parameter
Estuary area (km2)
Tidal fresh zone area (km2)
Mixing zone area (km2)
Saltwater zone area (km2)
Estuary volume (m3)
Estuary depth (m)
Estuary perimeter (km)
Percentage of estuary open (%)
Catchment area (km2)
Value
1,260
183
1,077
0
6.46 x 109
5.13
1,350
1.33
36,804
Metadata
Estuary area, calculated from NOAA shapefiles
Tidal fresh area, calculated from NOAA shapefiles
Mixing zone area, calculated from NOAA shapefiles
Saltwater area, calculated from NOAA shapefiles
Best estimate of volume from digital bathymetric chart
if available; otherwise, NOAA planimetry
From digital bathymetric chart if available; otherwise,
NOAA planimetry
Perimeter of estuary, based on shapefile; can be used to
calculate various aspect ratios
Percentage of the perimeter that is the "open" (or
oceanic) boundary; somewhat subjective
Not available
Final Risk and Exposure Assessment
Appendix 6-18
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Parameter
Catchment mean elevation (m)
Catchment maximum elevation
(m)
Catchment/estuary area ratio
Value
330
1,433
29.2
Metadata
Calculated from catchment shapefiles + Hydro IK (a
global 1 -km grid of elevation)
Calculated from catchment shapefiles + Hydro IK (a
global 1 -km grid of elevation)
Area ratio, based on catchment and area data given
above
Source: NEEA Estuaries Database
Note: m = meter
Table 1.2-6. Hydrological Characteristics of the Potomac Estuary
Parameter
Tide height (m)
Tide volume (m3)
Tides/day (#)
Tide volume/day (n^.d"1)
Tide ratio
Stratification ratio
Percent freshwater (%)
Percent mixed water (%)
Percent seawater (%)
Average salinity (psu)
Tidal exchange (days)
Tidal freshwater flush (days)
Daily freshwater/estuary
area (m.d"1)
Daily freshwater (m3/day)
(best)
Flow/estuary area (m/day)
(best)
Value
0.55
6.93 x 108
2
1,339,130,435
0.11
0.02649
14.5
85.5
0
11
121
36
27.063
34,100,000
27.063
Metadata
NOAA estimate of tide height, back-calculated
from tide volume; in some cases, guessed from
nearby systems
Tide height (m) x estuary area (km2) x 106
NOAA designation
Calculated from tide volume and tides per day
Tide height divided by estuary depth; a clean-up of
a NOAA variable
Total freshwater flux per day divided by tide
volume per day
Based on NOAA shapefiles of the three zones
according to their designation
Based on NOAA shapefiles of the three zones
according to their designation
Based on NOAA shapefiles of the three zones
according to their designation
Based on NOAA estimate of freshwater volume,
but scaled to "local coastal salinity," below
Exchange time as (Est V/net fw V per d) *
(coastal sal - avg sal)/coastal sal); a salinity-
based estimate of exchange
NOAA-based calculation, using (daily tide +
freshwater volume)/system volume
NOAA estimate of daily flow/estuary area
NOAA estimate above or (if not available) NCPDI
estimate
Best estimate/estuary area
Final Risk and Exposure Assessment
Appendix 6-19
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Parameter
Total freshwater Volume
(I/day)
Daily precipitation
(m3/day)
Daily evaporation (mVday)
Daily precipitation/estuary
area (mm/day)
Daily evaporation/estuary
area (mm/day)
Flow (nrVday)
Value
0.00549
3.64 x 106
2.26 x 106
2.889
1.794
2.33 x IQ7
Metadata
Best estimate/estuary volume (= hydraulic
exchange rate)
Direct precipitation on system, derived from
PRISM Shapefile
Direct evaporation from system, derived from
LOICZ 0.5 degree database, originally from
Wilmott
Daily precipitation/estuary area
Daily evaporation/estuary area
NCPDI_1 982- 1991
Source: NEEA Estuaries Database
Note: m = meter; psu = practical salinity unit; NCPDI = National Coastal Pollution Discharge
Inventory; PRISM = Parameter-elevation Regressions on Independent Slopes Model; LOICZ :
Land-Ocean Interactions in the Coastal Zone; mm = millimeters.
1.2.4 Neuse River and Neuse River Estuary
The Neuse River is the longest river in North Carolina, and the Neuse River watershed is
the third largest river watershed in the state (Figure 1.2-4). The Neuse River is a mainstem river
to the Pamlico Sound—one of the two largest estuaries on the Atlantic Coast. The river
originates in north-central North Carolina and flows southeasterly until it reaches tidal waters
upstream of New Bern, NC. At New Bern, the river broadens dramatically and changes from a
free-flowing river to a sound. While the Neuse River itself is 399 kilometers (km)_(248 miles)
long, there are 5,628 freshwater stream kilometers (3,497 miles), 6,643 hectares (16,414 acres)
of freshwater reservoirs and lakes, 149,724 estuarine hectares (369,977 acres), and 33.8
kilometers (21 miles) of Atlantic coastline within the entire Neuse River watershed. The drainage
area for the watershed is approximately 14,210 mi2 (36,804 km2). There are 19 major reservoirs
in the Neuse River watershed; most of these are located in the upper portion of the watershed.
The watershed starts in the eastern Piedmont physiographic region, with approximately two-
thirds of the watershed located in the Coastal Plain (NC DENR, 2002).
The Neuse River watershed encompasses all or portions of 18 counties and 74
municipalities. The watershed has a population of approximately 1,320,379 according to the
2000 census. Fifty-six percent of the land in the watershed is forested, and approximately 23% is
in cultivated cropland. Only 8% of the land falls into the urban/built-up category. Despite the
Final Risk and Exposure Assessment
Appendix 6-20
September 2009
-------
Aquatic Nutrient Enrichment Case Study
large amount of cultivated cropland and the relatively small amount of urban area, the basin has
seen a significant decrease (-72,800 hectares [-180,000 acres]) in cultivated cropland and forest
and an increase (+91,900 hectares [+227,000 acres]) in developed areas over the past 15 years
(NRCS, 2001). The Neuse River watershed is divided into 14 subbasins (6-digit North Carolina
Division of Water Quality [NC DWQ] subbasins) (NC DENR, 2002). Table 1.2-7 through Table
1.2-9, respectively, provide physical, land use and population, and hydrological characteristics of
the Neuse River and Neuse River Estuary.
There are 134,540 estuarine hectares (332,457 acres) classified for shellfish harvesting
(Class SA [shellfishing]) in the Neuse River Estuary. The Neuse River Estuary is important to
the commercial blue crab (Callinectes sapidus) fishery in the eastern United States and
accounted for approximately one-quarter of the blue crab harvest from 1994 to 2002 (Smith and
Crowder, 2005). Eutrophication became a water quality concern in the lower Neuse River and
Neuse River Estuary in the late 1970s and early 1980s. Nuisance algal blooms prevalent in the
upper estuary prompted investigations by the state. These investigations, as well as other studies,
indicated that algal growth was being stimulated by excess nutrients entering the estuarine waters
of the system. In 1988, a phosphate detergent ban was put in place, and the lower Neuse River
and Neuse River Estuary received the supplemental classification of nutrient-sensitive waters.
Phosphorus loading was greatly decreased, and algal blooms in the river and freshwater portions
of the system were lowered as a result of this action. However, the 1993 Neuse River Basin-wide
Water Quality Plan (NC DENR, 1993) recognized that eutrophication continued to be a water
quality problem in the estuary below New Bern. Extensive fish kills in 1995 prompted further
study of the problem. Low dissolved oxygen levels associated with algal blooms were
determined to be a probable cause of many of the fish kills. The algal blooms and
correspondingly high levels of chlorophyll a prompted the state to place the Neuse River and
Neuse River Estuary on the 1994, 1996, 1998, and 2000 303(d) List of Impaired Waters. It was
determined that control of nitrogen was needed to decrease the extent and duration of algal
blooms.
Atmospheric deposition is believed to play a role in nutrient loading to the Neuse River
and Neuse River Estuary. As excerpted from Whitall and Paerl (2001), the following discusses
the role of atmospheric deposition to nutrient loading for sensitive waterbodies:
Final Risk and Exposure Assessment September 2009
Appendix 6-21
-------
Aquatic Nutrient Enrichment Case Study
Excessive nitrogen loading to nitrogen-sensitive waters, such as the Neuse River
Estuary (North Carolina) has been shown to promote changes in microbial and
algal community composition and function (harmful algal blooms), hypoxia and
anoxia, and fish kills. Previous studies have estimated that wet atmospheric
deposit!on of nitrogen (WAD-N), as deposition of dissolved inorganic nitrogen
(DIN: NO's, NH3/NH+4) and dissolved organic nitrogen, may contribute at least
15% of the total externally supplied or "new" nitrogen flux to the coastal waters
of North Carolina. In a 3-year study from June 1996 to June 1999, Whitall and
Paerl calculated the weekly wet deposition of inorganic and organic nitrogen at 11
sites on a northwest-southeast transect in the watershed. The annual mean total
(wet DIN + wet organics) WAD-N flux for the Neuse River watershed was
calculated to be 956 mg N/m2/yr (15,026 Mg N/yr). Seasonally, the spring
(March-May) and summer (June-August) months contain the highest total weekly
nitrogen deposition; this pattern appears to be driven by nitrogen concentration in
precipitation. There is also spatial variability in WAD-N deposition; in general,
the upper portion of the watershed receives the lowest annual deposition and the
middle portion of the watershed receives the highest deposition. Based on a range
of watershed nitrogen retention and in-stream riverine processing values, we
estimate that this flux contributes approximately 24% of the total "new" nitrogen
flux to the estuary (Whitall and Paerl, 2001).
Of these atmospheric deposition measurements, it is
expected that the contributions will be greater from reduced
forms of nitrogen than from oxidized forms because of the large
amounts of agriculture within the watershed. One of the reasons
for selecting this case study area is to evaluate the impact of a
NOx-based standard on an area dominated by reduced forms of
nitrogen.
According to Whitall and
Paerl (2001), atmospheric
deposition accounts for
approximately 24% of the
Neuse River watershed's
total nitrogen loading, with
reduced forms of nitrogen
making up a larger portion of
the total than oxidized forms.
Final Risk and Exposure Assessment
Appendix 6-22
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Neuse River Watershed Study Area
Atlantic Ocean
Watershed boundary layers and mapping data was provided by USGS. The site
URL is: http://md.water.usgs.gov/gis/chesbay/sparrowa'doc/retv3.htm#section1
Atlantic
Ocean
Legend
o Calibration Stations
Watershed Stream
Segments
I Neuse/Tar Pamlico
' ' /Cape Fear
Watershed Area
Neuse River Basin
HUC8 Watersheds
^B 03020201
| | 03020202
| 03020203
| | 03020204
-i Segmented
-1 SPARROW
Watersheds
20 40
l Mile;
Figure 1.2-4. The Neuse River Watershed and Neuse River Estuary.
Table 1.2-7. Physical Characteristics of the Neuse River Estuary
Parameter
Estuary area (km2)
Tidal fresh zone area
(km2)
Mixing zone area (km2)
Saltwater zone area (km2)
Estuary volume (m3)
Estuary depth (m)
Estuary perimeter (km)
Percentage estuary open
(%)
Value
456
5
451
0
1.304 x 109
2.86
523
2.1
Metadata
Estuary area, calculated from NOAA shapefiles
Tidal fresh area, calculated from NOAA shapefiles
Mixing zone area, calculated from NOAA shapefiles
Saltwater area, calculated from NOAA shapefiles
Best estimate of volume from digital bathymetric chart
if available; otherwise, NOAA planimetry
From digital bathymetric chart if available; otherwise,
NOAA planimetry
Perimeter of estuary, based on shapefile; can be used to
calculate various aspect ratios
Percentage of the perimeter that is the "open" (or
oceanic) boundary; somewhat subjective
Final Risk and Exposure Assessment
Appendix 6-23
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Parameter
Catchment area (km2)
Catchment mean
elevation (m)
Catchment maximum
elevation (m)
Catchment/estuary area
ratio
Value
14,066
56
245
30.8
Metadata
Not available
Calculated from catchment shapefiles + Hydro IK (a
global 1-km grid of elevation)
Calculated from catchment shapefiles + Hydro IK (a
global 1-km grid of elevation)
Area ratio, based on catchment and area data given
above
Source: NEEA Estuaries Database
Note: m = meter
Table 1.2-8. Neuse River Watershed Land Use and Population
Parameter
Urban (km2)
Agriculture (km2)
Forest (km2)
Wetland (km2)
Range (km2)
Total (km2)
Population (#)
Population/estuary area
(#.krn2)
Value
1,328.66
(9.5%)
4,983.14
(35.6%)
6,648.5
(47.5%)
1,020.46
(7.3%)
5.17998(0%)
13,985.93998
1,015,059
2,226
Metadata
USGS (LUDA) for entire watershed 1972 with census
1990 information, base year early 1990s
USGS LUDA for entire watershed 1972 with census
1990 information, base year early 1990s
USGS LUDA for entire watershed 1972 with census
1990 information, base year early 1990s
USGS LUDA for entire watershed 1972 with census
1990 information, base year early 1990s
USGS LUDA for entire watershed 1972 with census
1990 information, base year early 1990s
USGS LUDA for entire watershed 1972 with census
1990 information, base year early 1990s
Based on gridded (1-km) U.S. 1990 census data,
corrected for catchments extending outside the United
States (with LANDSCAN)
Population based on gridded (1-km) U.S. 1990 census
data, corrected for catchments extending outside the
United States (with LANDSCAN). Estuary area,
calculated from NOAA shapefiles.
Source: NEEA Estuaries Database
Note: USGS = U.S. Geological Survey; LUDA = Land Use and Land Cover; LANDSCAN = a global
population database
Final Risk and Exposure Assessment
Appendix 6-24
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 1.2-9. Hydrological Characteristics of the Neuse River Estuary
Parameter
Tide height (m)
Tide volume (m3)
Tides/day (#)
Tide volume/day
(mVday)
Tide ratio
Stratification ratio
Percent freshwater (%)
Percent mixed water (%)
Percent seawater (%)
Average salinity (psu)
Tidal exchange (days)
Tidal freshwater flush
(days)
Daily freshwater/estuary
area (m/day)
Daily freshwater
(nrVday) (best)
Flow/estuary area
(m/day) (best)
Total freshwater volume
(I/day)
Daily precipitation
(mVday)
Daily evaporation
(mVday)
Daily precipitation/
estuary area (mm/day)
Daily evaporation/ estuary
area (mm/day)
Value
0.15
6.84 x 107
2
132,173,913
0.05
0.08318
1.1
98.9
0
13
74
73
22.368
10,200,000
22.368
0.00843
1.72 x 106
926,000
3.772
2.031
Metadata
NOAA estimate of tide height, back-calculated from tide
volume; in some cases, guessed from nearby systems
Tide height (m) x estuary area (km2) x 106
NOAA designation
Calculated from tide volume and tides per day
Tide height divided by estuary depth; a clean-up of a
NOAA variable
Total freshwater flux per day divided by tide volume per
day
Based on NOAA shape files of the three zones according
to their designation
Based on NOAA shape files of the three zones according
to their designation
Based on NOAA shape files of the three zones according
to their designation
Based on NOAA estimate of freshwater volume, but
scaled to "local coastal salinity," below
Exchange time as (Est V/netfw V per d)* (coastal sal -
avg_sal)/coastal_sal); a salinity-based estimate of
exchange
NOAA-based calculation, using (daily tide + freshwater
volume )/system volume
NOAA estimate of daily flow/estuary area
NOAA estimate above or (if not available) NCPDI
estimate
Best estimate/estuary area
Best estimate/estuary volume (= hydraulic exchange rate)
Direct precipitation on system, derived from PRISM
shapefile
Direct evaporation from system, derived from LOICZ 0.5
degree database, originally from Wilmott
Daily precipitation/estuary area
Daily evaporation/estuary area
Final Risk and Exposure Assessment
Appendix 6-25
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Parameter
Flow (mVday)
Value
7.95 x 106
Metadata
NCPDI_1 982- 1991
Source: NEEA Estuaries Database
Note: m = meter; psu = practical salinity unit; NCPDI = National Coastal Pollution Discharge
Inventory; PRISM = Parameter-elevation Regressions on Independent Slopes Model; LOICZ = Land-
Ocean Interactions in the Coastal Zone; mm = millimeters
Ammonia emissions from intensive livestock feeding operations are believed to
contribute to nitrogen deposition in eastern North Carolina watersheds. During a 10-year
legislative mandated moratorium on new operations, poultry populations increased in two Neuse
River watershed counties, according to the U.S. Department of Agriculture's 2002 Ag Census.
Statewide, the census reported an increase in poultry farms from 5,094 in 1997 to 6,251 in 2002
statewide (USD A, 2002). (As of this writing, the 2007 Ag Census is not complete.) The
continued contribution of poultry operations' growth to nitrogen deposition during the moratoria
has not been assessed, particularly in terms of its deposition in the Neuse River watershed.
2.0 APPROACH AND METHODS
Since it was necessary for this case study to span both terrestrial and aquatic systems to
accommodate indirect (i.e., to the watershed) and direct (i.e., to the water surface) deposition
effects, as well as a variety of indicators, a modeling approach was necessary to examine the
impacts due to aquatic nutrient enrichment from nitrogen deposition.
There are several complicating factors to carrying out an analysis of eutrophication in
waterbodies when one of the requirements is to include modeled output of atmospheric
deposition from a high-level, detailed atmospheric model. This analysis is considered a
multimedia analysis where the air, land, and water are involved. Typically, models or analysis
methods existing in the literature focus on only one of those components. Links between the
components with the desired output of eutrophi cation indicators are rare in the current literature
or modeling environments. Additionally, the few instances that are available in the literature tend
to focus on specific case study areas or on being highly empirical and difficult to scale or extend
to alternate locations. All these facts must be considered when developing a method to examine
the effects of Nr, including NOX, deposition on aquatic nutrient enrichment.
Final Risk and Exposure Assessment September 2009
Appendix 6-26
-------
Aquatic Nutrient Enrichment Case Study
2.1 MODELING
There are four basic steps necessary to undertake a modeling effort to examine the effects
of nitrogen deposition (RTI, 2007):
1. Choose the specific question/problem to address.
2. Choose the best models based on model formulation (e.g., are biological processes
considered?), desired output, study area, data availability, and necessary
uncertainty/sensitivity analyses for the models.
3. Determine and set up any processes/algorithms necessary to match atmospheric modeling
output (assumed to be from Community Multiscale Air Quality [CMAQ]) to the chosen
receiving water or terrestrial/watershed model.
4. Obtain the data needed for model parameterization.
The problem to be addressed in this analysis is assessment of the effects of deposition of
Nr, including NOX, on aquatic nutrient enrichment. The impacts of both direct (i.e., deposition on
the waterbody surface) and indirect (i.e., deposition within the watershed and transport to the
waterbody) deposition need to be identified. A method is needed to provide measures of the
eutrophication indicators that were previously described in Section 1.1.
A previous RTI International (RTI)1 report (RTI, 2007) detailed the difficulty, along with
the desire, to utilize atmospheric modeling in combination with the receiving-water and
terrestrial/watershed models for analyzing the effects of Nr, including NOX, deposition. The
multimedia approach to modeling is still in development; therefore, at this time, not many
models are set up to immediately accept the output from an atmospheric model such as CMAQ.
In the previous model investigation, RTI examined 35 receiving-water and terrestrial/watershed
models, which represent a wide diversity of types of ecosystems; history, location, and
spatial/temporal scales of application; scientific acceptance levels and organizational and agency
support; complexity and requirements; state variables and processes; and management uses.
Several existing models accept atmospheric concentration or flux data, but the time-step,
spatial resolution, and exact species required might all differ from the atmospheric model output.
The RTI report (2007) provided a list of models that could fulfill the multimedia approach while
using CMAQ output as input for the atmospheric component to the model. These models include
the Hydrologic Simulation Program-FORTRAN (HSPF), Regional Hydeo-Economic Simulation
System (RHESSys), the Georgia Tech Hydrologic Model/Multiple Element Limitation
(GT/MEL), Model of Acidification of Groundwater in Catchments (MAGIC), PnET-BGC,
1 RTI International is a trade name of Research Triangle Institute
Final Risk and Exposure Assessment September 2009
Appendix 6-27
-------
Aquatic Nutrient Enrichment Case Study
Integrated Nitrogen in Catchments (INCA), Spatially Referenced Regression on Watershed
(SPARROW) attributes, AQUATOX, Water Quality Analysis Simulation Program (WASP),
Enhanced Stream Water Quality Model (QUAL2K), CE-QUAL family of models, and Row
Column AESOP/Estuary and Coastal Ocean Model with Sediment Transport
(RCA/ECOMSED). These models are very different from one another in terms of the system
components included, process representations, data requirements, and output parameters (for
comprehensive details for each model, refer to the RTI report [2007]).
After determining which models could utilize CMAQ data, the ecosystem component
encompassed by the models was examined. The choice of case study areas that include estuaries
dictated that the model chosen must provide nutrient loads to an estuary waterbody and examine
the impacts of those loads within the estuary itself. Although AQUATOX and QUAL2K are
receiving-water models, they do not function for estuaries, nor do they account for indirect
deposition over the contributing watershed. The WASP, CE-QUAL family of models, and
RCA/ECOMSED are receiving-water models, which can be parameterized for estuaries, but they
do not simulate terrestrial processes. Several of the other models account for indirect deposition
and are strictly terrestrial models. These models include Regional Hydro-Economic Simulation
System (RHESSys) and GT/MEL. Other models include both the indirect deposition and direct
deposition, but only over streams and lakes within the watershed. These models are HSPF,
MAGIC, PnET-BGC, INCA, and SPARROW.
From this analysis, it was apparent that a multiple step (i.e., linked processes or
calculations) or model (i.e., separate but linked models) analysis would be optimal, including
both a step/model to examine the indirect deposition and a step/model to examine the estuarine
effects. The challenge was balancing analysis power against data, effort, and scalability
requirements. Higher-level modeling approaches could be used to evaluate the eutrophication
effects of interest if significant data resources, time, and expertise were available for a specific
site. An approach of this kind would not be scalable or applicable to wider regions, but it would
provide estimates with less uncertainty for a studied system.
The list of models above was used to identify several models that could be used to
produce nutrient loads to the estuary, the obvious critical component of an eutrophication
analysis. The best model for determining nitrogen loading to the estuary would track the
atmospheric deposition of nitrogen through the watershed and to the estuary. This requirement
Final Risk and Exposure Assessment September 2009
Appendix 6-28
-------
Aquatic Nutrient Enrichment Case Study
eliminated models that did not provide stream networking (i.e., PnET-BGC, MAGIC) or that
lumped land-use categories together (i.e., INCA). The remaining models of HSPF and
SPARROW are very different. HSPF is a highly parameterized, dynamic model that requires
extensive data inputs and calibration. SPARROW is a hybrid statistical and process-based,
steady-state model that requires much less data for parameterization, but still includes spatial
variation and source investigation. Therefore, the SPARROW model was chosen to estimate
nitrogen loadings to the estuaries.
Next, the most applicable method for examining eutrophication effects in an estuary was
assessed. The three identified models that could represent estuarine processes (i.e., WASP, CE-
QUAL family of models, and RCA/ECOMSED) were systematically ruled out as possibilities.
RCA/ECOMSED is a proprietary model with extensive data requirements and requires a high
level of expertise. The CE-QUAL family of models has primarily been used by the U.S. Army
Corps of Engineers. The various versions of CE-QUAL all have extensive data requirements,
and no indications of model integration have been uncovered in the literature. WASP provides
the output desired, but requires parameterization for each system of study. Considering that the
SPARROW model will provide TN loads to the estuary and the fact that that the chosen method
needs to be scalable and applicable to a variety of future case study areas, the SPARROW model
was selected for this case study.
With the elimination of the three identified dynamic modeling applications, a more
descriptive method of evaluation was sought. The method developed by NOAA and used in their
NEEA was identified as a likely candidate for eutrophication assessment.
The screening process that led to the decision to use SPARROW and a more descriptive
eutrophication evaluation technique considered the level of effort needed for an analysis of this
scale in the time available. Additionally, as summarized in recent literature (Howarth and
Marino, 2006), the complex processes that cause and express eutrophication within an estuary
are not greatly understood and could lead to under- and mis-representation within dynamic
models. The loss of a temporally varying analysis with the use of a steady-state or annual
average model results in the loss in detailing seasonal changes and some of the intricate
processes that may occur on a daily or even monthly time scale rather than over an entire year.
This trade-off allows for development toward the ultimate goal of providing a scalable
methodology that can be applied to various sites across the nation. With the acknowledged
Final Risk and Exposure Assessment September 2009
Appendix 6-29
-------
Aquatic Nutrient Enrichment Case Study
uncertainties in eutrophication process details across different systems, a more screening-level,
scalable approach was deemed appropriate for this initial study to link atmospheric nitrogen
deposition to eutrophicatic conditions.
2.2 CHOSEN METHOD
After examining several estuarine assessment options, the most comprehensive
evaluation technique that could be applied on a wide scale was revealed to be an assessment of
eutrophi cation as conducted in NO AA' s NEEA. This assessment method is titled Assessment of
Estuarine Trophic Status eutrophi cation index (ASSETS El) (Bricker et al., 2007a). NOAA's
ASSETS El results in an estimation of the likelihood that the estuary is experiencing
eutrophi cation or will experience eutrophication in the future.
The ASSETS El incorporates indirect deposition over the watershed through the
evaluation of nitrogen loading to the estuary. Thus, a decision was required on how to derive the
nitrogen load to the estuary based on the 2002 CMAQ-modeled deposition data. Because the
ASSETS El is more of a screening-level approach to assessing eutrophication, the nitrogen load
to the estuary is only required to be an annual estimate of TN loading. For these reasons, The
SPARROW model was chosen to provide the estimates of nitrogen loading to the estuary.
The combination of SPARROW modeling and the ASSETS method to developing an El
(Figure 2.2-1) provides a sound basis for conducting an eutrophication assessment. Both
SPARROW and the ASSETS El are supported by federal agencies and have been through
several improvement iterations. As shown in the following sections, the method provides a
screening-level approach that includes an appropriate level of detail for determining the impacts
on the degree of eutrophication in an estuary based on changes in atmospheric deposition
loadings.
Final Risk and Exposure Assessment September 2009
Appendix 6-30
-------
Aquatic Nutrient Enrichment Case Study
Atmos
Depo
Load
Land
Use
//
Streams /
& Catch- /----
merits /
SPARROW
________________________^ Nitrogen Load !_________
-~~~~~~~~~~~~~~~~~~~~~~~« by Reach h~~~~~~~~~~~
[Cumulative /
to Estuary \
Point
Sources
/Other
Non-
point
Sources /
Suscep-
tibility
Figure 2.2-1. Modeling methodology for case study.
Note: DO = dissolved oxygen; HAB = harmful algal bloom.
ASSETS El scores were available for both the Potomac and Neuse River estuaries, and
both estuaries were the subject of past and ongoing SPARROW modeling of point and nonpoint
sources, including atmospheric deposition.
2.2.1 SPARROW
2.2.1.1 Background and Description
SPARROW is a watershed modeling technique designed and supported by the U.S.
Geological Survey (USGS). The model relies on a nonlinear regression formulation to relate
water quality measurements throughout the watershed of interest to attributes of the watershed.
Both point and diffuse sources within the watershed are considered along with nonconservative
transport processes (i.e., loss and storage of contaminants within the watershed). SPARROW
Final Risk and Exposure Assessment
Appendix 6-31
September 2009
-------
Aquatic Nutrient Enrichment Case Study
follows the rules of mass balance while using a hybrid statistical and process-based approach
(Figure 2.2-2). "Because the dependent variable in SPARROW models (i.e., the mass of
contaminant that passes a specific stream location per unit time) is, in mathematical terms,
linearly related to all sources of contaminant mass in the model, all accounting rules relating to
the conservation of mass will apply" (Schwarz et al., 2006). Additionally, since SPARROW is a
statistical model at its core, it provides measures of uncertainty in model coefficient and water
quality predictions. Utilization of the SPARROW model results in estimates of long-term,
steady-state water quality in a stream. In most applications, SPARROW estimates represent
mean annual stream loadings of a contaminant.
Load generated within Load originating within the
Load leaving the _ upstream reaches and + reach's incremental watershed
reach transported to the reach via and delivered to the reach
the stream network segment
Figure 2.2-2. Mass balance description applied to the SPARROW model formulation.
A key component of SPARROW is its reliance on the spatial distribution of watershed
characteristics and sources. The stream reach network is spatially referenced against all
monitoring stations, GIS data for watershed properties, and source information. This structure
allows for the simulation of fate and transport of contaminants from sources to streams and
downstream ecological endpoints. "Spatial referencing and the mechanistic structure in
SPARROW have been shown to improve the accuracy and interpretability of model parameters
and the predictions of pollutant loadings as compared to those estimated in conventional linear
regression approaches" (e.g., Smith et al., 1997; Alexander et al., 2000) (Schwarz et al., 2006).
This spatially distributed model structure based on a defined stream network allows separate
statistical estimation of land and water parameters that quantify the rates of pollutant delivery
from sources to streams and the transport of pollutants to downstream locations within the stream
network (i.e., reaches, reservoirs, and estuaries) (Schwarz et al., 2006). Figure 2.2-3 shows how
each watershed and stream reach within the stream network defined for the SPARROW
application (represented by different colors in the figure) is processed separately and linked to
derive a final loading at a downstream location (the star labeled X). The SPARROW model is
calibrated at each monitoring station (represented by stars in Figure 2.2-3) by comparing the
modeled loads (i.e., a total of loads from each watershed segment and any upstream loads from
previous calibrations) against monitored data at the station. In this case, the modeled load at
Final Risk and Exposure Assessment September 2009
Appendix 6-32
-------
Aquatic Nutrient Enrichment Case Study
downstream monitoring station X would include loads from upstream monitoring station Y and
the five watershed segments between the two monitoring stations.
Stream
reach
segment
Downstream
monitoring
station, X
Upstream
monitoring
station, Y
Reservoir
Reach
contributing
area
1'oint source
Figure 2.2-3. Conceptual illustration of a reach network.
Within this Aquatic Nutrient Enrichment Case Study, the mathematical formulation of
the basic version of SPARROW presented by McMahon et al., (2003) is shown for consideration
in Equations 3 to 5. "The additive contaminant source components and multiplicative land and
water transport terms are conceptually consistent with the physical mechanisms that explain the
supply and movement of contaminants in watersheds" (Schwarz et al., 2006). Preservation of
mass, accounting for transport and decomposition at individual sources, is accomplished within
SPARROW through the spatial referencing of all processes with respect to the stream network
and the specific reach in which the process is carried out. Decomposition processes are
represented through losses in delivery to the stream and within the stream reach itself (Equation
4) or within a reservoir (Equation 5).
where
Load
nN
(3)
= Nitrogen load or flux in reach /', measured in metric tons
= Source index where TV is the total number of individual n sources
= Set of all reaches upstream, including reach /'
Final Risk and Exposure Assessment
Appendix 6-33
September 2009
-------
Aquatic Nutrient Enrichment Case Study
fin = Estimated source coefficient for source n
Sn.j = Nitrogen mass from source n drainage to reach y'
a = Estimated vector of land to water delivery coefficients
Zj = Land-surface characteristics associated with drainage to reach y'
fFjj = Fraction of nutrient mass present in water body y transported to water
body /' as a function of first-order loss process associated with stream
channels
HRtj = Fraction of nutrient mass present in water body y transported to water
body /' as a function of first-order loss process associated with lakes
and reservoirs
Si = Multiplicative error term assumed to be independent and identically
distributed across separate subbasins defined by intervening drainage
areas between monitoring stations.
(4)
where
m =
J-'i.j.m
First-order loss coefficient (km"1) (A A: value of 0.08, for example,
indicates that nitrogen is removed at a rate of approximately 8% per
km of channel length.)
Number of discrete flow classes
Length of the stream channel between water bodiesy and /' in flow
class m.
(5)
where
k = Estimated first-order loss rate (or settling velocity; units = m/yr)
qil = Reciprocal areal hydraulic load of lake or reservoir (ratio of water-
surface area to outflow discharge; units = yr/m) for each of the lakes
and reservoirs (/) located between water bodiesy and /'.
SPARROW has been designed to identify and quantify pollution sources that contribute
to the water quality conditions predicted by the model. Several different types of sources may be
examined, and sources may be for an individual stream location or summarized for a grouping of
stream locations. Examples of sources modeled within SPARROW include atmospheric
deposition, point sources, animal agriculture, or land use-based supply of contamination. "The
Final Risk and Exposure Assessment
Appendix 6-34
September 2009
-------
Aquatic Nutrient Enrichment Case Study
ability to develop quantitative information on pollution sources in SPARROW models stems
from the ability to trace, for each contaminant category, the predicted in-stream flux through a
given stream reach to the individual sources in each of the upstream reach watersheds
contributing contamination to that reach" (Schwarz et al., 2006). Figure 2.2-4 highlights some of
these sources in a conceptualization of the SPARROW model process.
Diffuse Sources
Industrial / Municipal
Point Sources
Landscape
Transport
SPARROW
Mode! Components
Aquatic Transport
Streams
Reservoirs
In-Stream Flux Prediction
Calibration minimizes differences
between predicted and calculated
mean-annual loads at the
monitoring stations
Water-Quality and
Flow Data
Periodic measurements
a! monitoring stations
Monitoring
Station Flux
Estimation
Rating Curve Model
of Pollutant Flux
Station calibration to
monitoring data
Mean-Annual Pollutant
Flux Estimation
Evaluation of Model
Parameters and Predictions
Figure 2.2-4. SPARROW model components (Schwarz et al., 2006).
Complete procedures, such as calculation of monitoring station flux estimation (Figure
2.2-4) and details on data formatting, will not be discussed here. The reader is pointed to the
documentation for the recently released SAS version of the SPARROW model available from
the USGS SPARROW Web site (http://water.usgs.gov/nawqa/sparrow/sparrow-mod.html) for
full details on the model. The reader may also review some of the previous SPARROW
applications presented in Table 2.1-1. The following sections describing SPARROW provide
basic definitions of terms that aid in understanding SPARROW inputs and outputs and discuss
some details that pertain to an application focused on atmospheric deposition inputs. Finally, an
alternate formulation of SPARROW is described that highlights contributions of ammonia to the
total Nr load for use in the Neuse River watershed.
Final Risk and Exposure Assessment
Appendix 6-35
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 2.1-1. Examples of SPARROW Applications
Location
National
Major estuaries of the United States
Chesapeake Bay
State of Kansas waters
Connecticut River Basin
State of New Jersey waters
New England waters
New Zealand river basins
North Carolina coastal watersheds
Tennessee and Kentucky watersheds
Citation
Smith and Alexander, 2000
Alexander et al., 2001
Preston and Brakebill, 1999; Brakebill and
2004
Preston,
Kansas Department of Health and Environment, 2004
NEIWPCC, 2004
Smith et al., 1994
Moore et al., 2004
Alexander et al., 2002; Elliott et al., 2005
McMahon et al., 2003
Hoos, 2005
2.2.1.2 Key Definitions for Understanding SPARROW Modeling
The following definitions have been summarized from the documentation accompanying
the SAS application of the SPARROW model available from the USGS (Schwarz et al., 2006).
Additional references are noted when used.
• Bootstrapping. This is the practice of estimating the properties associated with the model
coefficients by estimating those properties when sampling from a specified distribution using
replacement (e.g., the model coefficients are estimated a number of times until the best evaluation
properties of the coefficients are found).
• Delivered Yield (load per area). This is the amount of nutrients generated locally for each stream
reach and weighted by the amount of in-stream loss that would occur with transport from the reach
to the receiving water. The cumulative loss of nutrients from generation to delivery to the receiving
water is dependent on the travel time and in-stream loss rate of each individual reach (Preston and
Brakebill, 1999).
• Incremental Yield (load per area). This yield represents the local generation of nutrients. It is the
amount of nutrients generated locally (independent of upstream load) and contributed to the
downstream end of each stream reach. Each stream reach and associated watershed is treated as an
independent unit, quantifying the amount of nutrient generated (Preston and Brakebill, 1999).
• In-Stream Loss. This refers to stream attenuation processes that act on contaminant flux as it
travels along stream reaches. A first-order decay process implies that the rate of removal of the
contaminant from the water column per unit of time is proportional to the concentration or mass
Final Risk and Exposure Assessment
Appendix 6-36
September 2009
-------
Aquatic Nutrient Enrichment Case Study
that is present in a given volume of water. According to a first-order decay process, the fraction of
contaminant removed over a given stream distance is estimated as an exponential function of a
first-order reaction rate coefficient (expressed in reciprocal time units) and the cumulative water
time of travel over this distance. Within SPARROW, the in-stream loss rate is assumed to vary as a
function of stream channel length and various flow classes.
Landscape Variables. These variables describe properties of the landscape that relate to climatic,
or natural- or human-related terrestrial processes affecting contaminant transport. These typically
include properties for which there is (1) some conceptual or empirical basis for their importance in
controlling the rates of contaminant processing and transport, and (2) broad-scale availability of
continuous measurements of the properties for use in model estimation and prediction. Examples
include precipitation, evapotranspiration, soil properties like organic content or permeability,
topographic index, or slope. Particular types of land-use classes, such as wetlands or impervious
cover, may also potentially be used to describe transport properties of the landscape.
Land-to-Water Delivery Factor. This factor describes the influence of landscape characteristics
in the delivery of diffuse sources of contamination to the stream. The interaction of particular land-
to-water delivery factors with individual sources may also be important to consider in SPARROW
models.
Monitoring Station Flux Estimation. This refers to the estimates of long-term flux used as the
response variable in the model. Flux estimates at monitoring stations are derived from station-
specific models that relate contaminant concentrations from individual water quality samples to
continuous records of streamflow and time. These estimates are used to calibrate the model in each
application.
Non-linear Regression. The SPARROW model equation is a nonlinear function of its parameters.
As such, the model must be estimated using nonlinear techniques. The errors of the model are
assumed to be independent across observations and have zero mean; the variance of each
observation may be observation-specific. A general method commonly used for these types of
problems, one in which it is not necessary to assume the precise distribution of the residuals, is
nonlinear weighted least squares. This is the estimation method used by SPARROW.
Segmented Watershed Network. This network relates to the system of joined stream reaches that
define the watershed of interest. Previous SPARROW applications have relied on the River Reach
File 1 (RF1) hydrography developed by U.S. EPA (1996) and the 1:100,000 scale National
Hydrologic Dataset (USGS, 1999). These datasets may be used in their original form or modified
as needed depending on application requirements
Final Risk and Exposure Assessment September 2009
Appendix 6-37
-------
Aquatic Nutrient Enrichment Case Study
• Source. SPARROW distinguishes between source categories (e.g., point sources, atmospheric
sources, and animal agriculture) and individual sources (i.e., the rate of supply of contaminant of a
particular category originating in the watershed and draining to a specific stream reach). A variety
of sources based on knowledge of the watershed and inferences from literature may be examined
with SPARROW.
• Stream Reach. This is the most elemental spatial unit of the infrastructure used to estimate and
apply the basic SPARROW models. Stream reaches define the stream channel length that extends
from one stream tributary junction to another. Each reach has an associated contributing drainage
catchment.
• Total Yield (load per area). This is the amount of nutrients, including upstream load, contributed
to each stream reach. These estimates are calculated by stream reach and account for all potential
sources cumulatively and individually (Preston and Brakebill, 1999).
2.2.1.3 Concepts of Importance to Case Study—SPARROW Application
Previous SPARROW applications have typically relied on atmospheric deposition
measurements from NADP and have used wet NOs deposition as a surrogate for nitrogen
deposition over the watershed of interest. Within the case studies conducted, CMAQ-modeled
and NADP-monitored atmospheric deposition was used. Several differences in the final
parameterization of the SPARROW model will most likely result from this variation in input
data.
Expected rules of model coefficient estimation based on source type are described below.
When using direct measures of contaminant mass as a source estimate, "the source-specific
parameter (an) is expressed as a dimensionless coefficient that, together with standardized
expressions of the land-to-water delivery factor, describes the proportion or fraction of the source
input that is delivered to streams (note that source and land-to-water delivery coefficients that are
standardized in relation to the mean values of the land-to-water delivery variables are necessary
to compare and interpret the physical meaning of source coefficients). This fraction would be
expected to be < 1.0 but >0, reflecting the removal of contaminants in soils and ground water"
(Schwarz et al., 2006).
An example of a source of this type would include atmospheric deposition where the
model input would be the mass of nitrogen deposited over the watershed. When using only wet
deposition as an estimate of nitrogen deposition, the model would be expected to account
Final Risk and Exposure Assessment September 2009
Appendix 6-38
-------
Aquatic Nutrient Enrichment Case Study
for the additional nitrogen species (e.g., organic nitrogen, dry deposition of nitrate) to the extent
that they are correlated with the measured inputs of NOs (Alexander et al., 2001). This
accounting is revealed by estimation within the model application of a land-to-water delivery
fraction for wet NOs" deposition (i.e., product of the deposition coefficient and the exponential
land-to-water delivery function) that exceeds 1.0.
Although available estimates for the estuarine watersheds indicate that wet NOs
deposition is highly correlated with dry NOs^plus NH4+ and organic wet deposition, and
estimates of the ratio of (dry and wet TN) deposition to NCVwet deposition for the estuarine
watersheds range from 3.2 to 4.0 with an average of 3.6 (Alexander et al., 2001), the use of
NADP wet NOs measurements requires the assumption that the spatial distribution of the
various nitrogen species across a watershed does not vary. With the inclusion of explicit nitrogen
species in atmospheric deposition measures, this assumption will not be required, and the land-
to-water delivery fraction for the atmospheric deposition source term estimation is expected to be
<1.0. This variation was explored within the case studies as was the general model fit with the
improved atmospheric deposition inputs.
2.2.2 ASSETS Eutrophication Index
2.2.2.1 Background and Description
The NEEA Program defined and developed a Pressure-State-Response framework to
assess the potential for eutrophication termed the ASSETS El. It is categorical, where each of
three indices results in a score that, when combined, result in a final overall score, also known as
the ASSETS El score or rating, which is representative of the health of the estuary. The indices
are as follows:
• OHI. Physical, hydrologic, and anthropogenic factors that characterize the susceptibility of the
estuary to the influences of nutrient inputs (also quantified as part of the index) and eutrophication.
• OEC. An estimate of current eutrophic conditions derived from data for five symptoms linked to
eutrophication.
• DFO. A qualitative measure of expected changes in the system.
The following excerpt from Whitall et al., (2007) describes the objectives in applying the
ASSETS method:
Final Risk and Exposure Assessment September 2009
Appendix 6-39
-------
Aquatic Nutrient Enrichment Case Study
The ASSETS assessment method should be applied on a periodic basis to track
trends in nutrient-related water quality over time in order to test management
related hypotheses and provide a basis for more successful management. The null
hypothesis being tested in this approach is: The change in anthropogenic pressure
as a result of management response does not result in a change of state. The
hypothesis is tested, e.g., to verify whether decreased pressure improves State, or
whether increased pressure deteriorates State. In many cases, a reduction in
pressure will result in an improvement of State, but in some cases, such as
naturally occurring harmful algal bloom (HAB) advected from offshore, it will
not (Whitall et al., 2007).
Influencing Factors/Overall Human Influence
Influencing factors help to establish a link between a system's natural sensitivity to
eutrophication and the nutrient loading and eutrophic symptoms actually observed. This
understanding also helps to illustrate the relationship between eutrophic conditions and use
impairments (Bricker et al., 2007a). Influencing factors are determined by calculating two factors
of susceptibility and nitrogen load, where "susceptibility" provides a measure of a system's
nutrient retention based upon flushing and dilution, and "nitrogen loads" are a ratio between the
nitrogen input to the system from the oceans versus from the land (Figure 2.2-5).
Calculating influencing factors
Overall, the impact of influencing factors for
an estuarine system is determined by a matrix
(figure at right). Several calculations were made
to create the matrix. First, both susceptibility
and load were determined for each estuary and
placed in one of three categories: low, moderate,
or high. The load refers to a ratio of land-based
to oceanic nitrogen inputs, with a high rating
indicating primarily land-based inputs (Bricker
et al. 2003; Ferriera et al. 2007). The estuary's
susceptibility and nutrient loads were compared
in a matrix and given an influencing factors
rating. For example, an estuary with low
nutrient loads and moderate susceptibility is
moderately/slightly influenced. Each of the
systems in the survey can fall into one of five
categories: slightly influenced, moderately/
slightly influenced, moderately influenced,
highly/moderately influenced, and highly
influenced (see Bricker et af. 1999 for details).
Determination of influencing factors
•? >-
+ -g, c Moderately
42 IS" influenced
-iiiilirlv Moderately
iiiHiieiii^'i-]' influenced mn^'iK';.'!'
OQ. slightly (nfluented SMghtlyWuewced
I
low nitrogen
input
moderate nitrogen high nitrogen
input ^jnput
•*) Load (nitrogen ratio)
Due to the uncertainty in loading esrimaces. moderately/slightly anddigtidy influenced have
been ODmbined to both be slightly influenced, and highly/moderately and highly influenced are
combined to be highly influenced throughout the report (^colors indicate grouping^
Figure 2.2-5. Influencing factors/Overall Human Influence index description and decision
matrix (Bricker et al., 2007a).
Final Risk and Exposure Assessment
Appendix 6-40
September 2009
-------
Aquatic Nutrient Enrichment Case Study
The following factors take into account both the natural characteristics of and human
impacts to systems.
• Susceptibility. For a coastal system, susceptibility depends on the flow of water into and out of the
system. This flushing capability is determined by the physical properties (e.g., size, mouth) of the
system as well as the influence of tidal waters and inflow of freshwater from tributaries. When
water flows into and out of the system easily and quickly (i.e., there is a short residence time),
nutrients flush out of the system rapidly, and there is not enough time for eutrophic symptoms to
develop. Systems with short residence times have low susceptibility. The opposite also holds true.
When water, and therefore nutrients, does not flush quickly from the estuary or coastal system,
there is time for eutrophication effects to develop.
• Nitrogen Load. For this assessment, the loading component is estimated as the ratio of nitrogen
coming from the land (i.e., human-related) to that coming from the ocean and is given a rating of
low, moderate, or high (Bricker et al., 2003; Ferreira et al., 2007). For example, a high rating
means that >80% of the nutrient load comes from land, whereas a low rating signifies a land
percentage of <20%. This rating also provides insight into loading management because loads to
systems with primarily ocean-derived nitrogen are not easily controlled. Understanding the sizes of
current and expected future loads provides further insight into the application and success of
management measures.
Overall Eutrophic Condition
To assess the eutrophic conditions of a system, the NEEA relies on five symptoms. Each
of the five symptoms, divided into primary and secondary categories, is assessed based on a
combination of the following factors: concentration or occurrence, duration, spatial coverage,
frequency of occurrence, and confidence in the data (Figure 2.2-6). The two primary symptoms,
chlorophyll a and macroalgal abundance (Figure 2.2-7), were chosen as indicators of the first
possible stage in the process of water quality degradation leading to eutrophication. The
secondary symptoms, which in most coastal systems will develop from the primary symptoms,
include low dissolved oxygen levels, loss of SAV, and occurrences of nuisance/toxic algal
blooms (Figure 2.2-7). At times, the secondary symptoms may also be present or develop
without expression of primary symptoms. Nutrient concentrations are not employed as a
symptom indicator because concentrations may vary between low and high values based on a
number of factors, such as estuary susceptibility, which invalidates the use of nutrient
concentrations alone as an indicator. As stated by Bricker et al., "Through the use of a simple
Final Risk and Exposure Assessment September 2009
Appendix 6-41
-------
Aquatic Nutrient Enrichment Case Study
model, the current framework was established to help understand the sequence, processes, and
symptoms associated with nutrient enrichment. Despite its limitations, it represents an attempt to
synthesize enormous volumes of data and derive a single value for eutrophication in each
estuary, essentially representing a complex process in a simple way" (Bricker et al., 2007a).
Final Risk and Exposure Assessment September 2009
Appendix 6-42
-------
Aquatic Nutrient Enrichment Case Study
ruh ..... ;
Step 1: Determine expression value for each eutrophic symptom in each salinity zone.
Eu trophic symptom expression
values are determined for each
symptom in each salinity zone
(seawater. mixing, and tidal fresh),
resulting in a total of 15 calculations.
The expression is based on a set of
IF, AND, THEN, decision rules that
incorporate the symptom level (e.g.,
concentration), spatial coverage.
and frequency.
CanctMraOon
IF Mum AND
Lo*
AND
chin
Seep 1: Calculate estuary-wide symptom expressions (using chlorophyll a as an example).
Hie expression values are then used to _
EK)I symptom ra/ue re mukip/wd
by the estuary area ratio.
calculate estuary-wide symptom
expressions for each symptom. First,
each expression value is multiplied by
the area of the salinity zone and
divided by the entire area of the
system to establish the weighted
value. Then, the weighted expression
values in the tidal fresh, mixing, and
seawater zone for each symptom are
totaled to calculate the estuary-wide
symptom expression value. This
process Is repeated for all five
eutrophic symptoms. Note that "no
problem" is the rating assigned if the
value is 0, but that "no problem"
and low are combined for discussion
and tabulation throughout the report.
Lo*
Flag'
Value
0.25
0,5
X
weighted
expression value for
tidal fresh zone
for tack symptom, the weighted expression values for the three salinity zones are added.
tfj Mtiigi
— estuary-wide
express!on value for
chlorophyll a
Step 3: Assign categories For primary and secondary symptoms.
Primary and secondary estuary-wide symptom expression
values are determined in a two srcp process.'
Til e average of the primary
symptoms is calculated to represent
the estuary-wide primary symptom
value. The highest of the secondary
symptom values Is chosen to
represent the estuary-wide
secondary symptom expression
value and rating The highest value Is
chosen because an average might
obscure the severity of a symptom If
the other two have very low values
(a precautionary approach).
1)
Estuary-wide
primary symptom value
Estuary-wide symptom rating is determined;
Symptom expression value Symptom rating
ad co £0.3 (aiN
>0.3 to <0.6
Medium
or
or
— Estuary-wide
>0.6 to < 1
High
secondary symptom value
(Highest value is selected)
Step 4: Determine overall eutrophic condition.
i.e
A matrix is used to combine the High
estuary-wide primary and secondary Pnm"
symptom values Into an overall
eutrophic condition rating according
to the categories at right. Thresholds
between rating categories were
agreed on by the scientific advisory
committee and participants from the
1999 assessment (Brlckeretal. 1999),
Moderate
Moderate low Moderate high
Low Secondary 0.3 ModtrweSrajnciary 0* High Serarri in 1.C
'Flap are used to identify components for which data were inadequate or unknown. In these cases, assumptions were made based on conservative estimates
that unknown spatial coverage is at least io%of a zone, frequency at least episodic, and duration at lease days.
Figure 2.2-6. Overall Eutrophic Condition index description and decision matrix
(Bricker et al., 2007a).
Final Risk and Exposure Assessment
Appendix 6-43
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Symptom
Typical high concentration
(ujj L'1) in an annual cycle
determined as the 90th
percentile value.
/- — »,
[Ajfm\
Macroalgae \ '-^HP? \
'* ...
Causes a detrimental impact
on any natural resource.
Typical low concentration
(determined as the 10"
percentile value) in an
annual cycle.
/VllA
Submerged ; WJMf ]
aquatic vegetation **\JT\
- -
A change in SAV spatial area
observed since 1990.
•""*"-
Nuisance/toxic fatP'JH}}
bloom- * #^ •'
Causes detrimental impact
on any natural resources.
Parameter!
Spatial coverage: Frequency:
High >50* Episodic
Moderate 25-50% Periodic
Low 10-25% Persistent
Very low 0-10%
Concentration:
High >20usL-'
Medium 5-20 ug L1
Low 0- 5 ug L'1
Frequency of problem:
Episodic (occasion ali1 random)
Periodic (seasonal, annual.
predictable)
Persistent (always/continuous)
Spa tial co verage: Frequency:
High >50% Episodic
Moderate 25-50% Periodic
Low 10-25% Persistent
\ferylow 0-10%
State:
Anoxia 0 mg L1
Hypoxia 0-2 rng L'1
Biol. stress 2-5 mgL-1
Magnitude of change:
High >SO%
Moderate 25-50*
Low 10-25%
Very low 0-10%
Duration:
Persistent, seasonal. months, variable,
weeks, days, weeks to seasonal,
weeks to months, or days co weeks
Frequency:
Episodic, periodic, or persistent
Low
Low symptom expression:
Cone. Coverage frequency
tow any any
med lii rn mod. - v, low episod ic
high low • v. low episodic
No macroalgal bloom problems
have been observed.
Low symptom expression:
State Coverage Frequency
anoxia mod. - low episodic
anoxia very low periodic
hypoxia low -v. low periodic
hypoxia moderate episodic
stress any episodic
stress mod. - v. low periodic
The magnitude of SAV loss is
low to very tew.
Blooms are either a) short in
duration (days) and. periodic in
frequency; or b) moderate in
duration (days to weeks) and
episodic in frequency.
Expression
Moderate symptom expression:
Cone. Coverage frequency
medium high episodic
medium moderate periodic
high low -v. low periodic
high moderate episodic
Episodic macroalgal bloom
problems have been observed.
Moderate symptom expression:
State Co wag? Frequency
anoxia high episodic
anoxia low periodic
hypoxia moderate periodic
hypoxia high episodic
stress high periodic
The magnitude of SAV loss is
moderate.
Blooms are either a) moderate in
duration (days to weeks) and
periodic in frequency; or b) long
in duration (weeks to months)
and episodic in frequency.
High
High symptom expression:
Cow. Coverage Frequency
medium high periodic
high mod. - high periodic
high high episodic
Periodic or persistent macroalgal
bloom problems have been
observed.
Hjjgih symptom expression:
State Coverage Frequency
anoxia moderate- high periodic
hypoxia high periodic
The magnitude of SAV loss
is high.
Blooms are long in duration
(weeks, months, seasonal)
and periodic in frequency
'For further Technical dooumencation of the methods, tefer to BricKeretal. 1999 and Brickeretal. x>oj.
Figure 2.2-7. Detailed descriptions of primary and secondary indicators of eutrophication (Bricker et al., 2007a).
Final Risk and Exposure Assessment
Appendix 6-44
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Determined Future Outlook
The future outlook relies on a similar combination of factors as the influencing factors
(i.e., a rating of the system susceptibility and nutrient loading in the future). The aim of this
index is to estimate future changes in the system through a combination of any physical,
hydrologic, or pollutant loadings to the system itself or to its contributing watershed through
such actions as watershed management plans, development restrictions, or policy changes
resulting in nutrient decreases. The matrix in Figure 2.2-8 is used to determine the DFO index
rating.
f ^H Calculating future outlook
The analysis for future outlook is an attempt
to determine whether conditions in an estuary
will worsen, improve, or remain unchanged
over the next 20 years.
Li this analysis, expected nutrient input
changes were used to predict whether
eutrophic conditions will improve or worsen.
The system's susceptibility to nutrients is
then used to determine the magnitude of
this change. Population projections are used
as a primary indicator of the level of future
nutrient input changes. However, population
projections are unpredictable. Therefore,
experts at the NEEA update workshop were
asked to predict changes in nutrient load,
based on their knowledge of likely changes in
land use, management measures, and other
activities that affect nutrient loading.
Determination of the future outlook
md dilution
a
go
.5
-5
3
•5
43
'•£3
1
ck
'•>.
~ ImprO',-' high
\ g ~ SvTnpDMiis lil.i?!',' to
L .2 g- irripiwasubscanthll/
§ 1
tfi
Ol =
ftli
Hi
£ is
>.
M
+ ^1
\ .5?' 0.
- -c
-------
Aquatic Nutrient Enrichment Case Study
• Moderate: Any pressure, moderate-low to moderate-high eutrophic condition, and any expected
future change in eutrophic condition.
• Poor: Moderate-low to high pressure, moderate to moderate-high eutrophic condition, and any
expected future change in condition.
• Bad: Moderate to high pressure, moderate-high to high eutrophic condition, and any expected
future change in eutrophic condition.
• Unknown: Insufficient data for analysis.
2.2.2.2 Applications and Updates
The ASSETS El method developed out of the NEEA was first reported in 1999. Since
that time, it has been used in several assessments across the country and internationally and has
undergone revision and validation (Bricker et al., 1999, 2003, 2007a; Ferreira et al., 2007;
Whitall et al., 2007). The original NEEA ASSETS El assessment relied on questionnaires to
experts for each estuary considered (Bricker et al., 1999). Later assessments determined that
reliance on monitored data and less on reports from experts provided a more valid assessment
tool (Bricker et al., 2006, 2007a). With the NEEA Update in 2007 (Bricker et al., 2007a), an
online database was completed in which data users and data holders could access and input data.
Additional datasets have also been collected for smaller study areas (Bricker et al., 2006). These
data systems provide a wealth of information from which analyses may be conducted.
The original formulation of the ASSETS El within the NEEA used watershed nutrient
model estimates from SPARROW (Bricker et al., 1999). Although the updated ASSETS El
methodology has further apportioned nitrogen sources using the Watershed Assessment Tool for
Evaluating Reduction Strategies for Nitrogen (WATERSN) model (Whitall et al., 2007),
SPARROW is still appropriate for this study because atmospheric deposition inputs relative to
other nitrogen sources can be defined.
2.2.3 Assessments Using Linked SPARROW and ASSETS El
The link between the SPARROW model and the ASSETS El occurs when the
SPARROW output is used as the nitrogen load in the OHI index calculation of the ASSETS El
score. For the purposes of this study, a complete analysis from atmospheric deposition loading to
ecological endpoint of the ASSETS El score required an assessment of the relative changes in
the deposition load, the resulting instream nitrogen load to the estuary, and the change in
Final Risk and Exposure Assessment September 2009
Appendix 6-46
-------
Aquatic Nutrient Enrichment Case Study
ASSETS El score. An iterative assessment of the various possible ecological endpoints due to
changing nitrogen loads has not been previously undertaken. The methods that follow have been
designed to allow for set up of a process to link the SPARROW and ASSETS El assessment,
which includes uncertainty analysis. As will be detailed below, at this time the developed process
includes an uncertainty analysis for the ASSETS El assessment, whereas the uncertainty
surrounding the corresponding decreased atmospheric deposition loads predicted by SPARROW
has not yet been implemented. The process is described beginning with the individual
components of each assessment, followed by a description of the iterative processing designed to
sample among the assessments, producing a distribution of results.
The first step in setting up the linked analysis was to create a series of response curves.
The SPARROW model can be used to assess the different atmospheric deposition loads that will
result from changes in the NOX concentrations enforced with any new policy scenarios.
Therefore, a change in atmospheric deposition load produces a corresponding change in TN
loading at the outlet of the watershed, in this case, to the estuary. Converting the TN load to a
concentration value using the flow values employed in the SPARROW modeling allows for the
creation of a relationship between decrease in atmospheric deposition load from the current
condition scenario and instream total nitrogen concentration (TNS). A theoretical representation
of this relationship in an ideal situation is presented in Figure 2.2-9 (dotted lines represent
uncertainty bounds). The red "x" indicates the nitrogen concentration during the current
condition assessment where there is no load decrease in atmospheric deposition. The vertical
asymptote of the curve approaches the nitrogen concentration of the river if there were no
atmospheric inputs. In theory, the watershed response to changes in any loading would be
nonlinear because of various retention and loss processes occurring within the watershed.
However, because SPARROW is a statistical model, the empirical representation of these
processes does not produce a nonlinear response as shown in Figure 2.2-9. As discussed in the
results section, the response to changes in atmospheric deposition load is predicted to be linear
by SPARROW. This highlights one area of uncertainty due to the steady-state, statistical nature
of SPARROW. A benefit of using SPARROW is that the various atmospheric deposition levels
can be run through the model in a short amount of time.
A second response curve is set up for the ASSETS El based on the Influencing Factors/
OKI and OEC index scores (Section 2.2.2), which are functions of TN load. The ASSETS El
Final Risk and Exposure Assessment September 2009
Appendix 6-47
-------
Aquatic Nutrient Enrichment Case Study
assessment is essentially a pressure-state-response scenario (Figure 2.2-10) where the pressure is
the nitrogen load (represented by TNS) and the state is the current OEC index and ASSETS El
scores for the system. Response would be the change in state of the estuary (represented in the
ASSETS El by the DFO index). Bricker et al. (2007b), noted that the shape of the response curve
would vary based on the susceptibility of the system. Therefore, if the susceptibility is known
and held constant, a curve can be created.
T3
ro
o
8.
§
.E a-
<
o c
-------
Aquatic Nutrient Enrichment Case Study
Figure 2.2-10. Example of response for case study analysis (Bricker et al., 2007b).
With the case study analysis, response curves were created for two different estuaries
where the susceptibility is known. As previously described, the ASSETS El score is a
combination of OEC, OHI, and DFO index scores. It is possible to combine all three of these
scores with the ASSETS El into a single response curve when the susceptibility and DFO are
held constant. The DFO may be held constant when alternative effects levels are being evaluated
based on a current condition scenario, such as was done in this study. The susceptibility rating is
based on physical and hydrological conditions. Physical conditions are unlikely to change. The
hydrologic conditions may change because of extreme conditions, such as prolonged drought or
hurricane events, but overall the conditions should average to a steady value that can be used in
the analysis. Figure 2.2-11 highlights this combination of scores where the susceptibility is
"High" and the DFO is set at "Improve." Additionally, by holding the susceptibility constant, the
OHI index score becomes a function of the TNS. This is evident in the double x-axis. The state
response is the OEC index score along the y-axis. Underlying these combinations of OHI and
OEC index scores is the ASSETS El score.
The categorical nature of the assessment produces a mix of a continuous curve based on
TNS and blocks where different ASSETS El scores are valid. The shape of the curve can be
determined by fitting a logistic relationship through a series of points (red "x"s) within the
Final Risk and Exposure Assessment September 2009
Appendix 6-49
-------
Aquatic Nutrient Enrichment Case Study
pressure-state realm based on compilations of historical data. The logistic function is suggested
because of the pressure-state-response nature of the system meaning that initial, lower pressures
are thought to affect the system slowly, then changes become more rapid up until a point where
the pressure is so great that the state changes little because it has almost reached a maximum
level. A logistic response to inputs (e.g., nutrients) is a common pattern in biological systems and
is well documented. A variety of logistic (or sigmoidal) functions are available. In this case
study, the following four-parameter function was used:
OEC(TNS) = A + ^—-^ (6)
1 + exp
where
OEC(TNS) = OEC index score as a function of TNS (unitless)
TNS = TNs(mgN/L)
A, B, C, D = Parameters to be estimated.
Applying this logistic function as in Figure 2.2-11, parameter^ affects the OEC
intercept value, b in Figure 2.2-11; parameters A and B affect the OEC asymptote value; C
affects the "S" shape; and D shifts the "S" horizontally. The curve is fit through a nonlinear
optimization routine using a modified version of Box's algorithm (Box, 1965). Constraints on
the boundary conditions for the function are described below. At this time, 500 iterations of the
algorithm were used to define the function. A future assessment must complete a Monte Carlo
simulation analysis to determine a justifiable estimate of the number of iterations needed to bring
the algorithm to convergence for the problem as thus defined.
Final Risk and Exposure Assessment September 2009
Appendix 6-50
-------
Aquatic Nutrient Enrichment Case Study
o -S
O ro
(fl
O
LLJ
O
Low
Instream Total Nitrogen Concentration (TH5
1 1 V
Moderate Low Moderate Moderate High
OHI Score (assuming constant susceptibility)
High
Figure 2.2-11. ASSETS El response curve.
Within the analysis space created by both the OHI and OEC index scores, the axes are
limited to the scores of zero (really one) to five, but the corresponding TNS must be determined
separately. Point "a" represents the background nitrogen concentration that would occur in the
system with no anthropogenic inputs (assuming the system is not naturally eutrophic) or the
system at a pristine state. In almost all cases, this value will be unknown because of the extent to
which anthropogenic inputs have influenced the nation's ecosystems. A lower bound and upper
bound on this value, between which the algorithm randomly selects a different realization for
each iteration, were specified. The lower bound was specified as the offshore TN concentration
(used in the OHI index score) and the upper bound was specified at 1.0 milligrams per liter
(mg/L) for the Potomac River/Potomac Estuary Case Study Area based on best professional
judgment and at 0.1 mg/L for the Neuse River/Neuse River Estuary Case Study Area based on a
combination of the lowest monitored nitrogen values over a number of years and best
professional judgment. The upper bound of the TNS is the maximum nitrogen concentration at
which the stream is nitrogen-limited; above this point, the nitrogen inputs to the system no longer
Final Risk and Exposure Assessment
Appendix 6-51
September 2009
-------
Aquatic Nutrient Enrichment Case Study
affect the eutrophication condition. Several attempts to quantify this limit were made through
historical data analysis, examination of nitrogen to phosphorus (N:P) ratios within the estuary,
and consideration of the underlying eutrophi cation processes. For this first assessment, a lower
and upper bound on this maximum TNS between which the algorithm randomly selects a
different realization for each iteration were specified. The lower bound was specified as the
maximum observed TNS, whereas the upper bound was specified as 50% greater than this value
based on best professional judgment.
Further measures of uncertainty could be taken into account using upper and lower
bounds on the logistic curve. The horizontal spread at each OEC index score reflects the range of
nitrogen concentrations that may relate to that specific OEC index score. The vertical spread
relates to the range in OEC index scores that may occur at each nitrogen concentration. As with
the TNS axis, the lower and upper bounds on both extremes of the OEC scale were specified
again. At both ends of the scale (i.e., 0 and 5), the OEC maximum and OEC minimum values
could be randomly selected at each iteration within a range of 1 OEC unit or could be set at
constant values of 1 and 5. If randomly selected, the actual OEC minimum would be between 0
and 1 at the "0" end of the scale and between 4 and 5 at the "5" end of the scale. This would
account for the likelihood that an estuary may not physically be capable of reaching either a
"pure, pristine" condition (e.g., OEC = 5) or a "completely hypereutrophic" condition (e.g.,
OEC = 0). Consideration of this allowance and of holding the scores constant is discussed in the
uncertainty section.
2.2.3.1 Back Calculation Method
The creation of the two response curves allows an analyst to work backward from the
ecological endpoint to the source of the impairment; in this case from the ASSETS El score to
the atmospheric deposition loading of oxidized nitrogen. To accomplish this back calculation, a
computer program has been generated that processes through a series of user inputs, defined
boundary conditions, and set cases for determining the ASSETS El score to perform the iterative
calculations described in setting the logistic curve and determining the TNS. Currently, the
program is coded in the Visual Basic language and requires an input file defined by the user to
contain:
• The desired ASSETS El ecological endpoint (i.e., score)
Final Risk and Exposure Assessment September 2009
Appendix 6-52
-------
Aquatic Nutrient Enrichment Case Study
• Oceanic nitrogen and salinity values
• Instream salinity values
• The regression equation relating the decrease in atmospheric deposition of nitrogen to TNS
• The number of realizations on which to iterate the model calculations.
2.2.3.2 Uncertainty Bounds on TN& OEC Min/Max Values
Based on the degree of uncertainty, three different scenarios exist for the analysis (Figure
2.2-12a through Figure 2.2-12c). There are two relationships, or functions, involving uncertainty
in the proposed methodology. The first is the SPARROW-predicted TNS at the head of the
estuary given a total nitrogen atmospheric deposition load (TNatm). (Note that SPARROW model
predictions are actually provided in terms of instream TN loads [mass per time]; however,
concentrations [mass per volume] can be calculated by dividing the instream TN load by the
flow rate.) This functional relationship is denoted as TNs(TNatm) where the TNS is actually
evaluated on changes in the loading of oxidized nitrogen, not all nitrogen species, in the TN
atmospheric deposition load. This means that decreases are applied to the NOX load within the
atmospheric load, and a new TNatm is calculated). (For instance if the NOX load contribution is 10
kg N/yr and TNatm is 20 kg N/yr, then a 40% decrease would result in a TNatm equal to 16 kg
N/yr after decreasing the 10 kg N/yr by 40%.) Because of the SPARROW regression
uncertainties, this function is a probability distribution (i.e., given an TNatm, there are many
alternative TNS values). This distribution function is denoted as TNspdf(TNatm). Under the
standard assumptions of regression modeling, this is a normal distribution with a mean value
represented by the SPARROW estimate of TN load. The variability around the mean value
represents uncertainty in the SPARROW parameter estimates. However, the SPARROW
regression model does not give rise to normally distributed residuals, and some nonparameteric
methods have been developed by the SPARROW developers to estimate SPARROW'S
confidence limits on predictions (Schwarz et al., 2006). This measure of uncertainty has not yet
been incorporated into this methodology, and that is why the results for the Potomac
River/Potomac Estuary Case Study Area and the Neuse River/Neuse River Estuary Case Study
Area are based on uncertainty Scenario B (explained below; Figure 2.2-12b). Accordingly, the
uncertainty presented in those results will be an underestimate of the total uncertainty.
Final Risk and Exposure Assessment September 2009
Appendix 6-53
-------
Aquatic Nutrient Enrichment Case Study
The second uncertain function is the semi quantitative relationship between the ASSETS
El score and the TN load. The ASSETS El model yields a result based on three index "scores,"
the OEC, the OKI, and the DFO index scores as described in Section 2.2.2. The OEC and OKI
index scores are functions of TNS. Thus, the ASSETS El model can be written as ASSETS El =
f(OEC(TNs), OHI(TNS), DFO) where "f' is the functional relationship that is the ASSETS El
methodology. Based on the methodology developed for creation of the response curve, the
OEC(TNS) function also involves uncertainty, and the OEC index score resulting from any
particular TNS is also a probability distribution, which can be assumed to be a uniform
distribution for this first analysis. Incorporating this uncertainty, the ASSETS El model can be
expressed for this study as ASSETS El = f(OECpdf(TNs), OHI(TNS), DFO). Figure 2.2-12b and
Figure 2.2-12c present the iterative steps in which the probability distribution function of OEC
results is created in the coded model.
In setting up Scenario B for the evaluation of alternative effects levels, the goal was set to
determine the change in oxidized nitrogen load required to improve the ASSETS El score by
one, two, and three categories from its current level set in the 2002 current condition analysis.
Improvement in the ASSETS El score by categorical values means moving along the logistic
curve set to data points determined through the gathering of historical monitoring data and
reports. Figure 2.2-13 provides a visualization of what it means graphically to improve by one
ASSETS El score category. In reality, the exact results of improving by one, two, or three
categories will vary depending on the estuary and the baseline state of the estuary (i.e., where on
the logistic curve the analysis begins). In the example shown in Figure 2.2-13, improving the
ASSETS El score by one category also improves the OEC index score by one category but
allows for a decrease in the TNS, which results in the same OHI index score as the baseline. Also
in this example, the baseline for the estuary is an ASSETS El score of "bad," thus it is possible
to improve by three categories although, in doing so, the OEC and OHI/TNS would have to
improve to an almost pristine state. If a system begins with an ASSETS El score of "Moderate"
or "Good," the assessment will only be able to examine an improvement in two or one ASSETS
El score categories, respectively. Also noted in Figure 2.2-13 is the direction of the category
movement. In this example, improvement of the ASSETS El score along the determined logistic
curve is a vertical movement down rather than a horizontal move to the left (not illustrated with
this example); therefore, the OHI index score may remain in the same or similar state. If the
Final Risk and Exposure Assessment September 2009
Appendix 6-54
-------
Aquatic Nutrient Enrichment Case Study
movement were to the left, the OHI index score would have to improve, whereas the eutrophic
condition of the system (i.e., OEC index score) could remain the same or similar to the baseline
conditions.
ASSETS model: El = f(OEC(TNs), OHI(TNS), DFO)
SPARROW model response curve: TNs(TNatm) = a + b TNatm
Terms: El Eutrophication Indicator
f Denotes functional relationship between
parameters (i.e., f is the ASSETS methodology)
OEC Overall Eutrophic Condition
OHI Overall Human Influence
DFO Determined Future Outlook
El* Eutrophication Indicator of interest
TNS* Total Nitrogen concentration of interest
TNatm* Total atmospheric nitrogen deposition load evaluated
by decreasing the A/Ox contribution to deposition
a, b constants
i, j number of function interations
Cl Confidence interval
A) No uncertainty in OEC(TNS) or in TNs(TNatm)
Find TNS* to satisfy
El* = f(OEC(TNs*), OHI(TNS*), DFO)
Find
to satisfy TNS
From SPARROW response curve
atm
TN*
Result: TNatm*
(scalar value)
Figure 2.2-12a. Back calculation analysis scenario A: no uncertainty.
Final Risk and Exposure Assessment
Appendix 6-55
September 2009
-------
Aquatic Nutrient Enrichment Case Study
B) Uncertainty in OEC(TNS); No uncertainty in TNs(TNatm)
No
Select number of iterations to
sample OEC(TNS) function (n)
Select a OEC(TNS) function (i)
(Fit of a logistic function to OEC, TNS data pairs where
endpoints are randomly selected within a specified range)
Find TNS* to satisfy
El* = f(OEC(TNs*), OHI(TNS*), DFO)
Find TNatm* to satisfy TNS*
From SPARROW response curve
atm*
b
TN*
Present Results
TNatm*i fori = 1, ..., n
Figure 2.2-12b. Back calculation analysis scenario B: uncertainty in ASSETS El assessment.
Final Risk and Exposure Assessment
Appendix 6-56
September 2009
-------
Aquatic Nutrient Enrichment Case Study
C) Uncertainty in OEC(TNS) and TNs(TNatm)
No
Select number of iterations to
sample OEC(TNS) function (n)
Select number of iterations to sample
probability distribution of TNs*i(TNatm*i) (m)
Select a OEC(TNS) function (i)
(Fit of a logistic function to OEC, TNS data pairs where
endpoints are randomly selected within a specified range)
Find TNS* to satisfy
El* = f(OEC(TNs*), OHI(TNS*), DFO)
For a given percentile (j) of the TNs*i(TNatm*i) probability distribution
IN,*, = a + b TNatm*i+ Cl
Find the TNatm*j that satisfies TNS*
^No-
Present Results
TNatm*i,j for i = 1,...,n
j = 1, .... m
Figure 2.2-12c. Back calculation analysis scenario C: uncertainty in both ASSETS El
assessment and nitrogen loading assessment.
Final Risk and Exposure Assessment
Appendix 6-57
September 2009
-------
Aquatic Nutrient Enrichment Case Study
T3
o
o
u
V)
O
LLI
O
TJ
O
•g
r
.•"
..-••
<&-
• / •••"
S X
] Specified
ovements in
ETS Scores
(2) Resulting
Change in N
Needed to Meet
Enid points
In stream Total Nitrogen Concentration (TH5)
1
1 - 1
Low Moderate Low Moderate Moderate High
OHI Score (assuming constant susceptibility}
High
Figure 2.2-13. Example of an improvement by one ASSETS El score category in a back
calculation assessment.
In the second piece of the back calculation, the newly determined TNS needed to make an
improvement in the ASSETS El score is examined on the response curve of the TNS to TNatm.
Figure 2.2-14 presents an example similar to those found in the case studies where the response
is linear (note that the axes have been flipped from original representation in Figure 2.2-9 to
reflect how the response curve is handled within the coded program developed for this study).
While the linear nature of this curve during this preliminary assessment is a function of the use
of SPARROW, a statistical watershed model, valuable information on the required decrease in
TNatm, can still be determined. The scale of the axes of the response curve also provides a great
deal of information on the system of interest. For instance, if the TNS changes that result from a
small change in the TNatm are much greater (e.g., 5% change in atmospheric load versus 20%
change in instream concentration), then the system is highly influenced by the atmospheric
deposition of nitrogen. If the situation is reversed and large changes in TNatm result in only small
Final Risk and Exposure Assessment
Appendix 6-58
September 2009
-------
Aquatic Nutrient Enrichment Case Study
changes in the TNS, then the system is not greatly influenced by the atmospheric deposition of
nitrogen.
Specified change in instream
total nitrogen concentration from
improvement in ASSETS El
g
S
O
O
c
-------
Aquatic Nutrient Enrichment Case Study
reporting on various issues in the Chesapeake Bay, including the Chesapeake Bay Program
(CBP) and the Virginia Institute of Marine Science (VIMS).
3.1.1.1 SPARROW Assessment
The Version 3 Chesapeake Bay SPARROW application modeled the watershed for the
time period of the late 1990s. Stream nitrogen load estimates from 87 sites were used to calibrate
the model. The stream reach network used in this analysis relied on a modified version of the
RF1 used in previous Chesapeake Bay SPARROW applications, but included 68 reservoirs that
were not previously included. This analysis examined the sources of atmospheric deposition,
fertilizer and manure application, point sources, and land use. For details on the compilation of
each of these GIS-based datasets, see Brakebill and Preston (2004). Watershed characteristics
considered in the model as loss and decay variables include precipitation, temperature, slope, soil
permeability, and hydrogeomorphic regions.
The Version 3 calibrated model was selected to create the results for the 2002 current
condition analysis because of its temporal proximity to the desired base year. The 5-year
difference from 1997 to 2002 was not expected to result in a large change in the model if it were
recalibrated to more recent data (S. Preston, personal communication, 2008). Future updates to
this study should consider recalibrating the model to 2002 data to ensure that this assumption
holds true.
The SPARROW assessment for the 2002 current condition analysis used the same source
inputs and watershed characteristics, except in the case of atmospheric deposition. The Version 3
Chesapeake Bay SPARROW application relied on 1997 mean deposition values of wet-
deposition atmospheric NOs" using the 191-point measurements in the NADP program across the
country. As described in Section 2.2.1.3, relying on wet MV deposition as a surrogate for TN
deposition requires an assumption of spatial homogeneity between the nitrogen species. The
2002 current condition analysis used the CMAQ/NADP data that were prepared for the Risk and
Exposure Assessment as atmospheric deposition inputs to the model. Although the Version 3
Chesapeake Bay SPARROW model was calibrated against only wet NOs deposition loads, it
was decided that the 2002 current condition analysis should incorporate all available forms of
nitrogen deposition in order to fully reflect the result of changing oxidized nitrogen loads.
Final Risk and Exposure Assessment September 2009
Appendix 6-60
-------
Aquatic Nutrient Enrichment Case Study
The use of the 2002 atmospheric deposition loads with the 1997 Version 3 of SPARROW
introduces a source of uncertainty to the modeled results because the source parameter for
atmospheric deposition was calibrated to the spatial variability and magnitude of the original
atmospheric source. This degree of uncertainty will be omitted from any future analyses when
the time and data exist to recalibrate the SPARROW model to a full set of 2002 current condition
data. At this time, the study's goal, which is to reflect on the changes in magnitude of nitrogen
loads to the estuary due to changes in the deposition of oxidized nitrogen, can still be assessed
and described using this data compilation. The 1997 data used in the original Version 3
Chesapeake Bay SPARROW application shows wet NOsyields highest in the western,
mountainous region of the Potomac River watershed with lower values around the southern
central portion and areas surrounding the Potomac Estuary (Brakebill and Preston, 2004). There
is a clear trend of interpolation from the NADP data. In comparison, Figure 3.1-la through
Figure 3.1-lc reveal highly different spatial patterns in oxidized, reduced, and TN atmospheric
deposition yields across the Potomac River and Potomac Estuary watershed. Note that the scales
across the three figures use the same increments and colors so that they can be directly
compared. For the current condition 2002 analysis of the Potomac River and Potomac Estuary,
an estimated 40,770,000 kg of TN was deposited in the 3.2-million hectares (ha) Potomac River
watershed, for an average TN deposition of 12.9 kg N/ha/yr.
Application of the SPARROW model provides estimates of the incremental flux derived
within each catchment of the Potomac River watershed, as well as estimates of how much of that
incremental flux (i.e., delivered flux) ultimately reaches the estuary (Figure 3.1-2). By looking at
catchment-scale results, the spatial variability among flux/load contributions across the
watershed can be shown. Differences between the incremental and delivered yields/loads reflect
the instream losses that occur as the load from each catchment travels downstream to the target
estuary.
For this first application and analysis of the 2002 current condition case, SPARROW was
used to model the loads from the Potomac River and its watershed to the upper portions of the
Potomac Estuary. The most downstream modeled catchment in the analysis lies downstream of
several major point sources between Washington, DC, and the mixing zone of the estuary. These
point sources were major contributors of nutrients to the estuary, and by including them in the
analysis a more accurate load from the Potomac River watershed is defined than if the modeling
Final Risk and Exposure Assessment September 2009
Appendix 6-61
-------
Aquatic Nutrient Enrichment Case Study
stopped at the fall line of the river. Direct runoff from catchments surrounding the Potomac
Estuary and direct deposition to the estuary were not considered in this preliminary model
application. The majority of the nitrogen loading to the estuary was expected to derive within the
Potomac River watershed because of its overall larger land area and applications of fertilizer and
manure. Additionally, the major point sources to the Potomac Estuary were included in the most
downstream watersheds at the mouth of the estuary modeled in this application.
Potomac River Watershed: Atmospheric Deposition - Oxidized Nitrogen
Legend
| | Potomac River Watershed
I | NOAAUV* HUC8 Border
Atmospheric Deposition"
By Watershed Unit
(All units are in l\u li-'i vn
~U-6
|8-10
I 10- 12
** Oxidized nitrogen species and
sources: wet- nitrate (NADP); dry -
particulate nitrate, nitric acid,
nitrogen pentoxide, nitrous acid,
nitric oxide, nitrogen dioxide,
peroxycyl nitrate , and organic
nitrate (CMAQ)
0 10 20
40
60
uMiles
Watershed boundary layers
and mapping data were provided
byllSGS. The site URL is:
http://md.waterusgs.gov/gis/chesbay/
sparrow3/doc/retv3.htm#section1
Figure 3.1-la. Atmospheric deposition yields of oxidized nitrogen over the Potomac
River and Potomac Estuary watershed.
Final Risk and Exposure Assessment
Appendix 6-62
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Potomac River Watershed: Atmospheric Deposition - Reduced Nitrogen
Legend
Potomac River Vtetershed
I I NOAAU\Aft HUCB Border
Atmospheric Deposition "
By Watershed Unit
(All units are in kg.liayrt
| 2-4
| |4-6
^H B- 1°
H| 10- 12
^H 12- 14
"Reduced nitrogen species and
sources: wet- ammonium (NADP);
dry - ammonia and ammonium
(CMAQ).
0 10 20
40
60
ntvliles
Watershed boundary layers
and mapping data were provided
bylJSGS. The site URL is:
http://rnd.water.usgs.gov/gis/chesbay/
sparrow3/doc/retv3.htm#section1
Figure 3.1-lb. Atmospheric deposition yields of reduced nitrogen over the Potomac River and
Potomac Estuary watershed.
Final Risk and Exposure Assessment
Appendix 6-63
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Potomac River Watershed: Atmospheric Deposition - Total Nitrogen
Legend
Potomac River V^tershed
NOAAUWft HUC8 Border
Atmospheric Deposition "
By Watershed Unit
(All iinii; .neiii [HI h.i vn
f |e-s
I |8-10
**Total nitrogen is the sum of
oxidized and reduced nitrogen
species. Oxidized nitrogen
species and sources: wet-
nitrate (NADP); dry - particulate
nitrate, nitric acid, nitrogen
pentoxide, nitrous acid, nitric
oxide, nitrogen dioxide,
peroxycyl nitrate, and organic
nitrate (CMAQ). Reduced
nitrogen species and sources:
wet- ammonium (NADP); dry-
arnrnonia and ammonium (CMAQ).
0 10 20 40 60^
Watershed boundary layers end mapping
cMawsre provided by U SOS. The site URLis
tttp://md.v*ater .usgs.gov/gis/chesfaay/
stj.atTij'.^j.'d.'i./": ': i '[• .'r;:i-jdion1 _
Figure 3.1-lc. Atmospheric deposition yields of total nitrogen over the Potomac River and
Potomac Estuary watershed.
Final Risk and Exposure Assessment
Appendix 6-64
September 2009
-------
Aquatic Nutrient Enrichment Case Study
2002 Base Case Results for Potomac Watershed
Map #1: Local Nitrogen Generation
(Incremental Yield)
Map #2: Local Nitrogen Load at Receiving Estuary
(Delivered Yield)
Legend
| | Potomac River
Watershed
I Cathrnents not included
J in SPARROW Model ing
Nitrogen Yield
by Watershed Unit
(all units are in kg/ha/yr)
Map 81
Incremental Yield
I l«<
| | 4-6
liles
Watershed boundary layers
and mapping data were provided
by USGS. The site URL is:
http:// md. wate r usg s. g ov/g is/ c he sb ay/
sp arro w3 do c/retv 3. htm#se cti on 1
Figure 3.1-2. Total nitrogen yields from all sources as predicted using the Version 3 of the
Chesapeake Bay SPARROW application with updated 2002 atmospheric deposition inputs.
Final Risk and Exposure Assessment
Appendix 6-65
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Overall, the model produced an estimate
of TN loading to the Potomac Estuary of
36,660,000 kgN/yr. The atmospheric deposition
SPARROW modeling for 2002 predicts that
atmospheric deposition was 20% of the total
nitrogen loading to the Potomac River's
estuary, producing an TNS of 3.4 mg/L.
load was estimated at 7,380,000 kg N/yr or 20% of the total loading (Figure 3.1-3). These
modeling estimates are consistent with previous modeling estimates for the system (Preston and
Brakebill, 1999). The TNS resulting from this loading was approximately 3.4 mg/L.
Relative Contributions of Nitrogen Sources to Potomac Estuary
20%
25%
D Manure
D Fertilizer
• Urban Area
D Point Sources
D Atmospheric Deposition
46%
Total load to estaury:
36,660,000 kg N/yr
Figure 3.1-3. Source contributions to Potomac Estuary nitrogen load.
3.1.1.2 ASSETS El Assessment
An ASSETS El assessment was completed for the Potomac Estuary (Figure 3.1-4) in a
2006 NOAA project on the Gulf of Maine (Bricker et al., 2006). The data used to complete the
scoring were from 2002. That assessment showed that the system has a high susceptibility to
pressures and a high score for nutrient inputs, resulting in a score of High for influencing factors
(i.e., OHI index). Individual scores for the primary and secondary indicators varied, but resulted
in an overall score of High for the OEC (i.e., OEC index). The score of Improve Low for the
future outlook (i.e., DFO index) is based on the expectations that future nutrient pressures will
decrease and there will be significant population and development increases.
The ratings for the nutrient inputs and the OEC index were recreated and verified using
methods consistent with the 2007 NEEA Update (Table 3.1-1; Bricker et al., 2007a), which
included separate areal-weighted consideration of the tidal fresh, mixing, and saltwater zones
Final Risk and Exposure Assessment
Appendix 6-66
September 2009
-------
Aquatic Nutrient Enrichment Case Study
within the estuary (Bricker et al., 2007a). (Note that the Potomac Estuary is split approximately
between tidal fresh [14.5%] and mixing [85.5%] zones.) Data for dissolved oxygen and
chlorophyll a were downloaded from the CBP water quality database (CBP, 2008). Harmful
algal bloom (HAB) data were taken from data compilations available from VIMS. Conflicting
results were found for the SAV coverage where an unreasonably large increase was reported in
the original analysis (Bricker et al., 2006). The source of the data for the updated analysis was
the VIMS (VIMS, 2008), which was the same source cited for the original analysis. Possible
explanations for the discrepancy are different baseline periods for comparisons between the gains
and losses in SAV areas. The updated analysis compared only annual changes in areal growth of
SAV (in contrast to previous NEEA studies that used a baseline year of comparison [Bricker et
al., 1999; Bricker et al., 2007a]). Additionally, the areas of SAV measured for 2001 were listed
as partial measures, making the estimates in small gains and losses in the tidal fresh and mixing
areas, respectively, of the estuary uncertain. In the 2007 NEEA Update (Bricker et al., 2007a), a
value of 0.5 was used for any uncertain index score. A value of 0.5 was, therefore, given to the
uncertain loss estimated in the mixing zone of the Potomac Estuary for 2002.
Index scores for the updated analysis were compiled using the scoring methods and
matrices as shown in Figure 2.2-6 and Figure 2.2-7. Although there were uncertain/unknown
values for macroalgae in the primary symptoms and SAV in the secondary symptoms, the high
rankings of chlorophyll a within the primary symptoms and HAB in the secondary symptoms
overweighed the uncertainty of the other parameters in the final scores. Even if a value of 0.5
(denoting uncertainty) had been used for the macroalgae score, as was done with the mixing zone
index for SAV, the overall score for that indicator would have resulted in a primary score greater
than 0.6 and a High ranking. Combination of the primary and secondary scores (both High)
provided an overall OEC index score of High, which agreed with the original analysis.
The OHI index score (confirmed with the modeled nitrogen load from the 2002
SPARROW application) and the DFO index score remain the same as in the original analysis.
Therefore, the ASSETS El score for the 2002 current condition scenario is Bad.
Final Risk and Exposure Assessment September 2009
Appendix 6-67
-------
Aquatic Nutrient Enrichment Case Study
Indkes Methods Parameters/ Values /EAR Index category ASSETS grade
Pressure
OHI index
State
OEC index
Response
DFO index
Susceptibility
Nutrient inputs
Primary
Symptom
Method
Secondary
Symptom
Method
Future nutrient
pressures
Dilution potential
Fl ushi ng potential
High
Low
High
Susceptibility
High
Chlorophyll a
iVacroalgae
Di ssQhffidoaygen
Submerged
aquatic/egetation
Nuisance and
Toxic Blooms
High
NoProb
Low
Large Increase
Problem (1)
High
High
Futurenutrientpres5uresdecrease,5ignifi cant population/
development increases - Improve Low
High
High
Improve Low
OHI =1
OEC = »
DFO = 4
Bad
Figure 3.1-4. The ASSETS El scores for the Potomac Estuary (Bricker et al., 2006).
Final Risk and Exposure Assessment
Appendix 6-68
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 3.1-1. Potomac Estuary Current Condition Overall Human Influence Index Score
Year
2002
Parameter
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
HAB
Zone
MX
TF
ALL
MX
TF
MX
TF
ALL
Value
13.25
22.805
NA
3.6
5.8
NA
NA
NA
Concentration
MEDIUM
HIGH
UNKNOWN
BIO STRESS
NO
PROBLEM
LOSS
(Uncertain)
GAIN
(Uncertain)
NA
Spatial
Coverage
HIGH
LOW
UNKNOWN
HIGH
HIGH
NA
NA
NA
Frequency
PERSISTENT
PERIODIC
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
NA
Expression
HIGH
MODERATE
UNKNOWN
MODERATE
NO PROBLEM
UNCERTAIN
(0.5)
UNCERTAIN
(0)
HIGH
Score
0.9275
NA
0.4275
0.43
1
Primary/
Secondary
Scores
0.9275
1
OEC
Score
HIGH (1)
CHLA: Chlorophyll a
DO: Dissolved Oxygen
SAV: Submerged Aquatic Vegetation
HAB: Harmful/Toxic Algal Blooms
MX: Mixing Zone
TF: Tidal Fresh Zone
ALL: All Estuary Zones
NA: Not Applicable
Final Risk and Exposure Assessment
Appendix 6-69
September 2009
-------
Aquatic Nutrient Enrichment Case Study
3.1.2 Summary of Results for the Neuse River/Neuse River Estuary Case Study
Area
The 2002 current condition analysis of the Neuse River and Neuse River Estuary used
recently released data from the USGS to calibrate a new SPARROW application for 2002 to the
Neuse, Tar-Pamlico, and Cape Fear rivers' watersheds. Developing the ASSETS El score for the
Neuse River Estuary proved to be a greater challenge than for the Potomac Estuary because of
the availability of data sources that were less consolidated and more varied.
3.1.2.1 SPARROW Assessment
The release of digital data by the USGS for its Southeast Major River Basin SPARROW
Assessment (area includes all of the river basins draining to the south Atlantic and the eastern
Gulf of Mexico, as well as the Tennessee River basin [referred to collectively as the SAGT
area]) in the summer of 2008 provided the opportunity to calibrate a new SPARROW model for
the 2002 current condition analysis (Hoos et al., 2008). The SAGT data were compiled for 2002,
providing the necessary data inputs and calibration TN loads for model development. Because of
a limited number of calibration points within the Neuse River watershed itself, the SPARROW
model was expanded to include the Tar-Pamlico and Cape Fear river watersheds, providing a
total of 41 calibration points on which to base the SPARROW model. The river network within
these basins was again based on the RF1, with enhancements for calibration points. Source
variables investigated in the new model development included atmospheric deposition (modified
from the original SAGT dataset to use the CMAQ/NADP data developed for this study),
fertilizer application to farmland, manure from livestock production, point sources, and land
cover (urban, forest, and nonagriculture categories). Decay and loss terms considered in the
model included soil permeability, mean annual temperature, and slope.
Figure 3.1-5a through Figure 3.1-5c show the atmospheric deposition inputs used within
the modeling effort. The model was based on TN loads from deposition, but oxidized and total
Nr yields are also presented to highlight source information within the watershed. The Neuse
River watershed is the location of major agricultural operations focusing on swine facilities.
These operations are evident in the high levels of reduced nitrogen found within the south-central
catchments of the watershed (Figure 3.1-5b). For the current condition 2002 analysis of the
Final Risk and Exposure Assessment September 2009
Appendix 6-70
-------
Aquatic Nutrient Enrichment Case Study
Neuse River and Estuary, an estimated 18,340,000 kg of TN was deposited in the 1.3 million-ha
Neuse River watershed, for an TN average deposition of 14.0 kg N/ha/yr.
Development of the new SPARROW model used the USGS SAS SPARROW application
in predictive mode, with bootstrap analyses completed for additional analysis and exploratory
options (Schwarz et al., 2006). The model used two flow classes based on the work by McMahon
et al. (2003) and a reservoir decay factor. In model set up, the decay factors were constrained to
be nonnegative. Tables 3.1-2 and 3.1-3 present the calibrated model parameters and model
evaluation statistics, respectively. The model used point sources, atmospheric deposition,
fertilizer application, manure production, and urban land area. There was some lack of
significance in the parameter estimations for manure production and urban land area, but these
sources were deemed likely sources based on watershed knowledge and were left in the model.
The land-to-water delivery factor used in the model was the soil permeability factor. As
expected, the parameter estimate for this factor was less than one. The three decay terms (i.e.,
two flow classes and reservoir decay) all lacked significance in their parameter estimates but are
required for the model as it was configured. Future analyses should consider alternate flow
classes to investigate ways to improve statistical significance of these estimates. Overall, the
model evaluation criteria reveal a strong significance in the estimated model, with high
prediction values (i.e., R-squared values close to 1), little multicollinearity (i.e., Eigen value
spread less than 100), and normally distributed weighted model residuals (i.e., probability plot
correlation coefficient close to 1).
Final Risk and Exposure Assessment September 2009
Appendix 6-71
-------
Aquatic Nutrient Enrichment Case Study
Neuse River Watershed: Atmospheric Deposition - Oxidized Nitrogen
Legend
Atmospheric Deposition *
By Watershed Unit
|AII units are in ktj luvyrf
H4-6
8- 10
** Oxidized nitrogen species and
sources: wet - nitrate (NADP); dry -
participate nitrate, nitric acid,
nitrogen pentoxide, nitrous acid,
nitric oxide, nitrogen dioxide,
peroxycyl nitrate, and organic
nitrate (CMAQ)
0 5 10 20 30 40 50
Watershed boundary layers
and mapping data were provided
byUSGS. The site URL is:
http://md.water.usgs.gov/gis/chesbay/
sparrow3/doc/retv3.htm#sectionl
Figure 3.1-5a. Atmospheric deposition yields of oxidized nitrogen over the Neuse River
and Neuse River Estuary watershed.
Final Risk and Exposure Assessment
Appendix 6-72
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Neuse River Watershed: Atmospheric Deposition - Reduced Nitrogen
Legend
Atmospheric Deposition "
B/ Watershed Unit
jAII units are in ky/ha'yr)
** Reduced nitrogen species and
sources: wet- ammonium (NADP);
dry - ammonia and ammonium
(CMAQ).
16 24 32
40
Miles
Watershed boundary layers
and mapping data were provided
byUSGS. The site URL is
http://md.water.usgs.gov/gis/chesbay/
spa rrow3/do c/r etv3 .htm#sectio n1
Figure 3.1-5b. Atmospheric deposition yields of reduced nitrogen over the Neuse River
and Neuse River Estuary watershed.
Final Risk and Exposure Assessment
Appendix 6-73
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Neuse River Watershed: Atmospheric Deposition - Total Nitrogen
Atmospheric Deposition
By Watershed Unit
{All units (ire in ky/hayr
6-8
-10
- 12
12- 14
**Total nitrogen is the sum of
-n:-:idii:ed and reduced nitrogen
species. Oxidized nitrogen
ecies and sources: wet-
nitrate (NADP); dry - particufate
nitrate, nitric acid, nitrogen
pentoxide, nitrous acid, nitric
oxide, nitrogen dioxide,
peroxycyl nitrate, and organic
nitrate (CMAQ). Reduced
nitrogen species and sources:
wet- ammonium (NADP); dry-
ammonia and ammonium (CMAQ).
16 24 32 40
\ftfertershed boindaivlayers and mapping
re provided byUSGS. The siteURLis;
http: )7md .v\9ter.usgs.gQv7gis/chesbay/
span ov£'••••••. •. . _"' • ' •:•';j':-ection1
Figure 3.1-5c. Atmospheric deposition yields of total nitrogen over the Neuse River and
Neuse River Estuary watershed.
Table 3.1-2. Model Parameters for 2002 Current Condition SPARROW Application for the
Neuse River Watershed
Source Parameters
Point Sources
Atmospheric Deposition
Fertilizer
Manure
Urban Area
Parameter
Estimate
0.84
0.082
0.13
0.017
1.6
Standard
Error
0.121
0.045
0.038
0.019
1.7
t-statistic
6.95
1.80
3.38
0.91
0.96
p-value
<0.001
0.081
0.002
0.368
0.345
VIF
1.4
13.3
7.1
4.6
2.5
Decay and Loss Parameters
Soil Permeability
Reach Decay Group 1
Reach Decay Group 2
Reservoir Decay
-0.52
0.17
0.029
3.4
0.217
0.266
0.058
9.6
-2.41
0.64
0.49
0.35
0.022
0.528
0.627
0.728
4.8
2.3
2.3
1.3
Note: VIF = Variance Inflation Factor.
Final Risk and Exposure Assessment
Appendix 6-74
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 3.1-3. Model Evaluation Statistics for 2002 Current Condition SPARROW Application
for the Neuse River Watershed
Evaluation Criteria
Number of Observations
Degrees of Freedom — model error
Degrees of Freedom — coefficients
Sum of the Squared Errors
Mean Square Error
Root Mean Square Error
R Squared
Adjusted R Squared
Eigen Value Spread
Probability Plot Correlation Coefficient
Value
40
31
9
1.79
0.058
0.24
0.99
0.98
84.2
0.984
As with the Potomac River watershed results, the Neuse SPARROW application modeled
watershed loads to the upper edges of the estuary. Both the incremental and delivered yields are
presented in Figure 3.1-6. The TN load estimated to enter the estuary from the Neuse River is
4,380,000 kg N/yr, equating to a TNS of 1.11
mg/L. The load from atmospheric deposition
was estimated to be 1,150,000 kg N/yr, or 26%
SPARROW modeling for 2002 predicts that
atmospheric deposition was 26% of the total
nitrogen loading to the Neuse River's estuary,
producing a TNS of 1.1 mg/L
of the total load (Figure 3.1-7). These estimates fall in line with instream monitoring data and
previous loadings from the Neuse River, estimated at 9.61 million pounds or 4,359,000 kg N/yr
(Spruill et al., 2004).
Final Risk and Exposure Assessment
Appendix 6-75
September 2009
-------
Aquatic Nutrient Enrichment Case Study
2002 Base Case Results for Neuse River Watershed
Map #1: Local Nitrogen Generation
(Incremental Yield)
Map #2: Local Nitrogen Load at Receiving Estuary
(Delivered Yield)
\
Legend
Nitrogen Yield
by Watershed Unit
(all units are in kg/ha/yr)
Msptl
Incremental Yield
gz-4
I I 4-6
|8-10
Map #2
Delivered Yield
\2-4
[4-6
8-10
0 10 20 40 60
Watershed boundary layers
and mapping data were
provided byUSGS. The
site URL is:
http://md .water.usgs.gov/
gis/chesbay/sparrowS/
doc/retv3.htm#section1
Figure 3.1-6. Total nitrogen yields from all sources in the Neuse River watershed as
predicted by a SPARROW modeling application for the Neuse, Tar-Pamlico, and Cape
Fear rivers' watersheds with 2002 data inputs.
Final Risk and Exposure Assessment
Appendix 6-76
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Relative Contributions of Nitrogen Sources to Neuse Estuary
7%
D Point Sources
• Atmospheric Deposition
D Fertilizer
D Manure
• Urban Areas
48%
26%
Total load to estuary:
4,382,000 kg N/yr
Figure 3.1-7. Source contributions to Neuse River Estuary total nitrogen load.
3.1.2.2 ASSETS El Assessment
Previous work was completed using the ASSETS El assessment on the Neuse River
Estuary as part of the NEEA Update (Bricker et al., 2007a). The exact source of this load
estimate and the exact timeframe of the data used to calculate the ASSETS El score are still
unknown at this time, although the data should fall within the period of 2000 to 2002 (S. Bricker,
personal communication, 2008). That analysis revealed a Highly/Moderately Influenced or High
score for influencing factors (i.e., OHI index) where the nitrogen load was ranked as Moderate to
High and the ASSETS El score for the estuary was a Bad overall.
To develop an updated ASSETS El score specific to the 2002 current conditions, raw
data were compiled from several sources, including EPA's Storage and Retrieval System
(STORET), the Neuse River Estuary Modeling and Monitoring Project (MODMON), NC DWQ,
and journal articles. While there were a variety of sources of data, information on macroalgae
and SAV for the period of interest could not be obtained. Monitoring data for chlorophyll a,
dissolved oxygen, and nitrogen were obtained from MODMON records. NC DWQ has only
recently begun to measure SAV within defined areas and on defined intervals, so no data were
available for the 2002 time period. Monitoring experts within the Neuse River/Neuse River
Estuary Case Study Area could not identify any sources of macroalgae data (B. Peierls, personal
communication, 2008). HAB data were gleaned from notations in journal papers (Burkholder et
al., 2006) and in the reports beginning in 1997 tracked by NC DWQ (NC DWQ, 2008). The
Final Risk and Exposure Assessment
Appendix 6-77
September 2009
-------
Aquatic Nutrient Enrichment Case Study
available data were combined to form a 2002 OEC index score (Table 3.1-4). Because both the
chlorophyll a and HAB data were available and overwhelmingly pointed to a system with both
High primary and secondary scores, a rating of High is given to the OEC index with confidence
for 2002.
The High susceptibility ranking combined with the TN loads estimated by the
SPARROW assessment rank the OHI index as High as well. The DFO index score set during the
2007 NEEA Update remains unchanged, with a ranking of Worsen High due to nutrient
decreases from improved management practices in recent years being offset by increases in
human populations and factors related to swine production (Burkholder et al., 2006). Combining
the three index scores together results in an overall ASSETS El score of Bad for the Neuse
River/Neuse River Estuary Case Study Area for 2002.
Final Risk and Exposure Assessment September 2009
Appendix 6-78
-------
Aquatic Nutrient Enrichment Case Study
Table 3.1-4. Current Condition Overall Eutrophic Condition Index Score for the Neuse River/Neuse River Estuary Case Study Area
Year
2002
2002
2002
2002
2002
2002
2002
Parameter
CHLA
CHLA
Macroalgae
DO
DO
SAV
HAB
Zone
MX
TF
ALL
MX
TF
ALL
ALL
Value
35
9.0
NA
1.6
2.7
NA
NA
Concentration
HIGH
MEDIUM
UNKNOWN
HYPOXIA
BIO STRESS
NA
NA
Spatial
Coverage
HIGH
MODERATE
UNKNOWN
MODERATE
HIGH
NA
NA
Frequency
PERSISTENT
PERIODIC
UNKNOWN
PERIODIC
PERIODIC
NA
NA
Expression
HIGH
MODERATE
UNKNOWN
MODERATE
MODERATE
UNKNOWN
HIGH
Score
0.9945
NA
0.5
NA
1
Primary/
Secondary
Scores
0.9945
1
OEC
Score
HIGH (1)
CHLA: Chlorophyll a
DO: Dissolved Oxygen
SAV: Submerged Aquatic Vegetation
HAB: Harmful/Toxic Algal Blooms
MX: Mixing Zone
TF: Tidal Fresh Zone
ALL: All Estuary Zones
NA: Not Applicable
Final Risk and Exposure Assessment
Appendix 6-79
September 2009
-------
Aquatic Nutrient Enrichment Case Study
3.2 ALTERNATIVE EFFECTS LEVELS
Alternative effects levels were assessed for both the Potomac River/Potomac Estuary and
the Neuse River/Neuse River Estuary case study areas (separately) by applying percentage
decreases to the oxidized nitrogen loads in the estimated atmospheric deposition. Model
estimates then relied on the SPARROW models used (for the Potomac River/Potomac Estuary
Case Study Area) or developed (for the Neuse River/Neuse River Estuary Case Study Area) for
the 2002 current condition analysis to determine how the changing atmospheric inputs (i.e., TN
load evaluated with changes in oxidized nitrogen deposition, TNatm) affect the overall TN load to
the estuary of interest. These results were used to create the response curve relating TNS to
TNatm, as first described in Section 2.2.3. The second response curve described in Section 2.2.3
was defined for the alternative effects level analysis using historical data compilations of OEC
index scores and TNS, while holding the susceptibility portion of the OHI index (at its 2002
current condition level—in both cases a ranking of High) and the DFO index constant (at a
ranking of No Change [3]).
Upon creation of the two response curves, the back-calculation-coded program described
in Section 2.2.3 (referred to as BackCalculation through the remainder of this document) was
applied to the curves with the intent of defining the atmospheric loads that are needed to improve
the ASSETS El from a score of Bad (1) to Poor (2), Moderate (3), Good (4), or High (5). These
improvements represent improvements by 1, 2, 3, and 4 categories. The BackCalculation
program was run under Uncertainty Scenario B (Figure 2.2-12b) for both case studies.
The following sections describe the data used to create the two response curves and the
application of the BackCalculation program for each estuary.
3.2.1 Potomac River Watershed
Beginning with the data and model used for the current condition analysis, the
atmospheric deposition inputs derived from national coverage of CMAQ and NADP data were
altered to create various alternative effects levels by decreasing the oxidized nitrogen loads by
rates of 5%, 10%, 20%, 30%, and 40% from their original 2002 levels. A zero percent decrease
corresponds to the 2002 current condition analysis (Table 3.2-1). The remaining inputs to the
SPARROW model remained the same, and the model was rerun for each of these alternative
Final Risk and Exposure Assessment September 2009
Appendix 6-80
-------
Aquatic Nutrient Enrichment Case Study
effects level scenarios. The TN load to the estuary calculated from the model was then converted
to TNS using the annual average flow of the Potomac River. Plotting these concentrations against
the new TNatm incorporating the oxidized nitrogen decreases leads to the development of the
desired response curve and relationship (Figure 3.2-1).
Table 3.2-1. Potomac River Watershed Alternative Effects Levels
Percent
Decrease in
Oxidized
Nitrogen
0
5
10
20
30
40
Total Nitrogen
Atmospheric
Deposition Load,
TNatm (kg/yr)
40,770,000
39,450,000
38,130,000
35,480,000
32,840,000
30,200,000
Decrease in Oxidized
Nitrogen (NOX)
Atmospheric Deposition
Load (kg/yr)
0
1,320,000
2,640,000
5,290,000
7,930,000
10,580,000
Instream
Total
Nitrogen
Load (kg/yr)
36,660,000
36,420,000
36,180,000
35,700,000
35,220,000
34,740,000
Instream Total
Nitrogen
Concentration,
TNs(mg/L)
3.41
3.39
3.37
3.32
3.28
3.23
c
_o
'i3
5
+*
c
o
o
c
o
o
c _
0 j"
O) ^
O O>
t E
z """"
S
o
1-
E
n
S>
+*
(A
C
3.50 -.
3.45
3.40
3.35
3.30
3.25
3.20
3.15
3.10
3.05
3.00 -I
y = 1.69E-08x + 2.72 •
R2 =
= 1 jr
ff
jr
»
1 '
000
0 0
0 0
0 0
0 0
0 0
lO O
T~
1 1
0 0
0 0
0 0
0 0
0 0
0 0
lO O
•<- CXI
Total Nitrogen Atmospheric
I I I I
00000
00000
00000
00000
00000
00000
lO O lO O lO
cxi co co -
-------
Aquatic Nutrient Enrichment Case Study
Table 3.2-2 details the historical modeling data used to determine TN loads to the
Potomac Estuary, which are then combined with annual average flow values to calculate a final
TNS. These instream concentrations were then combined with the OEC index scores, which were
also determined from historical data (Table 3.2-3), to create the data points needed to create the
logistic response curve in the BackCalculation program. The years chosen for this analysis relate
to those years in which a modeled TN load to the estuary could be estimated. OEC data were
then gathered for those years to find the corresponding effects.
Table 3.2-2. Historical Potomac River Total Nitrogen Loads and Concentrations
Year
1985
1992
1997
2002
2005
Nitrogen Load to
Estuary (kg/yr)
32,110,000
49,750,000
39,380,000
31,160,000
23,790,000
Source
CBMPhase4.3
CBV2
SPARROW
CBV3
SPARROW
Model Run
CBMPhase4.3
Flow (m3/s)
332
340
340
196
307
Source
USGS gage
records
CBV2
SPARROW
Input Data
CBV3
SPARROW
Input Data
USGS gage
records
USGS gage
records
Concentration
(mg/L)
3.07
4.63
3.67
5.04
2.46
CBM Phase 4.3: Chesapeake Bay Model Phase 4.3 reported results
CB V2 SPARROW: Chesapeake Bay SPARROW Version 2 model results (Brakebill et al., 2001)
CB V3 SPARROW: Chesapeake Bay SPARROW Version 3 model results (Brakebill and Preston, 2004)
Model run: Current condition analysis results as determined using modeling efforts based on CB V3
SPARROW
Final Risk and Exposure Assessment
Appendix 6-82
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 3.2-3. Additional Potomac Estuary Overall Eutrophic Condition Index Scores for Alternative Effects Levels
Year
1985
1987
1992
Parameter
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
HAB
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
HAB
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
Zone
MX
TF
ALL
MX
TF
MX
TF
ALL
MX
TF
ALL
MX
TF
MX
TF
ALL
MX
TF
ALL
MX
TF
MX
TF
Value
26
32
NA
1.6
5.6
NA
NA
NA
26
20
NA
2.2
6.3
NA
NA
NA
12
14
NA
2.6
6.3
NA
NA
Concentration
HIGH
HIGH
UNKNOWN
HYPOXIA
NO PROBLEM
GAIN
GAIN
NA
HIGH
HIGH
UNKNOWN
BIO STRESS
NO PROBLEM
GAIN
LOSS
NA
MEDIUM
MEDIUM
UNKNOWN
BIO STRESS
NO PROBLEM
GAIN
LOSS
Spatial Coverage
MODERATE
HIGH
UNKNOWN
MODERATE
HIGH
NA
NA
NA
MODERATE/
HIGH
MODERATE
UNKNOWN
HIGH
HIGH
NA
NA
NA
HIGH
MODERATE
UNKNOWN
HIGH
HIGH
NA
NA
Frequency
PERIODIC
PERIODIC
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
NA
PERIODIC
PERIODIC
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
NA
PERSISTENT
PERIODIC/
PERSISTENT
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
Expression
HIGH
HIGH
UNKNOWN
MODERATE
NO PROBLEM
NO PROBLEM
NO PROBLEM
LOW/
UNKNOWN
HIGH
HIGH
UNKNOWN
MODERATE
NO PROBLEM
NO PROBLEM
VERY LOW
LOW/
UNKNOWN
HIGH
MODERATE
UNKNOWN
MODERATE
NO PROBLEM
NO PROBLEM
MODERATE
Score
1
NA
0.86
0
0.25
1
NA
0.86
0.04
0.25
0.93
NA
0.43
0.07
Primary/
Secondary
Scores
1
0.86
1
0.86
0.93
0.43
OEC Score
HIGH (1)
HIGH (1)
MODERATE
HIGH (2)
Final Risk and Exposure Assessment
Appendix 6-83
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Year
1997
2005
Parameter
HAB
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
HAB
CHLA
CHLA
Macroalgae
DO
DO
SAV
SAV
HAB
Zone
ALL
MX
TF
ALL
MX
TF
MX
TF
ALL
MX
TF
ALL
MX
TF
MX
TF
ALL
Value
NA
28
34
NA
3.5
6.6
NA
NA
NA
21
15
NA
2.2
5.7
NA
NA
NA
Concentration
NA
HIGH
HIGH
UNKNOWN
BIO STRESS
NO PROBLEM
GAIN
LOSS
NA
HIGH
MEDIUM
UNKNOWN
BIO STRESS
NO PROBLEM
GAIN
GAIN
NA
Spatial Coverage
NA
HIGH
HIGH
UNKNOWN
HIGH
HIGH
NA
NA
NA
MODERATE
HIGH
UNKNOWN
HIGH
HIGH
NA
NA
NA
Frequency
NA
PERIODIC/
PERSISTENT
PERIODIC/
PERSISTENT
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
NA
PERIODIC
PERSISTENT
UNKNOWN
PERIODIC
PERSISTENT
NA
NA
NA
Expression
LOW/
UNKNOWN
HIGH
HIGH
UNKNOWN
MODERATE
NO PROBLEM
NO PROBLEM
LOW
MODERATE/
UNKNOWN
HIGH
HIGH
UNKNOWN
MODERATE
NO PROBLEM
NO PROBLEM
NO PROBLEM
HIGH
Score
0.25
1
NA
0.43
0.04
0.5
1
NA
0.43
0
1
Primary/
Secondary
Scores
1
0.5
1
1
OEC Score
MODERATE
HIGH (2)
HIGH (1)
CHLA: Chlorophyll a
DO: Dissolved Oxygen
SAV: Submerged Aquatic Vegetation
HAB: Harmful/Toxic Algal Blooms
MX: Mixing Zone
TF: Tidal Fresh Zone
ALL: All Estuary Zones
NA: Not Applicable
Final Risk and Exposure Assessment
Appendix 6-84
September 2009
-------
Aquatic Nutrient Enrichment Case Study
The 2002 Potomac Estuary TNatm (evaluated in terms of decrease in oxidized nitrogen)
loading is estimated to be 4.08 x 107 kg/yr (Table 3.2-1). The response curve relationship
between atmospheric deposition and TNS (TNS [mg/L] = 2.72 + 1.69 x 10"8 xTNatm [kg/yr]) can
be found in Figure 3.2-1. Outside data specified for the model include the following:
• Mean salinity in estuary =11 (relative units)
• Mean salinity offshore = 33 (relative units)
• Mean offshore TN concentration = 0.028 mg/L.
There were five sets of observed OEC and TNS data points used to fit the OEC(TNS)
logistic model (not including ecological endpoints) from the data presented in Table 3.2-2 and
Table 3.2-3
For the purpose of estimating the OEC(TNS) function at each iteration, the range on the
OEC minimum value was specified as 0 to 0 (i.e., fixed at 0). In addition, the OEC maximum
value was fixed at 5. Notwithstanding the capability to vary OEC minimum and OEC maximum
from iteration to iteration, the decision was made to not do so. Varying OEC minimum and OEC
maximum would ideally be performed to reflect natural limitations on an estuary. For instance,
in its most pristine state, it may not fully attain OEC = 5 or, in its most degraded state, it may not
fully attain OEC = 1. However, because of the difficulty in knowing these natural limitations, the
option to vary these levels was not implemented, but rather left OEC minimum = 0 and OEC
maximum = 5. The range on the TNS minimum value was specified from 0.028 mg/L (i.e., the
offshore value) to 1.0 mg/L (i.e., a value determined by best professional judgment at this time
due to lack of supporting data). The TNS maximum value range was specified from 4.63 mg/L
(maximum observed above) to 1.5x4.63 = 6.95 mg/L.
For the purpose of illustrating the overall back calculation uncertainty analysis
methodology, each of the four ASSETS El scores constituting state improvements (i.e., Poor-2,
Moderate-It, Good-4, High-5) was treated as a "target" ASSETS El score, and 5002 iterations
were run under each target El scenario. As previously discussed, at each iteration, the
BackCalculation program first fits the four-parameter logistic function to the observed data,
including the randomly sampled ecological endpoints, using a weighted least squares criterion
2 A convergence test was performed to see how stable the various quantiles of the resulting NOx*i distribution were
to 500 iterations. There was still some variability, suggesting that future analyses should use more than 500
iterations (particularly if SPARROW model uncertainty is included).
Final Risk and Exposure Assessment September 2009
Appendix 6-85
-------
Aquatic Nutrient Enrichment Case Study
(Figure 3.2-2). (In this application, only the TN ecological endpoints were varied.) After the
nonlinear, least squares optimization is performed (see Figure 2.2-12b for an example), the
program then iteratively finds that TNS concentration, TNS*;, that satisfies the target ASSETS El
score. Once TNS ; is found, the SPARROW response model is then evaluated to find the TN
atmospheric deposition load, TNatm i, which results in TNS ;. That TNatm i is saved, and the
process is repeated 500 times. The percentage decrease in NOX deposition load to reach TNatm i is
calculated as 100 x (TNatm current load - TNatm ;)/ TNatm current load, and it is also saved. The
result is a distribution of 500 TNatm*i and percentage decrease values, each of which is equally
likely to result in the target El (because the OEC(TNS) function is assumed to be uniformly
distributed within its uncertain envelope). These 500 values for each result were then rank-
ordered (i.e., ascending order) to find statistics of interest (e.g., the mean, median, 5th percentile,
and 95th percentile).
It should be carefully noted that the TNatm i value is allowed to be determined at each
iteration without regard to what the current TNatm is. Thus, if a TNatm i value is greater than the
current TNatm (4.08 x 108 kg/yr), then the implication is that more TNatm would be required (i.e.,
added to the system) to attain the target ASSETS El score. Obviously, in this situation, the
estuary is currently at a better/higher ASSETS El score than the target ASSETS El score, and the
target score represents a more polluted scenario. The TNatm*i value is not prevented from
becoming negative. As can be seen from the general form of the linear, SPARROW response
model, TNS = a + b x TNatm, a TNS*; value less than parameter "a" would require more than a
decrease in TNatm to zero. It would actually require a "negative" TNatm load. Clearly, the lower
bound on TNatm load is zero, so the negative TNatm i value would represent additional nitrogen
(beyond decreasing the TNatm [i.e., total atmospheric nitrogen] deposition load, including NOX to
zero) in the system that must be removed to achieve the target ASSETS El score. (The actual
value of the negative load would be valid only if the additional nitrogen removed had the same
characteristics [e.g., spatial distribution, speciation, sources/sinks] as the TNatm. This is unlikely;
nonetheless, it is of interest to report the values with this caveat.)
Final Risk and Exposure Assessment September 2009
Appendix 6-86
-------
Aquatic Nutrient Enrichment Case Study
0
o
o
(/>
o
HI
o
0.0 1.0 2.0 3.0
TNS (mg/L)
4.0
5.0
6.0
Figure 3.2-2. Fitted Overall Eutrophic Condition curve for target ASSETS EI=2,
median TNatm*i (i = run 280)
The summary statistics of the 500 iterations for each target ASSETS El scenario for the
Potomac Estuary are presented in Table 3.2-4.
Table 3.2-4. Summary Statistics for Target ASSETS El Scenarios for the Potomac Estuary
Statistic
TNatmsi(kgN/yr)
% TNatm*i Decrease
ASSETS El = 2 (Poor)
Mean
Median
5th Percentile
95th Percentile
-1.78 x IO6
-1.46 x IO6
-3.67 x IO6
9.02 x IO6
104
104
109
78
ASSETS El = 3 (Moderate)
No feasible solutions found
ASSETS EI=4 (Good) and ASSETS El
All TNa1m*i = -1.61 x io8, i.e., TNS>
= 5 (High)
Omg/L
Target ASSETS El = 2 is the most interesting scenario and illustrates the power of the
uncertainty analysis. The mean and median TNatm ; values are negative, meaning again that not
only must all TNatm (including all NOX) be removed, but additional nitrogen as well. However,
Final Risk and Exposure Assessment
Appendix 6-87
September 2009
-------
Aquatic Nutrient Enrichment Case Study
there is a slim chance that ASSETS El = 2 can be attained only from TNatm decrease, as indicated
by the positive 95th percentile TNatm*i value of 9.02 x 106 kg N/yr (representing a 78%
decrease).
Target ASSETS El = 3 is a unique case because all solutions were infeasible. With a
TNS ; value of 0 mg/L, the other (i.e., fixed) components of the ASSETS El scoring methodology
(i.e., DFO index and Susceptibility Score) preclude satisfying any of the 95 combinations of
DFO, OEC, and OKI indices that comprise the ASSETS EI=3 combinations in the ASSETS El
lookup table. (At TNS*; = 0, an ASSETS El = 4 can be achieved, but not ASSETS El = 3,
according to the 95 score combinations defined by Bricker et al., (2003) for ASSETS El scores.)
Bricker et al. (2003) acknowledge that "not all combinations" were included in the lookup table
within their paper, so this scenario falls into that gap. Each of the intermediate scores (i.e., OEC,
OHI, and DFO index scores) can take on integer values of 1 to 5. Thus, there are 5x5x5=125
different possible combinations, yet the lookup tables presented by Bricker et al. (2003) include
only 95 combinations, so there are 30 "missing" combinations. (These are evenly distributed
among the 5 possible DFO index scores. Each DFO index score has associated with it 19
combinations of OEC and OHI index scores.) It is possible that these combinations were not seen
as likely combinations in nature by the experts that defined the scoring matrix (e.g., a TNS*; = 0
means there is no nitrogen coming in through the surface water and, hence, it is not feasible that
a system would score below Good in the ASSETS El score). Future assessments will explore the
"missing" combinations further with the ASSETS El experts.
Target ASSETS El = 4 and 5 had identical results. All 500 iterations returned a TNS*;= 0,
and a corresponding TNatm*i negative load equal to TNatm*i= (0 - 2.72)71.69 x 10"8= -1.61 x ICT8
kg/yr. Clearly, target ASSETS Els equaling 4 and 5 are very much unattainable when decreasing
the TNatm is the only policy option. To reach the target ASSETS El, all atmospheric nitrogen
(i.e., TNatm) must be removed plus an additional amount (represented by the negative resultant
load corresponding to TNS*;= 0) that is approximately equal to one order of magnitude greater
than the original atmospheric deposition load. These amounts could be compared to the other
nitrogen sources in the watershed (e.g., fertilizer and manure application or point sources) that
were used as inputs to the SPARROW model to determine the relative nature of the required
removal with other sources in the watershed. However, consideration must be given that this load
is a reflection of the characteristics of the source in the SPARROW model (e.g., spatial
Final Risk and Exposure Assessment September 2009
Appendix 6-88
-------
Aquatic Nutrient Enrichment Case Study
distribution, magnitude of loads, sources/sinks), and a decrease required in atmospheric load is
not equal to a decrease in another source. Relative proportions can be examined by comparing
the source characteristics and model parameters.
The SPARROW response curve can also be used to examine the role of atmospheric
nitrogen deposition in achieving specified decreases in TN estuarine load. For example, the
SPARROW modeling results predict that the 41 x 106 kg N/yr deposited (i.e., atmospheric
deposition input) over the Potomac River watershed in 2002 results in a loading of 7,380,000 kg
N/yr, or 20% of the annual TN load, to the Potomac Estuary. If a 30% decrease in annual TN
load to the estuary (i.e., a decrease of 11 x 106 kg N/yr) were desired, a decrease of 61 x io6 kg
N/yr in nitrogen inputs to the watershed would be required according to the SPARROW response
curve based on atmospheric deposition. This represents a 100% decrease in the atmospheric
deposition inputs (i.e., 41 x io6 kg N/yr) plus an additional 20 x IO6 kg N/yr removal of nitrogen
from other sources in the Potomac River watershed (i.e., point and nonpoint sources). Note that
this value of 20 x IO6 kg N/yr is an approximate value when applied to the other sources because
they differ in characteristics (e.g. spatial distribution and magnitude) from atmospheric
deposition which was used to estimate the loading.
3.2.2 Neuse River Watershed
The same methods for creating alternative effects levels were applied to the data from the
Neuse River/Neuse River Estuary Case Study Area as to data from the Potomac River/Potomac
Estuary Case Study Area. The oxidized nitrogen atmospheric deposition loads were decreased by
rates of 5%, 10%, 20%, 30%, and 40% from their original 2002 levels. A zero percent decrease
corresponds to the 2002 current condition analysis (Table 3.2-5). With the remaining inputs to
the SPARROW model kept the same, the SAS-developed model was rerun for each of these
alternative effects level scenarios. The TN load to the estuary calculated from the model was
then converted to TNS using the annual average flow of the Neuse River. Plotting these
concentrations against the new TNatm and incorporating the oxidized nitrogen decreases leads to
the development of the desired response curve and relationship (Figure 3.2-3). Note that the
instream concentration range presented in this figure is discussed at the end of this section.
Final Risk and Exposure Assessment September 2009
Appendix 6-89
-------
Aquatic Nutrient Enrichment Case Study
Table 3.2-5. Neuse River/Neuse River Estuary Case Study Area Alternative Effects Levels
Percent
Decrease
in
Oxidized
Nitrogen
0
5
10
20
30
40
Total Nitrogen
Atmospheric
Deposition Load,
TNatm (kg/yr)
18,340,000
17,920,000
17,510,000
16,680,000
15,850,000
15,020,000
Decrease in Oxidized
Nitrogen Atmospheric
Deposition Load
(kg/yr)
0
410,000
830,000
1,660,000
2,490,000
3,320,000
Instream Total
Nitrogen Load
(kg/yr)
4,382,000
4,378,000
4,374,000
4,366,000
4,358,000
4,351,000
Instream Total
Nitrogen
Concentration,
TNS (mg/L)
1.112
1.111
1.110
1.108
1.106
1.104
-2 1.115-1
Q)
O
o 1.110
Instream Total Nitrogen
(mg/L)
o o
O Oi
y = 2.0E-09x + 1.07
R2 = 1
/
/
/
000
000
000
o" o" o"
000
000
CN" •*" CD"
._. 0 0 0
S o o o
§ 0 0 0
.-.-ooo
S o o o
§ 0_ 0_ 0_
oo" ° £ J
Total Nitrogen Atmospheric Deposition
o
o
o
o"
o
o
CD"
Load
o o
o o
o o
o" o"
o o
o o
oo" o"
t- CN
(kg/yr)
Figure 3.2-3. Response curve relating instream total nitrogen concentration to total
nitrogen atmospheric deposition load for the Neuse River/Neuse River Estuary Case
Study Area.
Table 3.2-6 details the historical monitoring data used to determine TNS at the
downstream end of the Neuse River where the SPARROW model was used to determine current
condition and alternative effects levels nitrogen loads. The monitoring data were derived from
data downloaded from EPA's STORET Web site for monitoring location J8290000 from NC
DWQ. These instream concentrations were then combined with the OEC index scores, which
Final Risk and Exposure Assessment
Appendix 6-90
September 2009
-------
Aquatic Nutrient Enrichment Case Study
were also determined from historical data (Table 3.2-7), to create the data points needed to
create the logistic response curve in the BackCalculation program. Data from as many years as
possible were gathered for both the TNS and OEC index scores. However, because of the limited
amount of complete data from the various sources identified under the current condition analysis,
only three corresponding years of data were found. These years are highlighted in Table 3.2-6.
Table 3.2-6. Annual Average Instream Total Nitrogen Concentrations in the Neuse River
Year
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
Annual Average TNS (mg/L)
1.08
0.93
0.99
1.10
0.96
0.91
1.02
1.09
1.12
1.07
0.97
NC DWQ Station J8290000; Results from EPA's
STORET Summation of Total Kjeldahl Nitrogen and
Nitrate/Nitrite
Final Risk and Exposure Assessment
Appendix 6-91
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Table 3.2-7. Additional Neuse River Estuary Overall Eutrophic Condition Index Scores for Alternative Effects Levels
Year
1992
1993
2003
2007
Parameter
CHLA
Macroalgae
DO
SAV
HAB
CHLA
Macroalgae
DO
SAV
HAB
CHLA
CHLA
Macroalgae
DO
DO
SAV
HAB
CHLA
CHLA
Macroalgae
DO
DO
Zone
MX
ALL
MX
ALL
ALL
MX
ALL
MX
ALL
ALL
MX
TF
ALL
MX
TF
ALL
ALL
MX
TF
ALL
MX
TF
Value
28
NA
6.1
NA
NA
30
NA
3.4
NA
NA
46
7.2
NA
2.0
4.5
NA
NA
52
27
NA
2.6
3.1
Concentration
HIGH
UNKNOWN
NO PROBLEM
NA
NA
HIGH
UNKNOWN
BIO STRESS
NA
NA
HIGH
MEDIUM
UNKNOWN
HYPOXIA
BIO STRESS
NA
NA
HIGH
HIGH
UNKNOWN
BIO STRESS
BIO STRESS
Spatial
Coverage
VERY LOW
UNKNOWN
HIGH
NA
NA
HIGH
UNKNOWN
HIGH
NA
NA
HIGH
HIGH
UNKNOWN
HIGH
HIGH
NA
NA
HIGH
MODERATE
UNKNOWN
HIGH
HIGH
Frequency
PERIODIC
UNKNOWN
PERSISTENT
NA
NA
PERIODIC
UNKNOWN
PERIODIC
NA
NA
PERSISTENT
PERIODIC
UNKNOWN
PERIODIC
PERIODIC
NA
NA
PERSISTENT
PERIODIC
UNKNOWN
PERIODIC
PERIODIC
Expression
MODERATE
UNKNOWN
NO PROBLEM
UNKNOWN
HIGH
HIGH
UNKNOWN
MODERATE
UNKNOWN
HIGH
HIGH
HIGH
UNKNOWN
HIGH
MODERATE
UNKNOWN
MODERATE
HIGH
HIGH
UNKNOWN
MODERATE
MODERATE
Score
0.5
NA
0.25
NA
1
0.99
0.5
NA
1
1
NA
0.99
NA
0.5
1
NA
0.5
Primary/
Secondary
Scores
0.5
1
0.99
1
1
0.99
1
0.5
OEC Score
HIGH(l)
HIGH (1)
HIGH (1)
MODERATE
HIGH (2)
Final Risk and Exposure Assessment
Appendix 6-92
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Year
Parameter
SAV
HAB
Zone
ALL
ALL
Value
NA
NA
Concentration
NA
NA
Spatial
Coverage
NA
NA
Frequency
NA
NA
Expression
UNKNOWN
LOW/
MODERATE
Score
NA
0.5
Primary/
Secondary
Scores
OEC Score
CHLA: Chlorophyll a
DO: Dissolved Oxygen
SAV: Submerged Aquatic Vegetation
HAB: Harmful/Toxic Algal Blooms
MX: Mixing Zone
TF: Tidal Fresh Zone
ALL: All Estuary Zones
NA: Not Applicable
Final Risk and Exposure Assessment
Appendix 6-93
September 2009
-------
Aquatic Nutrient Enrichment Case Study
The current estuary TNatm (evaluated in terms of decrease in oxidized nitrogen) loading is
estimated to be 1.83 x io7 kg/yr (Table 3.2-5). The response curve relationship between TNatm
and TNS (TNS [mg/L] = 1.07 + 2.0 x 10'8 xTNatm [kg/yr]) can be found in Figure 3.2-3. Outside
data specified for the model include the following:
• Mean salinity in estuary = 13 (relative units)
• Mean salinity offshore = 35 (relative units)
• Mean offshore TN concentration = 0.014 mg/L.
There were three sets of observed OEC and TNS data points used to fit the OEC(TNS)
logistic model (not including ecological endpoints) from the data presented in Table 3.2-6 and
Table 3.2-7
For purposes of estimating the OEC(TNS) function at each iteration, the range on the
OEC minimum value was specified as 0 to 0 (i.e., fixed at 0). In addition, the OEC maximum
value was fixed at 5. Notwithstanding the capability to vary OEC minimum and OEC maximum
from iteration to iteration, the decision was made to not do so. The range on the TNS minimum
value was specified from 0.014 mg/L (i.e., the offshore value) to 0.1 mg/L (i.e., a value
determined by best professional judgment at this time due to lack of supporting data). The TNS
maximum value range was specified from 2.57 mg/L (i.e., maximum observed value from
STORET) to 1.5x2.57 = 3.86 mg/L.
As for the Potomac River/Potomac Estuary Case Study Area, each of the four ASSETS
El scores representing state improvements (i.e., Poor-2, Modemte-3, Good-4, High-5) was
treated as a "target" ASSETS El score, and 500 iterations were run under each target ASSETS El
scenario. Figure 3.2-4 shows one of the curve fits of the logistic function to the observed OEC
and TNS data, including the randomly sampled ecological endpoints for TNS.
The summary statistics of the 500 iterations for each target ASSETS El scenario are
presented in Table 3.2-8.
For target ASSETS El = 2, all decreases are positive, but exceed 100%, meaning that not
only must all TNatm be removed to meet ASSETS El = 2, but considerably more nitrogen from
other sources must be removed as well. Given these results, the Neuse River Estuary is clearly
currently somewhere between these two ASSETS El scores as was the Potomac Estuary. There
is some evidence that it is slightly more eutrophic than the Potomac Estuary because there was at
Final Risk and Exposure Assessment September 2009
Appendix 6-94
-------
Aquatic Nutrient Enrichment Case Study
least a slim chance for the Potomac Estuary (at the 95th percentile) that a decrease in TNatm (of
less than 100%) would achieve ASSETS El = 2.
Calibration Points
Fit Logistic Curve
0.0
0.5
1.0
1.5 2.0
TNS (mg/L)
2.5
3.0
3.5
Figure 3.2-4. Fitted Overall Eutrophic Condition curve for target ASSETS EI=2, median
TNatm*i (i = run 287).
Table 3.2-8. Summary Statistics for Target ASSETS El Scenarios for the
Neuse River/Neuse River Estuary Case Study Area
Statistic
TN.tm'iOtgN/yr)
% TNatm*i Decrease
ASSETS El = 2 (Poor)
Mean
Median
5th Percentile
95th Percentile
-1.43 x 108
-1.43 x 10'
-1.47 x 10'
-1.01 x 108
880
880
901
653
ASSETS El = 3 (Moderate)
No feasible solutions found
ASSETS EI=4 (Good) and ASSETS El = 5 (High)
All TN^ = -5.35 x 108, i.e. TN,*, = 0 mg/L
Target ASSETS El = 3 is again a unique case because all solutions were infeasible. With
a TNS ; value of 0 mg/L, the other (i.e., fixed) components of the ASSETS El scoring
methodology (i.e., DFO index and Susceptibility Scores) preclude satisfying any of the 95
combinations of DFO, OEC, and OHI index scores that comprise the ASSETS EI=3
Final Risk and Exposure Assessment
Appendix 6-95
September 2009
-------
Aquatic Nutrient Enrichment Case Study
combinations in the ASSETS El lookup table. This result again depends on the experts who set
up the ASSETS El scoring table, defining only 95 out of 125 possible combinations. The
likelihood that any of the 30 "missing" combinations are feasible in nature and could result in
reaching Target ASSETS El = 3 for this scenario will be examined in future analyses.
Target ASSETS El = 4 and 5 had identical results. All 500 iterations returned a TNS*; = 0
mg/L, and a corresponding TNatm*i negative load equal to TNatm*i = (0 - 1.07)72.0 x 10"9= -5.35 x
108 kg/yr. Clearly, target ASSETS El equal 4 and 5 are very unattainable when decreasing the
TNatm (including all NOX) is the only policy option. Again, the decrease required includes all of
the TNatm source plus an additional amount that is one order of magnitude greater than the
original atmospheric deposition load of 108 kg/yr). These amounts could be compared to the
other nitrogen sources in the watershed that were used as inputs to the SPARROW model, giving
consideration to the characteristics of each of these sources.
As with the Potomac River and Potomac Estuary watershed analysis, the SPARROW
response curve can be used to examine the role of nitrogen deposition in achieving desired
decreases in load to the Neuse River Estuary. In the Neuse River watershed, modeling results
indicate that 7 x 106 kg N/yr was deposited in 2002. SPARROW modeling predicts that this
deposition input results in a loading of 1.2 x 106 kg N/yr (i.e., 20% of the annual TN load) to the
Neuse River Estuary. Unlike the Potomac River and Potomac Estuary analysis, little change is
seen in the TN loading to the Neuse River Estuary with large decreases in the nitrogen
deposition. If all atmospheric nitrogen deposition inputs were eliminated (i.e., 100% decrease),
the total annual nitrogen load to the Neuse River Estuary would only decrease by 4%. There are
two apparent reasons for this lack of change in loadings. The first reason is a characteristic of the
Neuse River watershed. The second reason is an inherent characteristic of the SPARROW model
formulation.
First, the TN loadings to the Neuse River Estuary are highly dependent on the sources
other than atmospheric deposition within the SPARROW model. There are differences in
characteristics among the sources within the watershed, where fertilizer, in particular, has a
strong signature (i.e., indicating the large influence of agriculture within the watershed). This
result demonstrates that the SPARROW response curves of TN load to other sources would be
quite different, and the current response curve cannot be used to predict the relative magnitudes
of loads needed to produce decreases greater than this 4%. Figure 3.2-5 illustrates the theoretical
Final Risk and Exposure Assessment September 2009
Appendix 6-96
-------
Aquatic Nutrient Enrichment Case Study
response curves that may result when the SPARROW modeled loads are plotted against the other
TN source inputs. The green curve, or least influential source, displays the behavior of the
atmospheric deposition for the Neuse River Estuary. The red curve, or highly influential source,
likely corresponds to how agricultural sources within the watershed behave. These response
curves will depend on the source magnitudes, spatial distributions, and other characteristics.
I?
C D)
O ^
's ^
= 3
CD -P
o w
C LU
o o
o *-
z "°
j£ CD
I— O
E-1
CD Z
CD I-
Current conditions concentration or loading
Least influential source where a large
change in input loading produces a
small change in load to the estuary
Moderately influential source where a
large change in input loading produces
a large change in load to the estuary
Highly influential source where a small
change in input loading produces a
large change in load to the estuary
Source Input Load (kg N/yr)
Figure 3.2-5. Theoretical SPARROW response curves demonstrating relative influence
of sources on nitrogen loads to an estuary.
Second, the decay rates used by SPARROW play a part in the differing percentage
decreases seen. The changes in the loadings reaching the estuary in the alternate effects analyses
also occur because of the instream and reservoir decay terms. Within SPARROW, the decay
term is applied to the total runoff load from each catchment (i.e., the sum of the source loads) in
the upstream to downstream cumulative load calculations. When the atmospheric deposition
loads are decreased, the TOTAL load from a catchment is decreased, even though the other
sources remain constant. The same rate of decay is applied no matter the magnitude of the source
loads is because the decay rate is based on stream travel characteristics or reservoir
characteristics. So when the total load decreases, the amount of decay decreases, leading to a
proportionally higher delivered load. Thus, the same percentage decrease in atmospheric
deposition and fertilizer loads would not produce the same decrease in loads to the estuary.
Final Risk and Exposure Assessment
Appendix 6-97
September 2009
-------
Aquatic Nutrient Enrichment Case Study
4.0 IMPLICATIONS FOR OTHER SYSTEMS
Selection of the analysis method for aquatic nutrient enrichment considered applications
beyond a small number of case studies. The chosen method, consisting of a combination of
SPARROW modeling for nitrogen loads and assessment of estuary conditions under the NOAA
ASSETS El, provides a highly scalable and widely applicable analysis method. Both components
have been applied on a national scale—the national nutrient assessment using SPARROW
(Smith and Alexander, 2000) and the NEEA using the ASSETS El (Bricker et al., 1999, 2007a).
Additionally, both have been used on a smaller scale. These previous analyses supply a large
body of work—data, methods, and supporting experts—to draw from when conducting
additional analyses or updating past applications.
Requirements for applying this method to other systems include mandatory data inputs,
the ability to formulate a SPARROW application on a reliable stream network, and an estuary
that is likely to be subject to eutrophication. Data requirements and model formulations have
been described and detailed throughout this report.
The method is not currently designed to assess eutrophi cation impacts on inland waters.
SPARROW modeling can still be applied to determine nitrogen loadings to an inland waterway,
but the ASSETS El assessment would not apply, and as such, the indicators and overall
likelihood of eutrophication could not be assessed. For these inland waters, an alternate
methodology would be necessary to examine the effects of changing nitrogen loads within the
waterbody. A variety of methods could possibly be applied, including empirical relationships or
dynamic modeling. It is beyond the scope of this case study to further assess these inland waters
beyond the sensitive areas analysis in Section 1.2.1. An additional case study in this project
examines the effects of aquatic acidification on inland waters using dynamic modeling (See
Appendix 4).
The scalability of the methods and approaches taken in these case studies will rely on the
ability to group estuaries across the country into patterns of similar behavior either in terms of
nitrogen sources or eutrophication effects. In 2003 and 2004, NOAA and the Kansas Geological
Survey conducted a series of workshops to develop a type classification system for the 138
estuarine systems assessed in the original NEEA (Bricker et al., 1999). Participants considered
70 classification variables for grouping the estuarine systems. These variables included 51
physical characteristics (e.g., estuary depth and volume, tidal range, salinity, nitrogen and
Final Risk and Exposure Assessment September 2009
Appendix 6-98
-------
Aquatic Nutrient Enrichment Case Study
phosphorus concentrations, estimates of flushing time, evaporation), 10 perturbation
characteristics (e.g., population in watershed, estimates of nutrient loading), and nine response
characteristics (e.g., SAV loss, presence of nuisance and toxic blooms). Ultimately, the
workgroup selected five variables (i.e., depth, openness of estuary mouth, tidal range, mean
annual air temperature, and the log of freshwater inflow/estuarine area) deemed to be the most
critical physical and hydrological characteristics influencing nutrient processing and the
expression of eutrophic symptoms in a waterbody. Based on these five variables, the 138
estuarine systems were classified into 10 groups (Table 4-1; Figure 4-1). The two estuary
systems included in this case study, the Potomac and Neuse River estuary systems, were in
groups one and nine, respectively (Bricker et al., 2007a).
Table 4-1. Typology Group Categorizations
Group
Group 0
Group 1
Group 2
Group 3
Group 4
Group 5
Group 6
Group 7
Group 8
Group 9
Number of Systems
13
35
5
8
18
3
2
16
17
21
Overriding Characteristics
Low freshwater inflow:estuarine area ratio; low
depth; low estuary mouth openness
Medium depth, medium openness, high annual
air temperature
High depth, low annual air temperature
High estuary mouth openness; high depth
Low estuary mouth openness; high freshwater
inflow:estuarine area ratio; low annual air
temperature
High estuary mouth openness; high depth
High depth; high estuary mouth openness
High tidal range; medium estuary mouth
openness; low annual air temperature
High freshwater inflow:estuarine area ratio; low
depth
Low depth, medium estuary mouth openness;
high annual air temperature
Final Risk and Exposure Assessment
Appendix 6-99
September 2009
-------
Aquatic Nutrient Enrichment Case Study
Group
06
U7
• 9
Figure 4-1. Preliminary classifications of estuary typology across the nation (modified
from Bricker et al., 2007a).
Given that the response curve of the OEC index score to TNS is expected to change
shapes with different values of susceptibility, the typology classes thus defined in Table 4-1
provide an opportunity to assess the validity of this expectation. The first step in assessing this
statement would be to examine the nutrient loadings in other estuaries that fall within groups 1 or
9, the groups corresponding to the two case study areas. Once the shape and behavior of the
response curve for the estuary grouping is confirmed, work can begin to scale the results between
estuaries of that group. The ASSETS El score for an estuary may also be considered within this
analysis. For the 48 systems for which an ASSETS El score was developed in the 2007 NEEA
Update, only one system was rated as High (i.e., Connecticut River), whereas five were rated as
Good (i.e., Biscayne Bay, Pensacola Bay, Blue Hill Bay, Sabine Lake, Boston Harbor). Eighteen
systems were rated as Moderate, and 24 systems were rated as Poor or Bad. Those estuaries that
fall within groups 1 or 9 and are rated as Poor or Bad would be the most appropriate candidates
to start the scaling analysis.
Scaling of results will also need to account for the response of the watershed to
atmospheric nitrogen deposition inputs. If SPARROW is used, either through the in-development
Web-enabled national SPARROW application or through regional or site-specific applications,
Final Risk and Exposure Assessment
Appendix 6-100
September 2009
-------
Aquatic Nutrient Enrichment Case Study
the shape of the response curve will be determined by the model and its parameters. If a different
approach is taken to develop TN loadings, then the systems will need to be grouped according to
the shape and behavior of the response curve. Additional consideration should be given to the
magnitude of the percentage contributions of the atmospheric deposition to the TN load to the
watershed and the resulting TN load to the estuary.
5.0 UNCERTAINTY
There are several areas of uncertainty with this method of assessment for aquatic nutrient
enrichment, which are summarized below.
• Data inputs to SPARROW. The compilation of data needed for creation or application of
SPARROW relies on geographic and temporal analyses. For this study, the data used were
developed under separate studies and published by the USGS. Because the data were
independently verified before publication by the USGS, only quality checks were performed on the
data rather than full validation exercises. For any future analyses that require new compilations of
data, close attention should be paid to the source and geographic and temporal precision and
accuracy of the data because SPARROW relies on the distribution of sources across the watershed
to create model parameters on the annual average basis.
• Modeling uncertainty in SPARROW estimates. With any measured or modeled results, there is
a certain amount of uncertainty that should be quantified. Because SPARROW relies on a
nonlinear regression basis, a number of parameters can be used to assess the uncertainty within the
model and provide confidence intervals around the estimates. The Version 3 Chesapeake Bay
SPARROW application met evaluation criteria based on degrees of freedom, model error, and R-
squared values. The calibration of the Neuse River watershed SPARROW model using SAS
examined the standard deviation, t-statistics, p-values, and VIFs for each estimated parameter. The
overall model was evaluated based on minimizing model error, maximizing R-squared values, and
ensuring that the Eigen value range was below 100 while the probability plot correlation
coefficient was close to one. The model derived for the Neuse River/Neuse River Estuary Case
Study Area did produce some model parameters (i.e., manure production, urban area, and decay
terms) that did not reach desired statistical significance levels. The estimation of decay term
parameters may be rectified by adjusting the flow classes among which the parameters are split.
This should be examined in any future analyses. The manure production and urban area source
terms should also be examined as to their distribution throughout the watershed and overall
contributions to the load.
Final Risk and Exposure Assessment September 2009
Appendix 6-101
-------
Aquatic Nutrient Enrichment Case Study
* Sensitivity of SPARROW formulation due to atmospheric inputs in the Potomac
River/Potomac Estuary Case Study Area. The parameterization of a SPARROW application for
the Potomac River watershed is expected to change when recalibration is completed using the
atmospheric deposition of TN based on the combination of CMAQ and NADP data created for this
study, rather than the interpolated values of wet deposition of nitrate. As discussed in Section
3.1.1, the spatial gradient as well as the magnitude of the atmospheric deposition of different
nitrogen species varies across the watershed. While it is certain that the parameter estimated to
apply to the atmospheric deposition source will change, what is uncertain at this point is the extent
to which the other model parameters and the overall nitrogen load estimates will be affected by
using the CMAQ/NADP estimates in the model calibrated against the wet NO3 deposition values.
Sensitivity of the model parameters and nitrogen load estimate can be evaluated in future studies
where SPARROW is recalibrated against the 2002 data.
• Calibration data for SPARROW estimates. Monitoring data will be used to calibrate the
SPARROW model. By relying on data from federally recognized data systems, the aim is to use
data that has undergone quality assurance/quality control (QA/QC) procedures. Additionally,
collaboration has been completed with the researchers who have conducted the previous
SPARROW applications in each case study area to provide a rigorous check on the data used.
• Data inputs to the ASSETS El. Because of the numerous data requirements and sources required
to conduct a full ASSETS El analysis, there is a large range of uncertainty that can enter into the
calculations. For the water quality data evaluation of dissolved oxygen and chlorophyll a, the
numerical values of the 10th and 90th percentiles used in the evaluation were subjected to QA/QC
procedures as processed through regulated databases with checks. The frequency of occurrence of
these indicators and HABs events relied more on subjective judgment of temporal variations of
concentrations across the year. Best attempts were made to apply standardized evaluation methods
in order to minimize any uncertainties due to subjectivity or processing differences.
• Heuristic estimates of DFO index scores. The estimation of the DFO index score in the ASSETS
El assessment currently relies on heuristic estimates from systems experts. Future NOAA efforts
will seek to provide more scientifically structured estimates for this parameter, but at this time,
reliance must be on expert judgment on whether there will be increased or decreased pressures
because of nutrient loads, population growth, and land use change.
• Steady-state estimates/mean annual estimates. Both SPARROW and the ASSETS El methods
currently provide only longer-term estimates of the system conditions. There is the possibility of
conducting the analyses on a seasonal basis, which may be appropriate because the trends in
eutrophication indicators are likely to vary seasonally. Producing annual averages actually
Final Risk and Exposure Assessment September 2009
Appendix 6-102
-------
Aquatic Nutrient Enrichment Case Study
introduces some leverage to the uncertainty in the input data as previously discussed. Because the
ultimate values used to base the analysis on are averages, there is less reliance on the certainty of
individual measures.
• Use of a screening method. The methods used in this study are only of the screening level. The
screening level was more appropriate for a scalable, widely applicable set of case studies than for a
highly detailed modeling effort. Undoubtedly, details, such as the degree to which the soil-
groundwater system affects atmospherically deposited nitrogen, will be less quantified than
detailed processes using this method. However, for an initial approach to determining the aquatic
nutrient enrichment effects on a system, the screening method provides a response curve that can
be used in the evaluation of ecosystem services. Additionally, many of the complex concepts
linking the indicators of eutrophication to the effects of eutrophication are not highly developed or
understood at this time (Howarth and Marino, 2006). While some targeted studies may produce the
type of linked results from indicators to ecological endpoints that are the goal of this study, these
results cannot be readily expanded to multiple areas in multiple climate zones without great levels
of effort. As the base of literature and results expands, the concepts applied in this methodology
can be expanded to more deterministic, temporarily varying analyses.
• Use of a partially empirical framework. Because SPARROW is, at its core, an empirical
relationship, any model obtained using SPARROW is a function of the data used in the
calibration. Therefore, the predictions remain valid as long as there is no great change in the
conditions (in this case, the nitrogen loadings within each subbasin) underlying the model. This
aspect of the model introduces uncertainty into the alternative effects results because they are
calculated using a model calibrated under current conditions.
Uncertainties in Back Calculation Methods
• Missing ASSETS El scores per combinations of index scores. The combinations of OHI, OEC,
and DFO index scores provided by Bricker et al. (2003) leave out 30 of the possible 125
combinations that represent overall ASSETS El scores. In both case study analyses, these missing
scores have led to a conclusion of infeasible scenarios because an overall ASSETS El score could
not be determined from the resulting instream nitrogen load found during the back calculation
method. The methods used to determine the 95 combinations will be investigated, and the missing
30 scores pursued for future case study analyses.
• Better rationale for TNS minimum and maximum uncertainty range. The uncertainty about
how to best quantify the TNS ecological endpoint uncertainty is the biggest limitation of the current
analyses. This is particularly true for the TNS high end (i.e., the maximum TNS that would be
Final Risk and Exposure Assessment September 2009
Appendix 6-103
-------
Aquatic Nutrient Enrichment Case Study
expected to result in an OEC index score of 1). The assigned uncertainty ranges were based on best
professional judgment, but more research is needed. It is expected for the results of the back
calculation methodology to be very sensitive to these ecological endpoint ranges, especially on the
maximum TNS end. Because of this limitation, the results presented herein for the Potomac and
Neuse River estuaries should be interpreted as illustrative of the methodology, not strictly valid.
Methodology to incorporate uncertainty in the SPARROW model. Estimates of TNS at the
head of the estuary, predicted by SPARROW and driven by the TNatm (i.e., TN deposition
evaluated on decreases in NOX) over the watershed and other nitrogen sources, are uncertain. That
uncertainty was not considered in these two case studies; therefore, the probability distributions of
TNatm*! presented are artificially "tight" (i.e., the true distributions would exhibit more variability).
There is a need to explore the SPARROW literature more thoroughly to determine how to
incorporate the non-parametric confidence limits that have been developed for the SPARROW
model. Once such limits are incorporated, it is very unlikely that one would be able to explicitly
solve the SPARROW model, including these confidence limit terms explicitly for a TNatm*ij as a
function of the TNS ; value and the "jth" probability of confidence limit term. Some sort of implicit,
iterative method would be needed. An application of the Newton's method algorithm has already
been developed for these purposes and tested using an artificial confidence limit term. It seems to
work remarkably well (i.e., convergence within a few iterations to very tight convergence criteria),
and the researchers are optimistic that expanding the overall methodology to include SPARROW
uncertainty is very tractable.
More convergence testing to determine appropriate numbers of samples. As briefly
mentioned, some modest convergence testing was completed to determine how many samples of
the OEC(TNS) function need to be used for the statistics of interest for the resulting TNatm*!
distributions to be reasonably stable. The answer is something more than 500, which will
undoubtedly increase when SPARROW uncertainty is incorporated. More convergence testing is
needed.
Crossing of a categorical ranking system with a continuous nitrogen concentration scale.
Several assumptions and considerations had to be made in order to create and evaluate the logistic
response curve because the OEC index score is a categorical ranking of 1 through 5, whereas TNS
is a continuous variable. The functions evaluated in BackCalculation treat the OEC index score as
a continuous function. Until higher level models are developed to relate the nitrogen
concentrations in the system to eutrophication effects, these assumptions are necessary. Future
applications with additional data should be used to test and validate these assumptions and results.
Final Risk and Exposure Assessment September 2009
Appendix 6-104
-------
Aquatic Nutrient Enrichment Case Study
6.0 CONCLUSIONS
• A screening-level method has been determined to be an appropriate approach to assessing the
effects of atmospheric deposition of oxidized nitrogen on eutrophication/nutrient enrichment
because there is a lack of a generalized link development between these characteristics in the
literature.
• Both the Potomac and Neuse River estuaries have an ASSETS El score of Bad for 2002, meaning
that both systems are highly eutrophic and are not expected to improve greatly in the near future.
Atmospheric deposition over the watersheds account for approximately 24% and 26% of the
instream loads to the Potomac and Neuse River estuaries, respectively.
• The BackCalculation program designed and set up for this study succeeded in assessing the links
between TNS responding to changes in TNatm and the OEC index and ASSETS El scores. Results
of this assessment for the Potomac Estuary reveal that it is possible to improve the ASSETS El
score by only one category when the only change is in a decrease of the oxidized nitrogen
component of the atmospheric deposition (Table 3.2-4). This result showed that there was a 5%
chance (i.e., 95th Percentile of results) that decreasing the TNatm by 78% would result in the one
category improvement in the ASSETS El score. Within the Neuse River Estuary, this analysis
revealed that it would not be possible to improve the ASSETS El score by decreasing the oxidized
nitrogen in the atmospheric deposition loading to the estuary alone (Table 3.2-8; all percentage
decreases greater than 100). Additional source decreases would be necessary to produce OHI and
OEC index scores good enough to improve the ASSETS El score.
• Scaling of this methodology was a priority in development. Demonstration of the back calculation
methods was the first step to expanding the results to estuaries across the nation. Alternative
evaluation methods of eutrophication will be needed to assess nutrient enrichment in inland waters.
7.0 REFERENCES
Alexander, R.B., R.A. Smith, and G.E. Schwarz. 2000. Effect of stream channel size on the
delivery of nitrogen to the Gulf of Mexico. Nature 403:758-761.
Alexander, R.B., R.A. Smith, G.E. Schwarz, S.D. Preston, J.W. Brakebill, R. Srinivasan, and
P. A. Pacheco. 2001. Atmospheric nitrogen flux from the watersheds of major estuaries of
the United States—An application of the SPARROW watershed model. Pp. 119-170 in
Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective. Edited by R.A.
Final Risk and Exposure Assessment September 2009
Appendix 6-105
-------
Aquatic Nutrient Enrichment Case Study
Valigura, R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacey, and R.E.
Turner. American Geophysical Union, Coastal and Estuarine Studies, Washington, DC.
Alexander R.B., A.H. Elliott, U. Shankar, and G.B. McBride. 2002. Estimating the sources and
transport of nutrients in the Waikato River Basin, New Zealand. Water Resources
Research 38:1268-1290.
Atkins, W.A., andF. Anderson. 2009. Chesapeake Bay. In Water Encyclopedia. Online
information. Advameg, Inc., Flossmoor, IL. Available at
http://www.waterencyclopedia.com/Ce-Cr/Chesapeake-Bay.html (accessed February
2009).
Bergstrom, A.K., and M. Jansson. 2006. Atmospheric nitrogen deposition has caused nitrogen
enrichment and eutrophication of lakes in the northern hemisphere. Global Change
Biology 12:635-643.
Box, MJ. 1965. A new method of constrained optimization and a comparison with other
methods. The Computer Journal 5:42-52.
Boyer, E.W., C.L. Goodale, N.A. Jaworski, and R.W. Howarth. 2002. Anthropogenic nitrogen
sources and relationships to riverine nitrogen export in the northeastern U.S.A.
Biogeochemistry 5 7/58:13 7-169.
Boynton, W.R., J.H. Garber, R. Summers, and W.M. Kemp. 1995. Inputs, transformations, and
transport of nitrogen and phosphorus in Chesapeake Bay and selected tributaries.
Estuaries 75:285-314.
Brakebill, J.W., S.D. Preston, and S.K. Martucci. 2001. Digital Data Used to Relate Nutrient
Inputs to Water Quality in the Chesapeake Bay, Version 2.0. USGS Open File Report
OFR-01-251. U.S. Department of the Interior, U.S. Geological Survey, Baltimore, MD.
Available at http://md.usgs.gov/publications/ofr-01-25 I/index.htm.
Brakebill, J.W., and S.D. Preston. 2004. Digital Data Used to Relate Nutrient Inputs to Water
Quality in the Chesapeake Bay Watershed, Version 3.0. USGS Open File Report 2004-
Final Risk and Exposure Assessment September 2009
Appendix 6-106
-------
Aquatic Nutrient Enrichment Case Study
1433. U.S. Department of the Interior, U.S. Geological Survey, Baltimore, MD.
Available at http://pubs.usgs.gov/of/2004/1433/pdf/sparv3.pdf.
Bricker, S.B., C.G. Clement, D.E. Pirhalla, S.P. Orlando, and D.R.G. Farrow. 1999. National
Estuarine Eutrophication Assessment: Effects of Nutrient Enrichment in the Nation's
Estuaries. NOAA—NOS Special Projects Office. U.S. Department of Commerce,
National Oceanic and Atmospheric Administration, National Ocean Service, National
Centers for Coastal Ocean Science, Center for Coastal Monitoring and Assessment,
Silver Spring, MD.
Bricker, S.B., J.G. Ferreira, and T. Simas. 2003. An integrated methodology for assessment of
estuarine trophic status. Ecological Modelling 7(59:39-60.
Bricker, S., D. Lipton, A. Mason, M. Dionne, D. Keeley, C. Krahforst, J. Latimer, and J.
Pennock. 2006. Improving Methods and Indicators for Evaluating Coastal Water
Eutrophication: A Pilot Study in the Gulf of Maine. NOAA Technical Memorandum
NOS NCCOS 20. U.S. Department of Commerce, National Oceanic and Atmospheric
Administration, National Ocean Service, National Centers for Coastal Ocean Science,
Center for Coastal Monitoring and Assessment, Silver Spring, MD.
Bricker, S., B. Longstaff, W. Dennison, A. Jones, K. Boicourt, C. Wicks, and J. Woerner. 2007a.
Effects of Nutrient Enrichment in the Nation's Estuaries: A Decade of Change. NOAA
Coastal Ocean Program Decision Analysis Series No. 26. U.S. Department of Commerce,
National Oceanic and Atmospheric Administration, National Ocean Service, National
Centers for Coastal Ocean Science, Silver Spring, MD.
Bricker, S., J.G. Ferreira, and R. Pastres. 2007b. Eutrophication Assessment using ASSETS
Approach, Application and Results. U.S. Department of Commerce, National Oceanic
and Atmospheric Administration. Presented at the Venice Water Authority Meeting:
Exploring possibilities for collaborative work, December 3. Available at
http://www.eutro.org/presentations/VWA Seminar.pdf.
Final Risk and Exposure Assessment September 2009
Appendix 6-107
-------
Aquatic Nutrient Enrichment Case Study
Bricker, S. 2008. Personal communication from Suzanne Bricker, National Oceanic and
Atmospheric Administration, to Jennifer Schimek, RTI International, Research Triangle
Park, NC.
Bricker, S., B. Buddemeier, S. Smith, B. Maxwell. In prep. Results of Type Classification
Analyses from NOAA Workshop September, 2003 and KGS Work Session January, 2004.
White paper. U.S. Department of Commerce, National Oceanic and Atmospheric
Administration, National Ocean Service, Silver Spring, MD.
Burkholder, J.M., D.A. Dickey, C.A. Kinder, R.E. Reed, M.A. Mallin, M.R. Mclver, L.B.
Cahoon, G. Melia, C. Brownie, J. Smith, N. Deamer, J. Springer, H.B. Glasgow, and D.
Toms. 2006. Comprehensive trend analysis of nutrients and related variables in a large
eutrophic estuary: A decadal study of anthropogenic and climatic influences. Limnology
and Oceanography 57:463-487.
CBP (Chesapeake Bay Program). 2008. CBP Water Quality Database (1984-present). Online
information. Chesapeake Bay Program, Annapolis, MD. Available at
http://www.chesapeakebay.net/data_waterquality.aspx (accessed December 2008).
Eadie, B.J., B.A. McKee, M.B. Lansing, J.A. Robbins, S. Metz, and J.H. Trefry. 1994. Records
of nutrient-enhanced coastal productivity in sediments from the Louisiana continental
shelf. Estuaries 77:754-765.
Elliott, A.H., R.B. Alexander, G.E. Schwarz, U. Shankar, J.P.S. Sukias, and G.B. McBride.
2005. Estimation of nutrient sources and transport for New Zealand using the hybrid
physical-statistical model SPARROW. Journal of Hydrology 44:1-21.
Fact-Index.com. 2009. Potomac River. Online information. Fact-Index.com. Available at:
http://www.fact-index.com/p/po/potomac_river.html (accessed February 2009).
Ferreira, J.G, S.B. Bricker, and T.S. Castro. 2007. Application and sensitivity testing of a
eutrophication assessment method on coastal systems in the United States and European
Union. Journal of Environmental Management 52:433-445.
Final Risk and Exposure Assessment September 2009
Appendix 6-108
-------
Aquatic Nutrient Enrichment Case Study
Heinz Center for Science. 2007. Economics and the Environment: Indicators of Ecological
Effects of Air Quality. Analysis Subcommittee Meeting Summary. Heinz Center for
Science, Washington, DC. November.
Hoos, A.B. 2005. Evaluation of the SPARROW model for estimating transport of nitrogen and
phosphorus in streams of the Interior Low Plateau Ecoregion, Tennessee, Kentucky, and
Alabama, during 1992-2002. Pp. 2A-47 in Proceedings of the Fifteenth Tennessee Water
Resources Symposium., Burns, TN, April 13-15. Tennessee Section of the American
Water Resources Association.
Hoos, A.B., S. Terziotti, G. McMahon, K. Savvas, K.G. Tighe, andR. Alkons-Wolinsky. 2008.
Data to Support Statistical Modeling oflnstream Nutrient Load Based on Watershed
Attributes, Southeastern United States, 2002. Open File Report 2008-1163. U.S.
Department of the Interior, U.S. Geological Survey, Reston, VA.
Howarth, R.W. 2007. Atmospheric deposition and nitrogen pollution in coastal marine
ecosystems. In Acid in the Environment. Edited by G.R. Visgilio and D.M. Whitelaw.
New York: Springer.
Howarth, R.W., and R. Marino. 2006. Nitrogen as the limiting nutrient for eutrophication in
coastal marine ecosystems: Evolving views over three decades. Limnology and
Oceanography 51(1 part 2):364-376.
Howarth, R.W., G. Billen, D. Swaney, A. Townsend, N. Jarworski, K. Lajtha, J.A. Downing, R.
Elmgren, N. Caraco, T. Jordan, F. Berendse, J. Freney, V. Kueyarov, P. Murdoch, and Z.
Zhao-liang. 1996. Riverine inputs of nitrogen to the North Atlantic Ocean: Fluxes and
human influences. Biogeochemistry 35:75-139.
Kansas Department of Health and Environment. 2004. Surface Water Nutrient Reduction Plan.
Kansas Department of Health and Environment, Bureau of Water, Topeka, KS. Available
at http: //www. kdhe. state. ks. us/water/downl oad/ks_nutri ent_reducti on_pl an
_12_29_fmal.pdf (accessed March 1, 2005).
Final Risk and Exposure Assessment September 2009
Appendix 6-109
-------
Aquatic Nutrient Enrichment Case Study
McMahon, G., R.B. Alexander, and S. Qian. 2003. Support of total maximum daily load
programs using spatially reference regression models. Journal of Water Resources
Planning and Management 729(4): 315-329.
MEA (Millennium Ecosystem Assessment Board). 2005. Ecosystems and Human Well-being:
Current State and Trends, Volume 1. Edited by R. Hassan, R. Scholes, and N. Ash.
Washington: Island Press. Available at
http://www.millenniumassessment.org/documents/document.766.aspx.pdf.
Moore, R.B., C.M. Johnston, K.W. Robinson, and J.R. Deacon. 2004. Estimation of total
nitrogen and phosphorus in New England streams using spatially referenced regression
models. USGS Scientific Investigations Report 2004-5012. U.S. Department of the
Interior, U.S. Geological Survey, Pembroke, NH. Available at
http://water.usgs.gov/pubs/sir/2004/5012.
NC DENR (North Carolina Department of Environment and Natural Resources). 1993. Neuse
River Basinwide Water Quality Management Plan. North Carolina Department of
Environment and Natural Resources, Division of Environmental Management, Water
Quality Section, Raleigh, NC.
NC DENR (North Carolina Department of Environment and Natural Resources). 2002. Neuse
River Basin-wide Water Quality Plan. North Carolina Department of Environment and
Natural Resources, Division of Water Quality/Planning, Raleigh, NC. Available at
http://h2o.enr.state.nc.us/basinwide/Neuse/2002/Section A Chapter 2.pdf.
NC DWQ (Division of Water Quality). 2008. North Carolina Department of Environment and
Natural Resources. Environmental Sciences Section - Fish Kill Event Update. Online
information. North Carolina Department of Environment and Natural Resources,
Division of Water Quality, Raleigh, NC. Available at:
http://www.esb.enr.state.nc.us/Fishkill/fishkillmain.htm (accessed December 2008).
NEIWPCC (New England Interstate Water Pollution Control Commission). 2004. New England
SPARROW water quality model. Interstate Water Report 7(3):6-7. Available at
http://www.nei wpcc.org/PDF_Docs/i wr_s04.pdf
Final Risk and Exposure Assessment September 2009
Appendix 6-110
-------
Aquatic Nutrient Enrichment Case Study
NRCS (Natural Resources Conservation Service). 2001. 1997 National Resources Inventory,
Updated June 2001. Natural Resources Conservation Service, Washington, DC.
Officer, C.B., R.B., Biggs, J.L. Taft, L.E. Cronin, M.A. Tyler, and W.R. Boynton. 1984.
Chesapeake Bay anoxia: origin, development, and significance. Science 223:22-27'.
Paerl, H.W. 1995. Coastal eutrophication in relation to atmospheric nitrogen deposition: current
perspectives. Ophelia ¥7:237-259.
Paerl, H.W. 1997. Coastal eutrophi cation and harmful algal blooms: Importance of atmospheric
deposition and groundwater as "new" nitrogen and other nutrient sources. Limnology and
Oceanography ¥2:1154-1165.
Paerl, H. W. 2002. Connecting atmospheric nitrogen deposition to coastal eutrophi cation.
Environmental Science and Technology 36:323 A-326A.
Paerl, H., J. Pinckney, J. Fear, and B. Peierls. 1998. Ecosystem responses to internal and
watershed organic matter loading: Consequences for hypoxia in the eutrophying Neuse
River Estuary, NC, USA. Marine Ecology Progress Series 166:17-25.
Paerl, H.W., R.S. Fulton, P.H. Moisander, and J. Dyble. 2001. Harmful freshwater algal blooms,
with an emphasis on cyanobacteria. The Scientific World 7:76-113.
Paerl, H., J. Dyble, L. Twomey, J. Pinckney, J. Nelson, and L. Kerkhof. 2002. Characterizing
man-made and natural modifications of microbial diversity and activity in coastal
ecosystems. Antonie vanLeeuwenhoek81(l-4)A87-507.
Peierls, B. 2008. Personal communication.
Preston, S. 2008. Personal communication.
Preston, S.D., and J.W. Brakebill. 1999. Applications of Spatially Referenced Regression
Modeling for the Evaluation of Total Nitrogen Loading in the Chesapeake Bay
Watershed. USGS Water-Resources Investigations Report 99-4054. U.S. Department of
Final Risk and Exposure Assessment September 2009
Appendix 6-111
-------
Aquatic Nutrient Enrichment Case Study
the Interior, U.S. Geological Survey, MD-DE-DC Water Science Center, Baltimore, MD.
Available at http://md.usgs.gov/publications/wrir-99-4054.
RTI (RTI International). 2007. Review of Candidate Fate and Transport and Ecological Models.
Report prepared for the U.S. Environmental Protection Agency under Contract No.
EP-D-06-003. RTI International, Research Triangle Park, NC.
RTI (RTI International). 2008. Methodology Development for Linking Ecosystem Indicators to
Ecosystem Services. Report prepared for the U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards under Contract No. EP-D-06-003. RTI
International, Research Triangle Park, NC.
Schwarz, G.E., A.B. Hoos, R.B. Alexander, and R.A. Smith. 2006. The SPARROW Surface
Water-Quality Model—Theory, Applications and User Documentation. USGS
Techniques and Methods 6-B3, 248 p. and CD -ROM. U.S. Department of the Interior,
U.S. Geological Survey, Reston, VA.
Smith, R.A., and R.B. Alexander. 2000. Sources of nutrients in the nation's watersheds. In
Managing Nutrients and Pathogens from Animal Agriculture. Proceedings from the
Natural Resource, Agriculture, and Engineering Service Conference for Nutrient
Management Consultants, Extension Educators, and Producer Advisors, Camp Hill, PA,
March 28-30. Natural Resource, Agriculture, and Engineering Service, Cooperative
Extension, Ithaca, NY.
Smith, M.D., and L.B. Crowder. 2005. Valuing ecosystem services with fishery rents: a lumped
parameter approach to hypoxia in the Neuse River Estuary. Milan, Italy: Fondazione Eni
Enrico Mattel (FEEM), NRM Nota di Lavoro (Natural Resources Management Working
Papers), 115.05. Available at http://ssrn.com/abstract=825587 (accessed November 5,
2007).
Smith, R.A., R.B. Alexander, G.D. Tasker, C.V. Price, K.W. Robinson, andD.A. White. 1994.
Statistical modeling of water quality in regional watersheds. Pp. 751-754 in Proceedings
of Watershed '93—A National Conference on Water shed Management, Alexandria, VA,
Final Risk and Exposure Assessment September 2009
Appendix 6-112
-------
Aquatic Nutrient Enrichment Case Study
March 21-24, 1993. EPA 840-R-94-002. U.S. Environmental Protection Agency,
Washington, DC.
Smith, R.A., G.E. Schwarz, and R.B. Alexander. 1997. Regional interpretation of water-quality
monitoring data. Water Resources Research 33(12):2781-2798.
Spruill, T.B., AJ. Tesoriero, H.E. Mew, Jr., K.M. Farrell, S.L. Harden, A.B. Colosimo, and S.R.
Kraemer. 2004. Geochemistry and Characteristics of Nitrogen Transport at a Confined
Animal Feeding Operation in a Coastal Plain Agricultural Watershed, and Implications
for Nutrient Loading in the Neuse River Basin, North Carolina, 1999-2002. Scientific
Investigations Report 2004-5283. U.S. Department of the Interior, U.S. Geological
Survey, Reston, VA.
University of North Carolina. 2008. The Neuse Rover Estuary Modeling and Monitoring Project
(ModMon). Online information. University of North Carolina, Institute of Marine
Sciences, Morehead City, NC. Available at
http://www.unc.edu/ims/neuse/modmon/index.htm (accessed December 2008).
U.S. EPA (Environmental Protection Agency). 1996. Air quality criteria for ozone and related
photochemicaloxidants. EPA/600/AP-93/004aF-cF. 3v. U.S. Environmental Protection
Agency, Office of Research and Development, Research Triangle Park, NC. Available at
http ://cfpub2. epa.gov/ncea/.
U.S. EPA (Environmental Protection Agency). 2000. Deposition of air pollutants to the great
waters. Third report to Congress. EPA-453/R-00-005. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC.
Available at http://www.epa.gov/air/oaqps/gr8water/3rdrpt/ (accessed January 16, 2008).
U.S. EPA (Environmental Protection Agency). 2002. Summary Table for Nutrient Criteria
Documents. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
Available at:
http://www.epa.gov/waterscience/criteria/nutrient/ecoregions/files/sumtable.pdf.
Final Risk and Exposure Assessment September 2009
Appendix 6-113
-------
Aquatic Nutrient Enrichment Case Study
U.S. EPA (Environmental Protection Agency). 2005. Advisory on Plans for Ecological Effects
Analysis in the Analytical Plan for EPA 's Second Prospective Analysis—Benefits and
Costs of the Clean Air Act, 1990-2020. U.S. Environmental Protection Agency, Office of
the Administrator, Science Advisory Board, Washington, DC. June 23.
U.S. EPA (Environmental Protection Agency). 2008a. Integrated Science Assessment for Oxides
of Nitrogen and Sulfur—Ecological Criteria. Final Report. EPA/600/R-08/082F. U.S.
Environmental Protection Agency, National Center for Environmental Assessment-RTF
Division, Office of Research and Development, Research Triangle Park, NC. Available at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201485.
U.S. EPA (Environmental Protection Agency). 2008b. Draft Scope and Methods Plan for
Risk/Exposure Assessment: Secondary NAAQS Review for Oxides of Nitrogen and Oxides
of Sulfur. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, NC Available at
http://www.epa.gov/ttn/naaqs/standards/no2so2sec/cr_pd.html.
USDA (U.S. Department of Agriculture). 2002. 2002 Census of Agriculture - State Data: Table
13 "Poultry - Inventory and Sales: 2002 and 1997. " U. S. Department of Agriculture,
National Agricultural Statistics Service, Washington, DC. Available at
http://www.nass.usda.gov/census/census02/volumel/us/st99_2_013_013.pdf.
USGS (U.S. Geological Survey). 1999. Digital representation of "Atlas of United States Trees"
by ElbertL. Little, Jr. Digital Version 1.0. U.S. Department of the Interior, U.S.
Geological Survey, Denver, CO. Available at http://esp.cr.usgs.gov/data/atlas/little
(accessed July 25, 2008).
Valiela, I, and J. Costa. 1988. Eutrophication of Buttermilk Bay, a Cape Cod coastal
embayment: Concentrations of nutrients and watershed nutrient budgets. Environmental
Management 72:53 9-5 51.
Valiela, I, J. Costa, K. Foreman, J.M. Teal, B. Howes, and D. Aubrey. 1990. Transport of
groundwater-borne nutrients from watersheds and their effects on coastal waters.
Biogeochemistry 70:177-197'.
Final Risk and Exposure Assessment September 2009
Appendix 6-114
-------
Aquatic Nutrient Enrichment Case Study
VIMS (Virginia Institute of Marine Science). 2008. Submerged Aquatic Vegetation (SAV) in
Chesapeake Bay andDelmarva Peninsula Coastal Bays. Online information. Virginia
Institute of Marine Science, Gloucester Point, VA. Available at
http://web.vims.edu/bio/sav/?svr=www (accessed December 2008).
Whitall, D., and H.W. Paerl. 2001. Spatiotemporal variability of wet atmospheric nitrogen
deposition to the Neuse River Estuary, North Carolina. Journal of Environmental Quality
30:1508-1515.
Whitall, D., S. Bricker, J. Ferreira, A.M. Nobre, T. Simas, and M. Silva. 2007. Assessment of
eutrophication in estuaries: pressure-state-response and nitrogen source apportionment.
Environmental Management 40:678-690.
Final Risk and Exposure Assessment September 2009
Appendix 6-115
-------
-------
September 2009
Appendix 7
Terrestrial Nutrient Enrichment Case Study
Final
EPA Contract Number EP-D-06-003
Work Assignment 3-62
Project Number 0209897.003.062
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27709
Prepared by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
INTERNATIONAL
-------
-------
Terrestrial Nutrient Enrichment Case Study
TABLE OF CONTENTS
Acronyms and Abbreviations iv
Introduction 1
1.0 Background 2
1.1 Indicators, Ecological Endpoints, and Ecosystem Services 3
1.1.1 What Is Known 3
1.1.2 What Is Not Known 6
1.1.3 Benchmarks Selected for This Case Study 7
1.1.4 Ecosystem Services 8
1.2 Case Study Site Selection 9
1.2.1 National Overview of Sensitive Areas 9
1.2.1.1 Presence of Acidophytic Lichens 9
1.2.1.2 Anthropogenic Land Cover 9
1.2.1.3 Nitrogen-Sensitive Species Identified in Literature 10
1.2.1.4 Excluded Datasets 10
1.2.1.5 Overlay Results 10
1.2.2 Use of ISA Information and Rationale for Site Selection 12
1.3 Ecosystem Overview 14
1.3.1 Coastal Sage Scrub 14
1.3.2 Mixed Conifer Forest 18
1.4 Historical Trends 23
1.4.1 Coastal Sage Scrub 23
1.4.2 Mixed Conifer Forest 24
2.0 Approach and Methodology 25
2.1 Published Research 26
2.2 GIS Methodology 26
2.2.1 Overview 26
2.2.2 Available Data Inputs 26
2.2.2.1 Nitrogen Deposition 26
2.2.2.2 Range of Coastal Sage Scrub 27
2.2.2.3 Fire Threat 29
2.2.2.4 Changes in Coastal Sage Scrub Communities 29
2.2.2.5 Distribution of Invasive Species 29
2.2.2.6 Threatened and Endangered Species Habitat 29
2.2.2.7 Range of Mixed Conifer Forest 30
2.2.2.8 Distribution of Acid-Sensitive Lichens 30
3.0 Re suits 30
3.1 Literature Review Findings 31
3.1.1 Coastal Sage Scrub 31
3.1.1.1 Atmospheric Nitrogen Deposition 34
3.1.1.2 Nonnative Grasses 36
3.1.1.3 Mycorrhizae 36
3.1.1.4 Soil Nitrogen 37
Final Risk and Exposure Assessment September 2009
Appendix 7 - i
-------
Terrestrial Nutrient Enrichment Case Study
3.1.2 Fire 38
3.1.3 Coastal Sage Scrub Model 39
3.1.4 Mixed Conifer Forest Ecosystems 40
3.1.4.1 Nitrogen and Ozone Effects 40
3.1.4.2 Nitrogen Effects on Lichens 44
3.1.4.3 Nitrogen Saturation 46
3.2 Results Summary 48
4.0 Implications for Other Systems 50
5.0 Uncertainty 55
5.1 Coastal Sage Scrub 55
5.2 Mixed Conifer Forest 56
6.0 Conclusions 56
7.0 References 57
LIST OF FIGURES
Figure 1.1-1. Observed effects from ambient and experimental atmospheric nitrogen
deposition loads in relation to 2002 CMAQ/NADP deposition data.
Citations for effect results can be found in the ISA, Table 4-4 (U.S. EPA,
2008) 5
Figure 1.2-1. Areas of highest potential nutrient enrichment sensitivity. (Acidophytic
lichens, tree species, and the extent of the Mojave Desert are from data
obtained from the USFS. The extents of coastal sage scrub and California
mixed conifer are from the California Fire and Resource Assessment
Program. Grasslands are from the National Land Cover Dataset [USGS]) 11
Figure 1.3-1. Range of coastal sage scrub ecosystems 15
Figure 1.3-2. Presence of three threatened and endangered species in California's coastal
sage scrub ecosystem 16
Figure 1.3-3. Range of California's mixed conifer forests 19
Figure 1.3-4. Presence of two threatened and endangered species and mixed conifer
forests 19
Figure 1.4-1. Change in coastal sage scrub extent from 1977 to 2002 24
Figure 3.1-1. Coastal sage scrub range and total nitrogen deposition using CMAQ 2002
modeling results andNADP monitoring data 35
Figure 3.1-2. Current fire threats to coastal sage scrub ecosystems 39
Figure 3.1-3. Model of coastal sage scrub ecosystem in relation to fire and atmospheric
nitrogen deposition 40
Figure 3.1-4. Mixed conifer forest range and total atmospheric nitrogen deposition using
CMAQ 2002 modeling results andNADP monitoring data 42
Figure 3.1-5. Conceptual model for increased susceptibility to wildfire in mixed conifer
forests (Grulke et al., 2008) 44
Figure 3.1-6. Importance of lichens as an indicator of ecosystem health (Jovan, 2008) 44
Final Risk and Exposure Assessment September 2009
Appendix 7 - ii
-------
Terrestrial Nutrient Enrichment Case Study
Figure 3.1-7. Presence of acidophyte lichens and total nitrogen deposition in the
California mountain ranges using CMAQ 2002 modeling results and
NADP monitoring data 46
Figure 3.2-1. Illustration of the range of terrestrial ecosystem effects observed relative to
atmospheric nitrogen deposition 49
Figure 4.1-1. 2002 CMAQ-modeled and NADP monitoring data for deposition of total
nitrogen in the western United States 51
Figure 4.1-2. Benchmarks of atmospheric nitrogen deposition for several ecosystem
indicators with the inclusion of the diatom changes in the Rocky Mountain
lakes 52
Figure 4.1-3. Habitats that may experience ecological benchmarks similar to coastal sage
scrub and mixed conifer forest 53
LIST OF TABLES
Table 1.2-1. Potential Assessment Areas for Terrestrial Nutrient Enrichment Identified in
the ISA (U.S. EPA, 2008) 13
Table 1.3-1. Selected Flora and Fauna Associated with the Coastal Sage Scrub Habitat 17
Table 1.3-2. Selected Flora and Fauna Associated with the Mixed Conifer Forest Habitat 20
Table 1.3-3. List of Lichen Species Present in the Sierra Nevada Range and San
Bernardino Mountains (Jovan, 2008; Sigal and Nash, 1983) 22
Table 3.1-1. Summary of Selected Experimental Variables across Multiple Coastal Sage
Scrub Study Areasa 32
Table 3.1-2. Coastal Sage Scrub Ecosystem Area and Nitrogen Deposition 36
Table 3.1-3. Research Evidence of Ecosystem Responses to Nitrogen Relevant to Coastal
Sage Scrub 37
Table 3.1-4. Mixed Conifer Forest Ecosystem Area and Nitrogen Deposition 48
Table 4.1-1. Areas of Coastal Sage Scrub and Mixed Conifer Forest That Exceed
Benchmark Nitrogen Deposition Levels 53
Final Risk and Exposure Assessment September 2009
Appendix 7 - iii
-------
Terrestrial Nutrient Enrichment Case Study
ACRONYMS AND ABBREVIATIONS
AM arbuscular mycorrhizae
C:N carbon to nitrogen ratio
cm centimeter
CMAQ Community Multiscale Air Quality model
CC>2 carbon dioxide
CSS coastal sage scrub
FIA Forest Inventory and Analysis National Program
FRAP Fire and Resource Assessment Program
FWS U.S. Fish and Wildlife Service
GAP Gap Analysis Project
GIS geographic information systems
HNO3 nitric acid
ISA Integrated Science Assessment
kg kilogram
km kilometer
m meter
MCF mixed conifer forest
MEA Millennium Ecosystem Assessment
mm millimeter
NADP National Atmospheric Deposition Program
NH3 ammonia gas
NH4+ ammonium
NHX reduced nitrogen
NO nitric oxide
MV nitrate
NOX nitrogen oxides
O3 ozone
PNV Potential Natural Vegetation
8MB Simple Mass Balance
TM Thematic Mapper
ug/g micrograms per gram
USFS U.S. Forest Service
USGS U.S. Geological Survey
VTM Vegetation Type Map
Final Risk and Exposure Assessment September 2009
Appendix 7 - iv
-------
Terrestrial Nutrient Enrichment Case Study
INTRODUCTION
For the last half century, landscapes in the United States have been exposed to
atmospherically deposited nitrogen from anthropogenic activities. Some of the highest deposition
has occurred in Southern California, where researchers have documented measurable ecological
changes related to atmospheric deposition. This case study investigated the coastal sage scrub
(CSS) and the mixed conifer forest (MCF) ecosystems. Research was conducted on these
complex ecosystems to understand the relationships among the effects of nitrogen loads, fire
frequency and intensity, and invasive plants. The breadth of spatial and temporal data needed for
quantitative modeling of ecological response in the CSS and MCF ecosystems is not currently
available. However, biologically meaningful ecological endpoints were identified and compared
to ecological endpoints identified in the other case studies presented in the Risk and Exposure
Assessment (Chapters 4 and 5), as well as similar ecological endpoints from ecosystems in
different parts of the United States. The results in this case study report are based on geospatial
analysis and published empirical research.
Evidence from the two ecosystems discussed in this case study report supports the
finding that nitrogen alters the CSS and MCF ecosystems. For this analysis, the loss of the native
shrubs in the CSS and the increase in nonnative annual grasses were investigated. In MCF on the
slopes of the San Bernardino and Sierra Nevada Range, lichen communities associated with the
forest stands and nitrogen saturation were investigated to identify the effects of nitrogen
loadings. Changes in nitrogen loading may also affect the ecological services provided by the
CSS and MCF ecosystems, including regulation (e.g., water, habitat), cultural and aesthetic value
(e.g., recreation, natural landscape, sense of place), and provisioning (e.g., timber) (MEA, 2005).
In addition, these locations have the following characteristics that make them good candidates
for case studies:
• There is public interest
• Data were available to begin investigation (especially geographic information systems
[GIS] datasets)
• Effects observed have implications for other ecosystems and ecosystem services
• Ecological endpoints related to nitrogen deposition can be identified
Final Risk and Exposure Assessment September 2009
Appendix 7-1
-------
Terrestrial Nutrient Enrichment Case Study
• Observed effects, such as mycorrhizal responses, increase in nonnative annual grasses,
decrease in certain lichen species, and nitrate (N(V) leaching are considered by
researchers to be linked to atmospheric nitrogen deposition.
Section 3.3 of the Integrated Science Assessment (ISA) for Oxides of Nitrogen and
Sulfur-Ecological Criteria (FinalReport) (ISA) (U.S. EPA, 2008) describes the ecosystems and
species of concern, identifies trends in the ecosystems and the effects of these trends, and
discusses research efforts that investigated the variables and driving forces that may affect the
communities. The Community Multiscale Air Quality (CMAQ) 2002 modeling results and 2002
National Atmospheric Deposition Program (NADP) data were used to gain an understanding of
how atmospheric deposition of nitrogen is spatially distributed. GIS data on the spatial extent of
the habitat and changes in that extent, the location of fire threat (an important variable in both
CSS and MCF ecosystems), and the location of species of concern were used to compare these
patterns to the CMAQ/NADP data. In sum, spatial information and observed, experimental
effects were used to help identify the trends in these ecosystems and describe the past and current
spatial extent of the ecosystems.
The following ecological endpoints were identified for CSS:
• Loss of CSS native shrubs
• Mycorrhizal (a symbiotic association of fungi and plant roots) responses
• Nonnative annual grass biomass.
The following ecological endpoints were identified for MCF:
• Lichen community species
• NOs" leaching.
1.0 BACKGROUND
Current analysis of the effects of terrestrial nutrient enrichment from atmospheric
nitrogen deposition in both CSS and MCF seeks to improve scientific understanding of the
interactions among nitrogen deposition, fire events, and community dynamics. The available
scientific information is sufficient to identify ecological endpoints that are affected by nitrogen
deposition. These ecological endpoints can be compared to the ecological endpoints identified
Final Risk and Exposure Assessment September 2009
Appendix 7-2
-------
Terrestrial Nutrient Enrichment Case Study
from modeling conducted for other case studies in the Risk and Exposure Assessment (Chapters 4
and 5). These ecological endpoints can also be compared to similar ecological endpoints from
different ecosystems.
1.1 INDICATORS, ECOLOGICAL ENDPOINTS, AND ECOSYSTEM
SERVICES
1.1.1 What Is Known
Determining an acceptable ambient air concentration of nitrogen oxides (NOX) for this
case study required knowledge of ecosystem sensitivity to subsequent atmospheric deposition.
Terrestrial nutrient enrichment research has measured ecosystems' exposure to deposition of
various atmospheric nitrogen species, including nitrogen oxides, reduced nitrogen, and total
nitrogen. The ISA (U.S. EPA, 2008, Section 3.3) documents current understanding of the effects
of nitrogen nutrient enrichment on terrestrial ecosystems. The report concludes that there is
sufficient information to suggest a causal relationship between atmospheric nitrogen deposition
and biogeochemical cycling and fluxes of nitrogen in terrestrial systems. The ISA further
concludes that there is a causal relation between atmospheric nitrogen deposition and changes in
species richness, species composition, and biodiversity in terrestrial systems. These conclusions
are based on an extensive literature review that is summarized in Table 4-4 of the ISA. The
research involves both observational and experimental (e.g., nitrogen addition) projects. Alpine
ecosystems, grasslands (e.g., arid and semiarid ecosystems), forests, and deserts were included.
This extensive documentation was used to assist in selecting the case study areas to identify and
compare ecological endpoints from different habitats.
CSS is subject to several pressures, such as land conversion, grazing, fire, and pollution,
all of which have been observed to induce declines in other ecosystems (Allen et al., 1998).
Research suggests that both fire and increased nitrogen can enhance the growth of nonnative
grasses in established CSS ecosystems. It is hypothesized that many stands are no longer limited
by nitrogen and have instead become nitrogen-saturated due to atmospheric nitrogen deposition
(Allen et al., 1998; Westman, 1981a). Nitrogen availability may favor the germination and
growth of nonnative grasses, which can create a dense network of shallow roots that slow the
diffusion of water through soil, decrease the percolation depth of precipitation, and decrease the
water storage capability of the soil and underlying bedrock (Wood et al., 2006). Establishment of
Final Risk and Exposure Assessment September 2009
Appendix 7-3
-------
Terrestrial Nutrient Enrichment Case Study
CSS species may be decreased because of decreased water and nitrogen availability at depths
where more woody CSS tap roots are found (Keeler-Wolf, 1995; Wood et al., 2006).
The ISA (U.S. EPA, 2008, Section 3.3) notes that there are areas of CSS of Southern
California where dry nitrogen deposition approaches 30 kilograms (kg) N/ha/yr (Bytnerowicz
and Fenn, 1996). Seedlings of native shrubs and nonwoody plants in these areas of high nitrogen
deposition are unable to compete with dense stands of exotic grasses, and thus are gradually
replaced by grasses, particularly following disturbances, such as fire (Eliason and Allen, 1997;
Yoshida and Allen 2001; Clone et al., 2002). CSS has been declining in land area and in shrub
density for the past 60 years, and in many places is being replaced by nonnative annual grasses
(Allen et al., 1998; Padgett and Allen, 1999). Nitrogen deposition has been suggested as a
possible cause or factor in this ecosystem alteration (U.S. EPA, 2008, Section 3.3).
The ISA (U.S. EPA, 2008, Section 3.3) discusses the extensive land areas in the western
United States that receive low levels of atmospheric nitrogen deposition and which are
interspaced with areas of relatively higher atmospheric deposition downwind of large
metropolitan centers and agricultural areas. Fenn et al. (2008) determined empirical critical loads
for atmospheric nitrogen deposition in MCF, based on changes in leached N(V in receiving
waters and decreased fine-root biomass in Ponderosa pine (Pinusponderosa), and based on
changes in epiphytic lichen communities. An atmospheric nitrogen deposition of 17 kg N/ha/yr
was found to be associated with NCVleaching and an approximately 25% reduction in fine root
biomass. The study further noted that lichens are good early indicators of atmospheric nitrogen
deposition effects on other MCF species because lichens rely entirely on atmospheric nitrogen
and cannot regulate uptake. From the lichen data, Fenn et al. (2008) predicted that a critical load
of 3.1 kg N/ha/yr would be protective for all components of the forest ecosystem.
Figure 1.1-1 displays a map of observed effects from ambient and experimental
atmospheric nitrogen deposition loads in relation to 2002 CMAQ-modeled deposition results.
The map depicts the areas where empirical effects of terrestrial nutrient enrichment have been
observed and the area's proximity to elevated levels of nitrogen deposition. The ISA (U.S. EPA,
2008, Section 3.3) also identifies areas of the western United States where atmospheric nitrogen
deposition effects have been reported.
Final Risk and Exposure Assessment September 2009
Appendix 7-4
-------
Terrestrial Nutrient Enrichment Case Study
16.17
Legend
Total N Deposition
(kfl(haryr)
•i High 66.507 _>L°c.t»ns
•1 National Parks
_ LOW 0 761 M National Forests
V
1. Nitrogen enrichment or eutrophication of lakes (Loch Vale, CO: 0.5 to1.5 kg N/ha/yr; Niwot Ridge, CO: 4.71 kg N/ha/yr)
2. Alpine lakes increase shift in diatom species (Rocky Mountains, CO: 2 kg N/ha/yr)
3. Alpine meadows' elevated NO3" levels in runoff (Colorado Front Range: 20, 40, 60 kg N/ha/yr)
4. Alpine meadows' shift toward hairgrass (Niwot Ridge, CO: 25 kg N/ha/yr)
5. Nitrogen enrichment or nitrogen saturation (e.g., soil and foliar nitrogen concentration) (eastern slope of Rocky Mountains: 1.2;
3.6 kg/ha/yr; Fraser Forest, CO: 3.2 to 5.5 kg N/ha/yr)
6. Increased nitrogen mineralization rates and nitrification (Loch Vale, CO (spruce): 1.7 kg N/ha/yr)
7. Alpine tundra with increased plant foliage and decreased species richness (Niwot Ridge, CO: 50 kg N/ha/yr)
8. Nitrogen saturation, high NO3" in streamwater, soil, leaves; high nitric oxide (NO) emissions (Los Angeles, CA air basin:
saturation at 24 to 25 kg N/ha/yr (dry) and at 0.8 to 45 kg N/ha/yr (wet); northeastern U.S.: 3.3 to 12.7 kg N/ha/yr)
9. Nitrogen saturation, high NO3"in streamwater (San Bernardino Mountains, CA (coniferous): 2.9 and 18.8 kg N/ha/yr)
10. NO3" leaching (New England: Adirondack lakes: 8to10 kg N/ha/yr)
11. Nitrogen saturation, high dissolved inorganic nitrogen (San Bernardino Mountains, San Gabriel Mountains, CA, chaparral,
hardwood, coniferous): 11 to 40 kg N/ha/yr)
12. Increased tree mortality and beetle activity (San Bernardino Mountains, CA (Ponderosa): 8 and 82 kg N/ha/yr)
13. Enhanced growth of black cherry and yellow poplar; possible decline in red maple vigor; increased foliar nitrogen (Fernow
Forest, WV: 35.5 kg N/ha/yr)
14. Impacts on lichen communities (California MCF: 3.1 kg N/ha/yr; Columbia R. Gorge, OR/WA: 11.5 to 25.4 kg N/ha/yr)
15. Evidence that threatened and endangered species impacted San Francisco Bay, CA (checkerspot butterfly and serpentinitic
grass invasion: 10 to15 kg N/ha/yr; Jasper Ridge, CA: 70 kg N/ha/yr)
16. Decreased diversity of mycorrhizal communities (Southern California: ~10 kg N/ha/yr; Northern Michigan (maple/sugar maple): 5
to 9 kg N/ha/yr)
17. Decreased abundance of CSS (Southern California: 3.3 kg N/ha/yr)
18. Loss of grasslands (Cedar Creek, MN: 5.3 [1.3 to 9.8] kg N/ha/yr)
19. Decrease in abundance of desert creosote bush, increase in nonnative grasses (Mojave Desert and Chihuahuan Desert, CA:
1.7 kg N/ha/yr and up)
20. Decrease in pitcher plant (Hawley Bog, MA; Molly Bog, VA: 10 to 14 kg N/ha/yr)
Figure 1.1-1. Observed effects from ambient and experimental atmospheric
nitrogen deposition loads in relation to 2002 CMAQ/NADP deposition data.
Citations for effect results can be found in the ISA, Table 4-4 (U.S. EPA, 2008).
Final Risk and Exposure Assessment
Appendix 7-5
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
1.1.2 What Is Not Known
The ISA (U.S. EPA, 2008, Section 3.3) indicates that information is limited about the
spatial extent and distribution of terrestrial ecosystems most sensitive to nutrient enrichment
from atmospheric nitrogen deposition: "Effects are most likely to occur where areas of relatively
high atmospheric nitrogen deposition intersect with nitrogen-limited plant communities. The
factors that govern the sensitivity of terrestrial ecosystems to nutrient enrichment from
atmospheric nitrogen deposition include the degree of nitrogen limitation, rates and form of
atmospheric nitrogen deposition, elevation, species composition, length of growing season, and
soil nitrogen retention capacity." Examples of sensitive ecosystems include the following:
• Alpine tundra (low rates of primary production, short growing season, low temperature,
wide moisture variation, low nutrient supply).
• Western U.S. ecosystems, such as the alpine ecosystems of the Colorado Front Range,
chaparral watersheds of the Sierra Nevada Range, lichen communities in the San
Bernardino Mountains and the Pacific Northwest, and CSS communities in Southern
California.
• Eastern U.S. ecosystems where sensitivities are typically assessed in terms of the degree
of NOs" leaching from soils into ground and surface waters. These ecosystems are
expected to include hardwood forests, semiarid lands, and grassland ecosystems, but
effects on individual plant species have not been studied well.
Major indicators for nutrient enrichment to terrestrial systems from atmospheric
deposition of total reactive nitrogen, such as those described above, require measurements based
on available monitoring stations for wet deposition (NADP/National Trends Network) and
limited networks for dry deposition (Clean Air Status and Trends Network [CASTNet]).).
However, data have been limited, particularly at the spatial scale required for a more accurate
analysis. Wet deposition monitoring stations can provide more information on an extensive range
of nitrogen species than can dry deposition monitoring stations. In the Mediterranean systems of
Southern California where rainfall is concentrated during some months of the year, dry
deposition is particularly important. Individual studies measuring atmospheric nitrogen
deposition to terrestrial ecosystems that involve throughfall estimates for forested ecosystems
Final Risk and Exposure Assessment September 2009
Appendix 7-6
-------
Terrestrial Nutrient Enrichment Case Study
can provide better approximations for total atmospheric nitrogen deposition levels; however,
such estimates and related bioassessment data are not available for the entire country.
Finally, in the area of what is still unknown, the exact relationship between atmospheric
nitrogen loadings, fire frequency and intensity, and nonnative plants, particularly in the CSS
ecosystem, have not been quantified. Various conceptual models linking these factors have been
developed, but an understanding of cause and effect, seasonal influences, and benchmarks
remains undeveloped. These potential confounders are discussed at greater length in Section 3.
1.1.3 Benchmarks Selected for This Case Study
The data limitations on atmospheric nitrogen deposition (described above), along with
current data to describe the full extent and distribution of nitrogen-sensitive U.S. ecosystems,
presented a barrier to designing a case study that used quantitative monitoring and modeling
tools. Instead, this case study used published research results to identify meaningful ecological
endpoints associated with different levels of atmospheric nitrogen deposition.
The ecological endpoints that were identified for the CSS and the MCF are included in
the suite of ecological endpoints identified in the ISA (U.S. EPA, 2008, Section 3.3). There are
sufficient data to confidently relate the ecological effect to a loading of atmospheric nitrogen.
For the CSS community, the following ecological benchmarks were identified:
• 3.3 kg N/ha/yr—the amount of nitrogen uptake by a vigorous stand of CSS; above this
level, nitrogen may no longer be limiting
• 10 kg N/ha/yr—mycorrhizal community changes
For the MCF community, the following ecological benchmarks were identified:
• 3.1 kg N/ha/yr—shift from sensitive to tolerant lichen species
• 5.2 kg N/ha/yr—dominance of the tolerant lichen species
• 10.2 kg N/ha/yr—loss of sensitive lichen species
• 17kg N/ha/yr—leaching of N(V into streams.
Final Risk and Exposure Assessment September 2009
Appendix 7-7
-------
Terrestrial Nutrient Enrichment Case Study
1.1.4 Ecosystem Services
Ecosystem services are generally defined as the benefits individuals and organizations
obtain from ecosystems. In the 2005 Millennium Ecosystem Assessment (MEA), ecosystem
services are classified into four main categories:
• Provisioning—includes products obtained from ecosystems
• Regulating—includes benefits obtained from the regulation of ecosystem processes
• Cultural—includes the nonmaterial benefits people obtain from ecosystems through
spiritual enrichment, cognitive development, reflection, recreation, and aesthetic
experiences
• Supporting—includes those services necessary for the production of all other ecosystem
services (MEA, 2005).
Atmospheric nitrogen deposition affects CSS and MCF ecological processes that, in turn,
are related to ecosystem services. These processes include the following:
For CSS:
• Decline in CSS habitat, shrub abundance, species of concern—cultural and
regulating
• Increased abundance of nonnatives—cultural and regulating
• Increase in wildfires—cultural and regulating.
For MCF:
• Change in habitat suitability and increased tree mortality—cultural and regulating
• Decline in MCF aesthetics—cultural
• Increase in fire intensity, change in forest's nutrient cycling, other nutrients
becoming limiting—regulating
• Decline in surface water quality—regulating.
Terrestrial nutrient enrichment for CSS potentially affects ecosystem services, such as
biodiversity; threatened and endangered species and rare species (both national and state);
landscape view; water quality; and fire hazard mitigation. Linking ecological endpoint to
services involves the measurement of changes in biodiversity and abundance and distribution of
Final Risk and Exposure Assessment September 2009
Appendix 7-8
-------
Terrestrial Nutrient Enrichment Case Study
threatened and endangered species, comparison of past and present photography, and
measurement of the distribution of soil moisture with depth and possible N(V leaching. The
relationship between fire frequency, CSS ecosystem, and property values will be investigated in
the ecosystem services analysis.
The case study approach for MCF focused on ecosystem services, such as visual and
recreational aesthetics provided by the CSS community and water quality. Linking ecological
endpoints to services includes measurement of the density of stands, shifts in tree dominance,
shifts in lichen communities, foliar nitrogen increases, and increased NOs concentrations in
streams due to leaching.
1.2 CASE STUDY SITE SELECTION
1.2.1 National Overview of Sensitive Areas
The selection of case study areas specific to terrestrial nutrient enrichment began with
national GIS mapping. GIS datasets of physical, chemical, and biological properties that were
indicative of potential terrestrial nutrient enrichment were considered in order to identify
sensitive areas in the United States. The publicly available geospatial datasets outlined in the
following paragraphs have been identified as important contributors to terrestrial nutrient
enrichment and met the selection criteria.
1.2.1.1 Presence of A cidophytic Lichens
Acidophytic lichens are known to be sensitive to increased levels of nitrogen loading.
Other species are dependent upon lichens for both food and habitat. For this exercise, the list of
acidophytic species from Fenn et al. (2008) was used. Data on these species were available for
the years 2001 to 2006. Geospatial data were obtained from the U.S. Forest Service (USFS)
Forest Inventory and Analysis National Program (FIA) (USFS, 2008a). Locations where
acidophytic lichen were identified were defined as being sensitive.
1.2.1.2 A nthropogenic Land Cover
Urban and agricultural land covers were mapped to so that they could be used to exclude
areas that are not sensitive to terrestrial nutrient enrichment, such as agricultural areas and
urbanized areas. This information was obtained from the U.S. Geological Survey (USGS)
Final Risk and Exposure Assessment September 2009
Appendix 7-9
-------
Terrestrial Nutrient Enrichment Case Study
National Atlas of the United States (USGS, 2006) and covered the continental United States at a
spatial resolution of 1-km grid cells.
1.2.1.3 Nitrogen-Sensitive Species Identified in Literature
Although there is no known nationwide species that has shown range loss because of
additional nitrogen, it was possible to assemble a "patchwork quilt" of species and forest types
from across the United States that are identified as sensitive in the published literature. A range
was extracted from national datasets for each species or forest type where the range existed. The
cumulative extent of all ranges allowed for the definition of sensitive areas in the United States.
1.2.1.4 Excluded Datasets
The publicly available spatial datasets outlined below were considered for inclusion in
the national sensitivity assessment, but were not used.
• Soil Nitrogen Content. This pre-1980 dataset was requested but not received at the time
of this report's production. The quality of data is uncertain.
• Presence of Mountains. The physiographic provinces of the United States were
considered to provide leeward sides of mountains that tend to receive a greater amount of
atmospheric nitrogen deposition. Continental U.S. data identified were from USGS and
dated 1946. The spatial resolution was a scale of 1:7,000,000. If used, the benchmark
value would have been for mountain ranges only. However, this dataset was not used
because terrain is already taken into account by the CMAQ modeling.
1.2.1.5 Overlay Results
The extraction of the areas of greatest nutrient enrichment sensitivity was constrained by
the relative lack of available national datasets. Therefore, the review involved two steps within
the GIS. First, the ranges of sensitive species identified in the literature were combined with a
layer of acidophytic lichen distribution. Second, areas of human use (i.e., urban and agricultural
land covers) were removed. The resulting map illustrates the area of highest potential sensitivity
(see Figure 1.2-1)
Final Risk and Exposure Assessment September 2009
Appendix 7-10
-------
Terrestrial Nutrient Enrichment Case Study
Acidophytic Lichen | | Red Pine ^] Sugar Maple/Beech/Yellow Birch Coastal Sage Scrub
| Ponderosa Pine ^^| Black Cherry ^^| Engelmann Spruce Mojave Desert
| Red Maple/Oak | Pitch Pine ^] Engelmann Spruce/Subalpine Fir Grasslands
I CA Mixed Conifer
250 500 750 1,000km
Figure 1.2-1. Areas of highest potential nutrient enrichment sensitivity. (Acidophytic lichens, tree species, and the
extent of the Mojave Desert are from data obtained from the USFS. The extents of coastal sage scrub and California
mixed conifer are from the California Fire and Resource Assessment Program. Grasslands are from the National Land
Cover Dataset [USGS]).
Final Risk and Exposure Assessment
Appendix 7-11
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
1.2.2 Use of ISA Information and Rationale for Site Selection
Potential case study areas identified in the ISA (U.S. EPA, 2008) were considered in site
selection along with information gathered in the national GIS analysis. Table 1.2-1 contains the
relevant nutrient enrichment areas identified in the ISA.
After considering this information, California's CSS and MCF ecosystems were selected
for this case study analysis. The following selection factors supplement those listed in the
Introduction:
• Availability of atmospheric ambient and deposition data (monitored or modeled)
• Availability of digitized datasets of biotic communities; fire-prone areas; and sensitive,
rare species
• Scientific results of research on nitrogen effects for the case study area
• Representation of western U.S. ecosystems potentially impacted by atmospheric nitrogen
deposition
• Scalability and generalization opportunities for risk analysis results from the case studies.
California's CSS has been the subject of intensive research in the past 10 years, which
has provided the data needed for a first phase of GIS analysis of the role of atmospheric nitrogen
deposition in terrestrial ecosystems. California's MCF have an even longer record of study that
includes investigations into the effects of atmospheric pollution, changes to forest structure,
changes to the lichen communities, and measurements of nitrogen saturation. Another ecosystem
that was considered but not selected for this case study was the alpine ecosystem in the Rocky
Mountains. As noted in the ISA (U.S. EPA, 2008, Section 3.3), results from a number of studies
indicate that nitrates may be leaching from alpine catchments, and there appear to be changes in
plant communities related to the deposition of atmospheric nitrogen. The amount of data from
these alpine ecosystems is more limited than that from the CSS and MCF. However, the
ecological benchmarks suggested for alpine ecosystems were comparable to the benchmarks
from CSS and MCF ecosystems.
In summary, CSS and MCF were selected as case study areas for the following reasons:
• The two ecosystems have significant geographic coverage and are located where urban
areas interface with wilderness areas.
Final Risk and Exposure Assessment September 2009
Appendix 7-12
-------
Terrestrial Nutrient Enrichment Case Study
• Both sites are located in areas of sharp atmospheric nitrogen deposition gradients,
ranging from low background levels to some of the highest deposition levels recorded in
the United States.
• The two ecosystems have been researched for extended periods to understand the
interactive effects of deposition, climate change, fire, and other stressors.
• The results of these research investigations for CSS and MCF result are well documented
in the peer-reviewed literature.
Table 1.2-1. Potential Assessment Areas for Terrestrial Nutrient Enrichment Identified in the
IS A (U.S. EPA, 2008)
Area
Alpine and
subalpine
communities
of the eastern
slope of the
Rocky
Mountains,
CO
Fernow
Experimental
Forest near
Parsons, WV
Bear Brook,
ME
Harvard
Forest, MA
Southern
California
Indicator
Terrestrial
nutrient
enrichment
Terrestrial
nutrient
enrichment
Terrestrial
acidification
Terrestrial
nutrient
enrichment
Terrestrial
nutrient
enrichment
Detailed
Indicator
Biomass
production;
NO3
leaching;
species
richness
Forest
growth
Sugar
maple; red
spruce
Forest
growth —
species
Forest
growth —
species;
coastal sage
scrub
References in U.S. EPA, 2008
Baron et al., 1994; Baron et al., 2000;
Baron, 2006; Bowman, 2000; Bowman
and Steltzer, 1998; Bowman et al., 1993;
Bowman et al., 1995; Bowman et al.,
2006; Burns, 2004; Fenn et al., 2003a;
Fisk et al., 1998; Korb and Ranker, 2001;
Rueth et al., 2003; Seastedt and Vaccaro,
2001; Sherrod and Seastedt, 2001;
Steltzer and Bowman, 1998; Suding et
al., 2006; Williams and Tonnessen, 2000;
Williams et al.,1996a; Wolfe et al., 2001
Adams et al., 1997, 2000; DeWalle et al.,
2006; Edwards and Helvey, 1991;
Gilliam et al., 2006; Peterjohn, 1996
Elviretal., 2003
Magill et al., 2004; Magill, 2004
Bytnerowicz and Fenn, 1996, 2003a;
Takemoto et al., 2001
Source
ISA, Section
3.3,4.3,
Annex A,
Annex C, and
Annex D
ISA, Section
3.3,4.3,
Annex A,
Annex B
ISA, Section
3.3,4.3,
Annex A,
Annex B,
Annex C
ISA, Sections
2.8,2.10,3.2,
3.3,4.3,
Annex B,
Annex C
ISA, Section
3.2,3.3,3.4,
4.3, Annex B,
Annex C
Final Risk and Exposure Assessment
Appendix 7-13
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Area
Jasper Ridge
Biological
Preserve, CA
Loch Vale,
CO
Rocky
Mountain
National Park,
CO
Indicator
Terrestrial
nutrient
enrichment
Terrestrial
nutrient
enrichment
Terrestrial
nutrient
enrichment
Detailed
Indicator
Grasslands
Old-spruce
growth
Tundra
composition
switch
References in U.S. EPA, 2008
Zavaleta et al., 2003
Rueth et al., 2003
Interlandi and Kilham, 1998
Source
ISA, Sections
3.3,4.3
ISA, Section
3.3,4.3,
Annex B
ISA, Section
3.3,4.3,
Annex C
Source: U.S. EPA, 2008.
1.3 ECOSYSTEM OVERVIEW
1.3.1 Coastal Sage Scrub
CSS consists of more than 50 aromatic shrub and subshrub species, which range from
approximately 0.5 meters (m) to 2 m in height (Burger et al., 2003; Westman, 1981a). The range
of CSS extends from north of San Francisco down to Baja California in the lower elevation
coastal range of California (see Figure 1.3-1); however, the species composition may vary with
location (Westman, 1981b). According to the California Natural Diversity Database, there are 22
floristic alliances of CSS (e.g., Riversidian Sage Scrub, Venturan Sage Scrub, and Diegan Sage
Scrub). These alliances consist of similar species that help determine the significance, rarity, and
growth patterns of California vegetation types.
CSS grows in a warm Mediterranean climate and is characterized by approximately 300
millimeters (mm) of annual rainfall falling from December through March and little or no
rainfall from April through November (Egerton-Warburton and Allen, 2000; Westman, 1981b).
Underlying substrate types of CSS vary greatly across the CSS stands, although many CSS
floristic alliances are found on unconsolidated sand, sandstone, conglomerate, and shale
(Westman, 1981b).
CSS is also known as "soft chaparral" because of its semideciduousness, drought-tolerant
nature, and less-rigid leaves, respective to chaparral species (Westman, 1981b). CSS is
considered a fire-adapted community; meaning that although vegetation layers may be destroyed
in fires, CSS soil seed banks can withstand fire, and in some species, require fire to open the seed
cases. However, many CSS species can flourish and propagate in the absence of any fire (Keeler-
Final Risk and Exposure Assessment
Appendix 7-14
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Wolf, 1995). CSS has been observed to maintain a permanent cover without fire or other
disturbance regimes (e.g., land conversion, grazing) for at least a century (Westman, 1981a).
I CMital £a#e SoubHXQ
County
fl or CSS range i& me Gdibwia Dopvtmpnl
of Fwetsy ano Fn PiMfciion
Tim dm .« put*ihed rv 3O03 tinrrg bewi
•ar*«oo Wm seu"»* Ming MM*« i MO *xj 2402
Figure 1.3-1. Range of coastal sage scrub ecosystems.
The resprouting and competition of species post-fire is generally dependent upon fire
intensity, fire frequency, and seasonal timing (Keeler-Wolf, 1995). CSS species are generally
poor colonizers after a fire (Minnich and Dezzani, 1998). Annual forbs and any grass seedlings
present in the post-fire soils are usually dominant in the first few growth cycles. Significant
shrub growth is most likely to occur in later cycles, further disturbance not withstanding (Keeler-
Wolf, 1995).
The CSS ecosystem also supports the growth of more than 550 herbaceous annual and
perennial species between and beneath the shrub canopy. Of these herbs, nearly half are
endangered, sensitive, or of special status (Burger et al., 2003). Additionally, several avian,
Final Risk and Exposure Assessment
Appendix 7-15
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
arthropod, reptilian, amphibian, and mammalian species depend on CSS habitat for foraging,
breeding, and/or residence. These include several threatened and endangered species, such as the
coastal California gnatcatcher (Polioptila californica californica), the Stephens' kangaroo rat
(Dipodomys stephenst), and the Quino checkerspot butterfly (Euphydryas editha quino). Figure
1.3-2 presents the range of these three species. Table 1.3-1 presents a selected list of flora and
fauna species that are associated with CSS habitat.
olies
| ] County
Quino Checkerspot Butterfly
Kangaroo Rat
^^| Coastal CA Gnatcatcher
| Coastal Sage Scrub 199S
Source of CSS range is trie California Department
of Forestry and Fire Protection.
Source of critical habitats s the US Fish arxl Wildlife
Service Critical Wildlife Portal.
Figure 1.3-2. Presence of three threatened and endangered species in
California's coastal sage scrub ecosystem.
Final Risk and Exposure Assessment
Appendix 7-16
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Table 1.3-1. Selected Flora and Fauna Associated with the Coastal Sage Scrub Habitat
Scientific Name
Buteo swainsoni
Polioptila californica californica
Dipodomys merriami parvus
Dipodomys stephensi
Bufo microscaphus californicus
Euphydryas editha quino
Rhaphiomidas terminates
abdominalis
Allium munzii
Rosa minutifolia
Deinandra conjugens
Cordylanthus orcuttianus
Ambrosia pumila
Acanthomintha ilicifolia
Campylorhynchus brunneicapillus
couesi
Athene cunicularia
Cnemidophorus hyperythrus
Phrynosoma coronatum blainvillei
Masticophis lateralis euryxanthus
Common Name
Swainson's Hawk
Coastal California Gnatcatcher
San Bernardino Kangaroo Rat
Stephens' Kangaroo Rat
Arroyo Toad
Quino Checkerspot Butterfly
Delhi Sands Flower-Loving Fly
Munz's Onion
Small-Leaved Rose
Otay Tarplant
Orcutt's Bird's Beak
San Diego Ambrosia
San Diego Thorn-Mint
Coastal Cactus Wren
Burrowing Owl
Orange-Throated Whiptail
San Diego Horned Lizard
Alameda Whipsnake
Life Form
Bird
Bird
Mammal
Mammal
Amphibian
Insect
Insect
Perennial Forb
Shrub
Annual Forb
Annual Forb
Perennial Forb
Annual Forb
Bird
Bird
Reptile
Reptile
Reptile
Federal Listing*
Not listed
Threatened
Endangered
Endangered
Endangered
Endangered
Endangered
Endangered
Not listed
Threatened
Not listed
Proposed Endangered
Threatened
Not listed
Not listed
Not listed
Not listed
Threatened
State Listing*
Threatened
Not listed
Not listed
Threatened
Not listed
Not listed
Not listed
Threatened
Endangered
Endangered
Not listed
Not listed
Endangered
Not listed
Not listed
Not listed
Not listed
Threatened
* Status listed for threatened and endangered species only. Others may be species of concern, on federal watch lists, or state special status.
Final Risk and Exposure Assessment
Appendix 7-17
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
The principal source of nitrogen to the CSS ecosystem is atmospheric nitrogen (e.g.,
NOX, reduced nitrogen (NHX)). These nitrogen species are transported and deposited onto the
historically nitrogen-limited CSS soil in the form of nitrates and nitric acid. In the soil, these
nitrogen species are potentially available for plant uptake and nutrient cycles. The effects of
increased availability of nitrogen species in the CSS ecosystem were the focus of this case study.
1.3.2 Mixed Conifer Forest
MCF stand approximately 30 to 50 m tall and consist of conifer species that dominate
mid-range elevations (1300 to 2800 m) of the California San Bernardino and Sierra Nevada
mountain ranges. The San Bernardino Mountains lie east of the Los Angeles air basin, and the
Sierra Nevada Range span the majority of the state longitudinally. Figure 1.3-3 illustrates the
range of MCF in California. MCF have historically adapted to withstand fire at low, medium,
and even high intensities. As in the range of CSS, the climate is Mediterranean, where 80% of
rainfall occurs from October through March (Takemoto et al., 2001).
Dominant tree species shift along a precipitation gradient. Ponderosa pine (Pinus
ponderosa), white fir (Abies concolor), sugar pine (P. lambertiana\ and incense cedar
(Calocedrus decurrens) are the predominant species on moist windward slopes, whereas Jeffrey
pine (P. jeffreyf) and white fir are commonly found on leeward slopes and at higher elevations in
the mixed conifer elevation range. Important deciduous components of the MCF are canyon live
oak (Quercus chrysolepis), black oak (Quercus kelloggi), and quaking aspen (Popus
tremuloides). These stands support a number of shrubs, subshrubs, and annual and perennial
forbs, as well as mountain meadows (Minnich, 2007). Federal-listed species, Rana sierrae and
Rana muscosa (both called the mountain yellow-legged frog), and a number of state-listed
species, such as the Peninsular bighorn sheep (Ovis canadensis nelsoni), utilize MCF habitat.
The range of two of these selected species is illustrated in Figure 1.3-4. Table 1.3-2 shows
selected flora and fauna associated with MCF habitat.
Final Risk and Exposure Assessment September 2009
Appendix 7-18
-------
Terrestrial Nutrient Enrichment Case Study
Figure 1.3-3. Range of California's mixed conifer forests.
County
| Peninsular Bighorn Sheep
Mln Yellow-legged Frog
I Mixed Conifer
Source of the Mountain Yellow-legged Frog and Peninsular Bighorn Sheep ranges
is Ifte US Forest Service Critical Habitai Portal.
Source 01 Mixed Conifer is CADept.of Forestry and Fire Protection.
Figure 1.3-4. Presence of two threatened and
endangered species and mixed conifer forests.
Final Risk and Exposure Assessment
Appendix 7-19
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Table 1.3-2. Selected Flora and Fauna Associated with the Mixed Conifer Forest Habitat
Scientific Name
Abies concolor
Pinus ponderosa
Pinus lambertiana
Calocedrus decurrens
Rana sierras
Spea hammondii
Rana muscosa
Glaucomys sabrinus
Glaucomys sabrinus californicus
Ovis canadensis nelsoni
Odocoileus hemionus
Charina umbratica
Packera bernardina
Sidalcea pedata
Perideridia parishii ssp. parishii
Taraxacum californicum
Gilia leptantha ssp. leptantha
Piranga rubra
Haliaeetus leucocephalus
Strix occidentalis occidentalis
Strix nebulosa
Common Name
White Fir
Ponderosa Pine
Sugar Pine
Incense Cedar
Sierra Madre Yellow-Legged Frog
Western Spadefoot
Sierra Madre Yellow-Legged Frog
Northern Flying Squirrel
San Bernardino Flying Squirrel
Peninsular Bighorn Sheep
Black-Tailed Deer
Southern Rubber Boa
San Bernardino Ragwort
Bird-Foot Checkerbloom
Parish's Yampah
California Dandelion
San Bernardino Gilia
Summer Tanager
Bald Eagle
California Spotted Owl
Great Gray Owl
Life Form
Tree
Tree
Tree
Tree
Amphibian
Amphibian
Amphibian
Mammal
Mammal
Mammal
Mammal
Reptile
Perennial Forb
Perennial Forb
Perennial Forb
Perennial Forb
Shrub
Bird
Bird
Bird
Bird
Federal Listing*
Not listed
Not listed
Not listed
Not listed
Endangered
Not listed
Endangered
Not listed
Not listed
Endangered
Not listed
Not listed
Not listed
Endangered
Not listed
Endangered
Not listed
Not listed
Delisted
Not listed
Not listed
State Listing*
Not listed
Not listed
Not listed
Not listed
Not listed
Not listed
Not listed
Not listed
Not listed
Threatened
Not listed
Threatened
Not listed
Endangered
Not listed
Not listed
Not listed
Not listed
Endangered
Not listed
Endangered
* Status listed for Threatened and Endangered species only. Others may be species of concern, on federal watch lists, or state special status
Final Risk and Exposure Assessment
Appendix 7-20
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Additionally, several lichen species are associated with the MCF habitats. Lichens are
formed by a symbiotic relationship between fungus and algae or cyanobacterium. In the MCF
ecosystem, lichens are generally epiphytic, living on conifers and obtaining nutrients from the
atmosphere. Epiphytic lichens serve as food, habitat, and nesting material for various species in
the pine stands (Fenn et al., 2008). The presence of individual species is determined by the
amount of nitrogen present and the pH of the vegetation on which it grows; however, general
categories for lichens have been developed according to species' sensitivity to nitrogen. These
categories include nitrophytes, neutrophytes, and acidophytes (Jovan, 2008). Nitrophytes are
generally associated with ammonia and high pH environments. Neutrophytes tolerate increased
pH and ammonia, but exhibit slower growth patterns than nitrophytes when exposed to these
conditions. Acidophytes are sensitive to nitrogen species and deteriorate or die after relatively
small increments of exposure to nitrogen species (Fenn et al., 2008). Table 1.3-3 presents a list
of lichen species, classified by nitrogen sensitivity, that have been observed in the San
Bernardino Mountains and Sierra Nevada Range.
Final Risk and Exposure Assessment September 2009
Appendix 7-21
-------
Terrestrial Nutrient Enrichment Case Study
Table 1.3-3. List of Lichen Species Present in the Sierra Nevada Range and San Bernardino Mountains
(Jovan, 2008; Sigal and Nash, 1983)
Nitrophytes
Candelaria concolor
Flavopunctelia flaventiorb
Phaeophyscia orbicularis
Physcia adscendens
Physcia aipolia
Physcia dimidiate
Physcia stellaris
Physcia tenella
Physconia enteroxantha
Physconia perisidiosa
Xanthomendoza fallax
Xanthomendoza fulva
Xanthomendoza hasseana
Xanthomendoza oregano
Xanthoria candelaria
Xanthoria polycarpa
Potential Acidophytes
Bryoria fremontii
Cetraria canadensis
Cetraria chlorophylla
Cetraria merrillii
Cetraria orbata
Cetraria pallidula
Cetraria platyphylla
Evernia prunastri
Hypogymnia enteromorpha
Hypogymnia imshaugii
Hypogymnia occidentalis
Parmeliopsis ambigua
Platismatia glauca
Usnea filipendula
—
—
Potential Neutrophytes
Melanelia elegantula
Melanelia exasperatula
Melanelia glabra
Melanelia subargentifera
Melanelia subelegantula
Melanelia subolivacea
Parmelia hygrophilab
Parmelia sulcata
Ramalina subleptocarphab
—
—
—
—
—
—
—
Unknown
Ahtiana sphaerosporella
Alectoria sarmentosa
Collema furfuraceum
Esslingeriana idahoensis
Leptogium lichenoides
Letharia columbiana
Letharia vulpina
Nodobryoria abbreviata
Nodobryoria oregana
Parmelina quercina
Parmelina elegantula
Physcia biziana
Physconia americana
Physconia isidiigera
—
—
Final Risk and Exposure Assessment
Appendix 7-22
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
1.4 HISTORICAL TRENDS
1.4.1 Coastal Sage Scrub
The CSS ecosystem is a unique system that has experienced a significant decline in
coverage since vegetation types in Southern California were inventoried in 1929. Subsequently,
this community was designated for special status in California (CA DFG, 1993). This decline is
due to urban encroachment and sprawl, increased fire frequencies, and pollution (Minnich and
Dezzani, 1998). CSS is decreasing at a higher rate than habitat destruction alone would indicate
(Allen et al., 1998; Fenn et al., 2003; Minnich and Dezzani, 1998).
Nonnative grasses were introduced to California by explorer expeditions and Franciscan
missionaries arriving in the region prior to documentation of indigenous vegetation. However,
accounts of herbaceous vegetation in the coastal range exist from the late 1700s and throughout
the 1800s (Minnich and Dezzani, 1998). CSS was first scientifically inventoried during the
California Forest and Range Experiment Station Vegetation Type Map (VTM) Survey,
beginning in 1929. Recently, 54 of the VTM sites in Southern California that were dominated by
CSS cover in the 1930s were resampled (Talluto and Suding, 2008). Since the 1930s, CSS
declined 49% and was mainly replaced by nonnative annual grasses (Talluto and Suding, 2008).
Figure 1.4-1 illustrates the decline in CSS from 1977 to 2002.
Based on changes in CSS cover from VTM data since the early 1930s, it is estimated that
approximately 18% of the extent of Riverside County CSS had been completely converted to
nonnative grasses, and an additional 42% of the cover had nonnative grasses intermixed with
CSS. Therefore, only 40% of the original extent of CSS in Riverside County remained intact and
contiguous. The 2005 resampling of part of Riverside and Orange counties indicated that 15% of
the remaining CSS had not been invaded by annual grasses (Talluto and Suding, 2008). Across
the entire CSS range, Westman (198 la) estimated that only 10% to 15% of the historical CSS
extent remained in the late 1970s. This estimate is based upon the fraction of potential CSS land
cover (in the absence of pressures) in which CSS vegetation was actually observed at the time of
the study. The potential CSS land cover estimates may also be supported by the broad range in
which specimens of the Quino checkerspot butterfly have historically been observed and
collected (Mattoni et al., 1997). Therefore, the remaining extent of CSS is most likely 10% to
82% of the historical CSS coverage, depending on the development pressures and the spread of
Final Risk and Exposure Assessment September 2009
Appendix 7-23
-------
Terrestrial Nutrient Enrichment Case Study
nonnative grasses in each stand. Additionally, these nonnative grasses are less diverse and are
not likely to support the majority of the sensitive, threatened, and endangered species that
currently rely on CSS (Allen et al., 2005).
County
| CoasTsl Sage Sc«* 1977
S«»f<» gf CSS rang*
s tn» CsHVjirvfl DepjuiTOfll
or fittstrf anO Fn
J County
| Coastal Sage Scrub Hx)2
Swc9 or CSS ino*
« mo
or Fwrelry win fire
Figure 1.4-1. Change in coastal sage scrub extent from 1977 to 2002.
1.4.2 Mixed Conifer Forest
The major trends observed in MCF are "densification" and increased litter accumulation.
Densification occurs when aboveground biomass is stimulated, resulting in increased numbers of
needles, decreased average tree age, decreased overall trunk size, and increased branches (Grulke
et al 2008; Minnich et al., 1995; Takemoto et al., 2001). In a retrospective comparison of MCF
stands in the San Bernardino Mountains from 1932 to 1992, Minnich et al., (1995) noted
significant shifts in age distribution, stand density, and branch density. Tree density increased
approximately 77% according to the VTM surveys, and there were 3 to 10 times the number of
trees in the younger age brackets when compared to conifer populations 60 years earlier.
Final Risk and Exposure Assessment
Appendix 7-24
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Additionally, a 79% increase in the average number of tree branches was reported in the San
Bernardino conifer forests. Studies have indicated that increasing stand densities are also
occurring within the Sierra Nevada Range (Minnich et al., 1995).
Increased litter on the forest floor has also been observed across the conifer ecosystems,
particularly in the MCF stands in the San Bernardino Mountains. These MCF stands have been
observed to shed needles approximately six times faster than more remote northern Sierra
Nevada Range conifer stands (Takemoto et al., 2001). Additionally, litterfall depths up to 15
centimeters (cm) have been noted in MCF stands near Camp Paivika in the eastern San
Bernardino Mountains (Grulke et al., 2008).
Across the San Bernardino Mountains, a tree community composition shift was also
noted. In MCF stands where Ponderosa pine has been historically dominant, trees in the youngest
age bracket are now predominantly white fir and incense cedar. Additional research is needed to
determine if a shift in community composition is also occurring in the Sierra Nevada Range
MCF (Minnich et al., 1995). Although research on understory communities revealed no clear
trends with atmospheric nitrogen deposition and ozone (Os), it was noted that native diversity
had declined in those areas receiving the highest loads of atmospheric nitrogen (Allen et al.,
2007).
Lichen communities associated with the MCF habitat have also been dramatically altered
(Fenn et al., 2003, 2008; Sigal and Nash, 1983). Of the!6 lichen species reported to be associated
with the San Bernardino Mountains MCF in the early part of the 20th century, only 8 species
were found 60 years later. Additionally, deterioration was observed on some of the lichens,
particularly in the areas with the highest levels of air pollution (Sigal and Nash, 1983). Lichens
are significant members of the MCF community. They serve as forage for wildlife, and changes
in the lichen community are considered by some to be a warning signal for deteriorating
conditions in the rest of the forest.
2.0 APPROACH AND METHODOLOGY
Using the approach and methodology described below, a number of significant ecological
endpoints have been identified. These results come from empirical results and from spatial
databases. Dose/response relationships beyond benchmark values were investigated, but these
have not yet been well quantified. Nitrogen deposition data was available at a 12-km resolution,
Final Risk and Exposure Assessment September 2009
Appendix 7-25
-------
Terrestrial Nutrient Enrichment Case Study
and many of the ecosystems, especially CSS, are fragmented into smaller areas. The analysis is,
therefore, somewhat limited by the discrepancy between resolution of the nitrogen deposition
data and the distribution of habitats, as well as by the specific areas where ecological processes
were measured. Further, some models have been tested, but with limited results. For example,
the steady-state simple mass-balance model (UNECE, 2004) still has many unresolved
uncertainties. Uncertainty exists in establishing the linkage between soil and biological impacts
and the ability to account for forest management and wildfires (Fenn et al., 2008). The DayCent
biogeochemical model is not a water shed-scale model and may not represent N(V leaching
accurately. Although, application of DayCent yielded results more comparable to empirically
based findings than the steady-state model (Fenn et al., 2008).
For the above reasons, empirical data, in tandem with GIS analysis, was deemed more
suitable to develop potential correlations between atmospheric nitrogen deposition and
ecological endpoints.
2.1 PUBLISHED RESEARCH
The ISA (U.S. EPA 2008, Sections 3.3, 4.3) was used as the basis for identifying the
published scientific literature on CSS and MCF ecosystems.
2.2 GIS METHODOLOGY
2.2.1 Overview
For both the CSS and MCF ecosystems, spatially distributed data are available. Some of
the variables that are known to influence terrestrial nutrient enrichment and have been cited in
the literature are available as either state-level or national-level datasets. It is important that
spatial data are temporally and spatially compatible and have well-documented metadata. It is
also desirable that they possess the ability to be scaled-up for a national characterization.
2.2.2 Available Data Inputs
2.2.2.1 Nitrogen Deposition
Wet nitrogen deposition in the forms of NOs" and ammonium (NH4+) are available
nationally from the NADP. This national network of 321 sampling stations is the best wet
Final Risk and Exposure Assessment September 2009
Appendix 7-26
-------
Terrestrial Nutrient Enrichment Case Study
monitored data available. Scientists at the NADP have created continuous surfaces of deposition
values by using interpolation algorithms to estimate values between measurements at known
locations. Dry nitrogen deposition can be estimated using the output from the CMAQ 2002
modeling system. This model produced estimates of many nitrogen species aggregated to 12-
kilometer (km) squares. Although these data are fairly coarse spatially, they are the best that are
currently available.
2.2.2.2 Range of Coastal Sage Scrub
Several publicly available spatial vegetation datasets were examined for this analysis.
They range in dates from 1945 to 2002 and are compiled from a combination of field data and
remotely-sensed imagery.
The Wieslander VTM (USFS, 2008b) collection is a dataset published in 1945 that
brought together data recorded on photos, species inventories, plots maps, and vegetation maps.
This dataset was obtained from the California Spatial Information Library at the University of
California, Davis. It divided the entire state of California into polygons that were attributed with
23 different vegetation types (i.e., communities) such as "coastal sagebrush" or "chaparral."
Individual species were not recorded.
Another vegetation layer named CALVEG (California Vegetation) was created in 1977
from LANDS AT imagery that was used to create 1:1,000,000 scale maps. This dataset is
available from the California Department of Forestry and Fire Protection's Fire and Resource
Assessment Program (FRAP) Web site (California Department of Forestry and Fire Protection,
2007a). This dataset contains land cover/land use polygons for California digitized from the
1:1,000,000 scale maps. The minimum mapping unit is approximately 400 acres, and the data
contains vegetation attributes for series-level species groups only.
A land cover change dataset is also available from the FRAP Web site that uses Thematic
Mapper (TM) data from 1993 and 1997 to determine areas of change (California Department of
Forestry and Fire Protection, 2007b). This dataset also contains information on the cause of the
change. The spatial resolution of this dataset is 30-meter pixels. These data do not contain
species-level data; these contain only community-level data.
California Gap Analysis Project (GAP) data is available from the University of
California, Santa Barbara Biogeography Lab (Davis et al., 1998). It contains vegetation attributes
Final Risk and Exposure Assessment September 2009
Appendix 7-27
-------
Terrestrial Nutrient Enrichment Case Study
for landscape scale map units, including canopy dominant species, canopy density, presence of
regional endemic species, and inclusion of wetland habitats. These data were published in 1998
and used a variety of sources including TM data, aerial photography, Wieslander VTM data, and
field maps.
The most recent land cover data for the state of California is also available from the
FRAP website. It was published in 2002 and was created by compiling the best available land
cover data into a single data layer. This agency classified California's vegetation into 59
different categories, including CSS, at a spatial resolution of 100 m. Decision rules were
developed that controlled which layers were given priority in areas of overlap. Cross-walks were
used to compile the various sources into the common California Wildlife Habitat Relationships
system classification. No species specific data are available.
One of the central analytical tasks for this case study was to quantify the amount of CSS
and MCF extent loss and to see if loss corresponded spatially to areas of high nitrogen
deposition, fire threat, or both. The land cover change layer created by the California Department
of Forestry and Fire Protection was used for this case study analysis. While the temporal
difference for this layer depicting land cover change was fairly small (i.e., 5 years), the two
datasets used to create the change layer were fully compatible, and the results were verified by
field confirmation.
The feasibility of using the 1945 VTM, the 1977 CALVEG, the 1998 GAP, and the 2002
FRAP land cover data to determine changes in the extent of CSS and MCF ecosystems was
evaluated. In each case, the data sources, spatial resolution, and classification schemes were
different enough to prevent any meaningful measurement of change in these communities.
In addition to publicly available datasets, research datasets were obtained and plotted.
Field data were obtained directly from Talluto and Suding (2008). While these data provided
very detailed measurements of species distribution and percentage ground cover for their study
areas, they were not sufficiently spatially dispersed across the CSS range, nor were they
compatible with the very spatially coarse (i.e., 12-km grid size) CMAQ-modeled nitrogen
deposition data.
Additionally, The Kuchler Potential Natural Vegetation (PNV) Groups (Kuchler, 1988)
data layer that was created to show "climax" vegetation was not used because the intent of this
case study was to quantify known changes to the extent of CSS. The PNV data illustrates where
Final Risk and Exposure Assessment September 2009
Appendix 7-28
-------
Terrestrial Nutrient Enrichment Case Study
the vegetation might potentially be found without disturbance or climate change. PNV is an
expression of environmental factors, such as topography, soils, and climate across an area.
2.2.2.3 Fire Threat
The California Department of Forestry and Fire Protection's FRAP also compiles data
about fire threat. These data consider fire rotation (i.e., how frequently fire occurs) and potential
fire behavior, which take into account topography and potential vegetative fuels. Fire threat is
classified into four unique categories that range from moderate to extreme.
2.2.2.4 Changes in Coastal Sage Scrub Communities
Although spatial datasets mapping CSS communities exist for 1945, 1977, 1998, and
2002, none are compatible enough to calculate meaningful change (e.g., the methods used to
ascertain CSS extent and define ecosystems were not consistent across datasets). Therefore, a
spatial dataset published by the California Department of Forestry and Fire Protection was
chosen. This dataset documented change to CSS and other ecosystems between 1993 and 1997.
2.2.2.5 Distribution of Invasive Species
Two data sources for invasive species were found for California. The first is the PLANTS
program, which is part of the U.S. Department of Agriculture (USDA) (USD A, 2009;
http://plants.usda.gov/index.html). This resource posts maps that indicate whether a species is
present or not in a given county, but not the distribution of that species within the county. The
second is the California Invasive Plant Council (2008), (http://www.cal-
ipc.org/ip/mapping/statewide_maps/index.php), which lists the relative abundance by county of a
select number of species.
2.2.2.6 Threatened and Endangered Species Habitat
The U.S. Fish and Wildlife Service (FWS) publishes critical habitat range information for
threatened and endangered species by state, county, and species through the Critical Habitat
Portal (http://crithab.fws.gov/) (U.S. FWS, 2008). For example, the Critical Habitat Portal
locates 16 species for Riverside County, 5 of which are associated with CSS habitat.
Final Risk and Exposure Assessment September 2009
Appendix 7-29
-------
Terrestrial Nutrient Enrichment Case Study
2.2.2.7 Range of Mixed Conifer Forest
The most recent (2002) land cover dataset from the California Department of Forestry
and Fire Protection's FRAP Web site was used to extract the range of MCF. Research data were
also obtained from a series of sample plot locations documented in Fenn et al. (2008). The
locations of the field sites were listed as latitude and longitude coordinates, which were
converted into a GIS layer with atmospheric nitrogen deposition as an attribute.
2.2.2.8 Distribution of Acid-Sensitive Lichens
• The USFS FIA datasets were the source of lichen distributions.
3.0 RESULTS
Effects of elevated atmospheric nitrogen deposition on the CSS and MCF ecosystems are
the result of increased long-term chronic, rather than short-term pulsed, nitrogen deposition. It is
difficult to quantify effects in both ecosystems because of confounding stressors, such as fire and
Os. The literature available on long-term research and application of robust models on these
ecosystems is extremely limited.
The CSS analysis relies upon peer-reviewed literature and spatial analyses to derive
major conclusions regarding the effects of nitrogen. Spatial analyses were used to determine the
changes in the extent of CSS ecosystems and their associated habitat, as well as to investigate the
effects of nitrogen and fire, another driving component in the alteration of the CSS ecosystem.
The reviewed literature includes greenhouse experiments, field observations, and field
manipulation experiments that document the observed and measured effects of nitrogen.
The MCF analysis also contains a summary of the peer-review literature; however, this
case study focused on the empirical loading benchmarks derived from an analysis by Fenn et al.
(2008), which employed observational data and the Simple Mass Balance (8MB) model and the
DayCent simulation model to estimate critical loads. However, there are identified limitations to
both models (e.g., 8MB does not account for the effects of prescribed burns or wildfires on
nutrient uptake, and DayCent is not a watershed-scale model, and thus, does not accurately
represent N(V concentrations in surface and groundwater). Fenn et al. (2008) conclude that the
empirical approach is the most reliable source of information.
Final Risk and Exposure Assessment September 2009
Appendix 7-30
-------
Terrestrial Nutrient Enrichment Case Study
3.1 LITERATURE REVIEW FINDINGS
3.1.1 Coastal Sage Scrub
CSS is subject to several pressures, such as land conversion, grazing, fire, and pollution,
all of which have been observed to induce declines in other ecosystems (Allen et al., 1998). At
one extreme, development pressure (i.e., the conversion of CSS to residential and commercial
land uses) will simply eliminate acres of CSS. Other pressures will come into play in modifying
the remaining habitat. Research suggests that both fire and increased atmospheric nitrogen
deposition can enhance the growth of nonnative grasses in established CSS ecosystems.
Additionally, CSS declines have been observed when fire frequency is held constant and/or
nitrogen is held constant, suggesting that both fire and nitrogen play a role in CSS decline when
direct destructive factors are not an imminent threat. Table 3.1-1 contains a summary of selected
experimental variables across multiple CSS study areas.
Final Risk and Exposure Assessment September 2009
Appendix 7-31
-------
Terrestrial Nutrient Enrichment Case Study
Table 3.1-1. Summary of Selected Experimental Variables across Multiple Coastal Sage Scrub Study Areasa
Study Locations
Riverside-Ferris Plainb
Santa Margarita Ecological
Reserve
Santa Monica Mountains
Orange Countyb
Rancho Jamul Ecological
Reserve
Voorhis Ecological Reserve
Riverside-Ferris Plainb
Sedgwick Ranch Natural
Reserve
Southern California fuel
breaksb
Critical reviewb
Southern California burn
sitesb
Riverside-Ferris Plainb
Greenhouse experiment
Riverside-Ferris Plainb
University of California-
Riverside Agricultural
Research Station
Riverside-Ferris Plainb
Soil
Nitrogen
X
X
X
X
X
X
X
Atmospheric
Nitrogen
X
X
Vegetation
Change
X
X
X
X
X
X
X
X
X
X
X
X
Mycorrhizae
Change
X
X
Fire
Cycle
X
X
X
X
Author
Allen etal., 1998
Burger et al., 2003
Carrington and Keeley, 1999
Diffendorfer et al., 2007
Drus, 2004
Egerton-Warburton and Allen,
2000
Fierer and Gabet, 2002
Merriam et al., 2006
Keeley, 2001
Keeley et al., 2005
Minnich and Dezanni, 1998
Padgett et al., 1999
Padgett and Allen, 1999
Padgett et al., 2000
Siguenza et al., 2006
Final Risk and Exposure Assessment
Appendix 7-32
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Study Locations
Riverside-Ferris Plainb
Lake Skinner
Riverside-Ferris Plainb
67 sites across CSS rangeb
Riverside-Ferris Plainb
Lake Skinner Western
Riverside County Multi-
Species Reserve
Greenhouse experiment
Soil
Nitrogen
X
X
X
X
Atmospheric
Nitrogen
Vegetation
Change
X
X
X
Mycorrhizae
Change
X
X
X
Fire
Cycle
Author
Sirulnik et al., 2007a
Sirulnik et al., 2007b
Vourlitis et al., 2007
Westman, 1979, 1981a,b
Wood et al., 2006
Yoshida and Allen, 2001
a Empty cells indicate not studies identified.
b Multiple data sites within the study location.
Final Risk and Exposure Assessment
Appendix 7-33
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
3.1.1.1 Atmospheric Nitrogen Deposition
Increased atmospheric nitrogen deposition has been observed to alter vegetation types
when nitrogen is a limiting nutrient to growth. This has been observed in alpine plant
communities in the Colorado Front Range, as well as in lichen communities in the western Sierra
Nevada region (Fenn et al., 2003, 2008); however, in the case of CSS, it is hypothesized that
many stands are no longer limited by nitrogen and have instead become nitrogen-saturated due to
atmospheric nitrogen deposition (Allen et al., 1998; Westman, 1981a). This is supported by the
positive correlation between atmospheric nitrogen and soil nitrogen, increased long-term
mortality of CSS shrubs, and increased nitrogen-cycling rates in soil and litter and soil fertility
(Allen et al., 1998; Padgett et al., 1999; Sirulnik et al., 2007a; Vourlitis et al., 2007). Figure
3.1-1 illustrates the levels of atmospheric nitrogen deposition on CSS ecosystems using 2002
CMAQ/NADP data.
Wood et al. (2006) investigated the amount of nitrogen used by healthy and degraded
CSS ecosystems. In healthy stands, the authors estimated that 3.3 kg N/ha/yr was used for CSS
plant growth (Wood et al., 2006). It is assumed that 3.3 kg N/ha/yr is near the point where
nitrogen is no longer limiting in CSS. Therefore, this amount can be considered an ecological
benchmark for CSS. Figure 3.1-1 displays the spatial extent of CSS where nitrogen deposition is
above the ecological benchmark of 3.3 kg N/ha/yr. As shown in Figure 3.1-1 and Table 3.1-2,
almost all of CSS receive >3.3 kg N/ha/yr of nitrogen through atmospheric deposition. Note that
CSS is observed in areas receiving >3.3 kg N/ha/yr. This distribution may result from time lags
(i.e., years may be required for the CSS ecosystem to completely disappear), or it may indicate
that 3.3 kg nitrogen, although ecologically meaningful, may not be the benchmark value.
Final Risk and Exposure Assessment September 2009
Appendix 7-34
-------
Terrestrial Nutrient Enrichment Case Study
:
| | Counties
^^| Coastal Sage Scrub
Total N Deposition
kg/ha/yr
less than 3.3
33-99
10 or greater
\,
Source of CSS range is the California Department
of Forestry and Fire Protection.
Figure 3.1-1. Coastal sage scrub range and total nitrogen deposition using CMAQ
2002 modeling results and NADP monitoring data.
Final Risk and Exposure Assessment
Appendix 7-35
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Table 3.1-2. Coastal Sage Scrub Ecosystem Area and Nitrogen Deposition
N Deposition
(kg/ha/yr)
>3.3
>10
Area (hectares)
654,048
138,019
Percent of
CSS Area, %
93.51
19.73
3.1.1.2 Normative Grasses
The ecological effects of increased nitrogen are most easily explained by considering the
seasonal stages of a semiarid Mediterranean ecosystem. In the rainy, winter season, deposited
surface nitrogen is transported deeper into the soil and is rapidly mineralized by microbes, thus
making it available for plants. Faster nitrogen availability may favor the germination and growth
of nitrophylous colonizers, more specifically nonnative grasses (e.g., Bromus madritensis., Avena
fatua, and Hirschfeldia incand). This earlier flourishing of grasses can create a dense network of
shallow roots, which slows the diffusion of water through soil, decreases the percolation depth of
precipitation, and decreases the amount of water for soil and groundwater recharge (Wood et al.,
2006). Growth of CSS species, such as Artemisia californica, Eriogonumfasciculatum, and
Encelia farinose, may be decreased because of decreased water and nitrogen availability at the
deeper soil layers where more woody CSS tap roots are found (Keeler-Wolf, 1995; Wood et al.,
2006). Furthermore, an increased percentage of shrub species is established during wet years,
suggesting that percolation of nutrient-carrying water may be limited in years with average or
below average precipitation (Keeley et al., 2005).
3.1.1.3 Mycorrhizae
Elevated nitrogen may also play a role in altering the nutrient uptake of CSS plants by
decreasing the species richness and abundance of mutualistic fungal communities, such as
arbuscular mycorrhizae (AM) (Egerton-Warburton and Allen, 2000; Siguenza et al., 2006).
Although both CSS and nonnative grass species have AM and other mycorrhizal associations,
which increase the surface area and capacity for nutrient uptake, CSS is predominantly colonized
by a coarse AM species, and nonnative grasses are more likely mutualistic with finer AM
species. In the presence of elevated nitrogen, coarse AM colonizations were depressed in number
and volume. Egerton-Warburton and Allen (2000) documented shifts in AM species as well as
declines in spore abundance and colonization at approximately 10 kg N/ha/yr. In areas with the
highest levels of soil nitrogen tested (e.g., 57 micrograms per gram (ug/g) average annual soil
Final Risk and Exposure Assessment
Appendix 7-36
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
nitrogen present in Jurupa Hills, Riverside County), a shift in the timing of AM growth was also
observed. Therefore, it is suggested that these diminished mutualistic associations may
contribute to a decline in the overall health of CSS via a loss in nutrient uptake capacity and may
represent an ecological endpoint for the CSS ecosystem. Figure 3.1-1 displays the levels of
atmospheric nitrogen deposition on CSS ecosystems above the ecological benchmark of 10 kg
N/ha/yr using 2002 CMAQ/NADP data. The 12-km resolution CMAQ/NADP data indicate that
CSS within the Los Angeles and San Diego airsheds are likely to experience the noted effects at
the 10 kg N/ha/yr ecological benchmark.
3.1.1.4 Soil Nitrogen
In a greenhouse fertilization experiment, soil nitrogen levels of 50 ug/g ammonium N(V
had a 100% mortality rate after 9 months of continuous growth. The plants began to senesce at
approximately 6 months, whereas all lower-exposure individuals were still healthy and remained
healthy for more than 1 year (Allen et al., 1998). In the field, seasonal changes do not allow for
12 months of uninterrupted growth; therefore, the increased mortality shown in this study may be
realized over much longer periods of time in situ. Additionally, studies have suggested that soil
nitrogen may now be increasing because of soil fertility in conjunction with atmospheric
deposition, so that the soil itself becomes an intrinsic source (Padgett et al., 1999). In
combination with decreased establishment and the capacity for nutrient uptake, these responses
to elevated nitrogen levels may represent a detrimental and long-term pressure on CSS at varying
levels of nitrogen additions. Table 3.1-3 summarizes the various ecosystem responses to
nitrogen levels that affect CSS communities.
Table 3.1-3. Research Evidence of Ecosystem Responses to Nitrogen Relevant to
Coastal Sage Scrub
Environmental Impact
Enhanced growth of
nonnative species
Nutrient enrichment of
soil and plants
Decreased growth regulation
of shrubs
Decreased diversity of
mycorrhizal communities
Location
Southern California
Riverside-Perris Plain,
San Diego County
Greenhouse experiment
Riverside-Perris Plain
Reference
Minnich and Dezanni, 1998; Allen et al.,
1998; Weiss, 2006; Westman, 1981a,b
Sirulnik et al., 2007a; Allen et al., 1998;
Padgett et al., 1999; Vourlitis et al., 2007
Padgett and Allen, 1999
Egerton-Warburton and Allen, 2000;
Siguenza et al., 2006
Final Risk and Exposure Assessment
Appendix 7-37
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Environmental Impact
Increased runoff and nutrient
loss
Altered fire cycle
Increased dependent species
vulnerability
Increased erosion
Location
Santa Barbara
Riverside-Ferris Plain
All CSS; San Diego
County
California shrublands
Reference
Fierer and Gabet, 2002
Wood et al., 2006
Weiss, 2006; Weaver, 1998
Keeler-Wolf, 1995
3.1.2 Fire
Fire is also an inextricable and significant component in CSS losses. Although CSS
species are fire resilient, nonnative grass seeds are quick to establish in burned lands, reducing
the water and nutrient amounts available to CSS species for reestablishment (Keeler-Wolf,
1995). Additionally, when nonnative annual grasses have established dominance, these species
alter and increase the fire frequency by senescing earlier in the annual season and increasing the
dry, ignitable fuel availability (Keeley et al., 2005). With increased fire frequencies and faster
nonnative colonizations, CSS seed banks are eventually eradicated from the soil, and the
probability of re-establishment decreases significantly (Keeley et al., 2005). Figure 3.1-2
represents the fire threats to CSS ecosystems.
Final Risk and Exposure Assessment
Appendix 7-38
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
F/eSno
*
Oxnard XL
'l.os Angelti
Riverside
* cities
HI Coastal Sage Scrub 2002
Fire Threat
3 Moderate
| | High
I Extreme
*£$&
Source o( CSS range and fire mreat is ihe
California Departmerrt
of Forestry and Fire Protection
Figure 3.1-2. Current fire threats to coastal sage scrub ecosystems.
3.1.3 Coastal Sage Scrub Model
It appears that both atmospheric nitrogen deposition and fire are critical factors involved
in the decline of CSS. Figure 3.1-3 presents a model of CSS ecosystem response to nitrogen and
fire. Note that the model indicates that both nitrogen and fire play critical roles, and that there
may be positive feedback loops and possible synergies between fire and nitrogen loadings.
Final Risk and Exposure Assessment
Appendix 7-39
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Atmospheric
Nitrogen
Modified Fire
Cycle
-
^ Coastal
Sage Scrub
\
f Non-native
Grasses
h
\
Mycorrhizae
Associations
Modified
Nutrient and
Water
Retention
Figure 3.1-3. Model of coastal sage scrub ecosystem in relation to fire and
atmospheric nitrogen deposition.
3.1.4 Mixed Conifer Forest Ecosystems
The MCF ecosystem has been a subject of study for many years. There are a number of
important stressors on the community, including atmospheric fire, bark beetles, Os, particulates,
and nitrogen. Although fire suppression in the 20th century is probably the most significant
change that has led to alterations in morphology and perhaps to shifts in forest composition
(Minnich et al., 1995), stress from elevated levels of ambient atmospheric nitrogen
concentrations is the subject of increasing research.
3.1.4.1 Nitrogen and Ozone Effects
Measurements documenting increases in atmospheric nitrogen deposition have been
recorded with some regularity since the 1980s (Bytnerowicz and Fenn, 1996); however, the Los
Angeles area has seen elevated ambient atmospheric nitrogen concentrations for the last 50 years
(Bytnerowicz and Fenn, 1996). Also, some data have been published for the primary nitrogen
species of dry atmospheric nitrogen deposition in the San Bernardino Mountains (i.e., nitric acid
Final Risk and Exposure Assessment
Appendix 7-40
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
and ammonia gas [NH3]) from passive samplers (Bytnerowicz et al., 2007). The
pressures exerted on MCF ecosystems in California form a gradient across the Sierra Nevada
Range and San Bernardino Mountains. Nitrogen throughfall levels in the northern Sierra Nevada
Range are as low as 1.4 kg N/ha/yr, whereas forests in the western San Bernardino Mountains
experience measured throughfall nitrogen levels up to 33 to 71 kg N/ha/yr. (Note that the high
levels of nitrogen seen in some measured throughfall values are not reflected in the CMAQ
modeled results. This may be an artifact of using a 12-km grid.) The primary source of nitrogen
in the western San Bernardino Mountains stems from fossil fuels combustion, such as vehicle
exhaust. Other sources, such as agricultural processes, also play a prominent role in the western
portions of the San Bernardino Mountains and Sierra Nevada Range (Grulke et al., 2008).
Figure 3.1-4 illustrates the current total atmospheric nitrogen deposition on MCF in California.
At the individual tree level, elevated atmospheric nitrogen can shift the ratio of
aboveground to belowground biomass. Elevated pollution levels may result in increased uptake
of nutrients via the canopy, decreased nitrogen intake requirements on root structures, and
increased demand for carbon dioxide (CC^) uptake and photosynthetic structures to maintain the
carbon balances. Therefore, the increased nutrient availability stimulates aboveground growth
and increases foliar production, while reducing the demand for belowground nutrient uptake
(Fenn et al., 2000). Carbon allocation gradually shifts from root to shoot, and fine-root biomass
is decreased (Fenn and Bytnerowicz, 1997; U.S. EPA, 2008, Section 3.3). Grulke et al. (1998)
observed a 6- to 14-fold increase in fine-root mass in areas of low atmospheric nitrogen
deposition compared to areas of high deposition. Medium roots also declined at high levels
(Fenn et al., 2008).
At the stand level, elevated atmospheric nitrogen has been associated with increased
stand density, although other factors, such as fire suppression, also contribute to increased
density and can increase mortality rates (U.S. EPA, 2008, Section 3.3). As older trees die, they
are replaced with younger, smaller trees. Smaller trees allow more sunlight through the canopy
and, combined with an increased availability of nitrogen, may allow for more trees to be
established. Increased stand densities with younger-age classes are observed in the San
Bernardino Mountains, where air pollution levels are among the highest found in the California
MCF ranges studied (Minnich et al., 1995; Fenn et al., 2008). These shifts in stand density and
age distribution result in vegetation structure shifts which, in turn, may impact population and
Final Risk and Exposure Assessment September 2009
Appendix 7-41
-------
Terrestrial Nutrient Enrichment Case Study
community dynamics of understory plants and animals, including threatened and endangered
species.
| | Sierra Nevada
| | San Bernardino NF
| Mixed Conifer
CMAQ
kg/ha/yr
| | less than 3.1
[ | 3.1 -5.1
5.2- 10.1
10.2- 169
17 or greater
.'-. #
Yosemite National Park
jr Kings Canyon National Park
Sequoia National Park
® I
Source of Mixed Conifer" California Fire and
Resource Assessment Program
Figure 3.1-4. Mixed conifer forest range and total atmospheric nitrogen
deposition using CMAQ 2002 modeling results and NADP monitoring data.
Final Risk and Exposure Assessment
Appendix 7-42
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
It should be noted that the effects of 63 and atmospheric nitrogen are difficult to separate.
The atmospheric transformation of NOX can yield moderate concentrations of 63 as a byproduct
(Grulke et al., 2008). Therefore, since elevated nitrogen levels are generally correlated with O3
concentrations, researchers often report changes in tree growth and vigor as being the result of
both (i.e., Grulke and Balduman, 1999).
High concentrations of Os and atmospheric nitrogen can generate increased needle and
branch turnover. In areas subjected to low pollution, conifers may retain needles across 4 or 5
years; however, in areas of high pollution, such as Camp Paivika in the San Bernardino
Mountains, needle retention is generally less than 1 year (Grulke and Balduman, 1999; Grulke et
al, 2008). Needle turnover significantly increases litterfall. Litter biomass has been observed to
increase in areas with elevated atmospheric nitrogen deposition up to 15 times more than in areas
with low deposition, and the litter is seen to have higher concentrations of nitrogen (Fenn et al.,
2000; Grulke et al., 2008). Elevated nitrogen levels in litter may facilitate faster rates of
microbial decomposition initially, but over the long term, high nitrogen levels slow litter
decomposition, and litter accumulates on the forest floor (Grulke et al., 2008; U.S. EPA, 2008).
The increased litter depth may then affect subcanopy growth and stand regeneration over long
periods of time.
At the highest levels of nitrogen deposition, native understory species were seen to
decline (Allen et al., 2007). In addition to this decline in native understory diversity, changes in
decreased fine-root mass, increased needle turnover, and the associated chemostructural
alterations, MCF that are exposed to elevated pollutant levels have an increasing susceptibility to
drought and beetle attack (Grulke et al., 1998, 2001; Takemoto et al., 2001). These stressors
often result in the death of trees, producing an increased risk of wildfires. This complex model is
displayed in Figure 3.1-5 as a graphic developed by Grulke et al. (2008).
Final Risk and Exposure Assessment September 2009
Appendix 7-43
-------
Terrestrial Nutrient Enrichment Case Study
Rapid population increase,
Change In land use
Increased Os & N
Periodic drought
Increased tree
demffieatkm
Increased fee
susceptibility to
drought straw
of bark beetle,
Increased susceptibility
to fire
Increased fire starts.
Continued fire suppression
= WILDFIRE
Figure 3.1-5. Conceptual model for increased susceptibility to wildfire in mixed
conifer forests (Grulke et al., 2008).
3.1.4.2 Nitrogen Effects on Lichens
Lichens emerged as an indicator of nutrient enrichment from the research on the effects
of acid rain. Lichen species can be sensitive to air pollution; in particular, atmospheric nitrogen
deposition. Since the 1980s, information about lichen communities has been gathered, and
lichens have been used as indicators to detect changes in forest communities. Jovan (2008)
depicts how lichens might be considered as sentinels in the MCF community (Figure 3.1-6).
LICHEN
COMMUNITY
CAUSE-
EFFECT
INDICATES
CONDITION OF
RESOURCE
f'Orcsi |>rociucti\ i£s.
biodKcrsitx. health
VINDICATES
''CAUSE-
EFFECT
ENVIRONMENTAL
STRKSSORS
N- and S-bascd air pollutants:
duvet toxiciiy and acidifying and
fcnili?ing affects
Figure 3.1-6. Importance of lichens as an indicator of ecosystem health (Jovan, 2008).
Final Risk and Exposure Assessment
Appendix 7-44
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
As atmospheric nitrogen deposition increases, the relative abundance of acidophytic
lichens decreases, and the concentration of nitrogen in one of those species, Letharia vulpine.,
increases (Fenn et al., 2008). Fenn et al. (2008) were able to quantify the change in the lichen
community, noting that for every 1 kg N/ha/yr increase, the abundance of acidophytic lichens
declined by 5.6%. Figure 3.1-7 illustrates the presence of acidophyte lichens and the total
atmospheric nitrogen deposition in the California ranges.
In addition to abundance changes, species richness, cover, and health are affected in areas
of high Os and nitrogen concentrations. Fifty percent fewer lichen species were observed after 60
years of elevated air pollution in San Bernardino Mountains MCF, with the areas of highest
pollution levels exhibiting low species richness, decreased abundance and cover, and
morphological deterioration of existing lichens (Sigal and Nash, 1983).
Ecological endpoints relating to shifts in the abundance of acidophilic lichens were
identified by Fenn et al. (2008). They found that at 3.1 kg N/ha/yr, the community of lichens
begins to change from acidophilus to tolerant species; at 5.2 kg N/ha/yr, the typical dominance
by acidophilus species no longer occurs; and at 10.2 kg N/ha/yr, acidophilic lichens are totally
lost from the community. Additional studies in the Colorado Front Range of the Rocky Mountain
National Park support these findings and are summarized in Chapter 5 of the Risk and Exposure
Assessment. These three values are one set of ecologically meaningful benchmarks for the MCF.
As shown in Figure 3.1-7, much of the MCF receives nitrogen deposition levels above the 3.1
kg N/ha/yr ecological benchmark according to the 2002 CMAQ/NADP data, with the exception
of the easternmost Sierra Nevada Range. MCF in the southern portion of the Sierra Nevada
forests and nearly all MCF communities in the San Bernardino forests receive nitrogen
deposition levels above the 5.2 kg N/ha/yr ecological benchmark. Figure 3.1-7 also displays the
potential areas where acidophilic lichens are extirpated due to nitrogen deposition levels >10.2
kg N kg/ha/yr.
Final Risk and Exposure Assessment September 2009
Appendix 7-45
-------
Terrestrial Nutrient Enrichment Case Study
• Acidophyte Lichens
^ San Bernardino NF
Sierra Nevada
CMAQ Total N Dep
kg/ha/yr
^] less than 3.1
I 13-1-5.1
[ ] 5.2- 10.1
| 10 2 or greater
SsnFn
Figure 3.1-7. Presence of acidophyte lichens and total nitrogen deposition in the
California mountain ranges using CMAQ 2002 modeling results and NADP
monitoring data.
3.1.4.3 Nitrogen Saturation
The established signs of nitrogen saturation have been shown within the MCF ecosystem.
These symptoms include the following:
• Increased carbon and nitrogen cycling. The foliar turnover rates and changes in
microbial decomposition both suggest that carbon and nitrogen cycles have been altered
as a result of elevated nitrogen. Additionally, nitrogen fluxes in San Bernardino
Final Risk and Exposure Assessment
Appendix 7-46
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Mountains soils are elevated when compared to MCF in the northern Sierra Nevada
Range (Bytnerowicz and Fenn, 1996).
• Decreased nitrogen uptake efficiency of plants. Changes in root: shoot ratio
demonstrate structural alterations in response to increasing available nitrogen.
• Increased loss of forest nitrates to streamwater (i.e., NOs leachate). Elevated N(V
leachate levels are estimated to have begun in the late 1950s and have been observed
from the western MCF in the San Bernardino Mountains since 1979 (Fenn et al., 2008).
These losses are a result of high soil nitrogen driven by the combined litter, needle
turnover, and throughfall nitrogen exerted in these areas (Bytnerowicz and Fenn, 1996).
Changes in root biomass and stream leachate, in addition to lichen species compositional
shifts, have been used to develop benchmarks for nitrogen benchmarks in the MCF ecosystem.
These critical loading benchmarks, or empirical loads, are designed to estimate the levels at
which atmospheric nitrogen concentrations and subsequent deposition begin to affect selected
components of the ecosystem, such as forest growth, health, and composition. Some benchmarks
aim to estimate individual changes to an ecosystem, whereas others assess the levels at which the
entire ecosystem will not be altered because of atmospheric nitrogen deposition. The possibility
of using the MCF as a model for benchmarking is discussed below.
Fenn et al. (2008) established a critical loading benchmark of 17 kg throughfall N/ha/yr
in the San Bernardino Mountains and Sierra Nevada Range MCF ecosystems. This benchmark
represents the level of atmospheric nitrogen deposition at which elevated concentrations of
streamwater N(V leachate or potential nitrogen saturation may occur. At this deposition level, a
26% reduction in fine-root biomass is anticipated (Fenn et al., 2008). Rootshoot ratios are,
therefore, altered, and changes in nitrogen uptake efficiencies, litterfall biomass, and microbial
decomposition are anticipated to be present at this atmospheric nitrogen deposition level. This
benchmark is based on 30 to 60 years of exposure to elevated atmospheric concentrations. At
longer exposure levels, the benchmark is lower because of decreased nitrogen efficiencies of the
ecosystem. This benchmark is exceeded in areas of the western San Bernardino Mountains, such
as Camp Paivika.
MV leaching is a symptom that an ecosystem is saturated by nitrogen. NCV leaching is
also known to cause acidification in adjacent surface waters. The ecological benchmark of 17 kg
Final Risk and Exposure Assessment September 2009
Appendix 7-47
-------
Terrestrial Nutrient Enrichment Case Study
N/ha/yr is the last benchmark identified in this study. At this level of atmospheric nitrogen
deposition, N(V is observed in streams in the MCF (Fenn et al., 2008), denoting a change in
ecosystem function.
Table 3.1-4 displays the area in hectares of MCF experiencing different nitrogen
deposition levels.
Table 3.1-4. Mixed Conifer Forest Ecosystem Area and Nitrogen Deposition
N Deposition
(kg/ha/yr)
>3.1
>5.2
>10.2
>17
Area
(hectares)
1,099,133
130,538
11,963
0
Percent of MCF
Area, %
38.62
4.59
0.42
0.00
3.2 RESULTS SUMMARY
A range of ecological benchmarks were developed in the results. All benchmarks are tied
to a level of atmospheric nitrogen deposition, but include a number of different ecological
processes and ecological endpoints. All of the benchmarks are ecologically significant in that
changes are seen that are related to community structure and function. The benchmarks span a
range from 3.1 to 17 kg N/ha/yr and include:
• 3.1 kg N/ha/yr—shift in lichen communities are first observed in MCF
• 3.3 kg N/ha/yr—nitrogen no longer limiting in CSS
• 5.2 kg N/ha/yr—dominance of tolerant lichen species in MCF
• 10 kg N/ha/yr—AM community shift in C S S
• 10.2 kg N/ha/yr—loss of acidophilic lichen species from MCF
• 17 kg N/ha/yr—NO3" leaching in MCF.
This range of ecological benchmarks may be used to develop a "green line/red line"
schematic, similar to the forest screening model discussed in Lovett and Tear (2007) that
illustrates the levels at which ecosystem effects may occur or are known to occur. In Figure
3.2-1, the green area/line denotes that point at which there do not appear to be any effects and the
red line the point at which known negative effects occur.
Final Risk and Exposure Assessment
Appendix 7-48
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Nitrogen Deposition (kg/ha/yr)
-^ to
o o
0
High Probability of Negative Effects
— -^
— Moderate Probability of Negative Effects
~~ '-""""'- —
~~ Low Probability of Negative Effects
,---l Nitrogen Leaching to Streams
^^ Loss of Acidophyte Lichen
Shift in AM Community in CSS
^-^ Decline in Acidophyte Lichen
Healthy CSS Community
Figure 3.2-1. Illustration of the range of terrestrial ecosystem effects observed
relative to atmospheric nitrogen deposition.
For the benchmarks identified, effects may occur at the level of atmospheric nitrogen
deposition associated with the "green line" illustrated in Figure 3.2-1, so the "green line" may be
somewhat lower. Whereas, the higher level of atmospheric nitrogen deposition (i.e., both at 10.2
and 17 kg N/ha/yr) better resembles a "red line," where a known negative effect occurs.
The range of ecological benchmarks in CSS and MCF are not dissimilar from ecological
endpoints and benchmarks identified in other ecosystems with related characteristics, such as
arid systems, other forested systems, or grasslands. Egerton-Warburton et al. (2001) report that at
10 kg N/ha/yr, nitrogen changes in mycorrhizal communities/grass biomass are observed in
chaparral habitats. Nitrates are found to leach into streams from nitrogen-saturated forest soils at
deposition levels between 9 and 13 kg N/ha/yr (Aber et al., 2003). Results from several studies
suggest ecosystem changes that are related to nitrogen deposition. The capacity of alpine
catchments to sequester nitrogen is exceeded at input levels <10 kg N/ha/yr (Baron et al., 1994).
Changes in the Carex plant community were observed to occur at deposition levels near 10 kg
N/ha/yr (Bowman et al., 2006). Clark and Tilman (2008) predict that at 5.3 kg N/ha/yr, there is a
loss of species diversity in grasslands. In the Pacific Northwest and in Central California, a
number of investigators have observed declines in sensitive lichen species as air pollution
Final Risk and Exposure Assessment
Appendix 7-49
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
increases (Jovan and McCune, 2005; Geiser and Neitlich, 2007). In Europe, acidophyte decline
has been identified in regions with 8 to 10 kg N/ha/yr (Bobbink, 1998; Bobbink et al., 1998).
4.0 IMPLICATIONS FOR OTHER SYSTEMS
This Terrestrial Nutrient Enrichment Case Study examined the effects of atmospheric
nitrogen on two ecosystem types in California: CSS and MCF. Figure 4.1-1 presents the
coverage of 2002 CMAQ/NADP data for total nitrogen deposition in the western United States,
including California. Ecological effects have been documented across the United States where
elevated nitrogen deposition has been observed. Benchmarks documented in the literature for the
negative effects on ecosystems are summarized in Figure 4.1-2 and are discussed in this case
study report. Looking across the United States, Figure 4.1-3 illustrates the occurrence of these
ecosystems which are sensitive to nitrogen and/or have similar characteristics to the ecosystems
explored in this case study. These ecosystems may also experience levels of atmospheric
nitrogen deposition that exceed the benchmark levels identified in Figure 4.1-2. Table 4.1-1 lists
the area of CSS and MCF that exceed benchmark nitrogen levels.
In the western United States, other arid and forested ecosystems exposed to deposition at
levels discussed in this case study may experience altered effects. As noted in Section 3, research
on grasslands and chaparral habitats is underway. These arid systems may respond to
benchmarks similar to those observed for CSS, as was shown by Clark and Tilman (2008) for
bluestem grasslands in Minnesota. N(V leaching in forests with elevated deposition (similar to
the range found in this study) may result in nutrient enrichment in streams which can affect
aquatic ecosystems (Aber et al., 2003). Research is also being conducted on lichen species in the
Pacific Northwest and in Central California that are exposed to elevated levels of atmospheric
nitrogen deposition (Jovan, 2008). Extensive research on the eastern Front Range of the Rocky
Mountain National Park has been conducted in alpine and subalpine terrestrial and aquatic
systems at elevations above 3,300 m, where communities are typically adapted to low nutrient
availability but are now being exposed to >10 kg N/ha/yr in some study areas (Baron et al. 2000;
Baron, 2006). (Chapters 5 and 6 of the Risk and Exposure Assessment also provide discussion on
this topic.)
Final Risk and Exposure Assessment September 2009
Appendix 7-50
-------
Terrestrial Nutrient Enrichment Case Study
Sierra Nevada
Range
ocky Mountain
National Park
Transverse
„
Range
Total N Deposition
kg/ha/yr
^| 0.8 to < 1.5
^B >= 1.5 to < 3
| | >= 3 to < 6
| | >= 6 to < 9
| | >= 9 to < 12
j^| >= 12to< 18
^H >= 18 to 20
Deposition data is the result of
combining CMAQ (dry) and
NADP (wet) over 12-km grid cells
Figure 4.1-1. 2002 CMAQ-modeled and NADP monitoring data for deposition of
total nitrogen in the western United States.
Final Risk and Exposure Assessment
Appendix 7-51
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Rocky Mountain alpine lakes: shift in diatom community dominance (Baron, 2006)
• Southern California: CSS loss (Wood et al., 2006)
• San Bernardino Mountains and Sierra Nevada Mountains: acidophytic lichen
decline in MCF (Fenn et al., 2008)
• Eastern Rocky Mountain Slope: low carbon:nitrogen; low lignin:nitrogen (Baron et
al., 2000)
• Eastern Rocky Mountain Slope: increased foliar nitrogen; increased mineralization
{Baronetal., 2000)
• San Bernardino Mountains and Sierra Nevada Mountains: shift from acidophytic
to neutral or nitrogen-tolerant lichen in MCF (Fenn et al., 2008)
• Minnesota grasslands: decreased plant species {Clark and Tilman, 2008)
• Northeast U.S.: NO3 leaching (Aber et al., 2003)
Bay Area, CA: Increased cover of nonnative grasses; decreased native
grasses (Weiss, 1999)
San Bernardino Mountains and Sierra Nevada Mountains: loss of acidophytic
lichen in MCF (Fenn et al., 2008)
Southern California: shift in mycorrhizal species in CSS (Egerton-Warburton
and Allen, 2000)
Southern California: shift from native species to invasive grasses in CSS (Allen,
2008)
• San Bernardino Mountains: high dissolved organic nitrogen (Meixner
and Fenn, 2004)
• San Bernardino Mountains: nitrogen saturation (Fenn et al., 2000)
• Increased nitrogen in lichen (Fenn et al., 2007)
• MCF: NO3 leaching (Fenn et al., 2008)
• MCF: 25% decrease in fine-root biomass (Fenn et al., 2008)
* Southern California: NO3~ leaching (Fenn et al., 2003)
• Southern California: high foliar nitrogen (Bytnerowicz and
Fenn, 1996)
* Los Angeles Basin, California: High NO emissions
(Bytnerowicz and Fenn, 1996)
Fraser Experimental Forest, CO:
increased foliar nitrogen; increased
mineralization (Rueth et al., 2003)
0246 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50
Nitrogen Deposition, kg/ha/yr
Figure 4.1-2. Benchmarks of atmospheric nitrogen deposition for several
ecosystem indicators with the inclusion of the diatom changes in the Rocky
Mountain lakes.
Final Risk and Exposure Assessment
Appendix 7-52
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Coverage Areas of Interest
Alpine Areas
^ PNW and CA Regions with Lichen
^H NE Forests
| Bluestem Grasslands
^B CA Mixed-Conifer Forest
Coastal Sage Scrub
| California Grassland OverSerpentinite Bedrock
I California Grasslands
Note: PNW = Pacific Northwest; NE = Northeast.
Figure 4.1-3. Habitats that may experience ecological benchmarks similar to
coastal sage scrub and mixed conifer forest.
Table 4.1-1. Areas of Coastal Sage Scrub and Mixed Conifer Forest That Exceed Benchmark
Nitrogen Deposition Levels.
CSS:
N Deposition (kg/ha/yr)
>3.3
>10
Area (hectares)
654,048
138,019
Percentage of CSS Area, %
93.51
19.73
MCF
N Deposition (kg/ha/yr)
>3.1
>5.2
>10.2
>17
Area (hectares)
1099,133
130,538
11,963
0
Percentage of MCF Area, %
38.62
4.59
0.42
0.00
Final Risk and Exposure Assessment
Appendix 7-53
September 2009
-------
Terrestrial Nutrient Enrichment Case Study
Other systems with the following characteristics may also be found to be sensitive:
• Ecosystems with nitrogen-sensitive epiphytes, such as lichens or mycorrhizae. Such
systems may demonstrate shifts in community structure through changes in nutrient
availability or modified provisioning services.
• Ecosystems that may have been exposed to long periods of elevated atmospheric
nitrogen deposition. The established signs of nitrogen saturation are increased leaching
of N(V into streamwater, decreased nitrogen uptake efficiency of plants, and increased
carbon and nitrogen cycling. At prolonged elevated nitrogen levels, ecosystems are
generally less likely to use, retain, or recycle nitrogen species efficiently at both the
species and community levels.
• Critical habitats. Ecosystems that are necessary for endemic species or special
ecosystem services should be monitored for possible changes due to nitrogen.
• Locations where there are seasonal releases of nitrogen. In both the California CSS
and MCF ecosystems discussed in this case study report, a large portion of nitrogen is
dry-deposited and remains on the foliage and soil surface until the beginning of the
winter rainy season when nitrogen will be flushed into the soil.
Current analysis of the effects of terrestrial nutrient enrichment from atmospheric
nitrogen deposition in both CSS and MCF seeks to improve scientific understanding of the
interactions among nitrogen deposition, fire events, and community dynamics. The available
scientific information is sufficient to identify ecological thresholds that are affected by nitrogen
deposition, and ecological thresholds have been identified for CSS and MCF. This case study
report has examined the sensitivity and effects of nutrient enrichment on terrestrial ecosystems,
and although a diverse array of U.S. ecosystems exist, exposure levels and thresholds for effects
appear to be generally comparable to levels identified in other sensitive U.S. ecosystems (e.g.,
thresholds range from 3.1 to 30.5 kg N/ha/yr), including thresholds identified from modeling
conducted for other case studies in the Risk and Exposure Assessment (Chapters 4 and 5).
Knowledge and understanding of such relevant exposure levels can help inform decision makers.
Final Risk and Exposure Assessment September 2009
Appendix 7-54
-------
Terrestrial Nutrient Enrichment Case Study
5.0 UNCERTAINTY
5.1 COASTAL SAGE SCRUB
There are several areas of uncertainty associated with this case study of CSS.
• Although current research indicates that both atmospheric nitrogen deposition and fire
have contributed to the decline of CSS, the interaction between the variables and the
extent of their contributions requires further research. CSS declines have been observed
in the absence of fire when elevated nitrogen levels are present, and declines have also
been observed in the absence of elevated nitrogen, but due to fire. Therefore, there is still
a need for quantifiable and predictive results to indicate the pressure of each variable, as
well as the pressure of the combined variables (if synergism is present). Additional
studies are also required to test the proposed nitrogen-fire feedback loop and the
associated biogeochemical elements (e.g., changes in water availability and mycorrhizal
associations) that contribute to CSS decline.
• Many studies allude to a degradation of CSS by assessing species richness and
abundance, but it is not clear that indicators of CSS ecosystem health have been
adequately explored. Assessing the health of CSS ecosystems may help to identify a
response curve to the factors associated with CSS decline.
• Ongoing experiments are beginning to show changes in CSS in response to elevated
nitrogen over relatively long periods of time (Allen, personal communication, 2008). The
incremental process may be occurring slower than previous field research experiments
have lasted, making the reasons for the decline appear variable or imperceptible over the
duration of a typical study.
• At this point, CSS is fragmented into many relatively small parcels. The CMAQ/NADP
2002 data is being modeled at 4-km resolution. The availability of these 4-km resolution
data will provide a better sense of the relationship between the current distribution of
CSS and atmospheric nitrogen loads and fire threat.
• Very little research exists regarding the effects of 63 on CSS. Although there is some
support that 63 is negatively correlated with CSS, the role has yet to be quantified or
consistently studied (Westman, 198la).
Final Risk and Exposure Assessment September 2009
Appendix 7-55
-------
Terrestrial Nutrient Enrichment Case Study
• The last area of uncertainty is the relationship between current CSS distribution and the
changing climate.
5.2 MIXED CONIFER FOREST
The currently known areas of uncertainty for MCF are as follows:
• The long-term consequences of increased nitrogen on conifers are unclear. Although the
results indicate an increased susceptibility to wildfire and disease, the long-term health of
the stands and risk of cascading effects into the ecosystem require further investigation.
• The effects of Os for both MCF and lichens confound the effects of nitrogen.
• The intermingling of fire and nitrogen cycling require additional research.
• Research suggests that critical loading benchmarks can decrease over time if the nitrogen
benchmark is exceeded for long periods of time because of decreasing nitrogen
efficiencies within nitrogen-saturated ecosystems (Fenn et al., 2008). This may indicate
that a sliding-scale approach will be required when evaluating ecosystems of varying
nitrogen responses.
• There remains considerable uncertainty in the potential response of soil carbon to
increases in total reactive nitrogen additions.
6.0 CONCLUSIONS
Evidence from the two ecosystems discussed in this case study report supports the
finding that nitrogen alters CSS and MCF. For this analysis, the loss of the native shrubs in CSS
and the increase in nonnative annual grasses were investigated. In MCF on the slopes of the San
Bernardino Mountains and Sierra Nevada Range, lichen communities associated with the forest
stands and nitrogen saturation were investigated to identify the effects of nitrogen loadings.
California's CSS and MCF have important recreational value, protect water resources, and
provide habitats for many other species. In the CSS ecosystem, there is compelling evidence that
elevated atmospheric nitrogen deposition is a driving force in the degradation of CSS. A CSS
model was developed to help identify and parse the pressures and changes occurring within the
ecosystem. In the MCF ecosystem, lichen communities and nitrogen saturation can provide a
means to monitor and quantify the effects of nitrogen loadings.
Final Risk and Exposure Assessment September 2009
Appendix 7-56
-------
Terrestrial Nutrient Enrichment Case Study
Ecological benchmarks for a suite of indicators were identified in both ecosystems:
• 3.1 kg N/ha/yr—shift from sensitive to tolerant lichen species in MCF
• 3.3 kg N/ha/yr—the amount of nitrogen uptake by a vigorous stand of CSS; above this
level, nitrogen may no longer be limiting
• 5.2 kg N/ha/yr—dominance of tolerant lichen species in MCF
• 10 kg N/ha/yr—mycorrhizal community changes in CSS
• 10.2 kg N/ha/yr—loss of sensitive lichen species from MCF
• 17 kg N/ha/yr—NO3" leaching in MCF
Because these benchmarks are comparable to levels identified in other sensitive U.S.
ecosystems and are also comparable to modeled values found in the other case studies, this set of
ecological benchmarks supports the need for continued monitoring, research, and protection of
sensitive ecosystems and informs the decision-making process.
7.0 REFERENCES
Aber, J.D., C.L. Goodale, S.V. Ollinger, M.L. Smith, A.H. Magill, M.E. Martin, R.A. Hallett,
and J.L. Stoddard. 2003. Is nitrogen deposition altering the nitrogen status of northeastern
forests? Bioscience 53:375-389.
Allen, E.B. 2008. Personal communication from E.B. Allen, University of California, Riverside,
CA.
Allen, E.B., P.E. Padgett, A. Bytnerowicz, and R. Minnich. 1998. Nitrogen Deposition Effects on
Coastal Sage Vegetation of Southern California. General Tech. Rep. PSW-GTR-166.
U.S. Department of Agriculture, Forest Service, Albany, CA.
Allen, E.B., A.G. Sirulnik, L.M. Egerton-Warburton, S.N. Kee, A. Bytnerowicz, P.E. Padgett,
PJ. Temple, M.F. Fenn, M.A. Poth, and T. Meixner. 2005. Air Pollution and Vegetation
Change in Southern California Coastal Sage Scrub: A Comparison with Chaparral and
Coniferous Forest. General Tech. Rep. PSW-GTR-1995. U.S. Department of Agriculture,
Forest Service, Albany, CA.
Final Risk and Exposure Assessment September 2009
Appendix 7-57
-------
Terrestrial Nutrient Enrichment Case Study
Allen, E.B., PJ. Temple, A. Bytnerowicz, MJ. Arbaugh, A.G. Sirulnik, and L.E. Rao. 2007.
Patterns of understory diversity in mixed coniferous forests of Southern California
impacted by air pollution. TheScientificWorldJournal 7(Sl):247-263.
Baron, J.S. 2006. Hindcasting nitrogen deposition to determine an ecological critical load.
Ecological Applications 7(5:433-439.
Baron, J.S., H.M. Rueth, A.M. Wolfe, K.R. Nydick, EJ. Allstott, J.T. Minear, and B. Moraska.
2000. Ecosystem responses to nitrogen deposition in the Colorado Front Range.
Ecosystems 3:352-368.
Baron, J.S., D.S. Ojima, E.A. Holland, and WJ. Parton. 1994. Analysis of nitrogen saturation
potential in Rocky Mountain tundra and forest: Implications for aquatic systems.
Biogeochemistry 27:61-82.
Bobbink, R. 1998. Impacts of tropospheric ozone and airborne nitrogenous pollutants on natural
and semi-natural ecosystems: A commentary. New Phytologist 739:161-168.
Bobbink, R., M. Hornung, and J.G.M. Roelofs. 1998. The effects of air-borne nitrogen pollutants
on species diversity in natural and semi-natural European vegetation. Journal of Ecology
56:717-738.
Bobbink, R., M. Ashmore, S. Braun, W. Fliickiger, and IJJ.Van Den Wyngaert. 2003. Empirical
nitrogen critical loads for natural and semi-natural ecosystems: 2002 update. Pp. 43-170
in Empirical Critical Loads for Nitrogen. Edited by B. Achermann and R. Bobbink.
Berne: Swiss Agency for Environment, Forest and Landscape (SAEFL).
Bowman, W.D., J.R. Gartner, K. Holland, and M. Wiedermann. 2006. Nitrogen critical loads for
alpine vegetation and terrestrial ecosystem response: are we there yet? Ecological
Applications 16:1183-1193.
Burger, J.C., R.A. Redak, E.B. Allen, J.T. Rotenberry, and M. F. Allen. 2003. Restoring
Arthropod Communities in Coastal Sage Scrub. Conservation Biology 7 7(2):460-467.
Final Risk and Exposure Assessment September 2009
Appendix 7-58
-------
Terrestrial Nutrient Enrichment Case Study
Bytnerowicz, A., and M.E. Fenn. 1996. Nitrogen deposition in California forests: a review
Environmental Pollution 92(2): 127-146.
Bytnerowicz, A., M. Arbaugh, S. Schilling, W. Fraczek, D. Alexander, and P. Dawson. 2007. Air
pollution distribution patterns in the San Bernardino Mountains of Southern California: A
40-year perspective. TheScientificWorldJournal 7(S1):98-109.
CA DFG (California Department of Fish & Game). 1993. Southern California Coastal Sage
Scrub NCCP Conservation Guidelines. California Department of Fish & Game and
California Resources Agency, Sacramento, CA. Available at
http://www.dfg.ca.gov/habcon/nccp/cgindex.htm.
California Department of Forestry and Fire Protection. 2007a. Historic 1977 California
Vegetation (CALVEG). California Department of Forestry and Fire Protection's Fire and
Resource Assessment Program, Sacramento, CA. Available at
http ://frap. cdf. ca.gov/data/frapgisdata/select. asp.
California Department of Forestry and Fire Protection. 2007b. Thematic Mapper (TM) data from
1993 and 1997. California Department of Forestry and Fire Protection's Fire and
Resource Assessment Program, Sacramento, CA. Available at
http ://frap. cdf. ca.gov/data/frapgisdata/select. asp.
California Invasive Plant Council. 2008. Risk Mapping for Early Detection. California Invasive
Plant Council, Berkeley, CA. Available at http://www.cal-
ipc.org/ip/mapping/statewide_maps/index.php.
Carrington, M.E., and I.E. Keeley. 1999. Comparison of post-fire seedling establishment
between scrub communities in Mediterranean and non-Mediterranean climate
ecosystems. Journal of Ecology 57(6): 1025-1036.
Clone N.K., P.E. Padgett, and E.B. Allen. 2002. Restoration of a native shrubland impacted by
exotic grasses, frequent fire and nitrogen deposition in Southern California. Restoration
Ecology 70:376-384.
Final Risk and Exposure Assessment September 2009
Appendix 7-59
-------
Terrestrial Nutrient Enrichment Case Study
Clark, C.M., and D. Tilman. 2008. Loss of plant species after chronic low-level nitrogen
deposition to prairie grasslands. Nature ¥57:712-715.
Davis, F.W., D.M. Stoms, A.D. Hollander, K.A. Thomas, P.A. Stine, D. Odion, M.I. Borchert, J.
H. Thorne, M.V. Gray, R.E. Walker, K. Warner, and J. Graae. 1998. The California Gap
Analysis Project—Final Report. University of California, Santa Barbara, CA. Available
at http://www.biogeog.ucsb. edu/proj ects/gap/gap_home.html.
Diffendorfer, I.E., G.M. Fleming, J.M. Duggan, R.E. Chapman, M.E. Rahn, MJ. Mitrovich, and
R.N. Fisher. 2007. Developing terrestrial, multi-taxon indices of biological integrity: An
example from coastal sage scrub. Biological Conservation 140:130-141.
Drus, G.M. 2004. Species Richness Patterns in a Coastal Sage Scrub Bird Community. Master's
Thesis. California State Polytechnic University, Pomona, CA. Available at
http://rivrlab.msi.ucsb.edu/drus_thesis.pdf
Egerton-Warburton, L.M., and E.B. Allen. 2000. Shifts in arbuscular mycorrhizal communities
along an anthropogenic nitrogen deposition gradient. Ecological Applications 70(2):484-
496.
Egerton-Warburton, L.M., R.G. Graham, E.B. Allen, and M.F. Allen. 2001. Reconstruction of
the historical changes in mycorrhizal fungal communities under anthropogenic nitrogen
deposition. Proceedings of the Royal Society of London B 2(55:2479-2484.
Eliason S.A., and E.B. Allen. 1997. Exotic grass competition in suppressing native shrubland re-
establishment. Restoration Ecology 5:245-255.
Fenn, M.E., and A. Bytnerowicz. 1997. Summer throughfall and winter deposition in the San
Bernardino Mountains in southern California. Atmospheric Environment 37(5):673-683.
Fenn, M.E., M.A. Poth, S.L. Schilling, and D.B. Grainger. 2000. Throughfall and fog deposition
of nitrogen and sulfur at an N-limited and N-saturated site in the San Bernardino
Mountains, southern California. Canadian Journal of Forest Research 30:1476-1488.
Final Risk and Exposure Assessment September 2009
Appendix 7-60
-------
Terrestrial Nutrient Enrichment Case Study
Fenn, M.E., J.W. Baron, E.B. Allen, H.M. Rueth, K.R. Nydick, L. Geiser, W.D. Bowen, J.O.
Sickman, T. Meixner, D.W. Johnson, and P. Neitlich. 2003. Ecological effects of
nitrogen deposition in the western United States. Bioscience 53(4):404-420.
Fenn, M.E., L. Geiser, R. Bachman, TJ. Blubaugh, and A. Bytnerowicz. 2007. Atmospheric
deposition inputs and effects on lichen chemistry and indicator species in the Columbia
River Gorge, USA. Environmental Pollution 146:11-91.
Fenn, M.E., S. Jovan, F. Yuan, L. Geiser, T. Meixner, and B.S. Gimeno. 2008. Empirical and
simulated critical loads for nitrogen deposition in California mixed conifer forests.
Environmental Pollution 755(3): 492-511.
Fierer, N.G., and EJ. Gabet. 2002. Carbon and Nitrogen Losses by Surface Runoff following
Changes in Vegetation. Journal of Environmental Quality 31:1207-1213.
Geiser, L.H., and P.N. Neitlich. 2007. Air pollution and climate gradients in Western Oregon and
Washington indicated by epiphytic macrolichens. Environmental Pollution 7 ¥5:203-218.
Grulke, N.E., and L. Balduman. 1999. Deciduous conifers: high N deposition and Os exposure
effects on growth and biomass allocation in ponderosa pine. Water, Air, and Soil
Pollution 116:235-248.
Grulke, N.E., C.P. Andersen, M.E. Fenn, and P.R. Miller. 1998. Ozone exposure and nitrogen
deposition lowers root biomass of ponderosa pine in the San Bernardino Mountains,
California. Environmental Pollution 103:63-73.
Grulke N.E., C. Andersen, W.E. Hogsett. 2001. Seasonal changes in above- and belowground
carbohydrate concentrations of ponderosa pine along a pollution gradient. Tree
Physiology 27:175-184.
Grulke, N.E., R.A. Minnich, T.D. Paine, SJ. Seybold, D. Chavez, M.E Fenn, PJ. Riggan, and A.
Dunn. 2008. Air pollution increases forest susceptibility to wildfires: a case study for the
San Bernardino Mountains in southern California. In Wild Land Fires and Air Pollution,
8. Edited by A. Bytnerowicz, M. Arbaugh, A. Riebau, C. Andersen. Burlington, MA:
Elsevier.
Final Risk and Exposure Assessment September 2009
Appendix 7-61
-------
Terrestrial Nutrient Enrichment Case Study
Jovan, S. 2008. Lichen bioindication of biodiversity, air quality, and climate: baseline results
from monitoring in Washington, Oregon, and California. General Technical Report.
PNW-GTR-737. U.S. Department of Agriculture, Forest Service, Pacific Northwest
Research Station, Portland, OR.
Jovan S., and B. McCune. 2005. Air-quality bioindication in the greater central valley of
California, with epiphytic macrolichen communities. Ecological Applications 75:1712-
1726.
Keeler-Wolf, T. 1995. Post-fire emergency seeding and conservation in Southern California
shrublands. Pp. 127-139 in Brushfires in California Wildlands: Ecology and Resource
Management. Edited by I.E. Keeley and T. Scott. International Association of Wildland
Fire, Fairfield, WA
Keeley, I.E. 2001. Fire and invasive species in Mediterranean-climate ecosystems of California.
Pp. 81-94 in Proceedings of the Invasive Species Workshop: the Role of Fire in the
Control and Spread of Invasive Species. Edited by K.E.M. Galley and T.P. Wilson. Fire
Conference 2000: the First National Congress on Fire Ecology, Prevention, and
Management. Miscellaneous Publication No. 11, Tall Timbers Research Station,
Tallahassee, FL.
Keeley, I.E., M.B. Keeley, and CJ. Fotheringham. 2005. Alien plant patterns during postfire
succession in Mediterranean-climate California shrublands. Ecological Applications
75(6):2109-2125.
Kuchler, A.W. 1988. Potential natural vegetation of California. Pp. 909-938 and map in
Terrestrial Vegetation of California. Edited by M.G. Barbour and J. Major. New York:
Wiley-Interscience.
Lovett, G.M., and T.H. Tear. 2007. Effects of Atmospheric Deposition on Biological Diversity in
the Eastern United States. Workshop Report. Institute of Ecosystem Studies, Millbrook,
NY, and The Nature Conservancy, Albany, NY.
Final Risk and Exposure Assessment September 2009
Appendix 7-62
-------
Terrestrial Nutrient Enrichment Case Study
Mattoni, R., G.F. Pratt, T.R. Longcore, J.F. Emmel, and J.N. George. 1997. The endangered
quino Checkerspot butterfly, Euphydryas editha quino. Journal of Research on the
Lepidoptera 34:99-118. Available at
http://www.urbanwildlands.org/Resources/Mattonietall997.pdf.
Merriam, K.E., I.E. Keeley, and J.L. Beyers. 2006. Fuel breaks affect nonnative species
abundance in Californian plant communities. Ecological Applications 7(5:515-527.
Meixner, T., and M. Fenn. 2004. Biogeochemical budgets in a Mediterranean catchment with
high rates of atmospheric N deposition - Importance of scale and temporal asynchrony.
Biogeochemistry 70:3 31 -3 5 6.
MEA (Millennium Ecosystem Assessment). 2005. Ecosystems and human well-being: current
state and trends: findings of the Condition and Trends Working Group. Edited by R.
Hassan, R. Scholes, and N. Ash. Washington, DC: Island Press.
Minnich, R.A. 2007. Southern California coniferous forest. Chapter 18 in Terrestrial Vegetation
of California, 3rd Edition. Edited M.G. Barbour, T. Keeler-Wolf, and A.S. Schoenherr.
University of California Press, Berkeley, CA.
Minnich, R., and R. Dezzani. 1998. Historical decline of coastal sage scrub in the Riverside-
Ferris Plain. Western Birds 39:366-391.
Minnich, R.A., M.G. Barbour, J.H. Burk, and R.F. Fernav. 1995. Sixty years of change in conifer
forests of the San Bernardino Mountains: Reconstruction of California mixed conifer
forests prior to fire suppression. Conservation Biology 9:902-914.
Padgett, P.E., and E.B. Allen. 1999. Differential responses to nitrogen fertilization in native
shrubs and exotic annuals common to Mediterranean coastal sage scrub of California.
Plant Ecology J44:93-W\.
Padgett, P.E., E.B. Allen, A. Bytnerowicz, and R.A. Minnich. 1999. Changes in soil inorganic
nitrogen as related to atmospheric nitrogenous pollutants in southern California.
Atmospheric Environment 33:769-781.
Final Risk and Exposure Assessment September 2009
Appendix 7-63
-------
Terrestrial Nutrient Enrichment Case Study
Padgett, P.E., S.N. Kee, and E.B. Allen. 2000. The effects of irrigation on revegetation of
semiarid coastal sage scrub in southern California. Environmental Management 26:427-
435.
Rueth, H.M., J.S. Baron, and EJ. Allstott. 2003. Responses of Engelmann spruce forests to
nitrogen fertilization in the Colorado Rocky Mountains. Ecological Applications 13:664-
673.
Sigal, L.L., and T.HNash, III. 1983. Lichen communities on conifers in southern California: an
ecological survey relative to oxidant air pollution. Ecology 64:1343-1354.
Siguenza, C., D.E. Crowley, and E.B. Allen. 2006. Soil microorganisms of a native shrub and
exotic grasses along a nitrogen deposition gradient in southern California. Applied Soil
Ecology 32:13-26.
Sirulnik, A.G., E.B. Allen, T. Meixner, and M.F. Allen. 2007a. Impacts of anthropogenic N
additions on nitrogen mineralization from plant litter in exotic annual grasslands. Soil
Biology and Biochemistry 39:24-32.
Sirulnik, A.G., E.B. Allen, T. Meixner, M.F. Fenn, and M.F. Allen. 2007b. Changes in N cycling
and microbial N with elevated N in exotic annual grasslands of southern California.
Applied Soil Ecology 36:1-9.
Takemoto, B.K., A. Bytnerowicz, and M.E. Fenn. 2001. Current and future effects of ozone and
atmospheric nitrogen deposition on California's mixed conifer forests. Forest Ecology
and Management 144:159-173.
Talluto, M.V., and K.N. Suding. 2008. Historical change in coastal sage scrub in southern
California, USA in relation to fire frequency and air pollution. Landscape Ecology
23:803-815.
UNECE (United Nations Economic Commission for Europe). 2004. Manual on Methodologies
and Criteria for Modeling and Mapping Critical Loads and Levels and Air Pollution
Effects, Risks, and Trends. Convention on Long-Range Transboundary Air Pollution,
Final Risk and Exposure Assessment September 2009
Appendix 7-64
-------
Terrestrial Nutrient Enrichment Case Study
Geneva Switzerland. Available at http://www.icpmapping.org (accessed August 16,
2006).
USDA (U.S. Department of Agriculture). 2009. USD A, Natural Resource Conservation Service.
2009. The PLANTS Database (http://plants.usda.gov, 22 May 2009). National Plant Data
Center, Baton Rouge, LA 70874-4490 USA
U.S. EPA (Environmental Protection Agency). 2008. Integrated Science Assessment (ISA) for
Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report). EPA/600/R-
08/082F. U.S. Environmental Protection Agency, National Center for Environmental
Assessment-RTF Division, Office of Research and Development, Research Triangle
Park, NC. Available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201485.
USFS (U.S. Forest Service). 2008a. Forest Inventory and Analysis National Program. U.S.
Department of Agriculture, Forest Service, Forest Inventory and Analysis National
Program, Arlington, VA. Available at http://fiatools.fs.fed.us/fiadb-
downloads/datamart.html.
USFS (U.S. Forest Service). 2008b. Wieslander Vegetation Type Map (VTM). U.S. Department
of Agriculture, Forest Service, Pacific Northwest Research Station, Forest Inventory and
Analysis Program, Portland, OR. Available at http://vtm.berkeley.edu/data.
U.S. FWS (Fish and Wildlife Service). 2008. FWS Critical Habitatfor Threatened and
Endangered Species. U.S. Department of the Interior, U.S. Fish and Wildlife Service,
Endangered Species Program, Washington, DC. Available at http://crithab.fws.gov.
USGS (U.S. Geological Survey). 2006. National Atlas of the United States. GIS datalayer. U.S.
Department of the Interior, U.S. Geological Survey, Reston, VA. Available at
http: //nati onal atl as. gov.
Vourlitis, G.L., G. Zorba, S.C. Pasquini, and R. Mustard. 2007. Carbon and nitrogen storage in
soil and litter of southern Californian semi-arid shrublands. Journal of Arid Environments
70:164-173.
Final Risk and Exposure Assessment September 2009
Appendix 7-65
-------
Terrestrial Nutrient Enrichment Case Study
Weaver, K.L. 1998. Coastal sage scrub variations in San Diego county and their influence on the
distribution of the California gnatcatcher. Western Birds 29:392-405.
Weiss, S.B. 1999. Cars, cows, and checkerspot butterflies: Nitrogen deposition and management
of nutrient-poor grasslands for a threatened species. Conservation Biology 13:1476-1486.
Weiss, S.B. 2006. Impacts of Nitrogen Deposition on California Ecosystems and Biodiversity.
CEC-500-2005-165. California Energy Commission, PIER Energy-Related
Environmental Research, Sacramento, CA.
Westman, W.E. 198la. Diversity relations and succession in Californian coastal sage scrub.
Ecology (52:170-184.
Westman, W.E. 1981b. Factors influencing the distribution of species of Californian coastal sage
scrub. Ecology 62:439-455.
Westman, W. 1979. Oxidant effects on California coastal sage scrub. Science 205:1001-1003.
Wood, Y., T. Meixner, PJ. Shouse, and E.B. Allen. 2006. Altered Ecohydrologic response
drives native shrub loss under conditions of elevated N-deposition. Journal of
Environmental Quality 35:76-92.
Yoshida, L.C., and E.B. Allen. 2001. Response to ammonium and nitrate by a mycorrhizal
annual invasive grass and a native shrub in southern California. American Journal of
Botany 55:1430-1436.
Final Risk and Exposure Assessment September 2009
Appendix 7-66
-------
June 5, 2009
Appendix 8
Analysis of Ecosystem Services Impacts for the
NOX/SOX Secondary NAAQS Review
Final
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards (OAQPS)
Climate, International and Multi-Media Group (CIMG)
(MD-C504-04)
Research Triangle Park, NC 27711
Prepared by
RTI International
3040 Cornwall! s Road
Research Triangle Park, NC 27709-2194
Maura Flight
Industrial Economics, Inc.
EPA Contract Number EP-D-06-003
RTI Project Number 0209897.003.066
-------
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
TABLE OF CONTENTS
1.0 Introduction 5
1.1 Ecosystem Service Categories 7
1.1.1 Descriptions and Examples of MEA Ecosystem Services 9
1.2 References 11
2.0 Aquatic Acidification 12
2.1 Overview of Affected Ecosystem Services 12
2.1.1 Provisioning Services 12
2.1.2 Cultural Services 13
2.1.3 Regulating Services 13
2.2 Changes in Ecosystem Services Associated with Alternative Levels of
Ecological Indicators 14
2.2.1 Improvements in Recreational Fishing Services due to Increased Acid
Neutralizing Capacity Levels in Adirondack and Other New York Lakes 18
2.2.2 Improvements in Total Ecosystem Services due to Increased Acid
Neutralizing Capacity Levels in Adirondack Lakes 29
2.3 References 35
3.0 Terrestrial Acidification 37
3.1 Overview of Affected Ecosystem Services 37
3.1.1 Provisioning Services 37
3.1.2 Cultural Services 39
3.1.3 Regulating Services 43
3.2 Changes in Ecosystem Services Associated with Alternative Levels of
Ecological Indicators 44
3.2.1 Increased Provisioning Services from Sugar Maple Timber Harvests due
to Elimination of Critical Load Exceedances 44
3.3 References 57
4.0 Aquatic Enrichment 60
4.1 Overview of Affected Ecosystem Services 62
4.1.1 Provisioning Services 62
4.1.2 Cultural Services 66
4.1.3 Regulating Services 66
4.2 Changes in Ecosystem Services Associated with Alternative Levels of
Ecological Indicators 67
4.2.1 The Chesapeake Bay Estuary 69
4.2.2 Neuse River Estuary 93
4.3 References 98
5.0 Terrestrial Enrichment 104
5.1 Overview of Affected Ecosystem Services 104
5.1.1 Cultural 105
5.1.2 Regulating 116
5.2 Value of Coastal Sage Scrub and Mixed Conifer Forest Ecosystem Services 120
Final Risk and Exposure Assessment September 2009
Appendix 8 - i
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
5.3 References 121
6.0 Conclusion 123
6.1 Benefits from Enhanced Provisioning Services 123
6.2 Benefits from Enhanced Cultural Services 125
6.3 Benefits from Enhanced Regulating Services 126
Attachment A: Annual Recreational Fishing Benefit Estimates for Reductions in
New York Lake Acidification Levels, 2002-2100 A-l
LIST OF FIGURES
Figure 1.1-1. Conceptual Framework for Linking Changes in Ambient NOX and SOX
Levels to Changes in Ecosystem Services and Public Welfare 6
Figure 1.1-2. MEA Categorization of Ecosystem Services and their Links to Human
Weil-Being (Source: MEA, 2005b) 9
Figure 2.2-1. Summary of Acid Neutralizing Capacity Values Relevant for Lake and
Fish Health (Source: Industrial Economics, Inc., 2008) 17
Figure 3.1-1. Combined Nitrogen and Sulfur Deposition (from 2002 CMAQ Dry
Deposition and NADP Wet Deposition Estimates) and the Range of Sugar
Maple in the United States 38
Figure 3.1-2. Combined Nitrogen and Sulfur Deposition (from 2002 CMAQ Dry
Deposition and NADP Wet Deposition Estimates) and the Range of Red
Spruce in the United States 39
Figure 3.1-3. Annual Value of Sugar Maple and Red Spruce Harvests and Maple Syrup
Production, 2006 40
Figure 3.2-1. Estimated Time Path of Welfare Gains in the Forestry and Agricultural
Sector due to Increased Sugar Maple and Red Spruce Growth (2000-
2065) 55
Figure 4-1. Conceptual Model of Eutrophication Impacts in Estuaries (Source: Adapted
from Bricker et al. [2007] and Bricker, Ferreira, and Simas [2003]) 61
Figure 4.2-1. Chesapeake Bay Coastal Block Groups 86
Figure 5.1-1. Coastal Sage Scrub Areas and Population 106
Figure 5.1-2. Mixed Conifer Forest Areas and Population 107
Figure 5.1-3. Boundaries of the NCCP Region and Subregions for Coastal Sage Scrub
(Source: California Department of Fish and Game, n.d.) 108
Figure 5.1-4. Mixed Conifer Forest Areas and National and State Park Boundaries 110
Figure 5.1-5. Coastal Sage Scrub Areas and Housing Values 113
Figure 5.1-6. Presence of Three Threatened and Endangered Species in California's
Coastal Sage Scrub Ecosystem 114
Figure 5.1-7. Presence of Two Threatened and Endangered Species in California's
Mixed Conifer Forest 115
Figure 5.1-8. Coastal Sage Scrub Areas and Fire Threat 118
Figure 5.1-9. Mixed Conifer Forest Areas and Fire Threat 119
Figure 5.1-10. Mixed Conifer Forest Areas and Major Lakes and Rivers 120
Final Risk and Exposure Assessment September 2009
Appendix 8 - ii
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
LIST OF TABLES
Table 2.2-1. Participation in Freshwater Recreational Fishing in Northeastern States in
2006 14
Table 2.2-2. Acid Neutralizing Capacity Levels (in |ieq/L) at 44 MAGIC-Modeled Lakes
in the Adirondacks 15
Table 2.2-3. Random Effects Model Results 20
Table 2.2-4. Count of Impacted Lakes 23
Table 2.2-5. Per Capita Willingness to Pay (2007 $) 24
Table 2.2-6. Present Value and Annualized Benefits, Adirondack Region 26
Table 2.2-7. Present Value and Annualized Benefits, New York State 27
Table 2.2-8. Comparison of Resources for the Future Contingent Valuation Scenarios
and EPA Zero-Out Scenario 30
Table 2.2-9. Aggregate Benefit Estimates for the Zero-Out Scenario 33
Table 3.1-1. Participation in Hunting and Wildlife Viewing in Northeastern States in
2006 42
Table 3.2-1. Summary of Plot-Level Data on Sugar Maple Growth and Exceedances (for
Plots above the Glaciation Line) 46
Table 3.2-2. Summary of Plot Level Data on Sugar Maple Growth and Exceedances (for
Plots above the Glaciation Line) 47
Table 3.2-3. Linear Exposure-Response Model for Exceedances (above a Critical Load
Calculated with Bc/Al = 10) and Sugar Maple Tree Growth: OLS
Regression Results (for Plots above the Glaciation Line) 48
Table 3.2-4. Linear Exposure-Response Model for Exceedances (above a Critical Load
Calculated with Bc/Al = 10) and Red Spruce Tree Growth: OLS
Regression Results (for Plots above the Glaciation Line) 49
Table 3.2-5. Estimated Increments in Sugar Maple and Red Spruce Timber Volume
(Resulting from Elimination of Critical Load Exceedances), by
FASOMGHG Region 52
Table 3.2-6. Proportions of Hardwood in Sugar Maple Production and Proportions of
Softwood in Red Spruce Production, by FASOM Region 53
Table 3.2-7. Proportion of Timberland under Private and Public Ownership by FIA
Regiona:2002 54
Table 4.1-1. Annual Values of East Coast Commercial Landings (in millions) 63
Table 4.1-2. Value of Commercial Landings for Selected Species in 2007 (Chesapeake
Bay Region) 64
Table 4.2-1. Participation in Selected Marine Recreation Activities in East Coast States
in 1999-2000 68
Table 4.2-2. Regression Analysis of the Chesapeake Water Quality Index on Water
Quality Parameters 71
Table 4.2-3. Average Catch Rate per Fishing Trip in the Chesapeake Bay, by State and
Targeted Fish Species 73
Table 4.2-4. Aggregate Number of Fishing Trips to the Chesapeake Bay, by State and
Targeted Fish Species 75
Table 4.2-5. Input Estimates for the Chesapeake Bay Boating Benefit Transfer Model 78
Table 4.2-6. Input Estimates for the Chesapeake Bay Beach-Use Benefit Transfer Model 82
Final Risk and Exposure Assessment September 2009
Appendix 8 - iii
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table 4.2-7. Summary of Housing Unit Numbers and Average Prices in Chesapeake
Coastal Block Groups in 2007 87
Table 5.1-1. Recreational Activities in California in 2006 by Residents and Nonresidents Ill
Table 6-1. Summary of Aggregate Benefit Estimates for Selected Ecosystem Services
and Areas (Zero Out of Nitrogen and Sulfur Deposition)21 124
Final Risk and Exposure Assessment September 2009
Appendix 8 - iv
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
1.0 INTRODUCTION
The U.S. Environmental Protection Agency (EPA) is conducting a review of the
secondary National Ambient Air Quality Standards (NAAQS) for nitrogen oxides (NOX) and
sulfur oxides (SOX). As part of the review, EPA is interested in linking changes in NOX and SOX
ambient air concentrations to the changes in ecosystem services and ultimately to changes in
public welfare. This process of linking changes in ambient NOX and SOX levels to public welfare
through the effects on ecosystem services is illustrated in Figure 1.1-1. Reducing NOX and SOX
concentrations will reduce the stresses on aquatic and terrestrial ecosystems by reducing
atmospheric deposition of nitrogen and sulfur compounds. As shown in the figure, EPA has
identified four main categories of adverse ecosystem effects—aquatic acidification, terrestrial
acidification, aquatic nutrient enrichment, and terrestrial nutrient enrichment.1 For each of these
categories, EPA has identified key ecological indicators, which provide quantitative measures of
adverse impacts on the affected ecosystems.
The purpose of this report is to identify, characterize, and, to the extent possible, quantify
the ecosystem services that are primarily impacted by nitrogen and sulfur deposition (see Section
1.1 for the definition and categorization framework used to define ecosystem services) and the
changes in ecosystem services that are expected to result from changes in the ecological
indicators. By linking indicators of ecological function to ecosystem service provision through
risk and economic assessments, the objective is to inform decisions regarding the adequacy of
current NAAQS and the ecosystem protection afforded by potential revisions to the current
primary standards for NOX and SOX.
This report includes four main sections (after this one), each dedicated to one of the main
ecosystem effect categories defined above and in Figure 1.1-1. Section 2 focuses on aquatic
acidification and provides an overview of the main ecosystem services affected by acidification
of freshwater. The section then applies the results of the Aquatic Acidification Case Study of
Adirondack lakes to quantify specifically the value of improved recreational fishing and other
cultural services caused by reductions in lake acidification in this part of the country.
1 Although other effects exist, the magnitude and/or scientific evidence of these effects is much more limited.
Final Risk and Exposure Assessment September 2009
Appendix 8-5
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Ambient Air Quality
Stressor
Affected Ecosystems
,- . .-„ . ,
Ecosystem Effects/
,; .
Symptoms
Ecological Indicators
Affected Ecosystem
Services
NO*/SO>< Levels Under
Current Conditions
I
NOX/SOX Levels Under
Alternative Conditions
Atmospheric N and S Deposition
L _LJ
Aquatic
I
Terrestrial
• Acidification
. , . . . _ . .
• Nutrient Enrichment
m
• Acidification
• Nutrient Enrichment
• Lake ANC Levels
• Eutrophication Indicators
• Forest Soil Chemistry
• Lichen Community Changes
O_J
• Provisioning
• Cultural
• Regulating
Figure 1.1-1. Conceptual Framework for Linking Changes in Ambient NOX and
SOX Levels to Changes in Ecosystem Services and Public Welfare
Final Risk and Exposure Assessment
Appendix 8-6
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Section 3 focuses on terrestrial acidification, providing an overview of the main
ecosystem services affected by acidification of forest soils. It then applies the results of the
Terrestrial Acidification Case Study and additional analyses of impacts on sugar maple trees to
quantify the value of improved provisioning services associated with expected enhancements to
forest productivity.
Section 4 focuses on aquatic nutrient enrichment. It describes and characterizes the
ecosystem services that are primarily affected by the eutrophication processes in estuaries that
result from excess nitrogen loadings. It then applies the results of the Aquatic Nutrient
Enrichment Case Study of the Potomac River/Potomac Estuary Case Study Area and the Neuse
River/Neuse River Estuary Case Study Area to quantify improvements in provisioning and
cultural services associated with reduced nitrogen loadings and improvements in eutrophic
conditions in the Chesapeake Bay and Neuse estuaries.
Section 5 focuses on terrestrial nutrient enrichment. It provides an overview of the
ecosystem services that are primarily affected by excess nitrogen loadings in two main terrestrial
ecosystems—California coastal sage scrub (CSS) and mixed conifer forest (MCF) habitats. It
also applies the findings from the Terrestrial Nutrient Enrichment Case Study of these affected
ecosystems; however, like the case study, because of data limitations and current knowledge
gaps, Section 5 does not quantify expected changes due to reductions in nitrogen and sulfur
deposition.
1.1 ECOSYSTEM SERVICE CATEGORIES
Ecosystem services are generally defined as the benefits individuals and organizations
obtain from ecosystems. This report uses the classification framework for ecosystem services
developed by the Millennium Ecosystem Assessment (MEA) (2005a, 2005b). In the MEA,
ecosystem services are defined to include both natural and human-modified ecosystems. Services
are further defined to encompass both tangible and intangible benefits that individuals and
organizations derive from ecosystems. In the MEA, ecosystem services are classified into four
main categories:
• Provisioning: includes products obtained from ecosystems.
Final Risk and Exposure Assessment September 2009
Appendix 8-7
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
• Cultural: includes the nonmaterial benefits people obtain from ecosystems through
spiritual enrichment, cognitive development, reflection, recreation, and aesthetic
experiences.
• Regulating: includes benefits obtained from the regulation of ecosystem processes.
• Supporting: includes those services necessary for the production of all other ecosystem
services.
Figure 1.1-2, taken from the MEA, displays the impact of the ecosystem services on
human well-being. The first three categories directly affect human well-being and economic
measures of welfare change. Supporting services do not have a direct effect on human well-being
but are vital to the functioning of the ecosystem.2 While other authors have proposed
categorizing ecosystem services using different systems, the MEA framework was chosen
because it is a well-developed and widely accepted system.3 The valuation of ecosystem
services benefits, however, is based on a careful linking of the MEA framework with
neoclassical economics.
2 One of the criticisms of the MEA framework from the perspective of economic analysis is that even some of the
regulating services, such as climate regulation, are more like ecosystem functions/processes or "supporting
services" that are only indirectly related to welfare (Boyd and Banzhaf [2007], Wallace [2007]).
3 Alternatives to the MEA ecosystem service classifications include, for example, Daily et al. (1997), National
Research Council (2005), and Wallace (2007).
Final Risk and Exposure Assessment September 2009
Appendix 8-8
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
CONSTITUENTS OF WELL-BEING
ECOSYSTEM SERVICES
Provisioning
FOOD
FRESH WATER
WOOD AND FIBER
FUEL
Supporting
NUTRIENT CYCLING
SOIL FORMATION
PRIMARY PRODUCTION
Regulating
CLIMATE REGULATION
FLOOD REQULATION
DISEASE REGULATION
WATER PURIFICATION
Cultural
AESTHETIC
SPIRITUAL
EDUCATIONAL
RECREATIONAL
LIFE ON EARTH - BIODIVERSITY
Security
PERSONAL SAFETY
SECURE RESOURCE ACCESS
SECURITY FROM DISASTERS
Basic material
for good life
ADEQUATE LIVELIHOODS
SUFFICIENT NUTRITIOUS FOOD
SHELTER
ACCESS TO GOODS
Health
STRENGTH
FEELING WELL
ACCESS TO CLEAN AIR
AND WATER
Good social relations
SOCIAL COHESION
MUTUAL RESPECT
ABIUTY TO HELP OTHERS
Freedom
of choice
and action
OPPORTUNITY TO BE
ABLE TO ACHIEVE
WHAT AN INDIVIDUAL
VALUES DOING
ANDBEINQ
Source: Millennium Ecosystem Assessment
Figure 1.1-2. MEA Categorization of Ecosystem Services and their Links to
Human Well-Being (Source: MEA, 2005b).
1.1.1 Descriptions and Examples of MEA Ecosystem Services
For each service category, the MEA identifies a variety of subcategories. The list below
(adapted from MEA [2005b]) highlights the services that are most relevant to this report;
however, the MEA framework contains more services than those listed. Note that supporting
services, which do not link directly to welfare, are not included, and that there is some overlap
between the categories.
1.1.1.1 Provisioning Services
• Food and fiber: This includes the vast range of food products derived from plants, animals,
and microbes, as well as materials such as wood, jute, hemp, silk, and many other products
derived from ecosystems.
• Fuel: Wood, manure, and other biological materials serve as sources of energy.
• Genetic resources: This includes the genes and genetic information used for animal and
plant breeding and biotechnology.
• Biochemicals, natural medicines, and pharmaceuticals: Many medicines, biocides, food
additives such as alginates, and biological materials are derived from ecosystems.
• Fresh water: Fresh water is another example of linkages between categories—in this case,
between provisioning and regulating services.
Final Risk and Exposure Assessment
Appendix 8-9
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
1.1.1.2 Regulating Services
• Air quality maintenance: Ecosystems both contribute chemicals to and extract chemicals
from the atmosphere, influencing many aspects of air quality.
• Climate regulation: Ecosystems influence climate both locally and globally. For example,
on a local scale, changes in land cover can affect both temperature and precipitation. On a
global scale, ecosystems play an important role in climate regulation by either sequestering
or emitting greenhouse gases.
• Water regulation: The timing and magnitude of runoff, flooding, and aquifer recharge can
be strongly influenced by changes in land cover, including, in particular, alterations that
change the water storage potential of the system, such as the conversion of wetlands or the
replacement of forests with croplands or croplands with urban areas.
• Erosion control: Vegetative cover plays an important role in soil retention and the
prevention of landslides.
• Water purification and waste treatment: Ecosystems can be a source of impurities in
freshwater but also can help filter out and decompose organic wastes introduced into
inland waters and coastal and marine ecosystems.
• Biological control: Ecosystem changes affect the prevalence of crop and livestock pests
and diseases.
• Biological control—food chain: Ecosystem changes affect the availability of vegetation
and, in turn, animals that comprise and sustain delicate food chains within an ecosystem.
• Storm protection: The presence of coastal ecosystems such as mangroves and coral reefs
can dramatically reduce the damage caused by hurricanes or large waves.
1.1.1.3 Cultural Services
• Spiritual and religious values: Many religions attach spiritual and religious values to
ecosystems or their components.
• Educational values: Ecosystems and their components and processes provide the basis for
both formal and informal education in many societies.
• Inspiration: Ecosystems provide a rich source of inspiration for art, folklore, national
symbols, architecture, and advertising.
Final Risk and Exposure Assessment September 2009
Appendix 8-10
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
• Aesthetic values: Many people find beauty or aesthetic value in various aspects of
ecosystems, as reflected in the support for parks, "scenic drives," and the selection of
housing locations.
• Recreation and ecotourism: People often choose where to spend their leisure time based, in
part, on the characteristics of the natural or cultivated landscapes in a particular area.
Environmental economists also have identified a category of services associated with
ecological benefits termed "nonuse values" (also referred to as "existence values" or "passive
use values"). As the name implies, nonuse values capture those values people have for the
environment or natural resources separate from the direct or indirect use value the resources
provide. The value some individuals hold for wilderness areas that they will never visit is one
type of nonuse value. This report includes nonuse values as a subcategory of cultural services.
1.2 REFERENCES
Boyd, J., and S. Banzhaf. 2007. "What are Ecosystem Services? The Need for Standardized
Environmental Accounting Units." Ecological Economics 63:616-626.
Daily, G.C., S. Alexander, P.R. Ehrlich, L. Goulder, J. Lubchenco, P.A. Matson, H.A. Mooney,
S. Postel, S.H. Schneider, D. Tilman, and G.M. Woodwell. 1997. "Ecosystem Services:
Benefits Supplied to Human Societies by Natural Ecosystems." Issues in Ecology 2:1-16.
Millennium Ecosystem Assessment (MEA). 2005a. Ecosystems and Human Well-being: Current
State and Trends, Volume 1. R. Hassan, R. Scholes, and N. Ash, eds. Washington, DC:
Island Press. Available at http://www.millenniumassessment.org/documents/
document.766.aspx.pdf.
Millennium Ecosystem Assessment (MEA). 2005b. Ecosystems and Human Well-being:
Synthesis. Washington, DC: World Resources Institute.
National Research Council. 2005. Valuing Ecosystem Services: Toward Better Environmental
Decision-Making. Washington, DC: National Academies Press.
Wallace, KJ. 2007. "Classification of Ecosystem Services: Problems and Solutions." Ecological
Conservation 139:235-246.
Final Risk and Exposure Assessment September 2009
Appendix 8-11
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.0 AQUATIC ACIDIFICATION
High levels of nitrogen and sulfur deposition, particularly in areas with soils containing
relatively low levels of alkaline chemical bases such as calcium or magnesium ions, often lead to
acidification of surface waters such as lakes and streams. These processes contribute to low pH
levels and other chemical changes that can be toxic to fish and other aquatic life. Evidence of
both chronic and episodic acidification of surface waters is particularly evident in the Eastern
and northeastern United States, where levels of nitrogen and sulfur deposition have also been
relatively high in recent decades. These surface waters support a wide variety of ecosystem
services, many of which can be affected adversely by acidification.
2.1 OVERVIEW OF AFFECTED ECOSYSTEM SERVICES
Because acidification primarily affects the diversity and abundance of aquatic biota, it
also primarily affects the ecosystem services that are derived from the fish and other aquatic life
found in these surface waters.
2.1.1 Provisioning Services
Food and freshwater are generally the most important provisioning services provided by
inland surface waters (Millennium Ecosystem Assessment [MEA], 2005). Whereas acidification
is unlikely to have serious adverse effects on, for example, water supplies for municipal,
industrial, or agricultural uses, it can limit the productivity of surface waters as a source of food
(i.e., fish). In the northeastern United States, the surface waters affected by acidification are not a
major source of commercially raised or caught fish; however, they are a source of food for some
recreational and subsistence fishers and for other consumers. Although data and models are
available for examining the effects on recreational fishing (see Section 2.1.2), relatively little
data are available for measuring the effects on subsistence and other consumers. For example,
although there is evidence that certain population subgroups in the northeastern United States,
such as the Hmong and Chippewa ethnic groups, have particularly high rates of self-caught fish
consumption (Hutchison and Kraft, 1994; Peterson et al., 1994), it is not known if and how their
consumption patterns are affected by the reductions in available fish populations caused by
surface water acidification.
Final Risk and Exposure Assessment September 2009
Appendix 8-12
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.1.2 Cultural Services
Inland surface waters support several cultural services, such as aesthetic and educational
services; however, the type of service that is likely to be most widely and significantly affected
by aquatic acidification is recreational fishing, since it depends directly on the health and
abundance of aquatic wildlife. Other recreational activities such as hunting and birdwatching are
also likely to be affected, to the extent that fish eating birds and other wildlife are harmed by the
absence offish in acidic surface waters.
Recreational fishing in lakes and streams is among the most popular outdoor recreational
activities in the northeastern United States. Data from the 2006 National Survey of Fishing,
Hunting, and Wildlife Associated Recreation (FHWAR), as summarized in Table 2.2-1, indicate
that more than 9% of adults in this part of the country participate annually in freshwater
(excluding Great Lakes) fishing. The total number of freshwater fishing days occurring in those
states (by both residents and nonresidents) in 2006 was 140.8 million days. Roughly two-thirds
of these fishing days were at ponds, lakes, or reservoirs in these states, and the remaining one-
third were at rivers or streams. Based on studies conducted in the northeastern United States,
Kaval and Loomis (2003) estimated an average consumer surplus value per day of $35.91 for
recreational fishing (in 2007 dollars). Therefore, the implied total annual value of freshwater
fishing in the northeastern United States was $5.06 billion in 2006.
2.1.3 Regulating Services
In general, inland surface waters such as lakes, rivers, and streams provide a number of
regulating services, such as hydrological regime regulation and climate regulation. There is little
evidence that acidification of freshwaters in the northeastern United States has significantly
degraded these specific services; however, freshwater ecosystems also provide biological control
services by providing environments that sustain delicate aquatic food chains. The toxic effects of
acidification on fish and other aquatic life impair these services by disrupting the trophic
structure of surface waters (Driscoll et al., 2001). Although it is difficult to quantify these
services and how they are affected by acidification, it is worth noting that some of these services
may be captured through measures of provisioning and cultural services. For example, these
biological control services may serve as "intermediate" inputs that support the production of
"final" recreational fishing and other cultural services.
Final Risk and Exposure Assessment September 2009
Appendix 8-13
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.2 CHANGES IN ECOSYSTEM SERVICES ASSOCIATED WITH
ALTERNATIVE LEVELS OF ECOLOGICAL INDICATORS
This section estimates values for changes in ecosystem services associated with
reductions in lake acidification in Adirondack State Park and in other lakes in the state of New
York. Using the results of the Adirondack Case Study Area of aquatic acidification effects, the
value of reducing nitrogen and sulfur deposition in the affected areas to background levels (i.e., a
"zeroing out" of anthropogenic sources of nitrogen and sulfur) was estimated. Although this
scenario is not realistic from a policy perspective, it allows the examination of the upper bound
of ecosystem service gains that would result from reducing the number of acidified (i.e., low acid
neutralizing capacity [ANC]) lakes to the lowest possible level.
Table 2.2-1. Participation in Freshwater Recreational Fishing in Northeastern States in 2006
State
Connecticut
Delaware
Illinois
Indiana
Maine
Maryland
Massachusetts
Michigan
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
West Virginia
Wisconsin
Total
Participation
Rates by State
Residents"
7.3%
6.4%
9.5%
13.8%
19.9%
6.6%
6.0%
12.3%
10.0%
4.0%
4.6%
11.6%
8.3%
5.3%
13.2%
20.9%
21.4%
9.3%
Activity Days by Residents and Nonresidents (in
thousands)
Ponds, Lakes, or
Reservoirs
2,856
764
10,318
6,843
3,734
2,882
4,494
15,175
2,144
2,377
8,548
9,781
7,507
504
1,264
3,069
13,026
95,286
Rivers or
Streams
2,409
770
5,088
1,819
1,521
2,379
978
4,426
627
1,116
5,086
3,710
6,998
104
453
3,617
4,439
45,540
Total
5,265
1,534
15,406
8,662
5,255
5,261
5,472
19,601
2,771
3,493
13,634
13,491
14,505
608
1,717
6,686
17,465
140,826
a Ages 16 or older.
Source: U.S. Department of the Interior (DOI), Fish and Wildlife Service, and U.S. Department of Commerce, U.S.
Census Bureau, 2007.
Final Risk and Exposure Assessment
Appendix 8-14
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
The case study analysis focused on 44 lakes in the Adirondacks. It estimated ANC levels
at each of these lakes under the alternative scenarios shown in Table 2.2-2. Using the MAGIC
model, it predicted median ANC levels for the years 2005, 2020, 2050, and 2100 under
"business-as-usual" conditions (i.e., accounting for expected emission controls associated with
Title IV regulations but no additional measures to reduce nitrogen and sulfur deposition). In
contrast, the model run for the year 1860 represents ANC levels for "background" conditions by
simulating the effect of zeroing out anthropogenic sources of nitrogen and S.
Table 2.2-2. Acid Neutralizing Capacity Levels (in jieq/L) at 44 MAGIC-Modeled Lakes in the
Adirondacks
Year:
Lake Name
Clear Pond (61)
Long Pond (65)
Hope Pond
Second Pond
Squaw Lake
Indian Lake
Big Alderbed
Long Lake
Gull Pond
Little Lilly Pond
Upper Sister Lake
Dry Channel Pond
Bennett Lake
Effley Falls Pond
Parmeter Pond
North Lake
Razorback Pond
Snake Pond
South Lake
Boottree Pond
Horseshoe Pond
Rock Pond
Antediluvian Pond
Seven Sisters Pond
Observed
2002
218.1
66.1
62.6
71.3
19.5
-8.0
57.4
-32.1
160.9
47.0
31.0
23.8
33.0
50.8
75.6
-1.6
33.2
3.6
-7.7
53.2
51.4
86.7
66.2
-14.1
MAGIC Model Simulations
2005
233.0
73.5
72.9
75.8
25.6
1.4
67.5
-20.8
166.8
54.4
37.4
31.7
37.5
59.8
85.7
6.9
39.6
12.3
0.1
59.0
63.0
95.1
70.1
-9.1
2020 a
243.2
78.3
78.4
77.0
27.1
6.2
72.6
-15.4
170.7
57.9
39.9
34.3
39.2
64.2
91.7
10.9
42.4
15.5
3.7
63.2
70.0
98.8
72.0
-6.9
2050a
246.7
80.4
81.1
75.3
24.9
6.2
74.5
-16.0
173.0
58.5
39.4
33.2
37.8
64.2
94.3
10.0
40.5
14.5
2.3
65.3
73.4
99.5
71.4
-7.2
2100a
247.6
81.2
82.8
72.5
21.3
5.1
75.6
-17.6
174.6
58.6
38.0
31.6
35.0
63.7
95.4
8.1
37.4
13.4
-0.2
66.1
74.8
99.5
69.9
-8.1
I860"
("Background")
290.3
106.4
126.5
121.5
73.8
52.2
124.1
34.4
208.8
95.5
80.3
78.6
69.7
132.4
134.8
66.0
94.3
78.5
56.6
84.5
117.6
151.5
95.3
21.9
Final Risk and Exposure Assessment
Appendix 8-15
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Year:
Lake Name
Canada Lake
Bickford Pond
Wolf Pond
Blue Mountain Lake
Carry Falls Reservoir
Rocky Lake
Bog Pond
Clear Pond (82)
Seventh Lake
Trout Pond
Hitchins Pond
Piseco Lake
Mccuen Pond
Arbutus Pond
Witchhopple Lake
Willys Lake
Lower Beech Ridge
Pond
Dismal Pond
Payne Lake
Whitney Lake
Observed
2002
53.8
19.9
-5.7
118.0
121.2
43.7
88.2
85.9
198.4
39.9
150.2
98.2
37.6
88.0
27.2
-48.4
-19.6
-23.6
53.5
22.3
MAGIC Model Simulations
2005
69.4
33.6
4.8
126.7
133.2
58.6
107.0
97.1
217.3
53.4
162.7
114.7
46.0
101.6
35.7
-38.8
-10.8
-12.0
56.2
30.7
2020 a
77.6
41.3
9.8
129.3
140.8
66.3
117.2
104.1
223.4
61.9
170.0
123.7
50.2
108.6
39.4
-33.5
-6.9
-7.6
58.1
33.7
2050a
80.1
45.2
11.2
127.8
144.1
68.3
120.5
107.4
227.1
65.7
172.6
127.2
51.7
111.3
38.9
-33.3
-7.4
-7.3
59.0
32.9
2100a
81.4
46.9
11.9
125.2
145.8
68.8
121.6
108.2
229.1
67.6
173.8
128.6
52.4
113.1
37.6
-33.4
-8.8
-7.6
59.4
31.5
I860"
("Background")
151.2
101.3
58.3
184.3
205.8
113.7
178.1
145.6
317.6
127.2
214.7
186.2
90.0
187.1
91.7
47.5
41.5
40.4
75.1
84.3
a Based on predicted future scenarios for nitroj
b Represents background levels and levels that
of nitrogen+sulfur deposition.
;en+sulfur deposition, accounting for Title IV emissions controls.
would eventually result from a "zero-out" of anthropogenic sources
In the following subsections, ecosystem service gains associated with going from the
business-as-usual reference conditions to the zero-out condition are estimated. It was assumed
that the zero out of nitrogen and sulfur deposition would occur in 2010 and that it would take 10
years for the full effect of these reductions on lake ANC levels to occur. In other words, by the
year 2020, lake ANC levels would increase and fully return to their estimated 1860 background
levels, as shown in Table 2.2-2.
In Section 2.2.1, a model that focuses specifically on recreational fishing services is
applied, and the value of current and future enhancements to these services from Adirondack and
other New York lakes is estimated. In Section 2.2.2, a model that takes a broader perspective on
Final Risk and Exposure Assessment
Appendix 8-16
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
ecosystem services is applied, and the value of improving all of the ecosystem services that are
affected by acidification of Adirondack lakes is estimated.
In both cases, the analysis focuses on ANC and evaluates the sensitivity of different ANC
thresholds for aquatic functioning. In general, moderate shifts in ANC levels may result in
changes in species composition, where acid-sensitive species are replaced by less sensitive
species. At more extreme acidification levels, however, species richness, defined as the total
number of species occupying a system, may be affected. Research has shown that the number of
fish species present is positively correlated with ANC (Driscoll et al., 2003). In the Adirondacks,
recent research indicates that aquatic biota begin to exhibit effects at an ANC of 50
microequivalents per liter (|j,eq/L) (Chen and Driscoll, 2004). Uncertainty exists regarding
threshold levels of ANC: the levels at which predictable effects occur. Several ANC thresholds
have been observed, however, at which lakes and fish are affected, as summarized in Figure 2.2-
1. To account for the uncertainty in the threshold level of acidification above which Adirondack
lakes may support recreational fishing, this analysis considers three threshold levels: 20 |j,eq/L,
50 neq/L, and 100 ^eq/L.
ANC THRESHOLD VALUES (UEQ/L)
20
40 50
100
200
Species richness
increases between 50
and 100 |jeq/L
Lakes are not sensitive to
acidification at values approaching
200 |jeq/L
Level below which aquatic biota are affected
Lakes are sensitive to episodic acidification
Level below which all biota are affected
Lakes are chronically acidic and fishless
Figure 2.2-1. Summary of Acid Neutralizing Capacity Values Relevant for Lake
and Fish Health (Source: Industrial Economics, Inc., 2008).
Final Risk and Exposure Assessment
Appendix 8-17
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.2.1 Improvements in Recreational Fishing Services due to Increased Acid
Neutralizing Capacity Levels in Adirondack and Other New York Lakes
To estimate the value of improved services, this analysis relied on commonly accepted
economic models to relate the predicted changes in lake acidity to a change in recreational
fishing behavior throughout the study area. First, a random effects model was used to extrapolate
lake ANC levels from the ecological model forecast for a subset of lakes to a broader suite of
regional lakes. This random effect model does this by relating acidification levels to lake
characteristics and geographic location. That is, the forecast ANC levels of the lakes modeled in
MAGIC for each year in the study period are tied to explanatory variables so that the forecast
changes in ANC can be extrapolated to other potentially affected lakes in the region. This model
was first applied to forecast ANC levels at lakes in the Adirondack region and then repeated to
forecast ANC levels for lakes in New York State (with the exception of New York City). The
result of this effort is a full time-series dataset of ANC levels for Adirondack and New York
State lakes.
The second economic model applied describes changes in behavior of recreational fishers
in response to changes in lake acidification levels. This step of the process relies on the
assumption that below the specified ANC threshold (of 20 |j,eq/L, 50 |j,eq/L, or 100 |j,eq/L) lakes
are no longer fishable. The specific type of model applied here is a "discrete choice model."
Generally, a discrete choice model predicts a binary decision (which may be thought of as "yes"
or "no") regarding whether to fish at a given site made by an individual as a function of a
number of independent variables (Greene, 2003). The independent variable is the catch rate at
the water body (itself a function of lake acidity). Additional independent variables may include
travel time required to reach the site and the concentration of fisherman at the site, among others.
A specific form of discrete choice model called a "random utility model," or RUM is
applied. In the study of economics, utility is defined as a measure of the happiness or satisfaction
gained from a good or service. In keeping with the tenets of neoclassical economics, this utility is
sought to be maximized subject to a constraint (often represented by income or time). Put more
simply, the model assumes that the fisherman will seek the most happiness at the lowest cost.
Section 2 describes the application of these models and the results of this analysis.
Final Risk and Exposure Assessment September 2009
Appendix 8-18
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.2.1.1 Analytic Method
The following steps were followed to connect the modeled changes in lake ANC levels to
the benefits of improved recreational fishing services.
Step 1: Development of the Random Effects Model
To develop this model, it was first necessary to compare the subset of lakes considered in
the ecological model (see Table 2.2-2) with the subset of lakes included in the database of lake
characteristics contained within the RUM. Nine of the 44 lakes were not usable for the analysis
because they did not appear in the database of lake characteristics within the RUM.4 As a result,
the analysis relied on data for a subset of 35 Adirondack lakes.
Because forecasted ANC levels were provided for the 35 lakes only, the next step of the
analysis was extrapolating these forecasts to the broader suite of lakes within the Adirondack
region. To this end, a random effects model was developed to determine the statistical
relationship between the lakes' ANC levels and their characteristics. Significant uncertainty
exists regarding the relationships between lake characteristics and ANC levels. Ecologists at the
Environmental Protection Agency (EPA) are researching the characteristics that best explain a
lake's sensitivity to acidification.
The random effects model used in this analysis to forecasted lake ANC levels was also
limited by the lake characteristic data that are currently available; in this case, elevation, surface
area, shoreline, and county location were considered as potential explanatory variables in
forecasting ANC levels. The relationship between these characteristics and the forecast ANC
levels for the 35 lakes informed the extrapolation of the results from the MAGIC model to the
broader population of lakes, first in the Adirondack region and then in New York State. These
variables that describe the lake characteristic and geographic location are the explanatory
variables in the model. The random effects model helps identify the influences of these
explanatory variables, net of other factors that are unknown and cannot be controlled.5
4Those lakes excluded include the following: Bickford Pond, Bog Pond, Hope Pond, Little Lilly Pond, Lower Beech
Ridge, Razorback Pond, Seven Sisters Pond, Snake Pond, and Witchhopple Lake.
5 Several important conditions must be satisfied for the random effects model to be appropriate. In this case, these
conditions are met. For both models, the Breusch-Pagan test for random effects rejects the null hypothesis of no
random effects in the data. The Hausman specification test (against a fixed effects alternative) rejects the null
hypothesis of systematic differences between random and fixed effects models for the Clean Air Interstate Rule
(CAIR) and t variables, which indicates that omitted variables are not biasing the coefficients for those variables.
Final Risk and Exposure Assessment September 2009
Appendix 8-19
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
The model cannot perfectly predict ANC levels in lakes; there are not enough available
data and there is no existing knowledge about the best determinants of ANC levels. Given that
there is some uncertainty and limited information available to explain ANC levels, a method
must be used that can remove the net effects of the unknown data and identify the effects of the
available information. The random effects model generates estimates of the net effects of the
explanatory variables.
Furthermore, random effects models are appropriate for situations where the study
sample is a random sample of a larger universe and one wishes to make inferences about the
larger universe of data (Kennedy, 2003).6 In this case, the group of lakes analyzed is sampled
from the total number of lakes for which ANC levels are forecast (the lakes to which the ANC
levels are forecast is the universe).
The modeled ANC levels for the 35 aforementioned lakes, along with the lake
characteristic information, served as inputs for a random effects regression analysis to isolate the
impact of each variable on ANC. Table 2.2-3 details the results of the random effects model.
Table 2.2-3. Random Effects Model Results
Variable
constant
elevation
area
In(shoreline)
T
Hamilton
Essex
Fulton
Franklin
Herkimer
Lewis
Warren
Coefficient
-106.171
-0.047
0.125
-36.005
0.108
9.430
55.149
-16.793
49.538
-38.655
-19.160
24.924
Std. Error
75.050
0.128
0.074
18.802
0.013
27.760
46.894
80.273
39.176
40.142
45.899
66.423
6 This criterion assumes that there are no omitted variable effects present; the previous footnote explains that there is
no evidence of this.
Final Risk and Exposure Assessment September 2009
Appendix 8-20
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Variables describing elevation, total area, and shoreline length were included to capture
physical differences between lakes. While the coefficients are not statistically significant, the
variables do lend some explanatory power to the model. The variable identified as "T" is an
annual time trend included to capture changes through time manifested in the greater system and
not a specific lake. The final seven variables listed in the table are binary variables indicating the
counties in which the lakes occur. The omitted variable is for St. Lawrence County. These
variables are intended as a proxy for a host of location-specific factors, including subsurface
geology and degree of forest cover because data were not available for these variables.
Step 2: Extrapolation to All Lakes in Adirondack Region/New York State
The Montgomery-Needelman RUM includes lake characteristic data for a total of 2,586
lakes in New York State. As described previously, the MAGIC model predicts ANC levels for
35 lakes within the Adirondack region that could be included in the random effects model. These
35 lakes are located in Hamilton, Essex, Fulton, Franklin, Herkimer, Lewis, Warren, and St.
Lawrence counties. Their explanatory value for lakes outside of this eight-county region is
uncertain. Therefore, this study performed a "tiered" extrapolation, where the random effects
model results were first extrapolated only to lakes in the Adirondack region represented by the
modeled lakes; this exercise was then repeated for the full suite of New York State lakes.
For the first tier (for the Adirondack region), the analysis was limited by two dimensions:
(1) only including lakes within the eight counties containing the 35 modeled lakes and (2)
limiting the analysis to lakes within the size range of the modeled lakes. Because none of the 35
modeled lakes occur in Clinton, Saratoga, and Oneida counties (all within the Adirondack
region), this analysis did not apply the model to forecast lake acidification in these three
counties. This assumption may lead to an understatement of the total benefits associated with
decreased lake acidification in the Adirondack region, but it avoids some uncertainty associated
with extrapolating ANC outside of the scope of the modeled region.
The second tier of the analysis (for all of New York State except New York City) was
also limited to consider only lakes within the size range of the modeled lakes. This portion of the
analysis required consideration of lakes outside of the eight-county geographic scope, however.
Therefore, an average of the eight-county binary variable coefficients for all lakes outside of the
eight counties was used. Further, as with the first tier of the analysis, all lakes with an area
Final Risk and Exposure Assessment September 2009
Appendix 8-21
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
greater than the largest lake in the ecological subset of 35 were "hardwired" to be unimpaired,
because changes in their ANC levels are unlikely to be represented by the subset of modeled
lakes. ^ A total of 62 lake sites were determined to be too large to be represented by the sample
MAGIC data and were, therefore, hardwired.
Step 3: Application of ANC Thresholds
This analysis employs three ANC threshold assumptions—20 |j,eq/L, 50 |j,eq/L, and 100
|j,eq/L—to indicate whether a lake is fishable. A lake was deemed to be affected if it was above
the threshold (fishable) in the "zero-out" scenario and below the threshold (impaired) in the
baseline scenario. As previously described and shown in Table 2.2-2, zero-out conditions are
defined by lake ANC levels in the year 1860 as estimated by MAGIC. MAGIC provided these
data for the subset of 35 lakes within the Adirondack region. To determine zero-out conditions
for the broader suite of lakes in the Adirondacks and in New York State, a simple ordinary least
squares (OLS) regression was run to determine whether lake size is a reasonable indicator of the
difference between the observed ANC level in 2002 and the pristine condition in 1860 for the 35
lakes. This analysis determined that no statistically significant relationship existed. Therefore, an
average difference in ANC level between the 2002 observed level and the 1860 pristine
condition for the 35 lakes was calculated; the average difference is 64.6 |j,eq/L. This average
difference was then added to the 2002 ANC levels for each lake (forecast by extrapolation using
the random effects model), and the resulting value was considered to be the "pristine" ANC
value in 2020, 2050, and 2100.
Table 2.2-4 reports the number of "impacted" lakes in each year, where impact means
that the lake is predicted to be below the ANC threshold under business-as-usual conditions and
above the threshold under zero-out conditions. This definition of impacted lakes is needed
because the RUM framework only estimates benefits accruing from lakes that switch from
nonfishable to fishable status. The lake counts for 2005 are zero because in this year no change
7 Hardwired lakes (in order of decreasing size) include Lake Ontario, Lake Erie, Great Sacandaga Lake, Oneida
Lake, Seneca Lake, Lake Champlain, Cayuga Lake, Lake George, Canandaigua Lake, Ashokan Reservoir,
Cranberry Lake, Owasco Lake, Chautauqua Lake, Tupper Lake, Stillwater Reservoir, Keuka Lake, Pepacton
Reservoir, Allegheny Reservoir, Raquette Lake, Cannonsville Reservoir, Indian Lake, Skaneateles Lake, Black
Lake, Long Lake, Otsego Lake, Saratoga Lake, Mount Morris Reservoir, Salmon River Reservoir, Great Sodus
Bay, Conesus Lake, Whitney Point Reservoir, and Onondaga Lake.
Final Risk and Exposure Assessment September 2009
Appendix 8 - 22
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
occurs in ANC level relative to the baseline (i.e., the reduction in emissions beginning the return
to pristine conditions was not assumed to have occurred in those years). The zero-out scenario
was assumed to be implemented in 2010, with lakes reaching their pristine conditions by 2020. It
should be noted that the nature of this model allows for lakes to switch between impaired and
unimpaired between years. As a result, the lake counts reported in Table 2.2-4 are not cumulative
counts and, in fact, may reflect different subsets of lakes.
Step 4: Application of the Random Utility Model
The Montgomery-Needelman model applied in this analysis is a repeated discrete choice
RUM that describes lake fishing behavior of New York residents (Montgomery and Needelman,
1997). In particular, the model characterizes decisions regarding (1) the number of lake fishing
trips to take each season and (2) the specific lake sites to visit on each fishing trip. The model
can be used to develop estimates of economic losses or gains associated with changes in the set
of lakes available to anglers.
Table 2.2-4. Count of Impacted Lakes
ANC Threshold
(in jieq/L)
20
20
20
20
50
50
50
50
100
100
100
100
Year
2005
2020
2050
2100
2005
2020
2050
2100
2005
2020
2050
2100
Adirondack
0
107
95
74
0
244
222
200
0
430
404
354
Lake Count
Region New York State
0
110
97
75
0
365
316
254
0
1,500
1,399
1,228
Note: There are 1,076 lakes in the "Adirondack Region" and 2,586 lakes in New York State (less New York City).
The data used to estimate the RUM were obtained from a 1989 repeat-contact telephone
survey of New York residents conducted as part of the National Acid Precipitation and
Final Risk and Exposure Assessment September 2009
Appendix 8-23
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Assessment Program (NAPAP).8 This survey provided information on the destinations of
anglers' fishing trips (day trips only) taken during the 1989 fishing season. The survey data were
supplemented with lake characteristics data obtained from New York State Department of
Environmental Conservation's (NYSDEC's) Characteristics of New York State Lakes: Gazetteer
of Lakes and Ponds and Reservoirs, New York State's Fishing Guide, and New York's 305(b)
report for 1990. Travel distances between anglers' homes and lake fishing sites were calculated
using Hy ways/By ways. The model and data used in the present analysis are described in greater
detail in a 1997 journal article by Montgomery and Needelman.9
The list of affected lakes generated in the previous step serves as the primary input to the
RUM. The model estimates the economic welfare value of enhancements to recreational fishing
services derived from shifting specific lakes from nonfishable to fishable status.10 The economic
benefits estimates represent New York State residents' average willingness to pay (WTP) to
improve recreational fishing services by reducing lake acidification levels. Table 2.2-5 reports
the estimated per capita values generated by the RUM. These values have been adjusted from
1989 dollars to 2007 dollars using the Consumer Price Index-All Urban Consumers (CPI-U).
Note that the zero-out scenario is assumed to begin at the end of 2010; therefore, the benefits do
not begin to accrue until the following year, and they are zero in 2005 and 2010.
Table 2.2-5. Per Capita Willingness to Pay (2007 $)
ANC Threshold
(in jieq/L)
20
20
20
20
20
50
Per Capita Benefits of a Return to Pristine Conditions
by 2020
Year
2005
2010
2020
2050
2100
2005
Adirondack Region
$0.00
$0.00
$0.41
$0.34
$0.28
$0.00
New York State
$0.00
$0.00
$0.47
$0.38
$0.32
$0.00
8 New York City counties were excluded from the sampling frame.
9 The published version of the model has had several minor updates, all of which have been discussed with Mark
Montgomery.
10 Since the RUM uses travel distances and travel costs to infer economic values, the benefit estimates are sensitive
to the spatial locations and distributions of the impacted lakes (i.e., the benefit estimates do not depend only on
the number or percentage of lakes impacted).
Final Risk and Exposure Assessment September 2009
Appendix 8 - 24
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
ANC Threshold
(in (ieq/L)
50
50
50
50
100
100
100
100
100
Year
2010
2020
2050
2100
2005
2010
2020
2050
2100
Per Capita Benefits of a Return
by 2020
Adirondack Region
$0.00
$0.74
$0.73
$0.70
$0.00
$0.00
$0.79
$0.77
$0.68
to Pristine Conditions
New York State
$0.00
$2.55
$2.26
$1.47
$0.00
$0.00
$11.05
$10.61
$9.40
Step 5: Interpolation of RUM Output
The RUM provided per capita loss estimates (reported in 1989 nominal dollars) for 2010,
2020, 2050, and 2100.11 Rather than running this model separately for each year, estimates for
the intervening years (between the four point estimates provided by the RUM) were generated
via simple linear interpolations.
Step 6: Estimation of Aggregate Benefits through Application of Per Capita Results
to Affected Population
To estimate aggregate benefits for New York residents, the per capita benefit estimates
must be multiplied by the corresponding population of residents. To match the characteristics of
the population surveyed in developing the RUM, this analysis required estimating the population
of New York State that will be over 18 years old and reside outside of New York City for each
year from 2011 through 2100. The U.S. Census Bureau provides estimated population figures for
2002 through 2008 and projected population through 2030 at the state level. Absent projection
information, the population was held constant from 2030 through the period of the analysis
(through 2100). The ratio of the New York State population residing outside New York City
(that is, the five counties of Bronx County, Kings County, New York County, Queens County,
11 As mentioned previously, the CPI-U, provided by the U.S. Bureau of Labor Statistics, was used to inflate these
estimates to 2007 dollars.
Final Risk and Exposure Assessment September 2009
Appendix 8-25
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
and Richmond County) was calculated for 2006 and assumed to remain constant throughout the
analysis. The U.S. Census also estimates and projects the 18+ population at the state level
through 2030. The 18+ population was held constant from 2030 through the end of the analysis
in 2100. The ratio of adults (18+) to the entire population was calculated for New York State,
and that ratio was applied to the population residing outside New York City.
2.2.1.2 Results
Table 2.2-6 summarizes the estimated present value and annualized benefits for each
acidification threshold assumption applying discount rates of 3% and 7%. The estimated present
value of benefits in 2010 range from $60.1 million to $298.7 million depending on the threshold
and discount rate assumptions applied. In comparison, a previous study of the recreational
fishing benefits in the Adirondacks associated with the Clean Air Act Amendments (CAAA)
estimated benefits ranging from $13.7 million to $100.6 million (EPA, Office of Air and
Radiation, 1999).12 Annualizing these benefits over the period 2010 to 2100 results in annual
benefit estimates ranging from $3.9 million to $9.3 million per year. Six tables containing
detailed results for each scenario (threshold assumption and geographic scope) by year are
included in Attachment A.
Table 2.2-7 describes total present value and annualized benefits associated with reduced
lake acidification in all of New York State. Estimated present value benefits in 2010 range from
$68.3 million to $4.16 billion, depending on the threshold and discount rate assumptions applied.
Annualizing these benefits over the period 2010 to 2100 results in annual benefit estimates
ranging from $4.5 million to $130 million per year.
Table 2.2-6. Present Value and Annualized Benefits, Adirondack Region
Present Value Benefits8 Annualized Benefits'5
ANC
T, , - , (in million of 2007 dollars) (in million of 2007 dollars)
(in jieq/L)
20
50
3% Discount Rate
$142.59
$285.15
7% Discount Rate
$60.05
$114.18
3% Discount Rate
$4.46
$8.91
7% Discount Rate
$3.94
$7.49
12 For comparison to the results in our analysis, presented in 2007 dollars, the estimated benefits from the Clean Air
Act report were inflated from 1999 to 2007 dollars using the GDP deflator
(http://www.bea.gov/bea/dn/nipaweb/SelectTable.asp?Selected=Y).
Final Risk and Exposure Assessment September 2009
Appendix 8-26
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Present Value Benefits" Annualized Benefits'5
ANC
T, , - , (in million of 2007 dollars) (in million of 2007 dollars)
3% Discount Rate 7% Discount Rate 3% Discount Rate 7% Discount Rate
100 $298.67 $120.61 $9.33 $7.91
a Annual benefits for 2010 to 2100 discounted to 2010.
b Present value benefits annualized over 2009 to 2100.
Table 2.2-7. Present Value and Annualized Benefits, New York State
Present Value Benefits" Annualized Benefits'5
ANC
T, , - , (in million of 2007 dollars) (in million of 2007 dollars)
(in (ieq/L)
20
50
100
3% Discount Rate
$161.76
$897.20
$4,159.64
7% Discount Rate
$68.34
$378.00
$1,685.80
3% Discount Rate
$5.05
$28.04
$129.98
7% Discount Rate
$4.48
$24.78
$110.52
a Annual benefits for 2010 to 2100 discounted to 2010.
b Present value benefits annualized over 2010 to 2100.
2.2.1.3 Assumptions and Caveats
The following assumptions and caveats are particularly important for interpreting the
results and the application of the ecological model for lake acidification:
• This analysis assumed that the level of impairment is binary as applied to a specific lake:
that is, the ANC threshold indicates whether a lake is fishable.
• The available literature suggests that ANC levels between 20 and 100 cover the range
where ecological affects are realized. Three points within this range (20, 50, and 100) were
tested as point estimates at which the fishability of lakes is affected.
• This analysis assumed that the 35 modeled lakes are a representative subset of lakes in the
Adirondacks (for the first tier of the analysis) and in New York State (for the second tier of
the analysis).
• This analysis used the ANC levels of the 35 modeled lakes in the year 1860 as a proxy for
"pristine" acidification levels.
• In the first tier, the analysis is not used to forecast acidification effects in Clinton,
Saratoga, and Oneida counties, which are generally considered to be part of the
Final Risk and Exposure Assessment September 2009
Appendix 8-27
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Adirondack region because they are not represented by the subset of lakes subject to the
ecological model. This restriction contributes to an underestimation of total benefits.
• Pristine ANC levels for the full population of New York State lakes are estimated by first
finding the average difference between 2002 observed ANC levels and the 1860 ANC
values for the 35 lakes modeled by MAGIC and then adding this average difference to the
2002 ANC values for all lakes (as estimated by extrapolating using the random effects
model). The ANC levels assumed to represent pristine lake conditions are therefore subject
to significant uncertainty.
The following assumptions and caveats are particularly important for interpreting the
application of the RUM model for estimating recreational fishing benefits to New York
residents:
• The RUM only considers the behavior of New York State residents. It may be reasonable
to assume that residents of neighboring jurisdictions (the Canadian provinces of Ontario
and Quebec, along with the State of Vermont) may also take day trips to these lakes and
respond in a rational manner comparable to New York State residents. This restriction
contributes to an underestimation of benefits.
• The output of the RUM is on a per capita basis. The results are presented in terms of
impacts to the entire population. This requires an extrapolation of the population through
2100. Absent specific projection information beyond 2030, the population was held
constant beyond this year.
• This analysis assumed that the demand for fishing, in other words, an individual's
propensity to fish, has remained constant from the time of the survey underlying the RUM
to the present. That is, this analysis does not account for any potential change in interest in
both recreational fishing and park use since the survey was conducted in 1989. In the case
that general demand for recreational fishing has decreased, this analysis may overstate
benefits. This restriction contributes to an overestimation of benefits.
• This analysis did not take into account income adjustments through time. The RUM holds
income to be constant and a lack of detailed demand elasticity functions precludes the
incorporation of an adjustment. Other EPA analyses have shown that increases in real
income over time lead to increases in WTP for a wide range of health effects and some
Final Risk and Exposure Assessment September 2009
Appendix 8-28
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
welfare effects, such as recreational visibility. This restriction contributes to an
underestimation of benefits.
2.2.2 Improvements in Total Ecosystem Services due to Increased Acid
Neutralizing Capacity Levels in Adirondack Lakes
To develop estimates of the overarching ecological benefits associated with reducing lake
acidification levels in Adirondacks National Park, researchers at Resources for the Future (RFF)
conducted a detailed contingent valuation (CV) survey (Banzhaf et al., 2006). Unlike other
valuation studies described in this report, the RFF study did not identify the specific categories of
ecosystem services that would be enhanced by improving aquatic conditions. Rather, the survey
described and elicited values for specific improvements in acidification-related water quality and
ecological conditions in Adirondack lakes. For this reason, and because the survey was
administered to a random sample of New York households, in this section the benefit estimates
from the RFF study are interpreted as measures that incorporate values for all ecosystem services
adversely affected by lake acidification.
In this section, the RFF study results were adapted and transferred to estimate the
ecological benefits of the zero-out scenario for Adirondack lakes. The fundamental benefit
transfer model can be summarized as follows:
AeeB -WTP *N *A%/7 (2 ]}
^-66-^lAdr ~" ±± Adr ^ NY ^/OIJ^ , V^-1^
where
= aggregate annual benefits (in 2007 dollars) to New York households in 2010
due to lake ecosystem improvements resulting from the zero-out scenario,
= average annual household WTP (in 2007 dollars) per unit of long-term
change in the percentage of Adirondack lakes impaired by acidification,
A/NY = projected total number of households in New York in 2010, and
A%7L = long-term change in the percentage of Adirondack lakes impaired by
acidification as a result of the zero-out scenario.
To develop estimates of WTPAdr, the estimates from the RFF study were used with results
reported in Banzhaf et al. (2006). The CV survey for the study was distributed to a random
sample of nearly 6,000 New York residents in 2003 to 2004 through the Internet and mail. As
part of the design and development of the survey instrument, experts were interviewed on the
Final Risk and Exposure Assessment September 2009
Appendix 8 - 29
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
ecological damages, and a summary of the science was used as the foundation for the description
of the park's existing condition and the hypothetical changes to be valued. The scientific review
indicated that there was significant uncertainty regarding the future status of lakes in the Park in
the absence of specific programs to improve lake acidification conditions. To bracket the range
of uncertainty in the science as well as to test the sensitivity of respondents' WTP to the scale of
ecological improvements, two versions of the survey instrument were developed and randomly
administered to separate subsamples.
Table 2.2-8 summarizes key features of the two survey versions. In both survey versions,
respondents were provided with information on the current (circa 2004) condition of the 3,000
lakes in the Park. Both versions describe half (1,500) of them as "lakes of concern" (i.e.,
unhealthy lakes where "fish and other aquatic life have been reduced or eliminated because of air
pollution in the past"), and both versions propose policies that would improve the lakes over a
period of 10 years (using lime to neutralize the excess acidity).
Table 2.2-8. Comparison of Resources for the Future Contingent Valuation Scenarios and EPA
Zero-Out Scenario
Percentage
of Adirondack
Current No Program"
(A) (B)
RFF "Base" Scenario
RFF "Scope" Scenario
Year = 2004
50%
Year = 2004
50%
50%
55%
Lakes that Are "
Future
Unhealthy"
With Program15 Reduction
(C) (B)-(C)
Year = 20 14
30%
Year = 20 14
10%
20%
45%
EPA "Zero-Out" Scenario
ANC Threshold
20 (ieq/L
50 (ieq/L
100 (ieq/L
Year = 20 10
22%
43%
79%
22%
42%
77%
Year = 2020
0%
11%
51%
22%
31%
26%
a Business-as-usual conditions.
bLake liming program for the RFF survey scenarios and a zero-out policy for the EPA scenario.
Final Risk and Exposure Assessment
Appendix 8-30
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
The "base" version of the survey asserts that, in the absence of any direct policy
intervention, the condition of the 1,500 unhealthy lakes and 1,500 healthy lakes is expected to
remain unchanged over the next 10 years. However, if a liming program is undertaken, it would
improve 20% (600) of the lakes in the Park relative to their expected 2014 condition without the
program.
In contrast, the "scope" version describes a gradually worsening status quo without the
liming program, in which 5% (150) of the healthy lakes are expected to gradually become
unhealthy. In other words, without the program, 55% (1,650) of the lakes would be unhealthy in
2014. With the liming program, however, only 10% of the lakes would be unhealthy in 2014, so
the program improves 45% (1,350) of the lakes relative to their expected 2014 condition without
the program.
Although scientific evidence indicates that a liming policy would not significantly
improve the condition of birds and forests, pretesting of the survey indicated that respondents
nonetheless tended to assume that these other benefits would occur. Therefore, to make the
scenarios more acceptable to respondents, other nonlake effects were added to the two survey
versions. In the base case, the red spruce (covering 3% of the forests' area) and two aquatic bird
species (common loon and hooded merganser) are said to be affected. In this version, the health
of birds and forests is described as unchanged in the absence of intervention, and minor
improvements are said to result from the program. In the scope version, a broader range of
damages is associated with acid rain—two additional species of trees (sugar maple and white
ash, all together covering 10% of forest area) and two additional birds (wood thrush and tree
swallow) are said to be affected. The scope version describes a gradually worsening status quo
along with large improvements due to the program.
Each respondent was presented with one of these (base or scope) policy scenarios and
then asked how they would vote in a referendum on the program, if it were financed by an
increase in state taxes for 10 years. To estimate the distribution of WTP, the annual tax amounts
were randomly varied across respondents.
Based on a detailed analysis of the survey data, Banzhaf et al. (2006) defined a range of
best WTP estimates, which were converted from 10-year annual payments to permanent annual
payments using discount rates of 3% and 5%. For the base version, the best estimates ranged
Final Risk and Exposure Assessment September 2009
Appendix 8-31
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
from $48 to $107 per year per household (in 2004 dollars), and for the scope version they ranged
from $54 to $154.
To specify values for WTPAdr, these estimates were converted to 2007 dollars using the
CPI and each of them was divided by the corresponding change in the percentage of lakes that
are unhealthy (20% for the base version and 45% for the scope version). For the base version, the
WTPAdr estimates range from $2.63 to $5.87 per percentage decrease in unhealthy lakes, and for
the scope version they range from $1.32 to $3.76.
To estimate NNY, the Census population projection for New York for 2010 was used,
which is 19.26 million people, and this amount was divided by the ratio of population size to the
number of households in New York (2.69) in the year 2000 (assuming that this ratio stays
constant from 2000 to 2010).
Finally, to estimate A%7L the MAGIC model results reported in Table 2.2-2 were used,
and it was assumed that the distribution of ANC levels for these 44 lakes is representative of all
3,000 lakes in the Adirondacks Park. For each of the three ANC thresholds, column (A) of Table
2.2-8 reports the estimated percentage of "unhealthy" (below the ANC threshold) lakes in 2010.
In columns (B) and (C), it also reports the percentage of unhealthy lakes in 2020 for the
reference and zero-out conditions, respectively. In 2020, the reduction in the percentage of lakes
that are unhealthy in the zero-out condition compared to the reference condition is 22% for the
20 ueq/L threshold. For the 50 ueq/L, and 100 ueq/L thresholds, it is 31% and 26%, respectively.
These 3% reduction values were used as the main estimates of A%IL.
To estimate aggregate benefits for the zero-out scenario using the RFF survey results, it is
important to use the results from the survey version that most closely match this scenario. Table
2.2-8 provides direct comparisons of the percentage of lakes that are defined as unhealthy under
the different conditions and scenarios. Although both RFF survey versions use 2004 as the
"current" year instead of 2010, they both use a 10-year horizon, which corresponds to the zero-
out scenario. Although no direct matches exist, the closest correspondence is between the zero-
out scenario assuming a 50 ueq/L threshold and the RFF scope survey. Under current and future
conditions with no additional policy interventions, the RFF scope scenario assumes a small
increase in unhealthy lakes from 50% to 55%, whereas the 50 ueq/L threshold is expected to
result in a small decrease from 43% to 42%. With the program, the RFF scope survey describes a
45% decrease in unhealthy lakes, whereas the zero-out scenario projects a 31% decrease.
Final Risk and Exposure Assessment September 2009
Appendix 8-32
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Moreover, although the RFF survey does not specify ANC thresholds, the survey's description of
unhealthy lakes is arguably closest to what the science defines for a 50 ueq/L threshold (as
summarized in Figure 2.2-1).
2.2.2.1 Results: Aggregate Benefits from Reduced Acidification in Adirondack Lakes
Table 2.2-9 reports the aggregate benefit estimates for the zero-out scenario using the 50
ueq/L threshold. As described above, the projected long-term decrease in the percentage of
unhealthy lakes (A%7L) for this scenario is 31%. Using the range of WTPAdr values from the RFF
scope survey and the projected number of New York households in 2010 and applying
Table 2.2-9. Aggregate Benefit Estimates for the Zero-Out Scenario
ANC
Threshold
20 (ieq/L
50 (ieq/L
100 (ieq/L
Reduction in
Percentage of
Unhealthy
Lakes
/±%IL
22%
31%
26%
Range of Average
Household WTP per
Percentage
Reduction
WTPAdr
$2.63 $5.87
$1.32 $3.76
$2.63 $5.87
Number of NY
Households
(in millions)
NNY
7.162
7.162
7.162
Range of Aggregate
Benefits
(in millions of 2007
dollars)
AggBAdr
$410.6 $916.4
$291.2 $829.4
$491.6 $1,097.2
Equation (2.1), the aggregate annual benefits of the zero-out scenario are estimated to
range from $291 million to $829 million.
Table 2.2-9 also reports aggregate benefit estimates for the zero-out scenarios using the
20 ueq/L and 100 ueq/L thresholds for ANC. Neither of these scenarios corresponds well with
the baseline descriptions of either the base or scope version of the RFF survey. The baseline
percentage of unhealthy lakes using the 20 ueq/L threshold (22%) is much lower than in either
the survey version. In contrast, the percentage using the 100 ueq/L threshold (77%) is much
higher. Nevertheless, the future reductions in the percentage of unhealthy lakes (22% and 26%)
are closest to the reductions described in the base version of the RFF survey. Therefore, the
aggregate benefits of the zero-out scenario with these thresholds are evaluated using the range of
WTPAdr values from the RFF base survey. With the 20 ueq/L threshold, the aggregate benefits are
estimated to range from $411 million to $916 million per year. With the 100 ueq/L threshold, the
aggregate benefits are estimated to range from $492 million to $1.1 billion per year.
Final Risk and Exposure Assessment
Appendix 8-33
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2.2.2.2 Limitations and Uncertainties
The benefit transfer model summarized in Equation (2.1) estimates the aggregate benefits
to New York households in 2010 due to lake ecosystem improvements resulting from the zero-
out scenario. To do this, estimates from two different studies were linked and combined. The
measures of improvements in lake ecosystems were obtained from the MAGIC model (as
described in Table 2.2-2), and the value estimates were obtained from the RFF survey study.
Uncertainties are associated with the estimates drawn from each study, and additional
uncertainties arise when these estimates were combined in the analysis. Some of these main
uncertainties and limitations are described below.
First, uncertainties are associated with extrapolating results from the 44 MAGIC-modeled
lakes to all (roughly 3,000) Adirondack lakes. The 44 modeled lakes are drawn from a larger,
randomly drawn sample of lakes; however, the representativeness of these 44 lakes for the
Adirondacks as a whole is uncertain.
Second, the time frame required for the zero-out scenario to match 1860 conditions is
uncertain. It was assumed that it takes 10 years for lakes to fully adjust to the reductions in
nitrogen and sulfur deposition and that conditions equivalent to "background" 1860 conditions
are achieved in 2020. The present value and annualized benefits would be lower if a longer time
frame were assumed.
Third, there is also some uncertainty related to the exact types of ecosystem services that
are included in these RFF study values, particularly regarding provisioning and regulating
services, which survey respondents may have been less likely to consider when formulating
responses to the CV questions. Importantly though, the values estimated by the RFF study are
likely to include (1) recreational fishing services, which means they cannot be added to the RUM
results, and (2) other cultural services, in particular recreational and nonuse services.
Fourth, the inclusion of other ecosystem changes (trees, birds, etc.) in the RFF CV survey
scenarios implies that respondents' stated values will overstate WTP for just changes in lake
acidification. This feature, therefore, contributes to potential overestimation of benefits.
Fifth, the lack of direct correspondence between the RFF CV scenarios and the zero-out
scenario requires assumptions for making the benefit transfer. In particular, baseline and future
levels (percentage of unhealthy lakes) are very different from those in the RFF survey if one uses
a 20 or 100 ANC threshold. Although the percentage changes are reasonably close to the RFF
Final Risk and Exposure Assessment September 2009
Appendix 8-34
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
20% and 45% decline scenarios, they are not exact and may not be applicable when applied to a
different baseline (something that was not specifically tested in the CV survey). Reseating the
WTP estimates for different percentage changes in unhealthy lakes also requires the somewhat
strong assumption that there is a constant WTP per percentage decline in unhealthy lakes.
Finally, the reported results only apply to Adirondack lakes and to New York residents.
The Adirondack region is more sensitive to acidity in contrast to many other areas of New York
State, which have calcium-rich limestone deposits that neutralize the acid. The bedrock soil and
shallow soil deposits have a lower buffering capacity. These geological factors together with
high and acidic precipitation levels contribute to the vulnerability of this region to acidification.
The uniqueness of the Park makes simple extrapolations of ecological conditions and human
values to other lakes very uncertain. Similarly, residents of other states are likely to value
improved ecosystem services from Adirondack lakes, but the magnitude of these values is
difficult to assess and, therefore, not included in the reported benefit estimates.
2.3 REFERENCES
Banzhaf, S., D. Burtraw, D. Evans, and A. Krupnick. 2006. "Valuation of Natural Resource
Improvements in the Adirondacks." Land Economics 82:445-464.
Chen, L., and C.T. Driscoll. 2004. "Modeling the Response of Soil and Surface Waters in the
Adirondack and Catskill Regions of New York to Changes in Atmospheric Deposition
and Historical Land Disturbance." Atmospheric Environment 38:4099-4109.
Driscoll, C.T. et al. 2003. "Effects of Acidic Deposition on Forest and Aquatic Ecosystems in
New York State" Environmental Pollution 123:327-336.Driscoll, C.T., G.B. Lawrence,
AJ. Bulger, TJ. Butler, C.S. Cronan, C. Eagar, K.F. Lamber, G.E. Likens, J.L. Stoddard,
and K.C. Weathers. 2001. Acidic Deposition in the Northeastern United States: Sources
and Inputs, Ecosystem Effects and Management Strategies. BioScience 51:180-198.
Greene, W.H. 2003. Econometric Analysis, 5th Ed. New Jersey: Prentice Hall.
Hutchison, R., and C.E. Kraft. 1994. "Hmong Fishing Activity and Fish Consumption." Journal
of Great Lakes Research 20(2):471-487.
Final Risk and Exposure Assessment September 2009
Appendix 8-35
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Industrial Economics, Inc. June 2008. "The Economic Impact of the Clean Air Interstate Rule on
Recreational Fishing in the Adirondack Region of New York State." Prepared for the
Clean Air Markets Division, Office of Air and Radiation, U.S. EPA.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number EVIDE-02-0070.
Kennedy, P. 2003. A Guide to Econometrics, pp. 312-313. Cambridge, MA: MIT Press.
Millennium Ecosystem Assessment (MEA). 2005. Ecosystems and Human Well-being:
Wetlands and Water. Synthesis. A Report of the Millennium Ecosystem Assessment.
Washington, DC: World Resources Institute.
Montgomery, M. and M. Needelman. 1997. "The Welfare Effects of Toxic Contamination in
Freshwater Fish." Land Economics 73(2):211-223.
Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Diedrich, H.A. Anderson, P.L.
Remington, and T.B. Sheffy. 1994. "Fish Consumption Patterns and Blood Mercury
Levels in Wisconsin Chippewa Indians." Archives of Environmental Health 49(l):53-58.
U.S. Department of the Interior, Fish and Wildlife Service, and U.S. Department of Commerce,
U.S. Census Bureau. 2007. 2006 National Survey of Fishing, Hunting, and Wildlife-
Associated Recreation.
U.S. Environmental Protection Agency (EPA), Office of Air and Radiation. November 1999.
The Benefits and Costs of the Clean Air Act 1990 to 2010: EPA Report to Congress.
EPA-410-R-99-001. Washington, DC: U.S. Environmental Protection Agency.
Final Risk and Exposure Assessment September 2009
Appendix 8-36
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
3.0 TERRESTRIAL ACIDIFICATION
Terrestrial acidification is the result of natural processes and anthropogenic sources of
acidic deposition. Elevated levels of atmospheric deposition of nitrogen and sulfur can alter the
chemical composition of soils by accelerating rates of base cation (e.g., calcium and magnesium)
leaching, which depletes available plant nutrients, and by mobilizing and leaching aluminum,
which can be toxic to tree roots. Consequently, among the most visible and significant effects of
acid deposition are damages to forest health and resulting reductions in tree growth.
Evidence of adverse effects due to terrestrial acidification is particularly strong for two
common tree species in the northeastern United States where levels of nitrogen and sulfur
deposition have historically been relatively high—sugar maples and red spruce. Therefore, the
discussion of ecosystem service effects focuses on these two species; however, more widespread
impacts that include other tree species are also possible.
3.1 OVERVIEW OF AFFECTED ECOSYSTEM SERVICES
The existing ecosystem services that are primarily affected by the terrestrial acidification
resulting from nitrogen and sulfur deposition are described and, to the extent possible, quantified
using the classification system outlined in Section 1.
3.1.1 Provisioning Services
Forests in the northeastern United States provide several important and valuable
provisioning services, which are reflected in measures of production and sales of tree products.
Sugar maples (also referred to as hard maples) are a particularly important commercial
hardwood tree species in the United States. As shown in Figure 3.1-1, the main range of the
sugar maple covers most of the United States east of the Mississippi River and north of Alabama
and Georgia. This range is also the area with the highest levels of nitrogen and sulfur deposition
in the country, according to monitored estimates from the National Atmospheric Deposition
Network (NADP) and modeled estimates from the Community Multiscale Air Quality (CMAQ)
modeling system.
The two main types of products derived from sugar maples are wood products and maple
syrup. The wood from sugar maple trees is particularly hard, and its primary uses include
Final Risk and Exposure Assessment September 2009
Appendix 8-37
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
construction, furniture, and flooring (Luzadis and Gossett, 1996). According to data from the
U.S. Forest Service's National Forest Inventory and Analysis (FIA) databases
(http://199.128.173.26/fido/mastf/index.html), in 2006, the total removal of sugar maple saw
timber from timberland in the United States was almost 2.12 million cubic meters. Assuming an
average price of $169.5 per cubic meter, the total value of these removals in 2006 was
approximately $358 million.
__] Sugar Maple Distribution
Combined N and S
Value (kgrtia/yr)
• High 150.0
M Low 10
Source of Deposition 2002 CMAQ hybrid dry deposition plus 2002 NADP wet deposition
Source of Sugar Maple Distribution United Slates Forest Service 2002-2003
Figure 3.1-1. Combined Nitrogen and Sulfur Deposition (from 2002 CMAQ Dry
Deposition and NADP Wet Deposition Estimates) and the Range of Sugar Maple
in the United States
During winter and early spring (depending, in part, on location and diurnal temperature
differences), sugar maple trees also generate sap that is used to produce maple syrup. From 2005
to 2007, annual production of maple syrup in the United States varied between 1.2 million and
5.3 million liters, which accounted for roughly 19% of worldwide production. The total annual
value of U.S. production in these years varied between $157 million and $168 million (National
Agricultural Statistics Service [NASS], 2008).
Red spruce is a common commercial softwood species. As shown in Figure 3.1-2, its
range in the United States is much more limited than the sugar maple's range, but it also
primarily grows in areas with relatively high levels of nitrogen and sulfur deposition. Red spruce
Final Risk and Exposure Assessment
Appendix 8-38
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
is now mainly found in northern New England, New York, and in a few high-elevation areas of
the Appalachian Mountain range. Wood from red spruce is used in a variety of products
including lumber, pulpwood, poles, plywood, and musical instruments. According to FIA data, in
2006, the total removal of red spruce saw timber from timberland in the United States was 0.77
million cubic meters. Assuming an average price of $42.37 per cubic meter, the total value of
these removals in 2006 was approximately $33 million.
^J Red Spruce Distribution
Combined N and S
Value (kg/ha/yr)
• High 1500
_, Low: 1.0
Source of Deposition 2002 CMAQ hybrid dry deposition plus 2002 NADP wet deposition
Source of Red Spruce Distribution. United States Forest Service 3002-2006
Figure 3.1-2. Combined Nitrogen and Sulfur Deposition (from 2002 CMAQ Dry
Deposition and NADP Wet Deposition Estimates) and the Range of Red Spruce
in the United States
Figure 3.1-3 shows and compares the value of annual production of sugar maple and red
spruce wood products and of maple syrup in 2006. Across states in the northeastern United
States, sugar maple timber harvests consistently generated the highest total sales value of the
three products. Although total sales of red spruce saw timber and maple syrup were of roughly
the same magnitude in the United States as a whole, the red spruce harvest was concentrated in
Maine, whereas maple syrup production was largest in Vermont and New York.
3.1.2 Cultural Services
Forests in the northeastern United States are also an important source of cultural
ecosystem services—in particular recreational and aesthetic services. Forestlands support a wide
Final Risk and Exposure Assessment
Appendix 8-39
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
variety of outdoor recreational activities, including fishing, hiking, camping, off-road driving,
hunting, and wildlife viewing. Regional statistics on recreational activities that are specifically
forest based are not available; however, more general data on outdoor recreation provide some
insights into the overall level of recreational services provided by forests. For example, most
recent data from the National Survey on Recreation and the Environment (NSRE) indicate that,
from 2004 to 2007, 31% of the U.S. adult (16 or older) population visited a wilderness or
primitive area during the previous year, and 32% engaged in day hiking (Cordell et al., n.d.).
From 1999 to 2004, 16% of adults in the northeastern United States1 participated in off-road
vehicle recreation, for an average of 27 days per year (Cordell et al., 2005). Using the meta-
analysis results reported by Kaval and Loomis (2003), which found that the average consumer
surplus value per day of off-road driving in the United States was $25.25 (in 2007 dollars), the
implied total annual value of off-road driving recreation in the northeastern United States was
more than $9.25 billion.
Figure 3.1-3. Annual Value of Sugar Maple and Red Spruce Harvests and Maple
Syrup Production, 2006
1 This area includes Connecticut, Delaware, District of Columbia, Illinois, Indiana, Maine, Maryland,
Massachusetts, Michigan, New Hampshire, New Jersey, New York, Ohio, Pennsylvania, Rhode Island, Vermont,
West Virginia, and Wisconsin.
Final Risk and Exposure Assessment
Appendix 8-40
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
State-level data on other outdoor recreational activities associated with forests are also
available from the 2006 FHWAR (U.S. Department of the Interior [DOT], 2007). As summarized
in Table 3.1-1, 5.5% of adults in the northeastern United States participated in hunting, and the
total number of hunting days occurring in those states was 83.8 million. Data from the survey
also indicated that 10% of adults in northeastern states participated in wildlife viewing away
from home. The total number of away-from-home wildlife viewing days occurring in those states
was 122.2 million in 2006. For these recreational activities in the northeastern United States,
Kaval and Loomis (2003) estimated average consumer surplus values per day of $52.36 for
hunting and $34.46 for wildlife viewing (in 2007 dollars). The implied total annual value of
hunting and wildlife viewing in the northeastern United States was, therefore, $4.38 billion and
$4.21 billion, respectively, in 2006.
As previously mentioned, it is difficult to estimate the portion of these recreational
services that are specifically attributable to forests and to the health of specific tree species.
However, one recreational activity that is directly dependent on forest conditions is fall color
viewing. Sugar maple trees, in particular, are known for their bright colors and are, therefore, an
essential aesthetic component of most fall color landscapes. Thus, declines in sugar maple stocks
due to terrestrial acidification are expected to have detrimental effects on these landscapes.
Statistics on fall color viewing are much less available than for the other recreational and tourism
activities; however, a few studies have documented the extent and significance of this activity.
For example, based on a 1996 to 1998 telephone survey of residents in the Great Lakes area,
Spencer and Holecek (2007) found that roughly 30% of residents reported at least one trip in the
previous year involving fall color viewing. In a separate study conducted in Vermont, Brown
(2002) reported that more than 22% of households visiting Vermont in 2001 made the trip
primarily for the purpose of viewing fall colors. Unfortunately, data on the total number or value
of these trips are not available, although the high rates of participation suggest that numbers
might be similar to the wildlife viewing estimates reported above.
Although these statistics provide useful indicators of the total recreational and aesthetic
services derived from forests in the northeastern United States, they do not provide estimates of
how these services are affected by terrestrial and forest acidification. Very few empirical studies
have directly addressed this issue; however, two studies have estimated values for protecting
high-elevation spruce forests in the Southern Appalachians. Kramer, Holmes, and Haefele (2003)
Final Risk and Exposure Assessment September 2009
Appendix 8-41
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
conducted a CV study estimating households' WTP for programs to protect remaining high-
elevation spruce forests from damages associated with air pollution and insect infestation
(Haefele, Kramer, and Holmes, 1991; Holmes and Kramer, 1995). The study collected data from
486 households using a mail survey of residents living within 500 miles of Asheville, North
Carolina. The survey presented respondents with photographs representing three stages of forest
decline and explained that, without forest protection programs high-elevation spruce forests
would all decline to worst conditions (with severe tree mortality). The survey then presented two
Table 3.1-1. Participation in Hunting and Wildlife Viewing in Northeastern States in 2006
Participation Rates by State
Residents"
State
Connecticut
Delaware
Illinois
Indiana
Maine
Maryland
Massachusetts
Michigan
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
West Virginia
Wisconsin
Total
Hunting
1.2%
3.1%
2.8%
5.3%
13.6%
3.5%
1.3%
9.2%
5.0%
1.3%
3.3%
5.4%
9.5%
1.2%
11.3%
13.6%
15.0%
5.5%
Wildlife Viewing15
11.0%
7.0%
8.0%
13.0%
20.0%
7.0%
11.0%
11.0%
12.0%
8.0%
8.0%
13.0%
11.0%
11.0%
16.0%
9.0%
10.0%
10.0%
Activity Days by Residents and
Nonresidents (in thousands)
Hunting
509
654
4,688
4,808
2,283
2,262
1,149
11,905
1,057
1,457
10,289
10,633
16,863
155
1,111
3,939
10,059
83,821
Wildlife Viewing15
4,184
855
5,686
24,013
4,778
4,782
8,461
10,043
3,165
7,965
13,521
7,816
11,972
2,948
2,459
4,005
5,547
122,200
a Ages 16 or older.
b Wildlife viewing away from home.
Source: U.S. Department of the Interior (DOI), Fish and Wildlife Service, and U.S. Department of
Commerce, U.S. Census Bureau, 2007.
Final Risk and Exposure Assessment
Appendix 8 - 42
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
potential forest protection programs—one would prevent further decline in forests along roads
and trail corridors (one-third of the at-risk ecosystem) and the other would prevent decline in all
at-risk forests. Both programs would be funded by tax payments going to a conservation fund.
Median household WTP was estimated to be roughly $29 (in 2007 dollars) for the first program
and $44 for the more extensive program.
Jenkins, Sullivan, and Amacher (2002) conducted a very similar study in 1995 using a
mail survey of households in seven Southern Appalachian states. In this study, respondents were
presented with one potential program, which would maintain forest conditions at initial (status
quo) levels. It was explained that, without the program, forest conditions would decline to worst
conditions (with 75% dead trees). In contrast to the previously described study, in this survey the
initial level of forest condition was varied across respondent. In one version of the survey, the
initial condition was described and shown as 5% dead trees, while the other version described
and showed 30% dead trees. Household WTP was elicited from 232 respondents using a
dichotomous choice and tax payment format. The overall mean annual WTP for the forest
protection programs was $208 (in 2007 dollars), which is considerably larger than the WTP
estimates reported by Kramer, Holmes, and Haefele (2003). One possible reason for this
difference is that respondents to the Jenkins, Sullivan, and Amacher (2002) survey, on average,
lived much closer to the affected ecosystem. Multiplying the average WTP estimate from this
study by the total number of households in the seven-state Appalachian region results in an
aggregate annual value of $3.4 billion for avoiding a significant decline in the health of high-
elevation spruce forests in the Southern Appalachian region.
3.1.3 Regulating Services
Forests in the northeastern United States also support and provide a wide variety of
valuable regulating services, including soil stabilization and erosion control, water regulation,
and climate regulation (Krieger, 2001). As terrestrial acidification contributes to root damages,
reduced biomass growth, and tree mortality, all of these services are likely to be affected;
however, the magnitude of these impacts is very uncertain. Forest vegetation plays an important
role in maintaining soils in order to reduce erosion, runoff, and sedimentation that can adversely
impact surface waters. In addition to protecting the quality of water in this way, forests also help
store and regulate the quantity and flows of water in watersheds. Finally, forests help regulate
Final Risk and Exposure Assessment September 2009
Appendix 8-43
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
climate locally by trapping moisture and globally by sequestering carbon. The total value of
these ecosystem services is very difficult to quantify in a meaningful way, as is the reduction in
the value of these services associated with nitrogen and sulfur deposition.
3.2 CHANGES IN ECOSYSTEM SERVICES ASSOCIATED WITH
ALTERNATIVE LEVELS OF ECOLOGICAL INDICATORS
This section estimates values for changes in ecosystem services associated with
reductions in damages to commercial forests resulting from terrestrial acidification. With high
levels of acidifying nitrogen and sulfur deposition, trees may experience an increased
susceptibility to drought and pest damage, aluminum toxicity in roots, a reduced tolerance to
cold, and a greater propensity to frost injury (DeHayes et al., 1999; Driscoll et al., 2001; Fenn et
al., 2006). As a result, total stand volume and growth may be reduced. The tree growth response
and value of reducing nitrogen+sulfur deposition loads across the range of sugar maples and red
spruces (as shown in Figures 3.1-1 and 3.1-2, respectively) was estimated using a critical load
assessment methodology (described in the case study analysis) of terrestrial acidification. More
specifically, the beneficial effects of eliminating all exceedances of critical load for sugar maples
and red spruces in this range were estimated.
3.2.1 Increased Provisioning Services from Sugar Maple Timber Harvests due to
Elimination of Critical Load Exceedances
A three-stage approach was used to estimate the value of increased provisioning services
from sugar maple and red spruce timber harvests. In the first stage, exposure-response models2
for sugar maple and red spruce trees were estimated, which measure the empirical relationship
between exceedances of critical loads and growth in volume of live trees. In the second stage,
these exposure-response models were applied to estimate the average increase in sugar maple
and red spruce growth rates (in three regions) that would result from eliminating critical load
exceedances in the range of these tree species. In the third stage, these increased growth rates
were incorporated into an existing forest market model for North America and the value (i.e.,
See the case study report for alternative models of exposure-response.
Final Risk and Exposure Assessment September 2009
Appendix 8-44
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
increase in consumer and producer surplus) of expected future increases in sugar maple and red
spruce timber harvests and sales was estimated. Each of these stages is described below in detail.
3.2.1.1 Stage 1: Estimation of the Exposure-Response Model
The analysis of the relationship between critical load exceedances and sugar maple and
red spruce trees' growth was conducted using data from the USFS FIA database for 16 states in
the sugar maple range and 5 states in the red spruce range. Each data point in the analysis
corresponds with a permanent sampling plot location on classified forestland (timberland for
New York) covering 0.07 ha. Estimation of critical loads for each plot was based on a Simple
Mass Balance (8MB) modeling approach (described in the case study). FIA plots were only
included in the analysis if they (1) were nonunique3 permanent sampling plots; (2) provided
necessary soil, parent material, atmospheric deposition, and run-off data to apply the 8MB model
for critical load estimation; (3) were located to the north of the glaciation line (this line
represents the southernmost extension of the most recent glacial advancement)4; and (4) had
positive exceedances in deposition above the most protective critical load (Bc/Al = 10.0).5 With
these restrictions, 2,205 sugar maple plots and 187 red spruce plots were included in the analysis.
Tables 3.2-1 and 3.2-2 summarize the plot-level FIA sugar maple and red spruce data
used to model the exposure-response relationship. For each plot, exceedances of critical loads
were calculated by subtracting the results of the 8MB analysis (estimated critical loads
estimates) from corresponding 2002 CMAQ nitrogen+sulfur deposition estimates. Overall, 74%
of sugar maple plots above the glaciation line exceeded the critical loads, ranging from 21% in
Maine to over 95% in Connecticut, Massachusetts, New Jersey, New York, Pennsylvania, Ohio,
and Vermont. Thirty-one percent of red spruce above the glaciation line exceeded the critical
loads, ranging from 16% in Maine to 100% in Massachusetts and Vermont. For sugar maple
3 Nonunique permanent sampling plot locations are those that have maximum critical load attribute values (soils,
runoff, and atmospheric deposition) that are not distinct and are repeated within a 250-acre area of the plot
location. This "confidentiality" filter is a requirement of the USFS to prevent the disclosure of data that can be
directly linked to a location on private land. To comply with the necessary "confidentiality," full coverages of the
data required for the critical load calculations were given to the USFS, and the USFS matched and provided the
data for each nonunique permanent sampling plot.
4 This is because the base cation weathering term, one of the key components of critical load, may not have been
accurately estimated for plots from south of the glaciation line (see case study for details) that have older, more
weathered soils. Thus, using such plots in the analysis may potentially increase error in the data used.
5 For analysis using lower levels of protection, see the case study report.
Final Risk and Exposure Assessment September 2009
Appendix 8-45
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
plots with positive exceedances, the average exceedance ranged from less than 100 eq/ha/yr in
Missouri and Iowa to over 450 eq/ha/yr in Connecticut, Massachusetts, New Jersey, and Ohio.
For red spruce plots with positive exceedances, the average exceedance ranged from less than
150 eq/ha/yr in Maine to over 600 eq/ha/yr in Massachusetts.
Net annual individual tree volume growth and tree volumes for all live sugar maple and
red spruce trees6 (greater than 12.7 cm diameter at 1.3 m) were acquired from the USFS FIA
database for each plot. The volume growth calculations were based on the most recent
measurement period, and the time interval between measurements for the plots (to determine
annual growth rates) ranged from 1 to 11 years. These calculations included the influences of
growth and volume reductions or losses due to natural damage (pest, wind, frost) or natural
mortality. Average volume growth ranged from 0.1 in Massachusetts to 0.62 in Indiana for sugar
maple and from 0.14 in Massachusetts to 0.26 in Maine for red spruce. Volumes and volume
growth measures for the sugar maple and red spruce trees in each plot were averaged to produce
single values of each measure for each species.
Table 3.2-1. Summary of Plot-Level Data on Sugar Maple Growth and Exceedances (for Plots
above the Glaciation Line)
State
Alabama
Arkansas
Connecticut
Illinois
Indiana
Iowa
Kentucky
Maine
Maryland
Massachusetts
Total
Number
of Plots
12
8
33
25
266
8
14
242
4
27
Number of
Plots North
of
Glaciation
Line
0
0
33
20
234
8
0
242
0
27
Number of Number of
Plots with Plots with
Positive Positive CL
CL Exceedance
Exceed-
ance
Values
3
1
33
17
235
2
12
51
4
27
Values North
of Glaciation
Line
33
12
204
2
51
27
Average
CL
Exceed-
ance
(eq/ha/yr)
487.76
117.17
390.90
48.07
130.42
473.31
Average
Tree
Volume
Growth
(m3/yr)
0.009
0.007
0.018
0.005
0.011
0.003
Average
Tree
Volume
(m3)
0.279
0.227
0.397
0.123
0.323
0.366
6 All trees with reported volumes of "0" were excluded from the analyses.
Final Risk and Exposure Assessment
Appendix 8-46
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Number of Number of
Plots with Plots with
Number of Positive Positive CL
Plots North CL
State
Michigan
Minnesota
Missouri
New Hampshire
New Jersey
New York
North Carolina
Ohio
Pennsylvania
Tennessee
Vermont
Virginia
West Virginia
Wisconsin
TOTAL Observations
(used in calculations)
CL = critical load
Table 3.2-2. Summary
Total
Number
of Plots
596
257
122
72
6
280
13
55
270
264
162
104
337
870
4,047
of
Glaciation
Line
596
257
31
72
6
280
0
27
133
0
162
0
0
870
2,998
of Plot Level Data
Exceed-
ance
Values
418
79
58
60
6
264
9
54
263
132
160
63
318
719
2,988
Exceedance
Values North
of Glaciation
Line
418
79
18
60
6
264
26
126
160
719
2,205
Average
CL
Exceed-
ance
(eq/ha/yr)
242.17
156.06
84.02
378.40
601.39
437.94
452.60
387.35
301.67
185.16
2,205
Average
Tree
Volume
Growth
(m3/yr)
0.011
0.010
0.012
0.009
0.013
0.010
0.013
0.011
0.008
0.009
2,205
on Sugar Maple Growth and Exceedances (for
Average
Tree
Volume
(m3)
0.307
0.256
0.246
0.304
0.357
0.344
0.545
0.366
0.411
0.304
2,205
Plots
above the Glaciation Line)
State
Maine
Massachusetts
New Hampshire
New York
Tennessee
Vermont
West Virginia
TOTAL Observations
(used in calculations)
Total
Number
of Plots
483
3
42
18
1
60
6
613
Number
of Plots
North of
Glacia-
tion Line
483
3
42
18
0
60
0
606
Number of
Plots with
Positive
CL
Exceed-
ance
Values
78
3
32
14
1
60
6
194
Number of
Plots with
Positive CL
Exceedance
Values North
of Glaciation
Line
78
3
32
14
0
60
0
187
Average
CL
Exceed-
ance
(eq/ha/yr)
133.1048
628.5439
368.9527
282.5433
292.4329
187
Average
Tree
Volume
Growth
(m3/yr)
0.007
0.004
0.006
0.004
0.007
187
Average
Tree
Volume
(m3)
0.245
0.203
0.245
0.221
0.328
187
Final Risk and Exposure Assessment
Appendix 8-47
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
The results of a multivariate OLS regression, using average tree growth (measured in
cubic meters per year) as the dependent variable, are reported in Table 3.2-3 and 3.2-4. The
explanatory variables include the critical load exceedance (measured in eq/ha/year) for each plot,
linear and squared terms of average tree volumes (measured in cubic meters), and a categorical
(dummy) variable for each state (with Connecticut as the reference category for sugar maple and
Vermont for red spruce). The purpose of the state variables is to control for other unobserved
sources of variation in tree growth, which are related to a plot's general geographic location.
Examples of potential unobserved factors include differences in data collection methods and
measurements across reporting state, climatic factors, and geological characteristics. An F test
applied to the state categorical variables indicated that their coefficients are jointly significant at
the 5% level for sugar maple. In general, the growth of a tree rises with age but at a decreasing
rate. Because data on the age were unavailable, average tree volume was instead included as a
proxy variable in the regression to control for this relationship.
The coefficient of the critical load exceedance was negative for both species and was
statistically significant at the 5% level (p-value of 0.035) for red spruce and at the 10% (p-value
of 0.101) for sugar maple, thus supporting the theory that when critical loads are exceeded by
atmospheric nitrogen and sulfur deposition, tree health and growth can be impaired.
Table 3.2-3. Linear Exposure-Response Model for Exceedances (above a Critical Load
Calculated with Bc/Al = 10) and Sugar Maple Tree Growth: OLS Regression Results (for Plots
above the Glaciation Line)
Dependent
Variable:
Average Tree
Growth (m3/yr)
Explanatory Variables
Intercept
Critical load exceedance
Average tree volume
Square of average tree volume
Illinois
Indiana
Iowa
Maine
Massachusetts
Michigan
Coefficient
0.004875
-3.344E-06
0.021150
8.944E-04
-0.001884
0.005452
-0.002052
-0.000895
-0.008403
0.000222
t-statistic
1.48
-1.64
10.12
1.1
-0.31
1.63
-0.16
-0.22
-1.82
0.07
p-value
0.1385
0.1008
<0001
0.27
0.755
0.1029
0.8743
0.8245
0.0685
0.9456
Final Risk and Exposure Assessment
Appendix 8-48
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Dependent
Variable:
Average Tree
Growth (m3/yr)
Explanatory Variables
Minnesota
Missouri
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Vermont
Wisconsin
Number of Observations
Adjusted R2
Coefficient
0.000210
0.001850
-0.001647
0.001956
-0.000817
-0.002104
-0.000803
-0.005168
-0.002195
2,205
0.1722
t-statistic
0.06
0.35
-0.43
0.25
-0.25
-0.45
-0.23
-1.51
-0.68
p-value
0.9553
0.7255
0.6696
0.8042
0.8035
0.6522
0.8177
0.131
0.4958
3.2.1.2 Stage 2: Estimation of Average Increments in Tree Volume
In this stage of the analysis, the effect of eliminating all critical load exceedances in the
range of sugar maples and red spruce was simulated and the resulting average (at a region level)
percentage increase in tree volume was estimated.
Table 3.2-4. Linear Exposure-Response Model for Exceedances (above a Critical Load
Calculated with Bc/Al = 10) and Red Spruce Tree Growth: OLS Regression Results (for Plots
above the Glaciation Line)
Dependent Variable:
Average Tree
Growth (m3/yr)
Explanatory Variables
Intercept
Critical load exceedance
Tree volume
Square of tree volume
Maine
Massachusetts
New Hampshire
New York
Number of Observations
Adjusted R2
Coefficient
0.006034
-5.162E-06
0.005590
5.100E-03
0.000285
-0.000132
0.000435
-0.001805
187
0.1963
t-statistic
4.96
-2.12
1.26
1.23
0.32
-0.05
0.42
-1.32
p-value
<0001
0.0354
0.2093
0.2218
0.7489
0.9629
0.6736
0.1897
Final Risk and Exposure Assessment
Appendix 8-49
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Based on the results of the regression equation reported in Table 3.2-3 and 3.2-4, for each
plot /' with a positive critical loads exceedance, the following equation was first used to estimate
what tree volume growth would be under conditions with no critical loads exceedances:
(3.1)
where
CLEf = critical load exceedance at plot / under observed conditions
g° = annual tree volume growth on plot / under observed conditions (in m3/yr)
g] = annual tree volume growth on plot / under conditions with no exceedance of
critical load (CLEt = 0) (in m3/yr)
B = regression coefficient (slope) for critical load exceedance (from Table 3.2-3 and
3.2-4, equals -3.344E-06 for sugar maple and -5.162E-06 for red spruce)
Since this study calculated the effects of eliminating positive exceedances with the aim of
estimating reductions in damages to sugar maple and red spruce forests resulting from terrestrial
acidification, it was assumed that there is no change in growth for plots without positive critical
load exceedances. In practice, however, some reduced growth may be possible due to lower
nitrogen availability in plots that are below the critical load. Thus, the calculations should be
interpreted as an upper bound to the value of reducing nitrogen+sulfur deposition loads.
To apply these results in the market model used in the next stage of the analysis, these
volume growth estimates were then converted into an average percentage increment in volume.
In other words, in each period t, tree volume on plot / is expected to be greater (by a factor/)
under conditions with no critical loads exceedance, compared to conditions with currently
observed critical loads exceedances. In formal terms:
(g- I C-, Wl-i = (1 + (g\ I C-i )FS-i (3 .2)
where
Vft = average tree volume on plot / under observed conditions in period t (in m3)
Vl = average tree volume on plot / under conditions with no exceedance of critical load
in period t (in m3)
Final Risk and Exposure Assessment September 2009
Appendix 8-50
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Solving Equation (3.2) for/ results in
Using the plot-level estimates of g1, g°, and V°,ft for each plot in the data set was
estimated, and these estimates were then averaged across each region.
3. 2. 1. 3 Stage 3: Estimation of Increased Market-Based Benefits from Sugar Maple and Red
Spruce Timber Harvests
The next critical step in establishing the link between changes in nitrogen and sulfur
deposition and the changes in forest provisioning services is modeling the effect of the average
increase in tree growth (obtained in Stage 2) on public welfare. This section describes the
approach to obtaining valuation estimates for this incremental increase in the volume of
commercial sugar maple and red spruce stands.7 To implement this approach, the increase in the
percentage volume of timber was applied to all age categories, and FASOMGHG (Forest and
Agricultural Sector Optimization Model — Green House Gas version) was used to calculate the
resulting market-based welfare effects in the forest and agricultural sectors of the United States.
Data obtained from the FIA were used as inputs into FASOMGHG, which enabled the
adaptation of the model for this application. The different components of these input data are
described below.
FASOMGHG is a price-endogenous, dynamic, nonlinear programming model of the
forest and agricultural sectors in the United States (Adams et al., 2005). The model simulates the
allocation of land over time to competing activities in these two sectors and the resultant
consequences for the commodity markets supplied by these lands. It was developed to evaluate
the welfare and market impacts of public policies that cause changes in land use and activities
both between and within the two sectors. The results from this model yield a dynamic simulation
of prices, production, management, consumption, greenhouse gas (GHG) effects, and other
environmental and economic indicators within these two sectors. For this application,
7 Holmes (1992) describes a similar approach to estimate welfare effects for a decline in southern pine forest
productivity in the United States.
Final Risk and Exposure Assessment September 2009
Appendix 8-51
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
FASOMGHG's key outputs include economic welfare measures, such as changes in producer
and consumer surplus.8
The following discussion summarizes the other main features of FASOMGHG and
describes how they were used and adapted for this application:
• Temporal Frame: The time frame of this model is typically 70 to 100 years, and the
model is solved on a 5-year time-step basis. The base year for this model is 2002.
• Geographical Regions: FASOMGHG models forest and agricultural activity across the
conterminous United States, which is broken into 11 market regions. Forestry production
occurs in nine of these regions. The selection of FASOMGHG regions for this model
application was determined by comparing maps showing the regions where sugar maples
grow with a list of FASOM regions. Table 3.2-5 lists the states in each of the FASOM
regions used in this application. It also shows the average increase in tree growth (obtained
from Stage 2) for each of these regions.
Table 3.2-5. Estimated Increments in Sugar Maple and Red Spruce Timber Volume (Resulting
from Elimination of Critical Load Exceedances), by FASOMGHG Region
Key Region
States
Average
Percentage
Increment in Sugar
Maple Tree
Volume
(i.e., average/)
Average Percentage
Increment in Red
Spruce Tree Volume
(i.e., average/)
NE Northeast Connecticut, Delaware, Maine,
Maryland, Massachusetts, New
Hampshire, New Jersey, New
York, Pennsylvania, Rhode
Island, Vermont, West Virginia
0.59%
0015%
LS
CB
Lake
States
Corn Belt
Michigan, Minnesota, Wisconsin
Illinois, Indiana, Iowa, Missouri,
Ohio
0.28%
0. 57%
Source: Adams et al., 2005.
• Types of Forests: Two types of forests are considered when evaluating policy effects in
FASOMGHG—softwood and hardwood. To adapt these categories for the application,
8 For a detailed documentation of FASOMGHG, please see Adams et al. (2005).
Final Risk and Exposure Assessment
Appendix 8-52
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
sugar maples and red spruce trees needed to be expressed as a proportion of hardwoods
and softwoods, respectively. This was done for each of the regions modeled in this
application. These relevant data were obtained from FIA (Table 3.2-6) and are a
component of the input into FASOMGHG.
Table 3.2-6. Proportions of Hardwood in Sugar Maple Production and Proportions of Softwood
in Red Spruce Production, by FASOM Region
Sugar Maple
Red Spruce
FASOM Regions
NE
LS
CB
NE
Proportion of Hardwood/Softwood
13%
11%
14.5%9
Source: U.S. Department of Agriculture, Forest Service, 2002.
• Forestland: The FASOMGHG model does not track land under forest cover that produces
less than 0.57 m3/yr (called unproductive forestland) or on timberland that is reserved for
other uses, because these are not a part of the U.S. timber base. Endogenous land use
modeling is only done for privately held land, not publicly owned or managed timberlands.
The model assumes that the amount of public land in forests does not adjust to market
conditions but is set by the government. Thus, the average percentage increase in volume
is applied to only forests growing on private land. The proportions of the timberland under
private and public ownership are shown in Table 3.2-7 (obtained from FIA data).
• Welfare Measure: Mathematically, FASOMGHG solves an objective function to
maximize net market surplus. This is represented by the area under the product demand
function (an aggregate measure of consumer welfare) less the area under the factor supply
curves (an aggregate measure of producer costs). The value of the resultant objective
function is consumers' and producers' surplus. The welfare effects of a productivity
improvement are obtained from FASOMGHG as the difference in annual net market
surplus between a base case (without the policy in place) and a control case (with the
policy in place).
9 The RPA Assessment tables report the proportion of the spruce and balsam fir category as 29%. We assume that
half of this is due to red spruce.
Final Risk and Exposure Assessment September 2009
Appendix 8-53
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
To apply FASOMGHG for this analysis, the main input required for the model is the
annual percentage increase in total hardwood and total softwood volume by region. To address
this requirement, the estimate of the average percentage increment in sugar maple tree volume
(average/, shown in Table 3.2-5) was multiplied by the proportion of hardwoods in sugar maple
production (shown in Table 3.2-6) for each FASOM region, which ranges from 11% to 13%.
Similarly, the estimate of the average percentage increment in red spruce tree volume (average
ft, shown in Table 3.2-5) was multiplied by the proportion of softwoods in red spruce production
(shown in Table 3.2-6) for the NE region.
Table 3.2-7. Proportion of Timberland under Private and Public Ownership by FIA Regiona: 2002
FIA Region Private Timberland Public Timberland
Northeast 87% 13%
NorthCentral 28% 72%
a The states in the Northeast FIA region correspond exactly to states in NE in FASOM.
The states in the NorthCentral FIA region correspond exactly to states in LS and CB in FASOM.
Source: U.S. Department of Agriculture, Forest Service, 2002, Table 10.
3.2.1.4 Results: Aggregate Benefit Estimates
Figure 3.2-1 summarizes the FASOMGHG model results. These results are reported as
the present discounted values of future welfare changes in the forestry sector (in 5-year
increments from 2000 to 2065) due to increased tree growth as well as the future welfare changes
in the agricultural sector. Summing over this 65-year period, the value of gains to the forestry
sector is $17.1 million (in 2007 dollars, using a 4% discount rate). The agricultural sector has a
welfare loss of $1 million. This loss is possibly due to a shift in land use from agriculture to
forestry. The total present value of these welfare changes due to both the sectors is $16.09
million (in 2007 dollars, using a 4% discount rate). On an annualized basis (at 4%), this is
equivalent to $684,000 per year. Figure 3.2-1 presents the time path of these welfare estimates
for the forestry and agricultural sector as well as the total welfare estimates. The cyclical pattern
of the estimates is most likely driven by the fact that if more harvesting is done in any period,
this leads to less stock to harvest from in the next period.
Final Risk and Exposure Assessment September 2009
Appendix 8-54
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Welfare Changes in Forestry and Agriculture
4000
~ 3000
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
The estimated reduction in the forest damages, as explained in Section 3.2.1.2, should be
interpreted as an upper bound on the benefits of reducing nitrogen+sulfur deposition, since
it only includes the gains from reducing critical load exceedances. Nitrogen deposition
below the critical load may actually promote tree growth through fertilization effects;
therefore, reducing deposition may have potential counteracting effects on tree growth.
The current analysis does not estimate or include these counteracting effects.
Although this analysis of tree growth response is done for sugar maple and red spruce,
gains are also expected for other commercial species. Thus, we are underestimating the
total benefits of reducing nitrogen+sulfur deposition.
Because of data limitations, the exposure-response analysis does not control for other
factors that may affect tree growth, such as elevation, slope, density (to account for
sunlight and competition among trees for nutrients), age (though tree volume was used as a
crude proxy for this variable), different management practices, and climate. Also,
differences in measurement and reporting across plots and states may result in
discrepancies in the data. Although this study attempted to capture the differences in
measurement and in climate by using state dummies, this is not a perfect control, since, for
example, there is substantial variation in climate within a state. Inadequate controls for
these other factors could potentially lead to omitted variable bias. Other uncertainties and
limitations associated with the estimation of the exposure-response relationships are
discussed in the case study.
Linking changes in tree growth to economic welfare changes:
In applying the estimates from the exposure-response model to the FASOM model, it was
assumed that the plots used to calculate the percentage increase in tree volume are
representative of the FASOMGHG regions. However, because data on all plots in the
region are unavailable, there is some uncertainty regarding the representativeness of plots
used.
The tree volume growth estimates used in the exposure-response model were calculated
based on live trees on forestland, not timberland, which is what FASOMGHG uses. This
may potentially give rise to some uncertainty in applying the results to FASOMGHG
because the estimates of the slopes (b) may be different for timberland than for forestland.
Final Risk and Exposure Assessment September 2009
Appendix 8-56
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
• The exposure-response model uses data from both private and public lands, while in
FASOMGHG the growth is applied to private lands only. This is an additional source of
uncertainty because different management practices could potentially affect the
relationship between exposure and growth differently.
• The age structure (and consequently volume of trees) may not be the same. So the stands
to which the change in growth rates are applied in FASOMGHG may be different from the
ones used in the exposure-response model, and this study may be assuming a change in
growth rate that is not realistic for these stands.
• A general limitation when using FASOMGHG is that it is a very aggregated region-level
model; thus, effects pertaining to areas particularly vulnerable to acidification cannot be
identified. Also, to make future timber market projections, FASOMGHG requires several
assumptions regarding future product demands, production capacity, and timber inventory.
• It must also be emphasized that the economic welfare changes reported in this section are
only those associated with markets for sugar maple and red spruce timber. They do not
include potential gains associated with other provisioning services, such as sugar maple
syrup production or production of other hardwood or softwood species affected by
terrestrial acidification. They also do not include gains outside the United States (in
particular, Canada) or in other sectors of the U.S. economy.
3.3 REFERENCES
Adams, D., R. Alig, B.A. McCarl, and B.C. Murray. February 2005. FASOMGHG Conceptual
Structure, and Specification: Documentation. Available at
http://agecon2.tamu.edu/people/faculty/mccarl-bruce/FASOM.html. Accessed on
October 22, 2008.
Brown, L.H. 2002. Profile of the Annual Fall Foliage Tourist in Vermont: Travel Year 2001.
Report prepared for the Vermont Department of Tourism and Marketing. Burlington, VT:
Vermont Tourism Data Center, University of Vermont.
Cordell, H.K., CJ. Betz, G. Green, and M. Owens. 2005. Off-Highway Vehicle Recreation in the
United States, Regions and States: A National Report from the National Survey on
Final Risk and Exposure Assessment September 2009
Appendix 8-57
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Recreation and the Environment (NSRE). Report prepared for the Forest Service's
National OHV Policy & Implementation Teams. USDA Forest Service.
Cordell, K., B. Leeworthy, G.T. Green, C. Betz, and B. Stephens, n.d. The National Survey on
Recreation & the Environment. Research Work Unit 4953. Athens, GA: Pioneering
Research on Changing Forest Values in the South and Nation USDA Forest Service
Southern Research Station. Available at www.srs.fs.fed.us/trends.
DeHayes, D.H., P.G. Schaberg, GJ. Hawley, and G.R. Strimbeck, 1999. "Acidic Rain Impacts
on Calcium Nutrition and Forest Health." BioScience 49:789-800.
Driscoll, C.T., G.B. Lawrence, AJ. Bulger, TJ. Butler, C.S. Cronan, C. Eagar, K.F. Lamber,
G.E. Likens, J.L. Stoddard, and K.C. Weathers. 2001. "Acidic Deposition in the
Northeastern United States: Sources and Inputs, Ecosystem Effects and Management
Strategies." BioScience 51:180-198.
Fenn, M.E., T.G. Huntington, S.B. McLaughlin, C. Eager, A. Gomez, andR.B. Cook. 2006.
"Status of Soil Acidification in North America." Journal of Forest Science 52 (special
issue):3-13.
Haefele, M., R.A. Kramer, and T.P. Holmes. 1991. Estimating the Total Value of a Forest
Quality in High-Elevation Spruce-Fir Forests. The Economic Value of Wilderness:
Proceedings of the Conference. Gen. Tech. Rep. SE-78 (pp. 91-96). Southeastern For.
Exper. Station. Asheville, NC: USDA Forest Service.
Holmes, T.P. 1992. Economic Welfare Impacts of Air Pollution Damage to Forests in the
Southern United States. Asheville, NC: U.S. Dept. of Agriculture, Forest Service,
Southeastern Forest Experiment Station, pp. 19-26.
Holmes, T., and R. Kramer. 1995. "An Independent Sample Test of Yea-Saying and Starting
Point Bias in Dichotomous-Choice Contingent Valuation. " Journal of Environmental
Economics and Management 28:121-132.
Final Risk and Exposure Assessment September 2009
Appendix 8-58
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Jenkins, D.H., J. Sullivan, and G.S. Amacher. 2002. "Valuing High Altitude Spruce-Fir Forest
Improvements: Importance of Forest Condition and Recreation Activity." Journal of
Forest Economics 8:77-99.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number EVIDE-02-0070.
Kramer, A., T. Holmes, and M. Haefele. 2003. "Contingent Valuation of Forest Ecosystem
Protection." In Forests in a Market Economy., E.O. Sills and K.L. Abt, eds., pp. 303-320.
Dordrecht, The Netherlands: Kluwer Academic Publishers.
Krieger, DJ. 2001. Economic Value of Forest Ecosystem Services: A Review. Washington, DC:
The Wilderness Society.
Luzadis, V.A. and E.R. Gossett. 1996. "Sugar Maple." In Forest Trees of the Northeast, J.P.
Lassoie, V.A. Luzadis, and D.W. Grover, eds., pp. 157-166. Cooperative Extension
Bulletin 235. Cornell Media Services. Available at
http://maple.dnr.cornell.edu/pubs/trees.htm.
National Agricultural Statistics Service (NASS). June 12, 2008. "Maple Syrup Production Up 30
Percent Nationwide." New England Agricultural Statistics, NASS, USDA.
Spencer, D.M., and D.F. Holecek. 2007. "Basic Characteristics of the Fall Tourism Market."
Tourism Management 28:491-504.
U.S. Department of Agriculture, Forest Service. Forest Inventory and Analysis National
Program, RPA Assessment Tables. 2002. Available at
http://Ncrs2.Fs.Fed.Us/4801/Fiadb/Rpa_Tabler/Draft_RPA_2002_Forest_Resource_Tabl
es.pdf.
U.S. Department of the Interior, Fish and Wildlife Service, and U.S. Department of Commerce,
U.S. Census Bureau. 2007. 2006 National Survey of Fishing, Hunting, and Wildlife-
Associated Recreation.
Final Risk and Exposure Assessment September 2009
Appendix 8-59
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.0 AQUATIC ENRICHMENT
One of the main adverse ecological effects resulting from nitrogen deposition,
particularly in the Mid-Atlantic region of the United States, is the effect associated with nutrient
enrichment in estuarine waters. A recent assessment of 141 estuaries nationwide by the National
Oceanic and Atmospheric Administration (NOAA) concluded that 19 estuaries (13%) suffered
from moderately high or high levels of eutrophication due to excessive inputs of both nitrogen
and phosphorus, and a majority of these estuaries are located in the coastal area from North
Carolina to Massachusetts (NOAA, 2007). By several measures, the aquatic ecosystem of the
Chesapeake Bay estuary is particularly suffering from the effects of excessive nitrogen loads,
and roughly one-third of these loads are associated with atmospheric deposition of nitrogen in
the watershed (Sweeney, 2007).1 For other estuaries in the Mid-Atlantic region, the contribution
of atmospheric distribution to total nitrogen loads is estimated to range between 10% and 58%
(Valiguraetal., 2001).
Eutrophication in estuaries is associated with a range of adverse ecological effects. Using
the conceptual framework developed by NOAA, Figure 4-1 illustrates the main links between
nutrient loadings and ecological symptoms in estuaries. The framework emphasizes four main
types of eutrophication effects—low dissolved oxygen (DO), harmful algal blooms (HABs), loss
of submerged aquatic vegetation (SAV), and low water clarity.
Low DO (i.e., hypoxia) has become a chronic problem in several estuaries, particularly
during summer months. Five of the 22 estuaries evaluated by NOAA in the Mid-Atlantic region
suffer from serious DO problems. The mainstem of the Chesapeake Bay has been a particular
area of concern. For example, between 2005 and 2007, only about 12% of the Bay met DO
standards during the summer months (Chesapeake Bay Program, n.d.). Low DO disrupts aquatic
habitats, causing stress to fish and shellfish, which, in the short term, can lead to episodic fish
kills and, in the long term, can damage overall growth in fish and shellfish populations. Low DO
also degrades the aesthetic qualities of surface water.
1 Phosphorus loads, primarily from agricultural runoff and wastewater dischargers in the Chesapeake Bay
watershed, are the other main source of nutrients contributing to eutrophication in the Bay.
Final Risk and Exposure Assessment September 2009
Appendix 8-60
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
HABs were also rated by NOAA as a major problem in five Mid-Atlantic estuaries,
including the mainstem of the Chesapeake Bay and the Potomac River estuary. In addition to
often being toxic to fish and shellfish and leading to fish kills and aesthetic impairments of
estuaries, HABs can, in some instances, also be harmful to human health.
Nutrient Inputs to Estuary
Primary Symptoms of Secondary Symptoms
Eutroohication of Eutroohication
Ecosystem Endooints
Affected Ecosystem Services
Nitrogen Loadings from Direct
and Indirect Deposition
Nitrogen and Phosphorous
Loadings from Other Sources
Provisioning Services
productionof fish
production of shellfish
Cultural Services
• recreational fishing,
boating, swimming,
etc.
aesthetic enjoyment
Low Water
» Clarity and
Light
Availability
*
Loss of
Submerged
Aquatic
Vegetation
(SAV)
— *
Declines in
Fish/Shellfish
Abundance
Declines in
Shoreline
Quality
4 — 1
• nonuse services
T
Regulating services
• erosion control
• storm protection
Figure 4-1. Conceptual Model of Eutrophication Impacts in Estuaries (Source:
Adapted from Bricker et al. [2007] and Bricker, Ferreira, and Simas [2003]).
SAV provides critical habitat for many aquatic species in estuaries and, in some
instances, can also protect shorelines by reducing wave strength; therefore, declines in SAV due
to nutrient enrichment are an important source of concern. Although less prevalent than low DO
and HABs as a problematic symptom of eutrophication, it is nonetheless rated by NOAA as a
serious problem in the mainstem of the Chesapeake Bay and the New Jersey Inland Bays.
Low water clarity is the result of accumulations of both algae and sediments in estuarine
waters. In addition to contributing to declines in SAV, high levels of turbidity also degrade the
aesthetic qualities of the estuarine environment. Although NOAA's assessment of estuaries did
not focus on turbidity separately as an indicator of eutrophi cation, it is nonetheless a common
problem in the Mid-Atlantic region.
Final Risk and Exposure Assessment
Appendix 8-61
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.1 OVERVIEW OF AFFECTED ECOSYSTEM SERVICES
Figure 4-1 also extends the NOAA framework to include links to the main types of
ecosystem services that are affected by the primary and secondary symptoms of eutrophication.
The following sections provide a discussion and overview of the primarily affected provisioning,
cultural, and regulating services.
4.1.1 Provisioning Services
Estuaries in the eastern United States are an important source of food production, in
particular fish and shellfish production. The estuaries are capable of supporting large stocks of
resident commercial species, and they serve as the breeding grounds and interim habitat for
several migratory species.
To provide an indication of the magnitude of provisioning services associated with
coastal fisheries, Table 4.1-1 reports the annual value of commercial landings in recent years for
15 East Coast states. From 2005 to 2007, the average value of total catch was $1.5 billion per
year. It is not known, however, what percentage of this value is directly attributable to or
dependent upon the estuaries in these states. Table 4.1-2 focuses specifically on commercial
landings in Maryland and Virginia in 2007, and it reports values for the main commercial species
in these states. Although these values also include fish caught outside of the Chesapeake Bay, the
values for two key species—blue crab and striped bass—are predominantly from the estuary
itself. These data indicate that blue crab landings in 2007 totaled nearly $44 million in the Bay.
The value of striped bass and menhaden totaled about $9 million and $25 million, respectively.
To most accurately assess how eutrophication in East Coast estuaries is related to the
long-term provisioning services from their fishery resources requires bioeconomic models (i.e.,
models that combine biological models offish population dynamics with economic models
describing fish harvesting and consumption decisions). In most cases, these models address the
dynamic feedback effects between fish stocks and harvesting behavior, and they characterize
conditions for a "steady-state" equilibrium, where stocks and harvest levels are stabilized and
sustainable over time.
Section 4.2 describes one bioeconomic model linking blue crab harvests to nutrient loads
in the Neuse River estuary, and it applies the model to estimate how reductions in nitrogen loads
to the estuary would affect the societal value of future blue crab harvests. In practice, however,
Final Risk and Exposure Assessment September 2009
Appendix 8-62
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
very few other studies have developed empirical bioeconomic models to estimate how changes
in environmental quality affect fish harvests and the value of these services (Knowler, 2002).
One exception is Kahn and Kemp (1985), which estimated a bioeconomic model of commercial
and recreational striped bass fishing using annual data from 1965 to 1979, measuring the effects
of SAV levels on fish stocks, harvests, and social welfare. They estimated, for example, that a
50% reduction in SAV from levels existing in the late 1970s (similar to current levels
[Chesapeake Bay Program, 2008]) would decrease the net social benefits from striped bass by
roughly $16 million (in 2007 dollars).
In a separate analysis, Anderson (1989) developed an empirical dynamic simulation
model of the effects of SAV changes on commercial blue crab harvests in the Virginia portion of
the Chesapeake Bay. Applying the empirical model results, long-run (15-year) dynamic
equilibria were estimated under baseline conditions (assuming SAV area constant at 1987 levels)
and under conditions with "full restoration" of SAV (i.e., 284% increase). In equilibrium, the
increase in annual producer surplus and consumer surplus with full restoration of SAV was
estimated to be $3.5 million and $4.4 million (in 2007 dollars), respectively.
Table 4.1-1. Annual Values of East Coast Commercial Landings (in millions)
State
Connecticut
Delaware
Florida, East Coast
Georgia
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
North Carolina
Pennsylvania
Pvhode Island
2004
$33.40
$5.42
$39.98
$14.37
$367.09
$49.29
$326.00
$17.21
$145.86
$46.89
$79.70
$0.07
$76.25
2005
$37.57
$6.11
$35.49
$13.46
$391.90
$63.67
$427.07
$22.12
$159.01
$56.45
$64.89
$0.04
$91.58
2006
$36.89
$5.69
$42.00
$11.53
$361.85
$53.58
$437.05
$18.84
$136.05
$57.73
$70.12
$0.10
$98.58
2007
$42.08
$7.58
$42.74
$10.08
$319.52
$52.27
$457.18
$19.09
$152.46
$58.94
$82.31
$0.13
$76.79
Final Risk and Exposure Assessment September 2009
Appendix 8-63
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
State
South Carolina
Virginia
Total
2004
$18.54
$160.51
$1,380.60
2005
$17.57
$155.26
$1,542.20
2006
$17.03
$109.07
$1,456.11
2007
$15.57
$130.56
$1,467.31
Source: National Oceanic and Atmospheric Administration (NOAA), 2007.
Table 4.1-2. Value of Commercial Landings for Selected Species in 2007 (Chesapeake Bay
Region)
State Species
Maryland
Blue crab
Striped bass
Clams or bivalves
Sea scallop
Oyster, Eastern
Other
Total
Virginia
Sea scallop
Menhaden
Blue crab
Croaker, Atlantic
Striped bass
Clam, Northern Quahog
Summer flounder
Other
Total
Value
$30,433,777
$5,306,728
$5,007,952
$2,808,984
$2,524,045
$6,190,474
$52,271,960
$62,891,848
$25,350,740
$13,222,135
$4,615,924
$3,834,906
$3,691,319
$3,186,229
$16,954,893
$130,561,765
Source: National Oceanic and Atmospheric Administration (NOAA), 2007.
One study examining the short-term effects of DO levels on crab harvests is Mistiaen,
Strand, and Lipton (2003). Focusing on three Chesapeake Bay tributaries—the Patuxent,
Chester, and Choptank rivers—they estimated a "stress-availability" model measuring the effects
of DO levels on the availability of blue crabs for commercial harvest, given the stock levels and
number of fishing vessels. The model results indicated that, below a threshold of 5 mg/L,
Final Risk and Exposure Assessment September 2009
Appendix 8-64
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
reductions in DO cause a statistically significant reduction in commercial harvest and revenues.
For the Patuxent River alone, a simulated reduction of DO from 5.6 to 4.0 mg/L was estimated to
reduce crab harvests by 49% and reduce total annual earnings in the fishery by $275,000 (in
2007 dollars). However, this is an upper-bound estimate because it does not account for changes
in fishing effort that would likely occur, and if the measured changes are due to migration of crab
populations to other areas rather than to crab mortality, then the broader net effects on crab
harvests may also be considerably smaller.2
In addition to affecting provisioning services through commercial fish harvests,
eutrophication in estuaries may also affect these services through its effects on the demand for
seafood. For example, a well-publicized toxic pfiesteria bloom in the Maryland Eastern Shore in
1997, which involved thousands of dead and lesioned fish, led to an estimated $56 million (in
2007 dollars) in lost seafood sales for 360 seafood firms in Maryland in the months following the
outbreak (Lipton, 1999). Additional evidence regarding potential losses in provisioning services
due to eutrophication-related fish kills is provided by Whitehead, Haab, and Parsons (2003) and
Parsons et al. (2006). The survey used in both studies was conducted with more than 5,000
respondents in states bordering the Chesapeake Bay area and in North Carolina. The survey
asked respondents to consider how their consumption patterns would change in response to news
about a large fish kill caused by a toxic pfiesteria bloom. To address the fact that not all fish kills
are the same, the size and type of the described fish kill—either "major," involving more than
300,000 dead fish and 75% with pfiesteria lesions, or "minor," involving 10,000 dead fish and
50% with lesions—were randomized across respondents. Based on respondents' stated
behaviors, the studies estimated reductions in consumer surplus per seafood meal ranging from
$2 to $5.3 The survey also found that 42% of residents in the four-state area (Maryland, Virginia,
Delaware, and North Carolina) were seafood consumers and that the average number of seafood
meals per month among these consumers was between four and five. As a result, they estimated
2 The estimated relationship between harvest and DO is discontinuous at 5 mg/L. The size of the measured effect on
harvests is relatively small below 5 mg/L and zero above the 5 mg/L threshold; therefore, any sizable benefits
would require DO to cross the 5 mg/L threshold. Moreover, the 5 mg/L threshold was an assumption of the model
rather than a tested hypothesis, which raises additional questions about the accuracy of benefit estimates for
changes across the threshold.
3 Surprisingly, these estimates were not sensitive to whether the fish kill was described as major or minor or to the
different types of information included in the survey.
Final Risk and Exposure Assessment September 2009
Appendix 8-65
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
aggregate consumer surplus losses of $43 million to $84 million (in 2007 dollars) in the month
after a fish kill.
4.1.2 Cultural Services
Estuaries in the eastern United States also provide an important and substantial variety of
cultural ecosystem services, including water-based recreational and aesthetic services. One of the
difficulties with quantifying recreational services from estuaries is that much of the national and
regional statistics are jointly collected and reported for estuarine and other coastal areas.
Nevertheless, even these combined statistics provide several useful indicators of recreational
service flows. For example, data from the FHWAR indicate that, in 2006, 4.8% of the 16 or older
population in coastal states from North Carolina to Massachusetts participated in saltwater
fishing. The total number of days of saltwater fishing in these states was 26.1 million in 2006.
Based on estimates from Kaval and Loomis (2003), the average consumer surplus value for a
fishing day was $35.91 (in 2007 dollars) in the Northeast and $87.23 in the Southeast. Therefore,
the total recreational consumer surplus value from these saltwater fishing days was
approximately $1.28 billion (in 2007 dollars).
Recreational participation estimates for several other coastal recreational activities are
also available for 1999 to 2000 from the NSRE. These estimates are summarized in Table 4.2-1
based on data reported in Leeworthy and Wiley (2001). Almost 6 million individuals aged 16 or
older participated in motorboating in coastal states from North Carolina to Massachusetts, for a
total of nearly 63 million days annually during 1999-2000. Using a national daily value estimate
of $32.69 (in 2007 dollars) for motorboating from Kaval and Loomis (2003), the aggregate value
of these coastal motorboating outings was $2.08 billion per year. Almost 7 million people
participated in birdwatching, for a total of almost 175 million days per year, and more than 3
million participated in visits to nonbeach coastal waterside areas, for a total of more than 35
million days per year. In contrast, fewer than 1 million individuals per year participated in
canoeing, kayaking, or waterfowl hunting.
4.1.3 Regulating Services
Estuaries and marshes have the potential to support a wide range of regulating services,
including climate, biological, and water regulation; pollution detoxification; erosion prevention;
and protection against natural hazards (Millennium Ecosystem Assessment [MEA], 2005). It is
Final Risk and Exposure Assessment September 2009
Appendix 8-66
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
more difficult, however, to identify the specific regulating services that are significantly
impacted by changes in nutrient loadings. One potentially affected service is provided by SAV,
which can help reduce wave energy levels and thus protect shorelines against excessive erosion.
Declines in SAV may, therefore, also increase the risks of episodic flooding and associated
damages to near-shore properties or public infrastructure. In the extreme, these declines may
even contribute to shoreline retreat, such that land and structures are lost to the advancing
waterline.
4.2 CHANGES IN ECOSYSTEM SERVICES ASSOCIATED WITH
ALTERNATIVE LEVELS OF ECOLOGICAL INDICATORS
This section estimates values for changes in several ecosystem services associated with
reduced nutrient enrichment effects in the Chesapeake Bay and Neuse River estuaries. Using the
results of the Potomac River and Neuse River Case Studies, the value of removing all
atmospheric sources of nitrogen loadings to these estuaries was estimated. Although such a large
change represents an upper bound on possible loading reductions through controls on
atmospheric sources, it corresponds with the findings of the case studies, which indicate that
reductions of this magnitude are the minimum required to improve the eutrophication index (El)
score (based on NOAA's ASSETS framework) from current "bad" conditions (El = 1) in these
two estuaries to somewhat better "poor" conditions (El = 2).
Final Risk and Exposure Assessment September 2009
Appendix 8-67
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table 4.2-1. Participation
in Selected Marine Recreation Activities in East Coast States in 1999-2000
Visiting
Watersides besides
Beaches
State
Connecticut
Delaware
Maryland
Massachusetts
New Jersey
New York
North Carolina
Rhode Island
Virginia
Total
Na
0.18
0.08
0.47
0.47
0.45
0.56
0.44
0.27
0.48
3.41
Days"
2.41
C
5.89
2.93
4.58
3.74
4.16
3.31
8.27
35.29
Motorboating
Na
0.39
0.38
0.97
0.61
0.89
0.90
0.55
0.38
0.60
5.67
Days"
6.76
4.56
8.13
6.05
12.45
9.48
7.25
4.37
4.54
63.59
Canoeing
Na
0.05
0.04
0.16
0.07
0.07
0.07
0.04
0.15
0.15
0.79
Kayaking
Na
0.10
0.02
0.03
0.17
0.10
0.06
0.12
0.11
0.06
0.76
Bird Watching
Na
0.45
0.43
0.82
1.02
0.80
0.88
1.04
0.56
0.86
6.84
Days"
15.19
14.03
19.76
26.10
18.80
24.55
20.52
19.01
17.00
174.96
Waterfowl
Hunting
Na
0.00
0.02
0.03
0.00
0.01
0.00
0.03
0.00
0.04
0.13
a Number of resident and nonresident participants annually (in millions).
b Number of days by residents and nonresidents annually (in millions).
0 Insufficient data for estimate.
Source: Leeworthy and Wiley, 2001
Final Risk and Exposure Assessment
Appendix 8-68
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.2.1 The Chesapeake Bay Estuary
For the Chesapeake Bay analysis, the results of the Potomac River/Potomac Estuary Case
Study were applied. Other than the mainstem of the Bay (6,074 km2), the Potomac estuary is the
largest subestuary within the Chesapeake Bay estuary system (1,260 km2), and other than the
Susquehanna River, which flows directly into the mainstem, it contributes the largest portion of
freshwater (19%) to the Bay. Eutrophic conditions within the Potomac estuary are also reflective
of more widespread conditions in the Bay. For example, when assessing estuarine conditions
across the country in 2004, NOAA (2007) evaluated nine subestuaries of the Bay, including the
mainstem and the Potomac. Five subestuaries in the Bay, including the mainstem and the
Potomac, rated "high" with respect to overall eutrophic conditions (the worst level on a 5-point
scale from low to high). The remaining four subestuaries were all rated as "moderate high" (the
second worst level). Therefore, for this analysis, it was assumed that the results of the Potomac
River estuary case study are representative of the Chesapeake Bay as a whole.
According to the Aquatic Nutrient Enrichment Case Study, atmospheric deposition is
estimated to contribute 24% (7.38 million kg nitrogen/year) of total nitrogen loadings to the
Potomac estuary. This percentage falls within the range of the 23% to 33% that has been
estimated for the Chesapeake Bay as a whole (Valigura et al., 2001). The case study also
estimates that a reduction in nitrogen loadings roughly equivalent to the contribution from
atmospheric deposition in the Potomac River watershed would be required to improve the
Potomac estuary from "bad" to "poor" on the 5-point ASSETS EL4
For the Chesapeake Bay analysis, the change in selected ecosystem services associated
with a 24% reduction in loadings to the Chesapeake Bay as a whole was estimated, and it was
assumed that this reduction would also improve the Bay's overall El score from "bad" to "poor."
The selection of ecosystem services for this analysis, which includes recreational, aesthetic, and
nonuse services (i.e., specific cultural services), was based on availability of existing models,
data, and empirical results.
For each of the ecosystem service categories addressed in this section, the geographic
extent of aggregate benefits to residents and recreators in Maryland, Virginia, and Washington,
4 The mean and median estimate of required loading reductions is 104% of the annual atmospheric deposition
component (almost 25% of all N loadings).
Final Risk and Exposure Assessment September 2009
Appendix 8-69
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
DC (DC) was limited. Because these areas are directly adjacent to the Bay, this approach is
expected to include a large majority of the beneficiaries; however, this approach also will
unavoidably contribute to some underestimation of aggregate benefits. Other specific limitations
and uncertainties in the proposed methods are described in each of the subsections below.
4.2.1.1 Recreational Fishing Services
This section describes and applies a three-part "benefit transfer" framework for
estimating the recreational fishing benefits of improved eutrophic conditions in the Chesapeake
Bay. The first component translates changes in the 5-point El into equivalent changes in average
DO levels in the Bay. This step is required to link eutrophic conditions to existing recreational
catch rate models.
The second component predicts the effect of changes in average DO levels on
recreational fishing catch rates. These catch rates can be interpreted as indicators of the
recreational fishing services provided by the Bay. Two catch rate models are described: one
based on a study of striped bass fishing in the Bay and the other based on a study of summer
flounder fishing in the Maryland coastal bays.
The third component estimates the benefits of catch rate improvements using willingness
to pay (WTP) estimates derived from a meta-analysis study by Johnston et al. (2005) and annual
fishing trip estimates to the Bay using data from the Marine Recreation Fishing Statistics Survey
(MRFSS).
4.2.1.1.1 Converting Changes in El to Changes in DO
As described above, low DO is one of several ecosystem symptoms associated with
estuarine eutrophication; therefore, DO levels are one of several factors included in the ASSETS
framework to derive the composite 5-point EL
To derive changes in DO that are equivalent to a 1-unit change on the El, data for a
comparable water quality index were used. In collaboration with the University of Maryland's
Center for Environmental Science, NOAA has also developed a 100-point Chesapeake Water
Quality Index (CWQI) based on three main eutrophication symptom indicators—chlorophyll a,
water clarity, and DO (Chesapeake Eco-Check, 2007). Using annual data for the three water
quality parameters, a water quality index score was generated for 141 monitoring stations across
the Bay in 2007. Table 4.2-2 reports the results of regressing the CWQI score for each station
Final Risk and Exposure Assessment September 2009
Appendix 8-70
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
against four corresponding water quality measures—(1) average surface DO, (2) average bottom
DO, (3) average secchi depth, and (4) average chlorophyll a. Both DO measures have positive
and statistically significant effects (with a p-value less than 0.05) on the index score, although
the estimated effect of bottom DO is somewhat larger. A 1-unit change in both bottom and
surface DO is predicted to change the CWQI by a combined effect of 8.3 points. If it was
assumed that the 100-point CWQI and the 5-point El are directly proportional, then a 1-unit
change in both bottom and surface DO is predicted to change the El by 0.415 (= 8.3/20) points.
Alternatively, a 1-point increase in the El (e.g., from "bad" to "good") would be predicted to
result from a 2.41-unit increase in both surface and bottom DO.5 Therefore, going forward, the
effects on recreational fishing resulting from an increase in bottom and surface DO of 2.41 mg/L,
which is assumed to be equivalent to a 1-unit increase in the El, were estimated.
Table 4.2-2. Regression Analysis of the Chesapeake Water Quality Index on Water Quality
Parameters
Water Quality
Parameter
CWQI (dependent
variable)
Explanatory variables
Chlorophyll a (ug/L)
Secchi depth (m)
Bottom DO (mg/L)
Surface DO (mg/L)
Constant
Observations
R-squared
Regression
Coefficient
—
-0.556
0.163
5.069
3.235
6.296
141
0.357
Results
P-value
—
0.000
0.443
0.000
0.028
0.556
Summary Statistics
Mean Std. Dev. Min
48.85 22.30 9
15.99 14.09 2.0
2.32 7.44 0.1
5.01 2.00 0.1
6.91 0.97 4.5
— — —
Max
100
82.0
50.4
9.3
9.1
—
5 Based on the results in Table 4.2-2, other combinations of bottom DO and surface DO changes can also produce a
1-point increase in the El; however, for simplicity, equivalent changes in the two DO measures were considered.
Final Risk and Exposure Assessment
Appendix 8-71
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.2.1.1.2 Estimating Changes in Recreation Services (Catch Rates)
In two related papers, Lipton and Hicks (1999, 2003) reported the results of a travel cost
study of recreational striped bass fishing in the Chesapeake Bay. One of the main focuses of the
study was measuring the effect of DO levels on striped bass catch rates. The fishing data for this
study were drawn from the National Marine Fisheries Service's (NMFS's) 1994 MRFSS, which
included 407 intercept sites in the Bay and 1,806 striped bass angler respondents. The DO water
quality data were from biweekly summer sampling at 207 locations in the Bay.
The striped bass catch model assumes that the number offish caught per trip (in
logarithmic form) at a site is a linear function of several factors, including the hours spent by the
angler at the site on the trip, the angler's experience and skill in saltwater fishing, and water
quality conditions at the site. Water quality is characterized in the model by surface temperature
(57), bottom temperature (BT), surface DO (SDO), and bottom DO (EDO). According to the
functional form of the estimated model, the change in the expected striped bass catch rate per
trip due to a water quality change can be expressed as
Q) + In Q0 ) - Q0 , (4.1)
where Qt is the expected number of striped bass caught per trip under conditions /', such
that / = 0 represents reference conditions and / = 1 represents conditions after the water quality
change. The function fB(AWQ) represents the combined effect of changes in temperature and DO
on expected catch rates. Using the parameter estimates from the empirical catch rate model, this
function for striped bass can be expressed as
fB(kWQ) = In ft -In g0 = -0.2548GS7i -,STo) + 0.3225(57i -BT0)
- SDO0 ) + 02253(BDOl - BDO0 ) -
To quantify baseline catch rates (Qo), recent MRFSS data for the Bay, which are
summarized in Table 4.2-3, were used. The table reports average catch rates for striped bass and
other key recreational species for 2001 to 2005. Over the 5-year period, striped bass catch rates
averaged 1.65 fish per trip in Maryland and 0.59 fish per trip in Virginia.6
6 For comparison, Lipton and Hicks (1999) reported that average catch rates in 1994 were 0.71 in Maryland and
0.66 in Virginia.
Final Risk and Exposure Assessment September 2009
Appendix 8-72
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
With these baseline catch rate estimates, Equations (4.1) and (4.2) can be used to predict
the change in average catch rate (A(?) associated with specific changes in surface and bottom
temperature and DO levels. For example, if average surface and bottom DO levels in the Bay
both increase by 2.41 units (with no change in temperature), the striped bass catch rate is
predicted to increase by 1.57 in Maryland and by 0.56 in Virginia (a 94.9% increase).
Table 4.2-3. Average Catch Rate per Fishing Trip in the Chesapeake Bay, by State and Targeted
Fish Species
Fishing Trip
Maryland residents
Striped bass
Summer flounder
Other species
All species
Virginia residents
Striped bass
Summer flounder
Other species
All species
2001
1.20
0.09
0.29
0.34
0.42
0.96
0.34
0.37
2002
1.58
0.08
0.32
0.38
0.44
0.80
0.40
0.42
2003
1.99
0.09
0.32
0.40
0.52
0.91
0.26
0.29
2004
1.81
0.34
0.27
0.35
0.82
0.93
0.27
0.32
2005 Average 2001-2005
1.70
0.04
0.45
0.51
0.68
0.69
0.34
0.36
1.65
0.12
0.33
0.39
0.59
0.86
0.32
0.35
Source: National Oceanic and Atmospheric Administration (NOAA), 2009.
It is more difficult to develop catch rate predictions for other recreational species,
because of the apparent lack of any other empirical studies that have estimated the relationship
between water quality conditions and recreational catch rates in the Bay.7 One alternative is to
assume that the striped bass model described above is applicable to other species; however, the
resulting catch rate change estimates would inevitably have higher levels of uncertainty
associated with them.
A second approach is to use catch rate models developed in areas outside the Bay;
however, only one such study was found.8 Massey, Newbold, and Gentner (2006) used data from
7 Bricker et al. (2006) described similar models for the Potomac and Patuxent River estuaries and other East Coast
estuaries; however, they did not provide parameter estimates for these models.
8 Kaoru, Smith, and Liu (1995) also estimated the effects of estuarine water quality on recreational fishing in North
Carolina; however, rather than using ambient water quality measures, he used estimates of nutrient and
biochemical oxygen demand loadings as proxies for water quality conditions.
Final Risk and Exposure Assessment September 2009
Appendix 8-73
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
the Maryland coastal bays to estimate a catch rate model for recreational summer flounder
fishing. They found significant effects from DO, temperature (7), and water clarity (secchi depth
[SD]) on recreational catch. Using the parameter estimates from this model, the following
function summarizes the measured effects of water quality on summer flounder catch rates:
fF(AWQ) = Q.W(DOl -£>O0) + 0.126(7; -TQ) +1.392(SD, - SD0) . (4.3)
Applying this function to Equation (4. 1) in place offs(AWQ), a 2.41-unit increase in DO
(with no change in Tor SD) is predicted to increase summer flounder catch by an additional 0.04
fish per trip in Maryland and 0.28 fish per trip in Virginia (a 32.6% increase). Transferring this
model from the Maryland coastal bays to the Chesapeake Bay also contributes to the uncertainty
in catch rate predictions for summer flounder, although arguably less so than transferring models
from other species (i.e., striped bass) within the Bay.
4.2.1.1.3 Valuing Changes in Catch Rates
The second component of the proposed benefit transfer model for recreational fishing can
be summarized as follows:
AggBflsh = Z} (WTPfish x Tj) x A&, (4.4)
where
AggBfiSh = aggregate annual benefits (in 2007 dollars) to Chesapeake Bay anglers for
specified increases in species-specific average catch rates per trip (AQ-,
where y is the species indicator),
= predicted change in average catch rate per trip for speciesy in the
Chesapeake Bay (as described in Section 4.2.1.1.2),
= average WTP per additional fish caught per trip, and
TJ = total number of annual fishing trips (in 2007) targeting speciesy in the
Chesapeake Bay.
A large number of revealed- and stated-preference studies have estimated welfare
changes associated with changes in recreational fishing catch rates in the United States. Most of
these results have been synthesized in a meta-analysis study by Johnston et al. (2006), which
estimated meta-regression models controlling for differences across studies in type of water
resource, context, angler attributes, and in-study methods. Using these summary models, they
Final Risk and Exposure Assessment September 2009
Appendix 8-74
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
predicted average WTP per fish per trip for different species categories. For both Atlantic small
game (including striped bass) and Atlantic flatfish (including summer flounder), they predicted
WTP ranging from $3 to $11 in 2003 dollars. This meta-analysis study included one WTP
estimate from a Chesapeake Bay striped bass study (Bockstael, McConnell, and Strand, 1989),
which falls slightly below this range ($2.23), but it did not include a more recent striped bass
estimate from the Lipton and Hicks (1999) study, which falls within the upper end of the range
($10.91). Johnston et al.'s (2006) study also did not include the estimate for summer flounder in
Maryland coastal bays from Massey, Newbold, and Gentner (2006), which falls within the lower
end of the range ($4.22 in 2002 dollars).
Based on these WTP results from the literature, a value range of $2.50 to $12.50 for
WTPfish with a midpoint of $7.50, was selected.
To quantify annual trips by species (7/), recent MRFSS data for the Bay, which are
summarized in Table 4.2-4, were again used. The table reports total annual trips for striped bass
and other key recreational species from 2001 to 2005. To approximate trips in 2007, the average
number of trips from 2001 to 2005 by species were used. The same methodology was used to
approximate baseline catch rates for 2007.
Table 4.2-4. Aggregate Number of Fishing Trips to the Chesapeake Bay, by State and Targeted
Fish Species
Average
Fishing Trip 2001 2002 2003 2004 2005 2001-2005
Maryland
residents
Striped bass 2,594,971 2,014,818 2,579,771 2,176,824 2,351,145 2,343,506
Summer 2,106,810 1,268,048 1,598,484 1,486,154 1,734,101 1,638,719
flounder
Other species 33,457,937 31,349,971 48,352,248 39,740,106 34,503,965 37,480,845
All species 38,159,717 34,632,837 52,530,503 43,403,083 38,589,211 41,463,070
Virginia residents
Striped bass 2,043,025 1,911,180 2,369,576 2,525,057 2,549,248 2,279,617
Summer 2,285,628 1,982,130 2,300,633 2,556,902 2,549,248 2,334,908
flounder
Other species 49,915,214 47,535,158 67,839,883 65,345,054 58,036,434 57,734,349
All species 54,243,868 51,428,468 72,510,092 70,427,013 63,134,930 62,348,874
Final Risk and Exposure Assessment September 2009
Appendix 8-75
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Source: National Oceanic and Atmospheric Administration (NOAA), 2009.
4.2.1.1.4 Results: Aggregate Recreational Fishing Benefits
Combining the three model components, the aggregate recreational fishing benefits from
a 1-point El increase in Chesapeake Bay from "bad" to "poor" can now be estimated. Assuming
that this change is equivalent to a 2.41 mg/L increase in both surface and bottom DO, the result
is an annual benefit of $37.2 million to striped bass anglers in Maryland and Virginia and an
annual benefit of $5.4 million to summer flounder anglers.
Recognizing the uncertainties associated with transferring these models to other species,
the same benefit transfer framework can also be applied to other recreational fishing trips.
Striped bass and summer flounder fishing only accounted for 7.4% of the total number of trips
and 15.5% of the total catch from 2001 to 2005. If the striped bass catch rate model
(Equation [4.2]) is applied to all other types offish species, then a 2.41 mg/L increase in surface
and bottom DO would result in an estimated aggregate benefit of $217 million for recreational
anglers targeting these other species in the Bay.
4.2.1.1.5 Limitations and Uncertainties
Although the objective of the previously described approach is to make the best use of
existing research to quantify the relationship between changes in eutrophic conditions and
recreational fishing benefits in the Bay, the following limitations and uncertainties must also be
noted.
First, the conversion of changes in El to changes in DO requires several strong
assumptions. One key assumption is that the El and CWQI are directly proportionate to one
another. The reasonableness of this assumption rests on the fact that the two indexes use similar
symptom indicators (DO, SD, and chlorophyll a) and both have been designed and used by
NOAA as summary metrics of eutrophic conditions in estuaries. Another key assumption is that
the regression model in Table 4.2-2 can be used to generate equivalent changes in DO. Because
several water quality parameters besides DO are also measures of eutrophic symptoms, there is
no guarantee that a 1-unit change in El is uniquely associated with a 2.41 mg/L change in surface
and bottom DO (i.e., that the other factors are held constant).
Second, the catch rate models summarized in Equations (4.2) and (4.3) are most likely to
understate the effects of long-term changes (i.e., over several years) in water quality across the
Final Risk and Exposure Assessment September 2009
Appendix 8-76
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
entire Bay. Both models are based on analyses that use spatial and short-term (during a single
year's fishing season) temporal variation to measure the relationship between catch rates and
water quality conditions. Therefore, these measured relationships cannot be expected to capture
the dynamic effects of long-term changes in DO on the overall growth and abundance of the
striped bass and summer flounder populations in the Bay.
Third, as previously noted, empirical catch rate models are only available for striped bass
and summer flounder, and the model for the latter species is based on data from outside the Bay.
Although it is not difficult to apply these models to estimate catch rate changes for other species
within the Bay, the resulting estimates are subject to significant uncertainty, because there is
little evidence about how well these models transfer to other species.
Fourth, the valuation model summarized in Equation (4.4) uses a number of simplifying
assumptions. In particular, the value per fish caught is assumed to be constant, but within a large
range—$2.50 to $12.50—which can significantly affect the aggregate benefit estimates. In
addition, the total number of fishing trips is assumed to be unaffected by changes in catch rates.
This restriction is expected to understate the true aggregate benefits of increased catch rates,
because higher catch rates would most likely increase the number of fishing trips.
4.2.1.2 Boating
To estimate benefits to Chesapeake Bay boaters, a benefit transfer approach that uses
value estimates developed by Lipton (2004) is described. That study used a CV method and
survey data from 755 Maryland boaters in 2000 to estimate the individual and aggregate benefits
of a 1-unit improvement in respondents' water quality rating (on a 1 to 5 scale from "poor" to
"excellent") for the Bay. The benefit transfer model based on this study can be summarized as
follows:
AggBboat = Zt ZjfWTPboat.i >< Nij X bj) X &WQ5, (4.5)
where
AWQs = change in Chesapeake Bay water quality, expressed on a 5-point rating scale
(from "poor" to "excellent");
AggBboat = aggregate annual benefits (in 2007 dollars) to Maryland, Virginia, and DC
boat owners who use the Chesapeake Bay as their principal boating area for
a specified AWQs increase in water quality;
Final Risk and Exposure Assessment September 2009
Appendix 8-77
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
WTPboat,i = average annual WTP (in 2007 dollars) per boater for a 1-unit increase in
water quality on the WQs scale (/' = sailboat, trailered powerboat, or in-water
powerboat);
Nij = total number of boats by type /' and locationy (j = Maryland, Virginia, or
DC) of boat ownership in 2007; and
bj = the ratio of (1) registered boat owners whose principal boating area is the
Chesapeake Bay to (2) the total number of registered boats (by locationy).
Lipton (2004) reported estimates of average WTP by boat owners in three different
categories for a 1-unit increase in water quality (AWQs = 1) in the Chesapeake Bay. Sailboat
owners had the highest average WTP of $93.26 (in 2000 dollars). Trailered and in-water
powerboat owners had an average WTP of $30.25 and $77.98, respectively.
Converting these Lipton (2004) estimates to 2007 dollars with the consumer price index
(CPI) results in WTPboat estimates of $112.29, $36.42, and $93.89 for sailboat, trailered
powerboat, and in-water powerboat owners, respectively (Table 4.2-5).
Table 4.2-5. Input Estimates for the Chesapeake Bay Boating Benefit Transfer Model
Number of Registered Boats
Boat Type
Sailboat
Trailered powerboat
In-water powerboat
Total
NMD
8,200
93,300
23,600
125,100
NVA
9,200
104,600
26,400
140,300
NDC
100
1,100
300
1,500
Adjustment Factor
bMD bVA
60.76% 56.92%
60.77% 56.93%
60.77% 56.93%
boc
60.76%
60.77%
60.77%
WTPboal
$112.29
$36.42
$93.89
NMD was estimated for the three boater categories using data on Maryland boat ownership
from Lipton (2008) and Lipton (2006). The former data source quantifies sailboat and powerboat
ownership for 2007, but it does not break out powerboats according to whether they were
trailered or in-water boats. To develop separate estimates for these two subcategories, the
proportions reported for 2005 (Lipton, 2006), which indicated that 79.8% of powerboats in
Maryland were trailered, were applied. To estimate NVA and NDC, the total number of registered
boats in Virginia and DC in 2006 was obtained from the National Marine Manufacturers
Association (2008), and this number was augmented by the observed growth rate in Maryland
boat ownership from 2006 to 2007. To separate these total numbers into the three categories of
Final Risk and Exposure Assessment September 2009
Appendix 8-78
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
boat ownership, the same proportions estimated for Maryland registered boats in each category
were applied.
The value bMo represents a two-part adjustment to the total number of registered boats in
Maryland, as estimated by Lipton (2004). The first converts the total number of registered boats
to the total number of boat owners, because some boat owners own more than one boat. The
second adjusts for the fact that, for some Maryland boaters, the Chesapeake Bay is not their
principal boating area. Every 100 registered boats correspond to an estimated 60.8 boat owners
whose principal boating area is the Chesapeake Bay. The same adjustment factor for registered
boaters in DC was applied to estimate bDc.
To estimate bvA, the expected 6.3% of registered boats in Virginia Beach (Murray and
Lucy, 1981), which is the main Virginia coastal area outside the Bay, was first excluded; then the
same adjustment factor developed for Maryland and DC was applied. Thus, in Virginia, for
every 100 registered boats, there are 56.9 boat owners whose principal boating area is the
Chesapeake Bay.
4.2.1.2.1 Results: Aggregate Recreational Boating Benefits
To apply the previously described framework, it was first assumed that there is a direct
one-to-one correspondence between the 5-point El and the 5-point subjective WQ5 index. Based
on this assumption, a 1-unit increase in Chesapeake Bay water quality (AWQs = 1 and AE7 = 1)
was estimated to yield an annual aggregate benefit of $8.2 million for Maryland, Virginia, and
DC boat owners whose principal boating area is the Chesapeake Bay.
4.2.1.2.2 Limitations and Uncertainties
A potential limitation of the proposed benefit transfer model for boating services is the
uncertainty associated with directly translating the WQs index into the El index. Although both
metrics are summary 5-point measures of Chesapeake Bay water quality, the first is a subjective
index based on boaters' perceptions and experience. These perceptions may be based on
observations unrelated to eutrophic conditions (e.g., trash in the water or advisories based on
pathogen levels) or boaters may implicitly assign more or less importance to eutrophic
conditions than is assigned by the EL
The other main source of uncertainty is with the number of affected boaters. As in the
recreational fishing model, the affected number of recreators is assumed to be unaffected by the
Final Risk and Exposure Assessment September 2009
Appendix 8-79
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
change in water quality. This assumption is likely to lead to an underestimate of the aggregate
benefit to boaters of a water quality improvement.
One alternative approach is to use value estimates from Bockstael, McConnell, and
Strand (1989), who also estimated changes in consumer surplus for trailered boat owners in
Maryland resulting from a 20% decrease in the product of total nitrogen and phosphorus (TNP)
levels in the Bay. By reseating and updating their estimates to 2007 dollars, the implied average
WTP per Maryland trailered boat owner per 1% decrease in TNP is $5.38. Applying this value to
the estimated total number of trailered powerboat owners in Maryland, Virginia, and DC (see
Table 4.2-1), implies that the aggregate benefits to these boaters per 1% decrease in TNP in the
Bay would be $120,000. Assuming that a 24% decrease in nitrogen loadings would result in a
24% reduction in TNP levels in the Bay, the resulting estimate of annual aggregate benefits is
$2.9 million. The main advantage of this approach compared with the model summarized in
Equation (4.5) is that it is based on an objective measure of water quality. The fact that it is
based on values estimated through a revealed-preference travel cost model of actual boating
behavior, compared with a stated-preference CV approach, may be seen as an advantage.
However, this approach also has several drawbacks: (1) it is based on considerably older data
(from 1984), (2) it only includes direct estimates for trailered boaters, and (3) it includes a
potentially narrower measure of value than the Lipton (2004) study because it uses revealed-
rather than stated-preference data. This approach also requires the assumption that decreases in
nitrogen loads to the Bay are proportional to decreases in TNP levels in the Bay.
4.2.1.3 Beach Use
To estimate benefits to Chesapeake Bay beach users, the benefit transfer approaches
developed by Morgan and Owens (2001) and Krupnick (1988) were adapted and updated. Both
of these studies estimated the aggregate benefits to Maryland, Virginia, and DC households of
percentage reductions in levels in the Bay. The fundamental benefit transfer model can be
summarized as follows:
AggBbeach=(WTPbeach*(Nlbl + N2b2)*tbeachr&%WQTNP, (4.6)
where
&%WQTNP = percentage change in Chesapeake Bay water quality, expressed in terms
of the average TNP levels, each measured in parts per million (ppm);
Final Risk and Exposure Assessment September 2009
Appendix 8-80
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
ach = aggregate annual benefits (in 2007 dollars) to Maryland, Virginia, and
DC households for a specified t^/oWQ^p increase in water quality in the
Bay;
WTP beach = average annual household WTP (in 2007 dollars) per trip for a 1%
reduction in TNP levels in the Bay;
N\ = total number of households in the 1980 Baltimore and DC standard
metropolitan statistical areas (SMSA) in 2007;
N2 = total number of Maryland and Virginia households outside the SMSA in
2007;
b] = portion of SMSA households with at least one Chesapeake Bay beach trip
in the year;
b2 = portion of non-SMSA households in Maryland and Virginia with at least
one Chesapeake Bay beach trip in the year; and
tbeach = average number of Chesapeake Bay beach trips per year for beach-going
Maryland, Virginia, and DC households.
Table 4.2-6 summarizes value estimates for these model components. Values for
ach were derived using estimates from Bockstael, McConnell, and Strand (1988, 1989).
Using data from 408 summer beach users in 1984 at nine Maryland western shore beaches and
average county-level summer TNP values, they estimated a varying parameter travel cost model.
Based on the model results, they reported aggregate annual consumer surplus gains of $34.66
million (in 1987 dollars) for beachgoers residing in the SMSA associated with a 20% decrease in
TNP in the Bay. The study also reported that (1) 401,000 SMSA households per year (in the
early 1980s) visited Chesapeake Bay beaches and (2) the average number of trips per year for
these beach-going households was 4.35,9 which implies that there were an estimated 1,745,000
trips to the Bay by SMSA households in 1984. Dividing the aggregate benefit estimate by this
number of trips implies an average per-trip benefit of $19.86 (in 1987 dollars), for a 20%
reduction in TNP.
9 This number is actually inferred from a description of values Bockstael, McConnell, and Strand (1989) derived
from an alternate model. The value per household user ($4.70) was divided by the value per trip ($1.08) to get
trips per household (4.35).
Final Risk and Exposure Assessment September 2009
Appendix 8-81
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table 4.2-6. Input Estimates for the Chesapeake Bay Beach-Use Benefit Transfer Model
Beach Use
SMSA
Non-SMSA
Total
Number of
Households
(N)
2,744,217
2,540,214
5,284,431
Percentage of Bay
Beachgoers
(b)
21.00%
3.08%
12.38%
Average Beach
Trips per Year
(0
4.35
4.35
4.35
WTPbeach
$1.81
$1.81
$1.81
To estimate WTPbeach, the $19.86 estimate was divided by 20 (i.e., it was assumed that
each percentage reduction in TNP has the same value), and the estimate was converted to 2007
dollars using the CPI to adjust for inflation. The resulting estimate for WTPbeach is $1.81.
NI and A/2 were estimated using the Census estimates of population by county in 2007,
multiplied by the ratio of households to population by county in the 2000 U.S. Census. From this
calculation, it was estimated that a total of 5.28 million households are in Maryland, Virginia,
and DC, and 2.74 million of these are within the SMSA.
For bj, the Bockstael, McConnell, and Strand (1989) estimate that 21% of households in
the SMSA take at least one beach trip to the Chesapeake Bay a year was applied. To derive 62,
this estimate was combined with data from the 2006 Virginia Outdoors Survey (Virginia
Department of Conservation and Recreation, 2007), which reports that 8% of all the households
in Virginia take at least one beach trip to the Chesapeake Bay (or other tidal bays) per year.
Taken together, these estimates imply that approximately 3% ofnon-SMSA Virginia households
take at least one beach trip per year to the Bay. Applying this estimate to Maryland non-SMSA
households as well, it was assumed that b2 equals 3%.
To estimate tbeach, the Bockstael, McConnell, and Strand (1989) estimate of 4.35 trips per
year was applied, recognizing that it is most likely an overestimate for non-SMSA beach-going
households.
4.2.1.3.1 Results: Aggregate Beach Use Benefits
To apply this benefit estimation framework, it was assumed that changes in nitrogen
loads to the Bay are directly proportional to changes in average TNP concentrations in the Bay
(i.e., a 24% reduction in loadings results in a 24% decline in TNP). It was estimated that the
aggregate annual benefits to Maryland, Virginia, and DC Chesapeake Bay beachgoers
Final Risk and Exposure Assessment September 2009
Appendix 8-82
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
(AggBbeach) per 1% decrease in TNP is $5.16 million (in 2007 dollars); therefore, the benefit of a
24% decrease is $124 million.
4.2.1.3.2 Limitations and Uncertainties
One of the main limitations of the beach-use valuation model described above is that it is
based on value estimates that are from 1984 and, therefore, may be outdated. Beach conditions
and recreator preferences in the Bay may have changed significantly since then. In addition,
several uncertainties are associated with the estimated number of beach trips by Maryland,
Virginia, and DC households in 2007. These estimates are based on limited and, in some cases,
relatively old data regarding the percentage of households in each state that use the Bay's
beaches and the average number of annual beach trips for those who do. A second limitation is
that it, again, requires the assumption that decreases in nitrogen loads to the Bay are proportional
to decreases in TNP levels in the Bay.
4.2.1.4 Aesthetic Services
To estimate the benefits of improved aesthetic services due to improvements in
Chesapeake Bay water quality, a benefit transfer model that is based on estimates of near-shore
residents' values for small water-quality changes was developed and applied. The transfer
function has the following form:
AggBhome = ZkMWTPk x AZW* x Nk, (4.7)
where
= reduction in dissolved inorganic nitrogen (DIN) levels in the portion of the
Chesapeake Bay closest to coastal Census block group k;
= aggregate annual benefits (in 2007 dollars) to homeowners in all
Chesapeake Bay coastal block groups for specified ADINk changes in water
quality;
Nk = estimated number of specified owner-occupied homes in block group k in
2007; and
MWTPk = estimated annual marginal WTP (in 2007 dollars) for a 1-unit reduction in
water quality, ADINk = 1, in block group k.
To parameterize this function, results from a hedonic housing price study by Poor,
Pessagno, and Paul (2007) were used. Using data on 1,377 residential home sales from 1993 to
Final Risk and Exposure Assessment September 2009
Appendix 8-83
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
2003 in St. Mary's River watershed in Maryland, this study regressed the natural log of real
home prices (in 2003 dollars) against structural, neighborhood, and environmental water quality
characteristics. It specifically estimated the effect of differences in DIN (mg/L), as measured by
the annual average in the year of sale at the closest water monitoring station, on log home
prices.10 The study found a statistically significant effect with a model coefficient estimate of
-0.0878.
To convert this semielasticity coefficient, which measures the marginal effect of DIN on
the log of home price, ioMWTPk, which represents the annualized average dollar value of a
1-unit reduction in DIN for homes in block group k, the following conversion equation was used:
MWTPk = 0.0878*Pk *A(i,T), (4.8)
where
Pk = average price of specified owner-occupied homes in block group k and
A = annualization factor, which is a function of the assumed interest rate (r) and
average lifetime of homes in years (7).n For r = 0.05 and T = 50, A = 0.0522.
To implement the model, Chesapeake Bay coastal block groups were defined as those
block groups with a Chesapeake Bay coastline, as delineated by the Census block group
boundary files (Environmental Systems Research Institute, Inc. [ESRI], 2002), as well as those
block groups whose geographic centroids are located within 1 mile of the coast. This second
condition was added to ensure that a majority of the included properties are located within
roughly 2 miles of the coast. As shown in Figure 4.2-1, 1,066 block groups met these criteria.
Within these block groups, the study focused on Census "specified owner-occupied
housing units," which include only single-family houses on less than 10 acres without a business
or medical office on the property. These properties match best with the types of properties
analyzed in the hedonic study described above, and the decennial Census provides both count
and property value estimates for these homes. Thirty-six of the identified 1,066 block groups had
no specified owner-occupied homes and were excluded from the analysis.
10 A separate model reported in Poor, Pessagno, and Paul (2007) used total suspended solids (mg/L) instead of DIN
as the water quality measure. It was also found to have a statistically significant effect on home prices.
Final Risk and Exposure Assessment September 2009
Appendix 8-84
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
To estimate Nk, the number of specified owner-occupied homes in each block group in
2000 was augmented by the growth rate in housing units in the block group's county from 2000
to 2007 (U.S. Census Bureau, 2008b).
To estimate P/t, the average price of specified owner-occupied homes in 2000 in each
block group was adjusted to 2007 using the CPI-Shelter values for Washington-Baltimore, DC-
MD-VA-WV.12 Table 4.2-7 summarizes the estimated values for Nk and Pk.
4.2.1.4.1 Results: Aggregate Aesthetic Benefits to Near-Shore Residents
To approximate the aggregate annual benefits from a 24% reduction in nitrogen loadings
to the Bay, the benefit transfer model summarized in Equations (4.7) and (4.8) was applied,
assuming uniform 24% reductions in DIN across all Chesapeake Bay waters. Based on 2005
monitoring data for the Potomac River estuary, the 25th and 75th percentile values of average
DIN levels were 0.46 mg/L and 1.22 mg/L, respectively. Using this range of initial DIN values, a
reduction of 24% translates to DIN decreases of between 0.11 mg/L and 0.29 mg/L. It was
estimated that these changes would result in aggregate annual benefits of between $38.7 million
and $102.2 million to residents of specified owner-occupied homes in the Chesapeake coastal
block groups.
12 In the decennial Census, values of specified owner-occupied homes were grouped into ranges of values (e.g.,
from $250,000 to $300,000). With the exception of the highest range, which is $1,000,000 and greater (no upper
bound), the midpoint of each range was used to calculate the mean value for each block group. For the highest
range, a central value of $1,250,000 was selected.
Final Risk and Exposure Assessment September 2009
Appendix 8-85
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Legend
Ches Bay Coastal BGs
Sate BGs
Ches Bay Coastline
Miles
Figure 4.2-1. Chesapeake Bay Coastal Block Groups
Final Risk and Exposure Assessment
Appendix 8-86
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table 4.2-7. Summary of Housing Unit Numbers and Average Prices
State, County
Maryland
Anne Arundel County
Baltimore County
Calvert County
Caroline County
Cecil County
Charles County
Dorchester County
Harford County
Kent County
Prince George's County
Queen Anne's County
St. Mary's County
Somerset County
Talbot County
Wicomico County
Worcester County
Baltimore city
Virginia
Accomack County
Charles City County
Essex County
Gloucester County
Isle of Wight County
James City County
Number
of Coastal
Block
Groups
163
101
25
2
18
6
20
23
13
1
18
29
14
20
7
1
116
9
4
5
16
10
7
in Chesapeake Coastal Block Groups in 2007
Number of Specified Single-
Unit Dwellings per Block
Group (Nk)
Mean
470
339
583
346
408
466
263
397
294
175
584
458
258
431
304
148
168
247
270
255
438
515
817
Std.
Dev.
258
160
380
37
206
208
125
296
105
0
249
195
118
181
129
0
95
93
46
69
248
204
551
Min
13
16
190
310
195
302
46
20
55
175
158
92
81
192
163
148
10
128
219
135
174
292
85
Max
1,604
859
1,957
383
1,088
909
584
926
418
175
1,104
910
528
943
576
148
436
388
340
348
1,063
1,076
1,734
Average Value of Specified Units
per Block Group (Pk)
Mean
$298,348
$154,975
$257,582
$145,589
$206,114
$263,744
$167,317
$187,685
$225,531
$296,739
$277,352
$252,720
$130,391
$341,180
$145,763
$98,596
$106,483
$127,159
$163,344
$175,598
$194,628
$210,389
$306,926
Std. Dev.
$128,985
$54,654
$68,476
$3,499
$41,341
$46,708
$61,534
$39,578
$48,381
$0
$73,004
$41,564
$30,635
$162,586
$29,367
$0
$52,800
$56,731
$43,411
$52,691
$40,091
$42,265
$173,652
Min
$81,409
$65,209
$174,459
$142,091
$138,614
$196,436
$86,191
$116,187
$143,163
$296,739
$153,496
$174,343
$77,591
$128,864
$110,355
$98,596
$23,628
$69,539
$104,671
$132,168
$127,058
$136,806
$88,149
Max
$686,879
$343,455
$498,410
$149,088
$270,798
$319,093
$310,723
$303,797
$306,886
$296,739
$478,396
$317,879
$181,198
$658,874
$186,045
$98,596
$345,380
$236,974
$227,234
$278,618
$285,437
$273,345
$569,755
Final Risk and Exposure Assessment
Appendix 8-87
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
State, County
King and Queen County
King George County
King William County
Lancaster County
Mathews County
Middlesex County
New Kent County
Northampton County
Northumberland County
Prince George County
Richmond County
Surry County
Westmoreland County
York County
Chesapeake city
Hampton city
Newport News city
Norfolk city
Poquoson city
Portsmouth city
Suffolk city
Virginia Beach city
Number
of Coastal
Block
Groups
2
3
1
11
4
9
1
9
9
4
3
2
13
14
29
53
40
109
10
53
5
18
Number of Specified Single-
Unit Dwellings per Block
Group (Nk)
Mean
200
227
398
325
700
301
497
267
363
362
307
317
311
549
347
301
281
168
348
296
1,426
391
Std.
Dev.
7
175
0
78
335
88
0
48
124
220
33
67
83
257
200
149
287
118
150
242
644
210
Min
193
3
398
145
279
138
497
178
219
88
276
250
181
97
38
17
5
7
156
14
545
14
Max
206
431
398
447
1,131
415
497
365
670
591
353
384
458
1,114
747
810
1,374
528
733
1,031
2,198
829
Average Value of Specified Units
per Block Group (Pk)
Mean
$121,113
$193,030
$128,864
$255,256
$203,309
$205,784
$188,507
$151,509
$233,922
$197,887
$172,231
$172,904
$161,391
$239,291
$140,654
$134,117
$122,332
$168,857
$226,740
$115,948
$214,610
$218,105
Std. Dev.
$3,296
$31,483
$0
$51,809
$28,957
$36,392
$0
$34,176
$35,330
$46,549
$19,788
$12,214
$32,705
$54,067
$51,189
$43,962
$59,341
$98,567
$58,274
$40,479
$35,399
$81,187
Min
$117,817
$151,588
$128,864
$157,850
$153,167
$146,531
$188,507
$100,355
$182,242
$144,641
$157,244
$160,690
$98,923
$123,412
$61,354
$67,363
$37,130
$51,507
$132,643
$54,008
$168,310
$132,151
Max
$124,409
$227,847
$128,864
$339,499
$220,637
$249,250
$188,507
$214,862
$303,884
$269,254
$200,192
$185,119
$228,513
$333,513
$261,780
$269,833
$316,437
$590,748
$311,350
$243,782
$254,324
$396,073
Final Risk and Exposure Assessment
Appendix 8 -
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.2.1.4.2 Limitations and Uncertainties
Many of the limitations and uncertainties surrounding this benefit transfer model are
associated with the limitations and uncertainties inherent in the hedonic "implicit price" estimate,
MWTPk. From a strictly conceptual standpoint, the hedonic implicit price provides a correct
measure of the welfare gains to residents of relatively small and localized improvements in the
amenity, in this case changes in DIN water quality. However, caution is required when using this
implicit price to estimate the benefits of either a large water-quality change or a change that
affects many housing consumers. The accuracy of the benefit transfer model summarized by
Equation (4.7) will tend to decline as the value ofADINk increases and as ^increases. This is
because changes that are larger and that affect more consumers are also more likely to cause
shifts in the housing market, resulting in potentially large transaction (e.g., moving) costs and
changes in the market price equilibrium. Nevertheless, Bartik (1988) has shown that, under many
common conditions, models such as Equation (4.7) can be interpreted as providing an upper-
bound estimate of aggregate benefits.
From an empirical standpoint, there are other potential limitations and uncertainties. First,
there are potential errors in the hedonic parameter estimate. For example, DINmay be correlated
with other influential housing or neighborhood characteristics that are not included in the
hedonic model, in which case the parameter estimate is likely to overstate the implicit price of
DIN. Second, for this benefit transfer model, it was assumed that the Census block groups along
the Chesapeake Bay coast represent the areas in which the hedonic estimates can most
reasonably be applied; however, this spatial extrapolation has inherent limitations. In particular,
the implicit price estimates are expected to be less accurate as a measure of WTP in areas that are
farther from the hedonic study area (e.g., St. Mary's River watershed), particularly areas that are
more urban and densely populated. By excluding homes in other noncoastal Census block groups
that are also near the Bay, the benefit transfer model is also likely to exclude some beneficiaries
of improved aesthetic services and, therefore, underestimate aggregate benefits. Third, the
implicit price was measured using data on individual homes and water quality measures within at
most a few miles from these homes; however, the model summarized in Equation (4.7) uses
properties aggregated at the Census block group level and (most likely) more spatially averaged
Final Risk and Exposure Assessment September 2009
Appendix 8-89
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
water quality. These differences are likely to reduce the accuracy of applying Equation (4.7) to
estimate benefits.
It is also important to recognize the expected overlap in ecosystem services captured by
the hedonic implicit price estimates and the WTP estimates summarized in Section 4.2. In
principle, the hedonic price estimate includes residents' values for all of the use-related services
they receive that depend on water quality. Therefore, in addition to capturing the aesthetic
services received by living near the Bay, the hedonic implicit price should include values for
recreational services received by near-shore residents. Unfortunately, the hedonic estimates do
not provide separate value estimates for these different use-related services. Decomposing the
value estimates into separate use-related categories requires additional assumptions, data, or
analysis.
Finally, to specify reductions in DIN levels across the Bay resulting from a 24%
reduction in nitrogen loadings, strong assumptions were made that DIN levels decline by the
same percentage. In addition, to address variation in initial DIN levels across the Bay, this
percentage reduction was applied to the range (25th to 75th percentile) of recently observed
values in the Potomac River estuary.
4.2.1.5 Nonuse Services
Some of the ecosystem services provided by the Chesapeake Bay may be independent of
individuals' recreational or other specific uses of the estuary. Measuring values for these nonuse
services is more difficult and involves more uncertainty than for recreational and aesthetic
services. Nevertheless, several stated-preference studies have estimated water quality values
using sample populations that include nonusers. Evidence from these studies indicates that,
compared with users of water resources, nonusers have significantly lower but still positive WTP
for water quality improvements. Based on this evidence, the following simple benefit transfer
equation was specified for estimating nonuse benefits:
Aeefi -N *WTP (AWO } (49}
^-66-^NU ~ VM7 "' ±± ATfj \L*'' \£\Q ) •> \* •" )
where
= change in Chesapeake Bay water quality, expressed on a 10-point
rating scale;
Final Risk and Exposure Assessment September 2009
Appendix 8-90
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
AggBNU = aggregate annual benefits (in 2007 dollars) to nonusers of the
Chesapeake Bay in Maryland, Virginia, and DC for a specified
Qio increase in water quality;
WTPNU(AWQio) = average annual WTP (in 2007 dollars) per nonuser, as a function of
the AWQw increase in water quality; and
NNU = total number of nonusers in Maryland, Virginia, and DC in 2007.
To estimate the WTP^u function, results from two meta-analytic studies summarizing
evidence from the water quality valuation literature were used. The first, Johnston et al. (2005),
included 81 WTP estimates from 34 stated-preference studies. Although these studies addressed
a wide variety of water quality changes, for the meta-analysis, they were all converted to a 10-
point index (where 0 and 10 represent the worst and best possible water quality, respectively)
based on the "Resources for the Future (RFF) water quality ladder" (Vaughan, 1986). The meta-
analysis regressed average WTP estimates on water quality measures (baseline and change),
characteristics of the water resource and study population, and several study method descriptors.
The resulting WTP function can be simplified and summarized as follows:13
WTPNU = exp
2.45 + (0.6827 * \n(AWQw )) - (0.129 * WQmase )
_ • (0.005 */M7/CP02)
where
*CP02, (4.10)
= baseline Chesapeake Bay water quality, expressed on the 10-point rating
scale;
INC = average annual household income of Maryland, Virginia, and DC nonusers
in 2007; and
CP02 = price adjustment factor for 2002 to 2007.
The second study, Van Houtven, Powers, and Pattanayak (2007), conducted a similar
meta-analysis using a somewhat different sample of studies (18 studies, including 11 for
freshwater resources) and WTP estimates (131). A 10-point index based on the RFF ladder was
13 The function is a simplified version of the translog unweighted parameter estimate model (Model 2) in Johnston
et al. (2005). This model includes several explanatory variables and coefficients, which are summarized in the
constant term (2.45). To derive this constant, values were assigned to the other explanatory variables as follows:
year is 2007 (yearjndex = 37), study method is a dichotomous choice through a personal interview
(discrete_ch = 1 and interview = 1) to a nonuser-only population (nonusers = 1) with a high response rate
(hi_response = 1), protest and outlier bids are excluded (protest_bids = 1 and outlier_bids = 1), and the species
benefiting from the water quality change are unspecified (InWQnon = lnwq_change).
Final Risk and Exposure Assessment September 2009
Appendix 8-91
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
also used to convert water quality changes to a common scale. The resulting WTP function from
this study can also be simplified and summarized as follows:14
WTPNU = exp
+ (0.8969 *ln(/M:/CPOO))
*CPOO, (4.11)
where
CPOO = price adjustment factor for 2000 to 2007.
Using these functions, WTPNU can be estimated for selected values ofAWQio, WQ whose,
and INC. To estimate WTP^u for a 1-unit change in the 5-point El scale, it was assumed that the
El scale is directly proportional to the 10-point WQw scale. In other words, it was assumed that
El = 1 is equivalent to WQwbase = 2, and a 1-unit increase in El is equivalent to AWQio = 2. For
INC, U.S. average household income in 2007 of $67,610 was used (U.S. Census Bureau, 2008a).
Based on these inputs, WTPNU for a 2-unit change in water quality is estimated to be $16.33
using the Johnston et al. (2005) function and $27.75 using the Van Houtven, Powers, and
Pattanayak (2007) function.
Estimates of the percentage of Maryland, Virginia, and DC residents who are nonusers of
the Chesapeake Bay are not readily available; however, they can be roughly approximated from
recreational participation statistics for the area. For example, data from the 2006 Virginia
Outdoors Survey suggest that (1) 92% of households in Virginia did not take any beach trips to
the Chesapeake Bay, (2) 84% did not engage in saltwater fishing, and (3) 92% did not engage in
powerboating. Assuming that these proportions represent independent probabilities of nonuse,
then the combined probability (proportion) of nonuse for these primary activities is roughly 70%.
Applying this percentage to the Maryland, Virginia, and DC population in 2007, which was
13,918,727, suggests that the number of nonusers (NNU) is approximately 9,743,109.
14 This function is a simplified version of the parsimonious log-linear model in Table 5 in Van Houtven, Powers,
and Pattanayak (2007). This model also includes several explanatory variables and coefficients, which are
summarized in the constant term (-1.197). To derive this constant in a way that is consistent with the previous
function, values were assigned to the other explanatory variables as follows: year is 2007 (studyyr73 = 34), study
method is a personal interview (inperson = 1) to a nonuser-only population (pctuser = 0) with a high response rate
(responserate = 100), publication outlet is peer reviewed (dpubjrlbk = 1), and the water quality change is not
expressed in terms of recreational uses (InwqlOchru = 0).
Final Risk and Exposure Assessment September 2009
Appendix 8 - 92
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.2.1.5.1 Results: Aggregate Nonuse Benefits
Applying these estimates to the benefit transfer models summarized in Equations (4.9)
through (4. 1 1), the aggregate annual nonuse benefits of a 1-unit improvement in the El scale
Qw = 2) are estimated to range from $159.1 million to $270.4 million.
4.2.1.5.2 Limitations and Uncertainties
As with the recreational boating services model described in Section 4.2.1.2, one of the
main practical limitations of applying these meta-analysis models is the water quality index used.
Translating changes in the El scale to the WQw metric requires strong assumptions. Another
inherent limitation of using the meta-analytic models as benefit transfer functions is their lack of
sensitivity to the spatial scale of water quality changes.
In addition to the limitations that primarily contribute uncertainty in the WTPNU
estimates, there is also significant uncertainty associated with the measurement ofNmj. First,
defining criteria for distinguishing users and nonusers of the Bay is somewhat inherently
subjective. Second, statistics on overall rates of visitation and use of the Bay by Maryland,
Virginia, and DC households are not readily available.
A final caveat for this approach to estimating nonuse values for water quality
improvements in the Bay is that, by design, it only includes nonuse values for nonusers.
However, it is not unreasonable to suspect that users also benefit to some extent from nonuse
services from the Bay. Whereas these types of nonuse values are likely to be captured in, for
example, the Lipton (2004) WTP values for boaters used in Equation (4.5), they are not included
in the benefit estimates in Equations (4.4), (4.6), and (4.7) for recreational anglers, beach users,
and residents, respectively.
4.2.2 Neuse River Estuary
To analyze changes in ecosystem services for the Neuse River, the results of the Neuse
River/Neuse Estuary Case Study were applied. This case study concluded that atmospheric
deposition contributes 26% (1.15 million kg nitrogen/year) of total nitrogen loadings to the
estuary. In contrast to the Potomac River/Potomac Estuary Case Study, it estimates that a much
larger reduction in nitrogen loadings than this 26% would be required to improve the Neuse
River estuary from "bad" to "poor." Therefore, for this analysis, the change in selected
Final Risk and Exposure Assessment September 2009
Appendix 8-93
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
ecosystem services associated with a 26% reduction (100% of atmospheric deposition) in
nitrogen loadings to the Neuse estuary was estimated.
4.2.2.1 Provisioning Services from the Blue Crab Fishery
As discussed in Section 4.1.1, there are few examples of empirical bioeconomic models
that link changes in nutrient-related water quality to changes in productivity of commercial
fisheries; however, one exception is a study by Smith (2007). This study, which is applied to the
Neuse River estuary, estimated the dynamic effects of a 30% reduction in nitrogen loads to the
estuary on blue crab stocks, commercial catch levels, and the producer and consumer surplus
derived from this fishery.
Smith (2007) applied a two-patch predator-prey model that incorporated both direct and
indirect effects of hypoxia (i.e., low DO) on blue crab communities. Direct effects include the
movement of blue crab to water habitats with higher DO content. Indirect effects include the
dying off of blue crab prey. The model compares producer and consumer surplus changes under
the existing open-access institutional structure to a 30% reduction of nitrogen loadings in the
same structure. The model was parameterized using results and estimates derived from several
other studies. To address uncertainty, the values of three key parameters—economic speed of
adjustment under open-access conditions, biological spatial connectivity, and price elasticity of
demand—were each allowed to take on three different values. For a 30% reduction in nitrogen
loadings to the estuary, the present value (100-year time horizon and 4.5% discount rate) of
producer benefits ranged from $0.7 million to $5.9 million (in 2002 dollars), and the present
value of consumer surplus ranged from $3.15 million to $425.20 million. The combined present
value of producer and consumer surplus changes was estimated to range from $3.8 to $31.0
million.
To estimate the annual aggregate benefits from the blue crab fishery due to a 26%
reduction in nitrogen loads, (1) the results reported in Smith (2007) were rescaled by the
percentage difference between 30% and 26%, (2) the benefit estimates (using the 100-year
horizon and 4.5% discount rate) were annualized, and (3) the estimates were converted to 2007
dollars using the CPI.
Final Risk and Exposure Assessment September 2009
Appendix 8-94
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
4.2.2.1.1 Results: Aggregate Benefits from the Blue Crab Fishery
Applying this modeling framework, the aggregate annual benefits to Neuse River crab
fishers and consumers from a 26% reduction in nitrogen loadings is estimated to range from
$0.12 million to $1.01 million.
4.2.2.1.2 Limitations and Uncertainties
The large range of the benefit estimates reported above reflects uncertainty in three key
model parameters—economic speed of adjustment under open-access conditions, biological
spatial connectivity, and price elasticity of demand. However, the model includes at least 16
other parameters whose values are drawn from other studies; thus, the overall uncertainty in
these benefit estimates is most likely understated by this range. In addition, by simply reseating
the results reported in Smith (2007) to address a 26% rather than a 30% reduction in nitrogen
loads, it was assumed that benefits are directly proportional to the percentage reduction in
nitrogen loads. This assumption adds additional (albeit, most likely small) uncertainty to the
reported benefit estimates.
4.2.2.2 Recreational Fishing Services
To estimate the benefits from improvements in recreational fishing services due to
reductions in nitrogen loadings to the Neuse, a benefit transfer model originally developed to
assess the nutrient-reduction benefits of EPA's effluent guidelines for Consolidated Animal
Feeding Operations (CAFOs) (EPA, 2002) was applied. For that analysis, EPA conducted a case
study evaluating the potential economic benefits of a reduction in nutrient loadings via changes
in recreational fishing opportunities in North Carolina's Albemarle and Pamlico Sounds (APS)
estuary (Van Houtven and Sommer, 2002). The Neuse River estuary is a subestuary within the
APS system.
To estimate the value of reductions in nitrogen loads, the APS case study relied on
economic value estimates obtained from two related studies—Kaoru (1995) and Kaoru, Smith,
and Liu (1995). Both studies used recreational data obtained from a 1981-1982 intercept survey
of recreational fishermen conducted at 35 boat ramps or marinas within the APS estuary.
Kaoru (1995) used a three-level nested random utility model (RUM), which broke the
recreational fishing decision into three stages: a decision on the duration of the trip (1, 2, 3, or
more than 3 days), a decision on which of the five regions to visit, and a decision on which of the
Final Risk and Exposure Assessment September 2009
Appendix 8-95
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
individual sites within the region to visit. The impact of nitrogen (and phosphorus) loadings was
specifically investigated in the second stage of the decision process (regional choice). A 25%
reduction in nitrogen loadings for the entire APS estuary resulted in a benefit estimate of $4.70
(in 1982 dollars) per person-trip.
Kaoru, Smith, and Liu (1995) also used a RUM approach to estimate the value of
improving water quality. First, a household production function (HPF) was estimated to predict
expected catch rates for individuals based on variables such as equipment used; effort exerted;
and the physical characteristics of the fishing site, including pollutant loadings. Second, the HPF
model was used to predict the impact of a 36% reduction in nitrogen loadings on expected catch
rates. The estimated values ranged from $0.76 to $6.52 (in 1982 dollars) per person-trip.
Based on a systematic review of the value estimates reported in these studies, the CAFO
case study selected three estimates to include in the benefit transfer model — $4.70 per person-
trip, for a 25% reduction in nitrogen loads (Kaoru, 1995) and $3.95 and $6.52 per person, for a
36% reduction (Kaoru, Smith, and Liu, 1995).
To apply these estimates, they were converted to comparable units. First, they were
converted to 2007 dollars using the CPI. Second, they were rescaled to values per 1% reduction
in loadings (i.e., dividing by 25 and 36, respectively). The resulting three unit values are $0.40,
$0.24, and $0.39 per person-trip per 1% reduction in nitrogen loads to the APS.
A further adjustment is necessary to convert these values into per-ton units. According to
Kaoru (1995), the average nitrogen load to the APS estuary at the time the study was conducted
was 1,741 tons per bordering county per year, which translates to a total of 22,633 tons of
nitrogen loadings per year because of the 13 counties bordering the APS estuary in North
Carolina. The resulting three unit values are $0.0018, $0.0010, and $0.0017 per person-trip per
1-ton reduction in nitrogen loads to the APS.
To estimate the aggregate annual recreational fishing benefits of total reductions in
nitrogen loads to the APS estuary, the following benefit transfer equation was specified:
= FX AL x T, (4.12)
where
= the aggregate annual recreational fishing benefits from reductions in
nitrogen loads to the APS estuary (in 2007 dollars),
Final Risk and Exposure Assessment September 2009
Appendix 8-96
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
V = the annual per trip value per unit (either in tons per year or percentage)
reduction in nitrogen (in 2007 dollars),
AL = reduction in nitrogen loadings (either in tons per year or percentage) to
the APS estuary, and
T = the total number of annual fishing trips to the APS estuary (person-trips
per year).
Although the unit value (F) estimates derived from Kaoru (1995) and Kaoru, Smith, and
Liu (1995) are based on data only for boating anglers, it was assumed that they apply to all
recreational fishing trips (7) in the APS. Data on visitation rates for recreational anglers in the
APS estuary are available from the MRFSS, which contains information on the number, type,
and destination of recreational fishers for several coastal regions in the United States. For 2006,
the MRFSS data provide an estimate of 753,893 person-trips to the APS for recreational fishing.
4.2.2.2.1 Results: Aggregate Recreational Fishing Benefits
As noted above, the findings of the Neuse River/Neuse Estuary Case Study indicate that
eliminating atmospheric deposition of nitrogen to the Neuse watershed would reduce nitrogen
loads to the Neuse estuary (and, thus, the APS estuary as well) by 1.15 million kg per year,
which is equivalent to 1,268 tons of nitrogen per year. Assuming that annual recreational fishing
levels in the APS remain at 2006 levels and applying Equation (4.12), the resulting aggregate
annual benefits (AggBAPsftsh) of such a reduction are estimated to be between $1.0 million and
$1.7 million.
If the Neuse case study results regarding the portion of nitrogen loadings attributable to
atmospheric deposition (26%) are extended to the entire APS system, then this extrapolation
implies that eliminating all atmospheric nitrogen loads to the APS watershed would also reduce
annual nitrogen loads to the APS estuary by 26%. Applying Equation (4.11) to this scenario
suggests that the aggregate recreational fishing benefits of zeroing out nitrogen deposition in the
entire APS watershed would be between $4.6 million and $7.9 million.
4.2.2.2.2 Limitations and Uncertainties
The following limitations and uncertainties should be considered when interpreting these
recreational fishing benefit estimates. First, the value estimates are based on fishing activity data
Final Risk and Exposure Assessment September 2009
Appendix 8-97
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
that are more than 2 decades old. The analysis assumes that the benefits of water quality changes
have remained constant (in real terms) over this period.
Second, the value estimates obtained from the two existing studies were based on
percentage reductions in nutrients that were uniform across the APS estuary. By converting these
estimates into per-ton terms and applying them only to the Neuse River nitrogen load reductions,
the analysis implicitly assumes that average per-trip benefits do not vary with respect to the
spatial distribution of the loadings reductions.
Third, the original value estimates are based on data only from boat fishermen; however,
the analysis assumes that these values are appropriate for both boat and nonboat fishers.
4.3 REFERENCES
Anderson, E. 1989. "Economic Benefits of Habitat Restoration: Seagrass and the Virginia Hard-
Shell Blue Crab Fishery." North American Journal of Fisheries Management 9:140-149.
Bartik, TJ. 1988. "Measuring the Benefits of Amenity Improvements in Hedonic Price Models."
Land Economics 64(2): 172-183.
Bockstael, N.E., K.E. McConnell, and I.E. Strand. 1988. "Benefits from Improvements in
Chesapeake Bay Water Quality, Volume III." Environmental Protection Agency
Cooperative Agreement CR-811043-01-0.
Bockstael, N.E., K.E. McConnell, and I.E. Strand. 1989. "Measuring the Benefits of
Improvements in Water Quality: The Chesapeake Bay." Marine Resource Economics
6:1-18.
Bricker, S., D. Lipton, A. Mason, M. Dionne, D. Keeley, C. Krahforst, J. Latimer, and J.
Pennock. 2006. Improving Methods and Indicators for Evaluating Coastal Water
Eutrophication: A Pilot Study in the Gulf of Maine. NOAA Technical Memorandum
NOS NCCOS 20. Silver Spring, MD: National Ocean Service, National Centers for
Coastal Ocean Science, Center for Coastal Monitoring and Assessment.
Bricker, S., Longstaff, B., Dennison, W., Jones, A., Boicourt, K., Wicks, C., and Woerner, J.
2007. Effects of Nutrient Enrichment in the Nation's Estuaries: A Decade of Change.
Final Risk and Exposure Assessment September 2009
Appendix 8-98
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
NOAA Coastal Ocean Program Decision Analysis Series No. 26. National Centers for
Coastal Ocean Science, Silver Spring, MD.
Bricker, S.B., J.G. Ferreira, and T. Simas. 2003. "An Integrated Methodology for Assessment of
Estuarine Trophic Status." Ecological Modelling 169:39-60.
Chesapeake Bay Program, n.d. "Chesapeake Bay Program: A Watershed Partnership." Available
at http://www.chesapeakebay.net/.
Chesapeake Bay Program. "Bay Grass Abundance (Baywide)." Available at
http://www.chesapeakebay.net/ status_baygrasses.aspx?menuitem=19652. Last updated
May 21,2008.
Chesapeake Eco-Check. 2007. "Indicator Details: 2007. Water Quality Index." Available at
http://www.eco-check.org/reportcard/chesapeake/2007/indicators/water_quality_index/.
Environmental Systems Research Institute, Inc. (ESRI). 2002. "Census Block Groups: DC, DE,
MD, and VA." [CD-ROM]. ESRI Data & Maps 2002.
Johnston, R.J., E.Y. Besedin, R. lovanna, CJ. Miller, R.F. Wardwell, and M.H. Ranson. 2005.
"Systematic Variation in Willingness to Pay for Aquatic Resource Improvements and
Implications for Benefit Transfer: A Meta-Analysis." Canadian Journal of Agricultural
Economics 53:221-248.
Johnston, R.J., M.H. Ranson, E.Y. Besedin, and E.G. Helm. 2006. "What Determines
Willingness to Pay per Fish? A Meta-Analysis of Recreational Fishing Values." Marine
Resource Economics 21:1-32.
Kahn, J.R., and W.M. Kemp. 1985. "Economic Losses Associated with the Degradation of an
Ecosystem: The Case of Submerged Aquatic Vegetation in Chesapeake Bay." Journal of
Environmental Economics and Management 12:246-263.
Kaoru, Y. 1995. "Measuring Marine Recreation Benefits of Water Quality Improvements by the
Nested Random Utility Model." Resource and Energy Economics 17(2): 119-136.
Final Risk and Exposure Assessment September 2009
Appendix 8-99
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Kaoru, Y., V. Smith, and J.L. Liu. 1995. "Using Random Utility Models to Estimate the
Recreational Value of Estuarine Resources." American Journal of Agricultural
Economics 77:141-151.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number IMDE-02-0070.
Knowler, D. 2002. "A Review of Selected Bioeconomic Models: With Environmental Influences
in Fisheries." Journal of Bioeconomics 4:163-181.
Krupnick, A. 1988. "Reducing Bay Nutrients: An Economic Perspective."MarylandLaw
Review 47(2):453-480.
Leeworthy, V.R., and P.C. Wiley. 2001. Current Participation Patterns inMarine Recreation.
Silver Spring, MD: U.S. Department of Commerce, National Oceanic and Atmospheric
Administration, National Ocean Service, Special Projects.
Lipton, D. W. 1999. "Pfiesteria's Economic Impact on Seafood Industry Sales and Recreational
Fishing." In Proceedings of the Conference, Economics of Policy Options for Nutrient
Management andPfiesteria. B. L. Gardner and L. Koch, eds., pp. 35-38. College Park,
MD: Center for Agricultural and Natural Resource Policy, University of Maryland.
Lipton, D. 2004. "The Value of Improved Water Quality to Chesapeake Bay Boaters." Marine
Resource Economics 19:265-270.
Lipton, D. 2006. "Economic Impact of Maryland Boating in 2005." University of Maryland Sea
Grant Extension Program. Available at
ftp://ftp.mdsg.umd.edu/Public/MDSG/rec_boat05.pdf.
Lipton, D. 2008. "Economic Impact of Maryland Boating in 2007." University of Maryland Sea
Grant Extension Program. Available at
ftp://ftp.mdsg.umd.edu/Public/MDSG/rec_boat07.pdf.
Final Risk and Exposure Assessment September 2009
Appendix 8-100
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Lipton, D.W., and R.W. Hicks. 1999. "Linking Water Quality Improvements to Recreational
Fishing Values: The Case of Chesapeake Bay Striped Bass." In Evaluating the Benefits of
Recreational Fisheries. Fisheries Centre Research Reports 7(2), TJ. Pitcher, ed., pp.
105-110. Vancouver: University of British Columbia.
Lipton, D.W., and R.W. Hicks. 2003. "The Cost of Stress: Low Dissolved Oxygen and the
Economic Benefits of Recreational Striped Bass Fishing in the Patuxent River." Estuaries
26(2A):310-315.
Massey, M., S. Newbold, and B. Gentner. 2006. "Valuing Water Quality Changes Using a
Bioeconomic Model of a Coastal Recreational Fishery." Journal of Environmental
Economics and Management 52:482-500.
Millennium Ecosystem Assessment (MEA). 2005. Ecosystems and Human Well-being:
Wetlands and Water. Synthesis. A Report of the Millennium Ecosystem Assessment.
Washington, DC: World Resources Institute.
Mistiaen, J.A., I.E. Strand, and D. Lipton. 2003. "Effects of Environmental Stress on Blue Crab
(Callinectes sapidus) Harvests in Chesapeake Bay Tributaries." Estuaries 26(2a): 316-
322.
Morgan, C., and N. Owens. 2001. "Benefits of Water Quality Policies: The Chesapeake Bay."
Ecological Economics 39:271-284.
Murray, T., and J. Lucy. 1981. "Recreational Boating in Virginia: A Preliminary Analysis."
Special Report in Applied Marine Science and Ocean Engineering No. 251. Available at
http://www.vims.edu/GreyLit/VIMS/sramsoe251 .pdf
National Marine Manufacturers Association (NMMA). 2008. "2007 Recreational Boating
Statistical Abstract." Available at
http://www.nmma.org/facts/boatingstats/2007/files/Abstract.pdf.
National Oceanic and Atmospheric Administration (NOAA). (2007, August). "Annual
Commercial Landing Statistics." Available at http://www.st.nmfs.noaa.gov/
stl/commercial/landings/annual_landings.html.
Final Risk and Exposure Assessment September 2009
Appendix 8-101
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
National Oceanic and Atmospheric Administration (NOAA). "Recreational Fisheries."
Available at www.st.nmfs.noaa.gov/stl/recreational/overview/overview.html. Last
updated February 12, 2009.
Parsons, G., A.O. Morgan, J.C. Whitehead, and T.C. Haab. 2006. "The Welfare Effects of
Pfiesteria-Related Fish Kills: A Contingent Behavior Analysis of Seafood Consumers."
Agricultural and Resource Economics Review 35(2): 1-9.
Poor, P.J., K.L. Pessagno, and R.W. Paul. 2007. "Exploring the Hedonic Value of Ambient
Water Quality: A Local Watershed-Based Study." Ecological Economics 60:797-806.
Smith, M.D. 2007. "Generating Value in Habitat-Dependent Fisheries: The Importance of
Fishery Management Institutions." Land Economics 83(l):59-73.
Sweeney, J. 2007. "Impacts of CAMD 2020 CAIR on Nitrogen Loads to the Chesapeake Bay."
University of Maryland, Chesapeake Bay Program Office.
U.S. Census Bureau. 2008a. "Annual Social and Economic (ASEC) Supplement." Available at
http://pubdb3.census.gov/macro/032008/hhinc/new02_001.htm.
U.S. Census Bureau. 2008b. "Housing Unit Estimates for Counties of MD and VA: April 1/2000
to July 1/2007." Available at http://www.census.gov/popest/housing/HU-EST2007-
CO.html.
U.S. Environmental Protection Agency (EPA). 2002. December. Environmental and Economic
Benefit Analysis of Final Revisions to the National Pollutant Discharge Elimination
System Regulation and the Effluent Guidelines for Concentrated Animal Feeding
Operations (EPA-821-R-03-003). Washington, DC: U.S. Environmental Protection
Agency, Office of Water, Office of Science and Technology.
Valigura, R.A., R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacy, and R.E.
Turner. 2001. Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective.
Washington, DC: American Geophysical Union.
Final Risk and Exposure Assessment September 2009
Appendix 8-102
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Van Houtven, G. and A. Sommer. December 2002. Recreational Fishing Benefits: A Case Study
of Reductions in Nutrient Loads to the Albemarle-Pamlico Sounds Estuary. Final Report.
Prepared for the U.S. Environmental Protection Agency. Research Triangle Park, NC:
RTI International.
Van Houtven, G.L., J. Powers, and S.K. Pattanayak. 2007. "Valuing Water Quality
Improvements Using Meta-Analysis: Is the Glass Half-Full or Half-Empty for National
Policy Analysis?" Resource and Energy Economics 29:206-228.
Vaughan, WJ. 1986. "The Water Quality Ladder." Included as Appendix B in R.C. Mitchell,
and R.T. Carson, eds. The Use of Contingent Valuation Data for Benefit/Cost Analysis in
Water Pollution Control. CR-810224-02. Prepared for the U.S. Environmental Protection
Agency, Office of Policy, Planning, and Evaluation.
Virginia Department of Conservation and Recreation. 2007. "2006 Virginia Outdoors Survey."
Available at
http://www.dcr.virginia.gov/recreati onal_planning/documents/vopsurvey06.pdf
Whitehead, J.C., T.C. Haab, and G.R. Parsons. 2003. "Economic Effects of Pfiesteria." Ocean &
Coastal Management 46(9-10): 845 -8 5 8.
Final Risk and Exposure Assessment September 2009
Appendix 8-103
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
5.0 TERRESTRIAL ENRICHMENT
Terrestrial enrichment occurs when terrestrial ecosystems receive nitrogen loadings in
excess of natural background levels, either through atmospheric deposition or direct application.
Evidence presented in the Integrated Science Assessment (Environmental Protection agency
[EPA], 2008) supports a causal relationship between atmospheric nitrogen deposition and
biogeochemical cycling and fluxes of nitrogen and carbon in terrestrial systems. Furthermore,
evidence summarized in the report supports a causal link between atmospheric nitrogen
deposition and changes in the types and number of species and biodiversity in terrestrial systems.
The Terrestrial Nutrient Enrichment Case Study focuses on the coastal sage scrub (CSS)
ecosystem and San Bernardino and Sierra Nevada mixed conifer forests (MCF), both located in
California. CSS is a unique and endemic ecosystem that provides habitat to several threatened
and endangered species. Additionally, CSS is generally less fire prone than the nitrophyllous
species that tend to amass dominance in abundance and richness with increased nutrient
enrichment. MCF provide habitat for animals as well as contribute other ecosystem services such
as timber, recreation, and water cycling. Nitrogen enrichment occurs over a long time period; as
a result, it may take as much as 50 years or more to see changes in ecosystem conditions and
indicators. This long time scale also affects the timing of the ecosystem service changes.
The Terrestrial Nutrient Enrichment Case Study differs from the other case studies in that
it focuses on geographic information system (GIS) analyses and existing nitrogen loading
threshold investigations as the basis for describing endpoints. The CSS investigation analyzed
GIS data in conjunction with the results from Community Multiscale Air Quality (CMAQ) 2002
modeling to assess the relationship between atmospheric nitrogen deposition and changes in the
CSS ecosystem. In the San Bernardino and Sierra Nevada MCF, nitrogen loading thresholds
obtained in situ and through simulation modeling were investigated for potential endpoints and
applicability to the San Bernardino and Sierra Nevada MCF system.
5.1 OVERVIEW OF AFFECTED ECOSYSTEM SERVICES
The ecosystem service impacts of terrestrial nutrient enrichment include primarily
cultural and regulating services. In CSS, concerns focus on a decline in CSS and an increase in
nonnative grasses and other species, impacts on the viability of threatened and endangered
species associated with CSS, and an increase in fire frequency. Changes in MCF include changes
Final Risk and Exposure Assessment September 2009
Appendix 8-104
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
in habitat suitability and increased tree mortality, increased fire intensity, and a change in the
forest's nutrient cycling that may affect surface water quality through nitrate leaching (EPA,
2008).
Both CSS and MCF are located in areas of California valuable for housing, recreation,
and development. CSS runs along the coast through densely populated areas of California (see
Figure 5.1-1). From Figure 5.1-2, MCF covers less densely populated areas that are valuable for
recreation. The proximity of CSS and MCF to population centers and recreational areas and the
potential value of these landscape types in providing regulating ecosystem services suggest that
the value of preserving CSS and MCF to California could be quite high. The value that
California residents and the U.S. population as a whole place on CSS and MCF habitats is
reflected in the various federal, state, and local government measures that have been put in place
to protect these habitats. Threatened and endangered species are protected by the Endangered
Species Act. The State of California passed the Natural Communities Conservation Planning
Program (NCCP) in 1991, and CSS was the first habitat identified for protection under the
program (see www.dfg.ca.gov/habcon/nccp). Figure 5.1-3 shows the boundaries of the NCCP
region and subregions for CSS. Private organizations such as The Nature Conservancy, the
Audubon Society, and local land trusts also protect and restore CSS and MCF habitat. According
to the 2005 National Land Trust Census Report (Land Trust Alliance, 2006), California has the
most land trusts of any state with a total of 1,732,471 acres either owned, under conservation
easement, or conserved by other means.
5.1.1 Cultural
The primary cultural ecosystem services associated with CSS and MCF are recreation,
aesthetic, and nonuse values. The possible ecosystem service benefits from reducing nitrogen
enrichment in CSS and MCF are discussed below, and a general overview of the types and
relative magnitude of the benefits is provided.
CSS, once the dominant landscape type in the area, is a unique ecosystem that provides
cultural value to California and the nation as a whole. Culturally, the remaining patches of CSS
contain a number of threatened and endangered species, and patches of CSS are present in a
number of parks and recreation areas. More generally the patches of CSS represent the iconic
landscape type of Southern California and serve as a reminder of what the area looked like
Final Risk and Exposure Assessment September 2009
Appendix 8-105
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
predevelopment. Changes that might impact cultural ecosystem services in CSS resulting from
nutrient enrichment potentially include
• decline in CSS habitat, shrub abundance, and species of concern;
• increased abundance of nonnative grasses and other species; and
• increase in wildfires.
Legend
HH Coastal Sage Scrub 2002
Block Groups
Projected 2007 Population
0 - 1.500
1.501 - 3.000
3.001 - e.ooo
BH 6,001 - 15,000
^^B 15,001 -38.798
Figure 5.1-1. Coastal Sage Scrub Areas and Population
Final Risk and Exposure Assessment
Appendix 8-106
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
:
f,
T " •"
\
Legend
IB Mixed Conifer Forest
Block Groups
Projected 2007 Population
0-1,500
1,501 -3,000
3,001 - 6,000
^B 6,001 -15,000
^H 15,001 - 38,798
Figure 5.1-2. Mixed Conifer Forest Areas and Population
Final Risk and Exposure Assessment
Appendix 8-107
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
San Bernardino
Riverside .
Hem et
I I NCCP Region
Subregional Planning Areas
Camp Pendleton Resource Management Plan
Coastal/Central Orange County NCCP
Northern Orange Count/ Subregion
PalosVerdes Peninsula NCCP
San Bernardino Valley-wide Multi-Species Habitat Conservation Plan
San Diego Multiple Habitat Conservation and Open Space Program (MHCOSP)
San Diego Multiple Species Conservation Program (MSCP)
San Diego Multiple Habitat Conservation Program (MHCP)
San Diego Northern MSCP Subarea
Southern Orange County NCCP
Western Riverside County Multiple Species Habitat Conservation Plan
Figure 5.1-3. Boundaries of the NCCP Region and Subregions for Coastal Sage Scrub
(Source: California Department of Fish and Game, n.d.).
Final Risk and Exposure Assessment
Appendix 8-108
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
For MCF, the changes from nutrient enrichment that might impact cultural ecosystem
services include
• change in habitat suitability and increased tree mortality and
• decline in MCF aesthetics.
5.1.1.1 Recreation
CSS and MCF are found in numerous recreation areas in California. Three national parks
and monuments in California contain CSS, including Cabrillo National Monument, Channel
Islands National Park, and Santa Monica National Recreation Area. All three parks showcase
CSS habitat with educational programs and information provided to visitors, guided hikes, and
research projects focused on understanding and preserving CSS. Together a total of 1,456,879
visitors traveled through these three parks in 2008. MCF is highlighted in Sequoia and Kings
Canyon National Park, Yosemite National Park, and Lassen Volcanic National Park, where a
total of 5,313,754 people visited in 2008. Figure 5.1-4 maps national and state parkland against
MCF areas.
In addition, numerous state and county parks encompass CSS and MCF habitat. Visitors
to these parks engage in activities such as camping, hiking, attending educational programs,
horseback riding, wildlife viewing, water-based recreation, and fishing. For example,
California's Torrey Pines State Natural Reserve protects CSS habitat (see
http://www.torreypine.org/).
Table 5.1-1 reports the results from the 2006 National Survey of Fishing, Hunting, and
Wildlife Associated Recreation (FHWAR) for California (DOI, 2007) on the number of
individuals involved in fishing, hunting, and wildlife viewing in California. Millions of people
are involved in just these three activities each year. The quality of these trips depends in part on
the health of the ecosystems and their ability to support the diversity of plants and animals found
in important habitats. Based on estimates from Kaval and Loomis (2003), in the Pacific Coast
region of the United States, a day of fishing has an average value of $48.86 (in 2007 dollars),
based on 15 studies. For hunting and wildlife viewing in this region, average day values were
estimated to be $50.10 and $79.81 from 18 and 23 studies, respectively. Multiplying these
average values by the total participation days reported in Table 5.1-1, the total benefits in 2006
from fishing, hunting, and wildlife viewing away from home in California were approximately
$947 million, $169 million, and $3.59 billion, respectively.
Final Risk and Exposure Assessment September 2009
Appendix 8-109
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
In addition, data from California State Parks (2003) indicate that in 2002 68.7% of adult
residents participated in trail hiking for an average of 24.1 days per year. Applying these same
rates to Census estimates of the California adult population in 2007 suggests that residents in
California hiked roughly 453 million days in 2007. According to Kaval and Loomis (2003), the
average value of a hiking day in the Pacific Coast region is $25.59, based on a sample of 49
studies. Multiplying this average-day value by the total participation estimate indicates that the
aggregate annual benefit for California residents from trail hiking in 2007 was $11.59 billion.
| California Counties
Mixed Conifer Forest
Parks
National
State
Figure 5.1-4. Mixed Conifer Forest Areas and National and State Park
Boundaries
Final Risk and Exposure Assessment
Appendix 8-110
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table 5.1-1. Recreational Activities in California in 2006 by Residents and Nonresidents
Activities in California by Residents and Nonresidents
Fishing
Anglers 1,730,000
Days of fishing 19,394,000
Average days per angler 11
Hunting
Hunters 281,000
Days of hunting 3,376,000
Average days per hunter 12
Wildlife Watching
Total wildlife-watching participants 6,270,000
Away-from-home participants 2,894,000
Around-the-home participants 5,259,000
Days of participation away from home 45,010,000
Average days of participation away from home 16
Activities in California by Residents
Fishing
Anglers 1,578,000
Days of fishing 18,310,000
Average days per angler 12
Hunting
Hunters 274,000
Days of hunting 3,339,000
Average days per hunter 12
Wildlife Watching
Total wildlife-watching participants 5,704,000
Away-from-home participants 2,328,000
Around-the-home participants 5,259,000
Days of participation away from home 41,436,000
Average days of participation away from home 18
Source: U.S. Department of the Interior, Fish and Wildlife Service, and U.S. Department of Commerce, U.S. Census
Bureau, 2007.
The potential impacts of an increase in wildfires on recreation are discussed in Section
5.1.2, Regulating.
Final Risk and Exposure Assessment September 2009
Appendix 8-111
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
5.1.1.2 Aesthetic
Beyond the recreational value, the CSS landscape and MCF provide aesthetic services to
local residents and homeowners who live near CSS or MCF. Aesthetic services not related to
recreation include the view of the landscape from houses, as individuals commute, and as
individuals go about their daily routine in a nearby community. Studies find that scenic
landscapes are capitalized into the price of housing. Although no studies came to light that look
at the value of housing as a function of the view in landscapes that include CSS or MCF, other
studies document the existence of housing price premia associated with proximity to forest and
open space (Acharya and Bennett, 2001; Geoghegan, Wainger, and Bockstael, 1997; Irwin,
2002; Mansfield, et al., 2005; Smith, Poulos, and Kim, 2002; Tyrvainen and Miettinen, 2000).
The CSS landscape itself is closely associated with Southern California, which should increase
the aesthetic value of the landscape in general. Figure 5.1-5 presents home values in 2000 by
Census block and CSS areas. CSS areas border a number of areas along the coast near large
cities with very high home values, as well as areas between the cities where home values are
lower.
5.1.1.3 Nonuse Value
Nonuse value, also called existence value or preservation value, encompasses a variety of
motivations that lead individuals to place value on environmental goods or services that they do
not use. The values individuals place on protecting rare species, rare habitats, or landscape types
that they do not see or visit and that do not contribute to the pleasure they get from other
activities are examples of nonuse values.
While measuring the public's willingness to pay (WTP) to protect endangered species
poses theoretical and technical challenges, it is clear that the public places a value on preserving
endangered species and their habitat. Data on charitable donations, survey results, and the time
and effort different individuals or organizations devote to protecting species and habitat suggest
that endangered species have intrinsic value to people beyond the value derived from using the
resource (recreational viewing or aesthetic value). CSS and MCF are home to a number of
important and rare species and habitat types. CSS displays richness in biodiversity with more
than 550 herbaceous annual and perennial species. Of these herbs, nearly half are endangered,
Final Risk and Exposure Assessment September 2009
Appendix 8-112
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
sensitive, or of special status (Burger et al., 2003). Additionally, avian, arthropod, herpetofauna,
and mammalian species live in CSS habitat or use the habitat for breeding or foraging.
Legend
HH! Coastal Sage Scrub 2002
Home Value by Census Tract
Year 2000
$0.00 -$150,000
$150,000- $275.000
$275,000 - $450,000
^H $450,000 - $725.000
^^H $725,OOO - $1,OOO,OO1
Figure 5.1-5. Coastal Sage Scrub Areas and Housing Values
Figure 5.1-6 shows communities of CSS and three important federally endangered
species. MCF is home to one federally endangered species and a number of state-level sensitive
species. Figure 5.1-7 provides a map of MCF habitat and two threatened and endangered
Final Risk and Exposure Assessment
Appendix 8-113
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
species. The Audubon Society lists 28 important bird areas in CSS habitat and at least 5 in MCF
in California (http://ca.audubon.org/iba/index.shtml). 1
t*
VjmaiSp
Coastal Sage Scrub
Quino Checkerspot
Kangaroo Ral
Coaslal Cfk Gnatcataher
[ | County
—
Source of CSS range is inc. California Department
of Forestry anrj Fire PrtHidion.
Source of critical habitats is Itie US Fish and Wild life
sanrice Cfiteal wadlite Portal.
Figure 5.1-6. Presence of Three Threatened and Endangered Species in
California's Coastal Sage Scrub Ecosystem
1 Important bird areas are sites that provide essential habitat for one or more species of bird.
Final Risk and Exposure Assessment
Appendix 8-114
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Can&fpmiT tCslj^wst*1/ *\
/fs i_San jQjquin >
S^n IhiioHiniriJ^"' YT«BHffllW\
Source or the MountMi "VSIbw-te^ged Frog ana
Peninsular Big Horn Sheep range Is ihe US Fare si
5*>r.-ee Ore leal na&:iai Poral
Source or Mixed CM iler MCA. Ospt ti( Foresiry and Fir
Figure 5.1-7. Presence of Two Threatened and Endangered Species in
California's Mixed Conifer Forest
To the authors' knowledge, only one study has specifically estimated values for
protecting CSS habitat in California. Stanley (2005) uses a contingent valuation (CV) survey to
measure WTP to support recovery plans for endangered species in Southern California. The
survey of Orange County, California, residents asked respondents to value the recovery of a
single species (the Riverdale fairy shrimp) and a larger bundle of 32 species found in the county.
The acquisition of critical habitat and implementation of the recovery plan were the specific
goods being valued in the WTP question and the programs would be financed by an annual tax
payment. The average WTP for fairy shrimp recovery was roughly $29 (in 2007 dollars) and for
Final Risk and Exposure Assessment
Appendix 8-115
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
all 32 species was $61 per household, depending on the model used. Aggregating benefits
(multiplying average household WTP by the number of households in the county) results in total
estimated WTP of over $27 million annually for protecting fairy shrimp and $57 million
annually for all 32 species.
In a more general study valuing endangered species protection, Loomis and White (1996)
synthesize key results from 20 threatened and endangered species valuation studies using meta-
analysis methods. They find that annual WTP estimates range from a low of $11 for the Striped
Shiner fish to a high of $178 for the Northern Spotted Owl (in 2007 dollars). None of the studies
summarized by Loomis and White are found in CSS or MCF, but the study provides another
indication of the value that the public places on preserving endangered species in general.
5.1.2 Regulating
Excessive nitrogen deposition upsets the balance between CSS and nonnative plants,
changing the ability of an area to support the biodiversity found in CSS. The composition of
species in CSS changes fire frequency and intensity, as nonnative grasses fuel more frequent and
more intense wildfires. More frequent and intense fires also reduce the ability of CSS to
regenerate after a fire and increase the proportion of nonnative grasses (EPA, 2008). A healthy
MCF ecosystem supports native species, promotes water quality, and helps regulate fire
intensity. Excess nitrogen deposition leads to changes in the forest structure, such as increased
density and loss of root biomass, which in turn can result in more intense fires and water quality
problems related to nitrate leaching (EPA, 2008).
The importance of CSS and MCF as homes for sensitive species and their aesthetic
services are discussed in Section 5.1.1. Here the contribution of CSS and MCF to fire regulation
and water quality are discussed.
5.1.2.1 Fire Regulation
The Terrestrial Nutrient Enrichment Case Study identified fire regulation as a service that
could be affected by enrichment of the CSS and MCF ecosystems. Wildfires represent a serious
threat in California and cause billions of dollars in damage. Over the 5-year period from 2004 to
2008, Southern California experienced, on average, over 4,000 fires a year burning, on average,
over 400,000 acres (National Association of State Foresters [NASF], 2009). Improved fire
regulation leads to short-term and long-term benefits. The short-term benefits include the value
Final Risk and Exposure Assessment September 2009
Appendix 8-116
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
of avoided residential property damages, avoided damages to timber, rangeland, and wildlife
resources; avoided losses from fire-related air quality impairments; avoided deaths and injury
due to fire; improved outdoor recreation opportunities; and savings in costs associated with
fighting the fires and protecting lives and property. For example, the California Department of
Forestry and Fire Protection (CAL FIRE) estimated that average annual losses to homes due to
wildfire from 1984 to 1994 were $163 million per year (CAL FIRE, 1996) and were over $250
million in 2007 (CAL FIRE, 2008). In fiscal year 2008, CAL FIRE's costs for fire suppression
activities were nearly $300 million (CAL FIRE, 2008). Therefore, even a 1% reduction in these
damages and costs would imply benefits of over $5 million per year.
Figure 5.1-8 is a map of the overlap between fire threat and CSS habitat. CSS overlaps
with areas of very to extremely high fire threat. MCF is found in some areas closer to the coast
with extremely high fire threat and in areas up in the mountains also under very high fire threat,
as seen in Figure 5.1-9.
In the long term, decreased frequency of fires could result in an increase in property
values in fire-prone areas. Mueller, Loomis, and Gonzalez-Caban (2007) conducted a hedonic
pricing study to determine whether increasing numbers of wildfires affect house prices in
southern California. They estimated that house prices would decrease 9.71% ($30,693 in 2007
dollars) after one fire and 22.7% ($71,722; $102,417 cumulative) after a second wildfire within
1.75 miles of a house in their study area. After the second fire, the housing prices took between 5
and 7 years to recover. The results come from a sample of 2,520 single-family homes located
within 1.75 miles of one of five fires during the 1990s.
Long-term decreases in wildfire risks are also expected to provide outdoor recreation
benefits. The empirical literature contains several articles measuring the relationship between
wildfires and recreation values; however, very few address fires in California, particularly in
CSS areas. One exception is Loomis et al. (2002), which estimates the changes in deer harvest
and deer hunting benefits resulting from controlled burns or prescribed fire in the San Bernardino
National Forest in Southern California. Using a CV survey of deer hunters in California, they
estimated that the net economic value of an additional deer harvested is on average $122 (in
2007 dollars). Based on predicted changes in deer harvest in response to a prescribed fire, they
estimated annual economic benefits for an additional 1,000 acres of prescribed burning ranges
from $3,328 to $3,893.
Final Risk and Exposure Assessment September 2009
Appendix 8-117
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Fresno
\ --•
\
**
V
Bakersfield
Santa'Maria
®
-,^ .
~ *. ~ .•'' '••
.IV^^v^^JS^
® cities
| California Counties
^^| Coastal Sage Scrub 2002
Fire Threat
| | Moderate
| | High
| ^| Very High
I Extreme
Riverside" -•
^.fe^
V ,-',,<•:"
^>C
-/
^rii Diego
Figure 5.1-8. Coastal Sage Scrub Areas and Fire Threat
Final Risk and Exposure Assessment
Appendix 8-118
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
®
Santa Maria
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Legend
Major Rivers
Major Lakes
Mixed Conifer Forest
Figure 5.1-10. Mixed Conifer Forest Areas and Major Lakes and Rivers
5.2 VALUE OF COASTAL SAGE SCRUB AND MIXED CONIFER
FOREST ECOSYSTEM SERVICES
The CSS and MCF were selected as case studies for terrestrial enrichment because of the
potential that these areas could be adversely affected by excessive nitrogen deposition. To date,
the detailed studies needed to identify the magnitude of the adverse impacts due to nitrogen
deposition have not been completed. Based on available data, this report provides a qualitative
discussion of the services offered by CSS and MCF and a sense of the scale of benefits
associated with these services. California is famous for its recreational opportunities and
beautiful landscapes. CSS and MCF are an integral part of the California landscape, and together
the ranges of these habitats include the densely populated and valuable coastline and the
mountain areas. Through recreation and scenic value, these habitats affect the lives of millions of
California residents and tourists. Numerous threatened and endangered species at both the state
and federal levels reside in CSS and MCF. Both habitats may play an important role in wildfire
Final Risk and Exposure Assessment
Appendix 8-120
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
frequency and intensity, an extremely important problem for California. The potentially high
value of the ecosystem services provided by CSS and MCF justify careful attention to the long-
term viability of these habitats.
5.3 REFERENCES
Acharya, G., and L.L. Bennett. 2001. "Valuing Open Space and Land-Use Patterns in Urban
Watersheds." Journal of Real Estate Finance and Economics 22(2/3): 221-23 7.
Burger, J.C., R.A. Redak, E.B. Allen, J.T. Rotenberry, and M.F. Allen. 2003. "Restoring
Arthropod Communities in Coastal Sage Scrub." Conservation Biology 17(2):460-467.
CAL FIRE (California Department of Forestry and Fire Protection). 1996. California Fire Plan.
Available at http://cdfdata.fire.ca.gov/fire_er/fpp_planning_cafireplan.
CAL FIRE (California Department of Forestry and Fire Protection). 2008. CAL FIRE 2007
Wildland Fire Summary.
California Department of Fish and Game. n.d.
http://www.dfg.ca.gov/habcon/nccp/images/region.gif
Geoghegan, J., L.A. Wainger, and N.E. Bockstael. 1997. "Spatial Landscape Indices in a
Hedonic Framework: An Ecological Economics Analysis Using GIS." Ecological
Economics 23:251 -264.
Irwin, E.G. 2002. "The Effects of Open Space on Residential Property Values." Land Economics
78(4):465-480.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number EVIDE-02-0070.
Land Trust Alliance. 2006. The 2005 National Land Trust Census Report. Washington, D.C.:
Land Trust Alliance, November 30, 2006.
Final Risk and Exposure Assessment September 2009
Appendix 8-121
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Loomis, J., D. Griffin, E. Wu, and A. Gonzalez-Caban. 2002. "Estimating the Economic Value
of Big Game Habitat Production from Prescribed Fire Using a Time Series Approach."
Journal of Forest Economics 2:119-29.
Loomis, J.B., and D.S. White. 1996. "Economic Benefits of Rare and Endangered Species:
Summary and Meta-Analysis." Ecological Economics 18(3): 197-206.
Mansfield, C.A., S.K. Pattanayak, W. McDow, R. MacDonald, and P. Halpin. 2005. "Shades of
Green: Measuring the Value of Urban Forests in the Housing Market." Journal of Forest
Economics 11 (3): 177-199.
Mueller, J., J. Loomis, and A. Gonzalez-Caban. 2007. "Do Repeated Wildfires Change
Homebuyers' Demand for Homes in High-Risk Areas? A Hedonic Analysis of the Short
and Long-Term Effects of Repeated Wildfires on House Prices in Southern California."
Journal of Real Estate Finance and Economics, 1-18.
National Association of State Foresters (NASF). 2009. Quadrennial Fire Review
2009.Washington, DC: NASF. Quadrennial Fire and Fuel Review Final Report 2009.
National Wildfire Coordinating Group Executive Board January 2009.
Smith, V.K., C. Poulos, and H. Kim. 2002. "Treating Open Space as an Urban Amenity."
Resource and Energy Economics 24:107-129.
Tyrvainen, L., and A. Miettinen. 2000. "Property Prices and Urban Forest Amenities." Journal of
Economics and Environmental Management 39:205-223.
U.S. Department of the Interior, Fish and Wildlife Service, and U.S. Department of Commerce,
U.S. Census Bureau. 2007. 2006 National Survey of Fishing, Hunting, and Wildlife-
Associated Recreation.
U.S. Environmental Protection Agency (EPA). 2008. Integrated Science Assessment for Oxides
of Nitrogen and Sulfur-Environmental Criteria.EPA/600/R-08/082. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment - RTF Division, Research Triangle Park, NC.
Final Risk and Exposure Assessment September 2009
Appendix 8-122
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
6.0 CONCLUSION
This report has identified, characterized, and, to the extent possible, quantified the
ecosystem services that are primarily affected by changes in nitrogen and sulfur deposition and
associated ecological indicators. The discussion has focused on four main categories of
ecosystem effects—aquatic and terrestrial acidification and aquatic and terrestrial nutrient
enrichment—and on three main categories of ecosystem services—provisioning, cultural, and
regulating services.
The report demonstrates that nitrogen and sulfur deposition have wide-ranging
detrimental effects on the services provided by ecosystems across the United States; however,
there continues to be significant uncertainty regarding the overall magnitude of these effects. To
partially address this uncertainty, where data and scientific evidence permit, this study has
estimated how reducing nitrogen and sulfur deposition in specific areas would affect the value of
selected ecosystem services. These estimates are summarized in Table 6-1.
6.1 BENEFITS FROM ENHANCED PROVISIONING SERVICES
Provisioning services are derived from goods and commodities whose production
depends directly on inputs from healthy ecosystems. Two main examples of provisioning
services that are constrained by nitrogen and sulfur deposition are the production and
consumption of forest products and seafood.
Terrestrial acidification has been shown to cause forest damages, and much of the
specific evidence has focused on two tree species—sugar maples and red spruce. The value of
commercial harvests from these two species in 2006 was roughly $400 million, but the more
relevant question is how much would the value of these services increase with reductions in
nitrogen and sulfur deposition? This study estimates that eliminating the growth suppression
effects of terrestrial acidification on sugar maples and red spruce would generate market benefits
of about $684,000 per year.
Aquatic enrichment resulting from excess inputs of nitrogen is also known to contribute
to eutrophic conditions in surface waters, which limits the growth and abundance of commercial
fish species. Evidence regarding the magnitude of these effects is limited, largely due to the
complexities involved in modeling the dynamic ecosystem processes and links between fish
Final Risk and Exposure Assessment September 2009
Appendix 8-123
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
stocks and commercial fishing behaviors. One exception is a model of commercial blue crab
fishing in the Neuse River estuary (Smith, 2007). Using the results of this study, it is estimated
that eliminating the contribution of atmospheric nitrogen deposition to the Neuse would generate
market benefits from blue crab fishery ranging between $0.1 million and $1 million per year.
Table 6-1. Summary of Aggregate Benefit Estimates for Selected Ecosystem Services and Areas
(Zero Out of Nitrogen and Sulfur Deposition)21
Ecosystem Effect
Ecosystem Service Range (in millions of 2007 dollars/year)
Area Low High
Aquatic Acidification
Recreational Fishing
Adirondack Lakes 3.9 9.3
New York Lakes 4.5 130.0
General
Adirondack Lakes 291.2 1097.2
Terrestrial Acidification
Commercial Timber (Sugar Maple and Red
Spruce)
Northeastern U.S. 0.684 0.684
Aquatic Enrichment
Commercial Fishing (Blue Crab)
Neuse 0.1 1.0
Recreational Fishing
Chesapeake 42.6 217.0
APS 1.0 7.9
Recreational Boating
Chesapeake 2.9 8.2
Recreational Beach Use
Chesapeake 124.0 124.0
Aesthetic (Nearshore Residents)
Chesapeake 38.7 102.2
Nonuse
Chesapeake 159.1 270.4
a Because of overlaps in the services covered, the value estimates reported in the table should not be added together.
Final Risk and Exposure Assessment September 2009
Appendix 8-124
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
6.2 BENEFITS FROM ENHANCED CULTURAL SERVICES
Cultural services are derived from the nonmaterial benefits that individuals receive from
ecosystems, including spiritual enrichment, cognitive development, reflection, recreation, and
aesthetic experiences. As this report discusses, acidification and enrichment effects from
nitrogen and sulfur deposition have the potential to affect a wide variety of these services;
however, most of the available evidence concerns recreation and aesthetic services.
As discussed in Sections 3 and 5, much of the evidence of adverse terrestrial ecosystem
impacts from acidification and enrichment centers on forests in the northeastern portion of the
United States and coastal sage scrub (CSS) and mixed conifer forest (MCF) ecosystems in the
west. These ecosystems support a wide variety of land-based outdoor recreational activities,
including hunting, wildlife viewing, and hiking, worth several billions of dollars each year to the
general public. Unfortunately, relatively little evidence is available to quantity how the benefits
of these recreational services are affected by terrestrial acidification or enrichment due to
nitrogen and sulfur deposition.
As discussed in Sections 2 and 4, aquatic ecosystem impacts due to the acidification and
nutrient enrichment of surface waters also adversely affect a broad and valuable range of outdoor
recreation services. In contrast to terrestrial effects, however, the impacts of these aquatic effects
on recreational services are relatively easier to quantify (at least for selected activities and
geographic areas). For example, based on the results of an Aquatic Acidification Case Study of
the Adirondacks, the benefits to recreational anglers in New York from zeroing out the
acidification effects of nitrogen and sulfur deposition on Adirondack lakes are estimated to be
roughly equivalent to $4 million to $9 million per year. If the zero-out conditions are extended to
all New York lakes, the annual benefits could be an order of magnitude higher.
This study also used the results of the aquatic enrichment case studies to estimate the
value of enhancements to several recreational activities in the Chesapeake and Albemarle-
Pamlico estuaries, as a result of zeroing out nitrogen deposition to their watersheds. In the
Chesapeake, the benefits to striped bass and summer flounder anglers were estimated to be
roughly $43 million per year. Extending the estimation methodology to all recreational species
implies benefits of nearly $220 million, but with a much higher degree of uncertainty. For
Chesapeake boaters and beach users, the main benefit estimates were $8 million and $124
million per year, respectively. In the Albemarle and Pamlico Sounds (APS), the benefits to
Final Risk and Exposure Assessment September 2009
Appendix 8-125
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
recreational anglers were $1 million if the nitrogen loading reductions only occurred in the
Neuse and almost $8 million if they applied to the entire APS.
In addition to the benefits from enhanced recreational services, this report also examines
benefits to other cultural ecosystem services as a result of reduced aquatic acidification and
nutrient enrichment effects. Using the results of the Resources for the Future (RFF) contingent
valuation ( CV) study of New York residents, total annual benefits (assumed to primarily be for
improved cultural services, including recreational fishing services) of between $291 million and
$1.1 billion per year for a zero out of nitrogen and sulfur deposition were estimated. In the
Chesapeake Bay, benefits to nearshore residents (assumed to be mainly from improved aesthetic
and recreation services) of $39 million to $102 million were estimated. Total nonuse benefits of
between $159 million and $271 million per year were also estimated.
6.3 BENEFITS FROM ENHANCED REGULATING SERVICES
Terrestrial and aquatic ecosystems provide a variety of regulating services, such as fire,
flood, and erosion control and hydrological and climate regulation; however, there is relatively
little evidence regarding the magnitude of impairments to these services due to the effects of
nitrogen and sulfur deposition. Therefore, this report provides more of a qualitative assessment
of these services. It describes how aquatic acidification and enrichment can affect biological food
chain control services through their effects on the growth and mortality offish species. It also
describes the regulating services provided by forests, including erosion and sedimentation
control, water storage, and carbon sequestration, which may be adversely affected by nitrogen
and sulfur deposition. Finally, it describes potential changes in wildfire risks and fire regulation
services as a result of changes in CSS ecosystems that have been altered through nitrogen
enrichment.
Final Risk and Exposure Assessment September 2009
Appendix 8-126
-------
ATTACHMENT A
ANNUAL RECREATIONAL FISHING BENEFIT ESTIMATES
FOR REDUCTIONS IN NEW YORK LAKE ACIDIFICATION
LEVELS, 2002-2100
-------
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-l. Adirondack Region—20 |ieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0680
0.1360
0.2039
0.2719
0.3399
0.4079
0.4759
0.5438
0.6118
0.6798
0.6760
0.6723
0.6685
0.6647
0.6610
0.6572
0.6534
0.6497
0.6459
0.6421
0.6384
0.6346
0.6309
0.6271
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$588,805
$1,180,951
$1,775,593
$2,371,578
$2,967,035
$3,561,548
$4,155,069
$4,747,921
$5,343,426
$5,937,266
$5,903,042
$5,868,169
$5,832,830
$5,797,195
$5,761,304
$5,725,873
$5,691,082
$5,656,945
$5,623,210
$5,589,950
$5,557,169
$5,524,387
$5,491,606
$5,458,824
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A-l
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-l. Adirondack Region—20 |ieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
0.6233
0.6196
0.6158
0.6120
0.6083
0.6045
0.6007
0.5970
0.5932
0.5894
0.5857
0.5819
0.5781
0.5744
0.5706
0.5668
0.5647
0.5625
0.5603
0.5581
0.5559
0.5538
0.5516
0.5494
0.5472
0.5450
0.5429
0.5407
0.5385
0.5363
0.5341
0.5320
0.5298
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$5,426,042
$5,393,261
$5,360,479
$5,327,698
$5,294,916
$5,262,134
$5,229,353
$5,196,571
$5,163,790
$5,131,008
$5,098,226
$5,065,445
$5,032,663
$4,999,882
$4,967,100
$4,934,318
$4,915,346
$4,896,373
$4,877,400
$4,858,428
$4,839,455
$4,820,482
$4,801,510
$4,782,537
$4,763,565
$4,744,592
$4,725,619
$4,706,647
$4,687,674
$4,668,701
$4,649,729
$4,630,756
$4,611,783
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 2
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-l. Adirondack Region—20 |ieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
0.5276
0.5254
0.5232
0.5211
0.5189
0.5167
0.5145
0.5123
0.5102
0.5080
0.5058
0.5036
0.5014
0.4993
0.4971
0.4949
0.4927
0.4906
0.4884
0.4862
0.4840
0.4818
0.4797
0.4775
0.4753
0.4731
0.4709
0.4688
0.4666
0.4644
0.4622
0.4600
0.4579
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$4,592,811
$4,573,838
$4,554,865
$4,535,893
$4,516,920
$4,497,947
$4,478,975
$4,460,002
$4,441,029
$4,422,057
$4,403,084
$4,384,111
$4,365,139
$4,346,166
$4,327,193
$4,308,221
$4,289,248
$4,270,275
$4,251,303
$4,232,330
$4,213,357
$4,194,385
$4,175,412
$4,156,439
$4,137,467
$4,118,494
$4,099,521
$4,080,549
$4,061,576
$4,042,603
$4,023,631
$4,004,658
$3,985,685
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 3
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-2. Adirondack Region—50 |ieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.1232
0.2463
0.3695
0.4927
0.6159
0.7390
0.8622
0.9854
1.1086
1.2317
1.2310
1.2302
1.2294
1.2287
1.2279
1.2271
1.2264
1.2256
1.2248
1.2241
1.2233
1.2226
1.2218
1.2210
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$1,066,847
$2,139,747
$3,217,170
$4,297,025
$5,375,925
$6,453,113
$7,528,505
$8,602,686
$9,681,672
$10,757,642
$10,748,533
$10,738,213
$10,727,000
$10,715,194
$10,702,859
$10,691,327
$10,680,945
$10,671,759
$10,663,309
$10,655,746
$10,649,090
$10,642,433
$10,635,777
$10,629,121
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 4
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-2. Adirondack Region—50 |ieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
1.2203
1.2195
1.2187
1.2180
1.2172
1.2164
1.2157
1.2149
1.2141
1.2134
1.2126
1.2118
1.2111
1.2103
1.2096
1.2088
1.2079
1.2070
1.2061
1.2051
1.2042
1.2033
1.2024
1.2015
1.2006
1.1997
1.1987
1.1978
1.1969
1.1960
1.1951
1.1942
1.1933
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$10,622,464
$10,615,808
$10,609,151
$10,602,495
$10,595,839
$10,589,182
$10,582,526
$10,575,869
$10,569,213
$10,562,557
$10,555,900
$10,549,244
$10,542,587
$10,535,931
$10,529,275
$10,522,618
$10,514,670
$10,506,723
$10,498,775
$10,490,827
$10,482,879
$10,474,931
$10,466,983
$10,459,035
$10,451,087
$10,443,139
$10,435,191
$10,427,243
$10,419,296
$10,411,348
$10,403,400
$10,395,452
$10,387,504
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 5
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-2. Adirondack Region—50 |ieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
1.1924
1.1914
1.1905
1.1896
1.1887
1.1878
1.1869
1.1860
1.1851
1.1841
1.1832
1.1823
1.1814
1.1805
1.1796
1.1787
1.1777
1.1768
1.1759
1.1750
1.1741
1.1732
1.1723
1.1714
1.1704
1.1695
1.1686
1.1677
1.1668
1.1659
1.1650
1.1641
1.1631
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$10,379,556
$10,371,608
$10,363,660
$10,355,712
$10,347,764
$10,339,817
$10,331,869
$10,323,921
$10,315,973
$10,308,025
$10,300,077
$10,292,129
$10,284,181
$10,276,233
$10,268,285
$10,260,338
$10,252,390
$10,244,442
$10,236,494
$10,228,546
$10,220,598
$10,212,650
$10,204,702
$10,196,754
$10,188,806
$10,180,859
$10,172,911
$10,164,963
$10,157,015
$10,149,067
$10,141,119
$10,133,171
$10,125,223
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 6
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-3. Adirondack Region—100 |ieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.1307
0.2614
0.3920
0.5227
0.6534
0.7841
0.9148
1.0454
1.1761
1.3068
1.3056
1.3044
1.3031
1.3019
1.3007
1.2995
1.2982
1.2970
1.2958
1.2946
1.2934
1.2921
1.2909
1.2897
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$1,131,871
$2,270,164
$3,413,256
$4,558,928
$5,703,587
$6,846,430
$7,987,367
$9,127,018
$10,271,769
$11,413,319
$11,400,066
$11,385,532
$11,370,052
$11,353,948
$11,337,287
$11,321,479
$11,306,893
$11,293,575
$11,281,036
$11,269,438
$11,258,798
$11,248,159
$11,237,519
$11,226,879
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 7
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-3. Adirondack Region—100 |ieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
1.2885
1.2872
1.2860
1.2848
1.2836
1.2824
1.2811
1.2799
1.2787
1.2775
1.2762
1.2750
1.2738
1.2726
1.2714
1.2701
1.2674
1.2647
1.2621
1.2594
1.2567
1.2540
1.2513
1.2486
1.2459
1.2432
1.2405
1.2378
1.2351
1.2324
1.2297
1.2270
1.2243
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$11,216,240
$11,205,600
$11,194,961
$11,184,321
$11,173,681
$11,163,042
$11,152,402
$11,141,763
$11,131,123
$11,120,484
$11,109,844
$11,099,204
$11,088,565
$11,077,925
$11,067,286
$11,056,646
$11,033,188
$11,009,730
$10,986,272
$10,962,814
$10,939,356
$10,915,897
$10,892,439
$10,868,981
$10,845,523
$10,822,065
$10,798,607
$10,775,149
$10,751,691
$10,728,233
$10,704,775
$10,681,317
$10,657,859
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 8
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-3. Adirondack Region—100 |ieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
1.2216
1.2189
1.2162
1.2135
1.2109
1.2082
1.2055
1.2028
1.2001
1.1974
1.1947
1.1920
1.1893
1.1866
1.1839
1.1812
1.1785
1.1758
1.1731
1.1704
1.1677
1.1650
1.1623
1.1597
1.1570
1.1543
1.1516
1.1489
1.1462
1.1435
1.1408
1.1381
1.1354
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$10,634,400
$10,610,942
$10,587,484
$10,564,026
$10,540,568
$10,517,110
$10,493,652
$10,470,194
$10,446,736
$10,423,278
$10,399,820
$10,376,362
$10,352,903
$10,329,445
$10,305,987
$10,282,529
$10,259,071
$10,235,613
$10,212,155
$10,188,697
$10,165,239
$10,141,781
$10,118,323
$10,094,865
$10,071,406
$10,047,948
$10,024,490
$10,001,032
$9,977,574
$9,954,116
$9,930,658
$9,907,200
$9,883,742
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 9
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-4. New York State—20 jieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0777
0.1555
0.2332
0.3110
0.3887
0.4665
0.5442
0.6220
0.6997
0.7775
0.7727
0.7680
0.7632
0.7585
0.7537
0.7490
0.7442
0.7395
0.7347
0.7300
0.7252
0.7205
0.7157
0.7110
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$673,415
$1,350,651
$2,030,742
$2,712,368
$3,393,391
$4,073,334
$4,752,142
$5,430,186
$6,111,264
$6,790,438
$6,747,422
$6,703,667
$6,659,383
$6,614,765
$6,569,857
$6,525,479
$6,481,834
$6,438,935
$6,396,496
$6,354,599
$6,313,245
$6,271,891
$6,230,538
$6,189,184
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 10
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-4. New York State—20 |ieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
0.7062
0.7015
0.6967
0.6920
0.6872
0.6825
0.6777
0.6730
0.6682
0.6635
0.6587
0.6540
0.6492
0.6445
0.6397
0.6350
0.6327
0.6305
0.6283
0.6261
0.6238
0.6216
0.6194
0.6171
0.6149
0.6127
0.6105
0.6082
0.6060
0.6038
0.6015
0.5993
0.5971
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$6,147,830
$6,106,477
$6,065,123
$6,023,769
$5,982,416
$5,941,062
$5,899,708
$5,858,355
$5,817,001
$5,775,647
$5,734,294
$5,692,940
$5,651,586
$5,610,233
$5,568,879
$5,527,525
$5,508,117
$5,488,710
$5,469,302
$5,449,894
$5,430,487
$5,411,079
$5,391,671
$5,372,264
$5,352,856
$5,333,448
$5,314,041
$5,294,633
$5,275,225
$5,255,818
$5,236,410
$5,217,002
$5,197,595
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 11
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-4. New York State—20 |ieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
0.5948
0.5926
0.5904
0.5882
0.5859
0.5837
0.5815
0.5792
0.5770
0.5748
0.5726
0.5703
0.5681
0.5659
0.5636
0.5614
0.5592
0.5569
0.5547
0.5525
0.5503
0.5480
0.5458
0.5436
0.5413
0.5391
0.5369
0.5347
0.5324
0.5302
0.5280
0.5257
0.5235
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$5,178,187
$5,158,779
$5,139,372
$5,119,964
$5,100,556
$5,081,149
$5,061,741
$5,042,333
$5,022,926
$5,003,518
$4,984,110
$4,964,703
$4,945,295
$4,925,887
$4,906,480
$4,887,072
$4,867,664
$4,848,257
$4,828,849
$4,809,441
$4,790,034
$4,770,626
$4,751,218
$4,731,811
$4,712,403
$4,692,995
$4,673,588
$4,654,180
$4,634,772
$4,615,365
$4,595,957
$4,576,549
$4,557,142
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 12
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-5. New York State—50 jieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.4224
0.8447
1.2671
1.6894
2.1118
2.5341
2.9565
3.3788
3.8012
4.2235
4.2077
4.1918
4.1759
4.1601
4.1442
4.1284
4.1125
4.0967
4.0808
4.0649
4.0491
4.0332
4.0174
4.0015
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$3,658,144
$7,337,045
$11,031,452
$14,734,201
$18,433,674
$22,127,279
$25,814,723
$29,498,013
$33,197,783
$36,887,209
$36,740,406
$36,589,530
$36,435,696
$36,279,955
$36,122,540
$35,967,946
$35,817,317
$35,670,765
$35,526,718
$35,385,658
$35,247,618
$35,109,578
$34,971,538
$34,833,498
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 13
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-5. New York State—50 |ieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
3.9857
3.9698
3.9539
3.9381
3.9222
3.9064
3.8905
3.8747
3.8588
3.8429
3.8271
3.8112
3.7954
3.7795
3.7637
3.7478
3.7216
3.6955
3.6693
3.6432
3.6170
3.5909
3.5647
3.5386
3.5124
3.4863
3.4601
3.4340
3.4078
3.3816
3.3555
3.3293
3.3032
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$34,695,458
$34,557,418
$34,419,378
$34,281,337
$34,143,297
$34,005,257
$33,867,217
$33,729,177
$33,591,137
$33,453,097
$33,315,057
$33,177,016
$33,038,976
$32,900,936
$32,762,896
$32,624,856
$32,397,187
$32,169,517
$31,941,848
$31,714,179
$31,486,509
$31,258,840
$31,031,171
$30,803,502
$30,575,832
$30,348,163
$30,120,494
$29,892,824
$29,665,155
$29,437,486
$29,209,816
$28,982,147
$28,754,478
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 14
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-5. New York State—50 |ieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
3.2770
3.2509
3.2247
3.1986
3.1724
3.1463
3.1201
3.0940
3.0678
3.0416
3.0155
2.9893
2.9632
2.9370
2.9109
2.8847
2.8586
2.8324
2.8063
2.7801
2.7540
2.7278
2.7017
2.6755
2.6493
2.6232
2.5970
2.5709
2.5447
2.5186
2.4924
2.4663
2.4401
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$28,526,808
$28,299,139
$28,071,470
$27,843,800
$27,616,131
$27,388,462
$27,160,793
$26,933,123
$26,705,454
$26,477,785
$26,250,115
$26,022,446
$25,794,777
$25,567,107
$25,339,438
$25,111,769
$24,884,099
$24,656,430
$24,428,761
$24,201,092
$23,973,422
$23,745,753
$23,518,084
$23,290,414
$23,062,745
$22,835,076
$22,607,406
$22,379,737
$22,152,068
$21,924,398
$21,696,729
$21,469,060
$21,241,390
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 15
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-6. New York State—100 jieq/L Threshold
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Per Capita Benefit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
1.8336
3.6671
5.5007
7.3342
9.1678
11.0014
12.8349
14.6685
16.5021
18.3356
18.3111
18.2865
18.2620
18.2374
18.2129
18.1883
18.1638
18.1392
18.1147
18.0901
18.0656
18.0410
18.0165
17.9919
Population
8,333,023
8,391,727
8,440,400
8,471,658
8,511,465
8,551,313
8,665,048
8,595,000
8,630,872
8,661,363
8,685,935
8,706,365
8,721,517
8,729,057
8,731,768
8,731,621
8,730,281
8,733,574
8,733,759
8,731,785
8,728,824
8,725,132
8,720,953
8,716,339
8,712,372
8,709,340
8,707,278
8,705,815
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$0
$0
$0
$0
$0
$0
$0
$0
$0
$15,881,141
$31,852,394
$47,890,966
$63,965,753
$80,026,316
$96,061,405
$112,069,745
$128,060,056
$144,121,906
$160,138,854
$159,888,275
$159,619,755
$159,338,035
$159,047,621
$158,749,471
$158,463,334
$158,194,349
$157,943,136
$157,702,858
$157,475,737
$157,262,020
$157,048,303
$156,834,587
$156,620,870
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 16
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-6. New York State—100 jieq/L Threshold (continued)
Year
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
Per Capita Benefit
17.9674
17.9428
17.9183
17.8937
17.8691
17.8446
17.8200
17.7955
17.7709
17.7464
17.7218
17.6973
17.6727
17.6482
17.6236
17.5991
17.5589
17.5187
17.4786
17.4384
17.3982
17.3580
17.3179
17.2777
17.2375
17.1973
17.1572
17.1170
17.0768
17.0366
16.9965
16.9563
16.9161
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$156,407,153
$156,193,436
$155,979,720
$155,766,003
$155,552,286
$155,338,569
$155,124,852
$154,911,136
$154,697,419
$154,483,702
$154,269,985
$154,056,269
$153,842,552
$153,628,835
$153,415,118
$153,201,401
$152,851,672
$152,501,943
$152,152,214
$151,802,485
$151,452,756
$151,103,027
$150,753,298
$150,403,569
$150,053,840
$149,704,111
$149,354,382
$149,004,652
$148,654,923
$148,305,194
$147,955,465
$147,605,736
$147,256,007
(continued)
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 17
September 2009
-------
Analysis of Ecosystem Services Impacts for the NOX/SOX Secondary NAAQS Review
Table A-6. New York State—100 jieq/L Threshold (continued)
Year
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
Per Capita Benefit
16.8759
16.8358
16.7956
16.7554
16.7152
16.6751
16.6349
16.5947
16.5545
16.5144
16.4742
16.4340
16.3938
16.3537
16.3135
16.2733
16.2331
16.1930
16.1528
16.1126
16.0724
16.0323
15.9921
15.9519
15.9117
15.8716
15.8314
15.7912
15.7510
15.7109
15.6707
15.6305
15.5903
Population
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
8,705,075
Undiscounted Benefit
$146,906,278
$146,556,549
$146,206,820
$145,857,091
$145,507,362
$145,157,633
$144,807,903
$144,458,174
$144,108,445
$143,758,716
$143,408,987
$143,059,258
$142,709,529
$142,359,800
$142,010,071
$141,660,342
$141,310,613
$140,960,883
$140,611,154
$140,261,425
$139,911,696
$139,561,967
$139,212,238
$138,862,509
$138,512,780
$138,163,051
$137,813,322
$137,463,593
$137,113,864
$136,764,134
$136,414,405
$136,064,676
$135,714,947
Final Risk and Exposure Assessment
Appendix 8, Attachment A - 18
September 2009
-------
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-09-008b
Environmental Protection Health and Environmental Impacts Division September 2009
Agency Research Triangle Park, NC
------- |