&EPA
United States
Environmental Protection
Agency
CLASSIFICATION FRAMEWORK
FOR COASTAL SYSTEMS
\
LjflflBMt* _.
•H . ^••IKJ
T. ',
Research and Development
t .
mm
^mm
••
v
a
oK
gp
' i.r.
flH
• Jl • '
-------
EPA 600/R-04/061
May 2004
Classification Framework
for Coastal Systems
U.S. Environmental Protection Agency
Office of Research and Development
National Health and Environmental Effects
Research Laboratory
Research Triangle Park, NC 27711
-------
Notice
The information in this document has been funded wholly by the U.S. Environmental Protection
Agency. This document has been prepared jointly by co-authors at the EPA National Health and
Environmental Effects Research Laboratory, Atlantic Ecology Division, Gulf Ecology Division, and
Mid-Continent Ecology Division, with support from FAIR II contract 68W01032 to CSC Corpora-
tion and Interagency Agreement #DW14980101 to U.S. Geological Survey. It has been subjected to
review by the National Health and Environmental Effects Research Laboratory and approved for
publication. Approval does not signify that the contents reflect the views of the Agency, nor does
mention of trade names or commercial products constitute endorsement or recommendation for
use.
-------
LIST OF AUTHORS
Robert Burgess1 Michael Lewis2
Cynthia Chancy2 Janis Kurtz2
Daniel Campbell1 Teresa Norberg-King3
Naomi Detenbeck3 Peg Pelletier1
Virginia Engle2 Kenneth Perez1
Brian Hill3 Lisa Smith
Kay Ho1 Virginia Snarski3
1 Atlantic Ecology Division (AED), Narrangansett, RI
2Gulf Ecology Division (GED), Gulf Breeze, FL
3Mid-Continent Ecology Division (MED), Duluth, MN
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC
-------
List of Acronyms
303(d) Section of the federal Clean Water Act requiring states to periodically prepare a
list of all surface waters impaired by pollutants
30 5 (b) Section of the federal Clean Water Act requiring each state to prepare a water
quality assessment report every two years.
ANCOVA Analysis of Covariance
AVS Acid Volatile Sulfides
BASINS Better Assessment Science Integrating Point and Non-Point Sources
CA&DS Coastal Assessment and Data Synthesis
CAP Coastal Assessment Framework
CART Classification and Regression Tree
CDA Coastal Drainage Area
CHAID Chi-Square Automatic Interaction Detector
DCP Dissolved Concentration Potential
EDA Estuarine Drainage Area
EMAP Environmental Monitoring and Assessment Program
EPA United States Environmental Protection Agency
EROS Earth Resources Observing Systems
FEMA Federal Emergency Management Agency
GAP Gap Analysis Program
GIS Geographic Information Systems
GLEAMS Groundwater Loading Effects from Agricultural Management Systems model
GLEI Great Lakes Environmental Indicators
HUC Hydrological Unit Code
HUMUS Hydrologic Unit Model for the United States
IBI Index of Biotic Integrity
IWI Index of Watershed Indicators
LIDAR Light Detection and Ranging (remote sensing technology)
LTER Long Term Ecological Research
MATC Maximum Acceptable Toxicant Concentration
MRPP Multi-response Permutation procedures
NAO North Atlantic Oscillation
NCA National Coastal Assessment
NEP National Estuaries Program
NHD National Hydrography Database
NHEERL National Health and Environmental Effects Research Laboratory
NLCD National Land Cover Data
NMDS Non-Metric Dimensional Scaling
NOAA National Oceanic and Atmospheric Association
NOS National Ocean Service
NPDES National Pollutant Discharge Elimination System
NPLD National Pesticide Loss Database
m
-------
List of Acronyms (continued)
NRC National Research Council
NRCS Natural Resources Conservation Service
NRI National Resources Inventory
NSI National Sediment Inventory
NWI National Wetland Inventory
OHDNR Ohio Department of Natural Resources
PAH Polycyclic Aromatic Hydrocarbon
PCI First Principal Component
PC2 Second Principal Component
PC3 Third Principal Component
PCA Principal Components Analysis
PCB Polychlorinated Biphenyls
PCS Permit Compliance System
PRE Particle Retention Efficiency
PRISM Parameter-elevation Regressions on Independent Slopes Model
Q2 Peak discharge with 2-year Recurrence Interval
R-EMAP Regional Environmental Monitoring Assessment Program
RF1 Reach File 1
SPARROW Spatially Referenced Regression of Watershed Attributes
SCS Soil Conservation Service
STAR Science to Achieve Results
STATSGO State Soil Geographic database
STORET Storage and Retrieval (EPA's Largest Computerized Environmental Data System)
SWAT Soil and Water Assessment Tool
SYSTAT Statistics software
TAES Texas Agricultural Experiment Station
TEU Toxicity Equivalent Unit
TM Thematic Mapper
TMDL Total Maximum Daily Load
TSS Total Suspended Solids
USDA United States Department of Agriculture
USFWS United States Fish and Wildlife Service
USGS United States Geological Survey
WIDNR Wisconsin Department of Natural Resources
-------
Table of Contents
List of Acronyms iii
List of Tables vi
List of Figures viii
List of Appendices x
EXECUTIVE SUMMARY ES-1
I. INTRODUCTION 1
Aquatic Stressors Framework 1
Stressors of Concern 1
Diagnostics Objectives 1
Need for a Coastal Classification System 2
Risk to Coastal Systems 2
Classification by Sensitivity to Common Stressors 3
Client Needs 3
Properties and Limitations of Existing Classification Systems 5
Conceptual Models 7
Context for Model Development and Classification 7
The Conceptual Model Used as aBasis for Classification 9
System Properties Controlling the Effects of Stressors 10
II. APPROACH 14
Methods Applied for Stage I Classification Database 14
Database structure 14
Data Sources for Stressor Exposure(s) 18
Data Sources for System Properties Affecting detention Time 25
Methods for Developing and Testing Classification System 28
Data deduction Methods 28
Methods for A. Priori Development of Classes 29
Tests of Classification Approaches for Coastal Watersheds 30
III. RESULTS 31
Estuaries 31
Cluster Analysis ofEstuarine Systems 31
Derivation ofHydrologic Regime Classes for Watersheds in Coastal and Great Lake States 34
Initial Testing of Classification Frameworks 36
Hydrologic Thresholds for Subwatersheds in Lake Michigan Basin 36
Relationship between Watershed Flashiness Indicators Derived for Lake Michigan Basin 36
-------
Hydrologic Thresholds andFlashiness Indices as Predictors of Water Quality andBiotic Condition in Lake
Michigan Coastal Riverine Wetlands 36
IV. STAGE II PLANS 43
Geographic Coverage 43
Extent 43
Scale of Units 43
Parameter Improvements 44
Loadings 44
detention Time Estimates 44
Modifying Factors 45
System Processing Capacity 46
Data Sources on Coastal System Condition 47
National Coastal Assessment 47
Great Lakes Regional Environmental Monitoring and Assessment Program 47
Testing of Estuarine Classification System 48
Improvements in A Priori Testing 48
Development of Model-Based Classes for Testing 48
Approaches for A Posteriori Development of Classes 49
Spatio-temporal classification 50
V. ACKNOWLEDGMENTS 52
VI. REFERENCES CITED 53
-------
List of Tables
Table 1. Modifying factors and processing rates relevant for major aquatic stressors 13
Table 2. Aggregation of land-cover categories for Great Lakes and estuarine watershed databases.
19
Table 3. Categories of risk for groundwater nitrate inputs 20
Table 4. Effects level for diagnostic screening 23
Table 5. Contaminant loads available from Permit Compliance System 29
Table 6. Frequency of missing values for physical and hydrologic parameters 32
Table 7. Ranges of values for classification variables, used to describe clusters 32
Table 8. Estuarine classes resulting from cluster analysis of physical and hydrologic variables 35
Table 9. Frequency of variable inclusion by state and hydrologic region in equations predicting
peak discharge for coastal and Great Lakes states 35
Table 10. Categories of variables identified through CART analysis by state, region, and urban
area that best discriminate among area-normalized peak flow classes for coastal and Great
Lake states 39
-------
List of Figures
Figure ES-1. Conceptual energy systems model of the factors controlling the action of stressors in
aquatic ecosystems ES-2
Figure ES-2. Watersheds associated with 110 R-EMAP Great Lakes coastal riverine wetlands,
characterized by flow responsiveness index. Flow responsiveness index is defined as 2-year
peak flood volume and watershed depressional storage volume ES-4
Figure ES-3. Estuarine areas classified through cluster analysis of physical and hydrological
variables ES-5
Figure 1. Generalized critical path for Diagnostics Research showing role of classification and
conceptual models. PIE = Pollutant Identification Evaluation 8
Figure 2. Network of ecosystem units arranged on the landscape within fundamental
watersheds A, B, C, and D 8
Figure 3. Emergy signatures for a) top: a micro-tidal estuary (Cobscook Bay, ME), b)
middle: a fluvial estuary (York River, VA), and c) bottom: a lagoon (Mosquito Lagoon,
Florida), from Campbell 2000) 11
Figure 4. Conceptual canonical energy system models of the factors controlling the action of
stressors in a system with a) top left: unidirectional water flow, b) top right: unidirectional
flows and two different processing capacities, e.g., stratified system (water column or water
plus sediment column); c) bottom left: bi-directional flows, or d) bottom right: bi-directional
flows and two different processing capacities 12
vin
-------
Figure 5. a) HUG 8090302 (West Central Louisiana Coastal) overlapped two ED As (G210x-
Terrebonne/Timbalier Bays and G220x-Atchafalaya/Vermillion Bays) resulting in two
unique spatial units for classification (8090302_G210x and 8090302_G220x), b) EDA
G310x (Corpus Chnsti Bay) overlapped four HUCs (12110111-Lower Nueces, 12110201-
North Corpus Chnsti Bay, 12110202-South Corpus Chnsti Bay, and 12100405-Aransas Bay)
resulting in four unique spatial units for classification (12110111_G310x, 12110201_G310x,
12110202_G310x, and 12100405_G310x). Yellow and green areas = EDA and red outline
= HUC 16
Figure 6. Breakdown of NLCD land cover classes for coastal ED As and CD As 19
Figure 7. Estuarine classes resulting from cluster analysis of physical and hydrological variables 38
Figure 8. Peak 2-year flow classes identified through CART analysis of data from USGS gauging
stations. Top left: Great Lakes states, top right: Atlantic coastal states, bottom left: Gulf states,
bottom right: Pacific coast states 40
Figure 9. Flow responsiveness watershed index (peak 2-year flood volume and watershed
depressional storage volume) for watersheds associated with Great Lakes coastal riverine
wetlands 41
Figure 10. Average total toxic equivalent units for PAHs (top) and metals (bottom) within estuarine
sediments of Gulf of Mexico, color-coded for associated ED As and CDAs 42
Figure 11. A classification tree to group estuaries by effective exposure regimes based on our
conceptual model of the controlling factors 51
XI
-------
List of Appendices
A-l.l Estuarine Drainage Area physical and hydrological characteristics: metadata
A-1.2 Estuarine Drainage Area physical and hydrological characteristics: data
A-2.1 EDA/CDA land-use and land-cover: metadata
A-2.2 EDA/CDA land-use and land-cover: data
A-3.1 EDA stressor loadings: metadata
A-3.2 EDA stressor loadings: data
A-4.1 EDA stressor exposure: metadata
A-4.2 EDA stressor exposure: data
A-5.1 EDA modifying factors: metadata
A-5.2 EDA modifying factors: data
B-l.l Great Lakes coastal riverine wetland watersheds: metadata
B-1.2 Great Lakes coastal riverine wetland watersheds: data
C-l.l Marine and Great Lakes coastal watersheds: equations for peak flow predictions:
metadata
C-1.2 Marine and Great Lakes coastal watersheds: equations for peak flow predictions: data
C-1.3 Marine and Great Lakes coastal watersheds: equations for peak flow predictions:
references
C-2 1 Marine and Great Lakes coastal watersheds: peak flow classes identified by CART
analysis: metadata
C-2 2 Marine and Great Lakes coastal watersheds: peak flow classes identified by CART
analysis: data
C-3 Hydrologic regions for marine and Great Lakes coastal states
D-l Classification of ED As by cluster analysis
E-l Matrix of properties of existing classification schemes
F-l Average total toxic equivalent units by chemical class and region for ED As.
-------
EXECUTIVE SUMMARY
This report contains initial results from the
Diagnostics Committee, produced under
the U.S. Environmental Protection Agency
(EPA) Aquatic Stressors Framework (USEPA,
2002a). The goal of Diagnostics Research is
to provide tools to simplify diagnosis of the
causes of biological impairment, in support of
State and Tribe 303(d) impaired waters lists.
The Diagnostics Workgroup has developed
conceptual models for four major aquatic
stressors that cause impairment: nutrients,
suspended and bedded sediments, toxics, and
altered habitat. The conceptual models form
the basis for classification of aquatic systems
according to their sensitivity to these stress-
ors. The proposed classification framework
should enable a more refined approach for
quantifying stressor-response relationships
over broad geographical scales.
A coastal classification framework was con-
structed which encompasses watersheds and
coastal wetlands in both Great Lakes and ma-
rine coastal states in the conterminous U.S.
This report provides an overview of the com-
ponents of the classification framework: 1) a
review of existing classification schemes and
examination of their relevance for different
management goals, 2) a conceptual model for
classification based on risk from stressors, 3)
coastal classification databases for both Great
Lakes and marine coastal states, 4) a descrip-
tion of potential approaches to classification,
5) application of an empirical approach for
classification to coastal estuarine systems, 6) a
regional test of a watershed classification
framework based on data from Lake Michigan
coastal riverine wetlands, and 7) plans for
Stage II of the coastal classification frame-
work.
As the most developed areas in the nation,
coastal areas are valuable ecological and eco-
nomic resources affected by multiple, interact-
ing stressors. A classification framework is
required to describe and inventory near-
coastal communities, understand stressor im-
pacts, predict which systems are most sensi-
tive to stressors, and manage and protect
ecosystem resources. Numerous approaches
have been proposed to classify aquatic re-
sources. Classification schemes have included
geographic, hydro-dynamic, and habitat-based
characteristics and have been applied to wet-
lands, fluvial systems, near coastal waters, and
estuaries. With the exception of classification
systems developed to explain differences in
estuarine susceptibility to eutrophication, few
existing classification schemes address system
response or susceptibility to multiple stress-
ors.
Three primary factors control the action of
pollutants in aquatic ecosystems: 1) the resi-
dence time of water and pollutant in the sys-
tem, 2) the natural processing capacity of the
system for the pollutant including the path-
ways that decompose, bind, bioaccumulate, or
sequester the material, and 3) ancillary factors
that modify the form of a pollutant, the rate
of processing, or the kind of action the pol-
lutant exerts within the ecosystem (Figure ES-
1). We can evaluate these factors in a manner
that quantitatively determines the effective
dose of a pollutant for different types of eco-
systems. Characteristic properties related to
residence time, processing capacity, and modi-
fying factors can be used to differentiate
classes of ecosystems that develop different
biologically effective concentrations of a ma-
terial when loaded with a given quantity of
that pollutant. The problem can be further
simplified by grouping pollutants according to
ES- 1
-------
their mode of action such that an ecosystem
processes all compounds in a class in a similar
manner. In this case, we can express the
bioeffective concentration in aggregate units
(i.e., standard toxicity units).
Residence Time
M = Modifying Factors
P = Pollutant
Figure ES- 1. Conceptual energy systems model
of the factors controlling the action of stressors in
aquatic ecosystems.
To test the conceptual model for risk-based
classification, we constructed databases for 1)
hydrologic regimes in coastal Great Lakes and
marine coastal states, 2) a subset of 155
coastal riverine wetlands in the Great Lakes
and their associated watersheds, and 3) estua-
rine systems along the marine coast of the
conterminous U.S. To support prediction of
flow responsiveness of coastal watersheds for
the Great Lakes, we constructed databases
from U.S. Geological Survey (USGS) reports
containing nonlinear regression parameters
for 2-year peak flow magnitudes. We com-
piled raw data on peak flows and watershed
characteristics for all drainage basins associ-
ated with USGS gauging stations chosen for
flood frequency analysis from these same re-
ports.
Simon et al. (2003) had identified a subset of
155 Great Lakes coastal riverine wetlands for
monitoring using a probability-based survey
design for a Regional Environmental Moni-
toring and Assessment Program (R-EMAP)
project. We delineated watersheds associated
with these coastal wetlands and characterized
these with respect to land cover, soil proper-
ties, climatic variables, watershed and soil
storage indicators, and indicators of the
"flashiness" of hydrologic regimes (Figure
ES-2).
All 145 Estuarine Drainage Areas (EDAs) and
58 associated Coastal Drainage Areas (CDAs)
in the conterminous U.S. within the National
Oceanic and Atmospheric Association's
(NOAA) Coastal Assessment and Data Syn-
thesis (CA&DS) database were included in
estuarine databases (NOAA, 2003a). Estua-
rine databases include data on physical and
hydrologic characteristics of estuaries and
both indirect and direct indicators of exposure
(loadings or concentration, land cover, risk
indices) for nutrients, suspended sediments,
and toxics, and modifying factors (Figure ES-
3).
Classification approaches can be applied ei-
ther a priori or a posteriori. We base a priori clas-
sification on a conceptual model or
hypothesis concerning expected differences in
behavior of ecological response along stressor
gradients as a function of watershed or water
body characteristics. We have developed and
tested a priori classification strategies based on
conceptual models of watershed hydrology,
determining discriminating factors for classifi-
cation based on hydrological endpoints as in-
tegrators of expected ecological effects
(Detenbeck et al., 2000; Detenbeck et al.,
2003a or 2003b). In future work, we will ap-
ply simple canonical models of stressor effects
and interactions to determine discontinuities
in stressor-response surfaces for estuaries as a
function of water-body retention time, modi-
fying factors, and processing capacity (Camp-
bell et al., 2003; Stefan et al., 1995; Stefan et
al., 1996). A posteriori classification is based on
analysis and interpretation of available data.
Water-body classes can be derived empirically
both through indirect and factor-based meth-
ods, using cluster analysis of water-body and
watershed characteristics, and through direct
ES-2
-------
and response-based approaches, using Bayes-
ian approaches to determine natural break-
points in assessment endpoints as a function
of stressor gradients and classification factors
(Breiman et al, 1984; Kass, 1980).
Using subsets of the classification databases,
we applied and tested three approaches to
classification of coastal systems. We identi-
fied classification factors and breakpoints as-
sociated with watersheds of varying flow-
responsiveness levels through Classification
and Regression Tree (CART) analysis of the
gauging station watershed database. We used
the magnitude of two-year peak flows normal-
ized to watershed area as an indicator of flow-
responsiveness of watersheds or "flashiness"
of hydrologic regimes (Figure ES-2). We ap-
plied flow-regime classes derived for water-
sheds in the Lake Michigan basin to
watersheds associated with coastal riverine
wetlands monitoring in the Region 5 R-
EMAP project. We successfully used flow-
regime types to explain differences among
classes in response of nutrients, chlorophyll,
thermal regime, and periphyton communities
in coastal wetlands along land cover gradients.
Finally, we developed an empirical classifica-
tion of estuarine systems using cluster analysis
on physical and hydrological data (Figure ES-
3).
Stage II of the classification framework will
include improvements in spatial extent and
resolution of coastal units, incorporation of
additional classification factors, application of
empirical and model-based approaches for
classification to coastal systems (both water-
sheds and coastal wetlands), and testing of
classification systems using data gathered
through the Environmental Monitoring and
Assessment Program (EMAP), National
Coastal Assessment (NCA), and R-EMAP in
the Great Lakes. Initial classifications will be
refined using data from intensively monitored
systems from different classes. In addition,
we will explore model-based approaches to
classification of coastal systems.
ES-3
-------
Map Legend
Great Lakes Shoreline
j i State Boundaries
REMAP Wateisheds
Flow Responsiveness Index
I""7". 0.04-0.50
, 'JO 51 -1 .06
[_ ] 1.07 -1 .7!
[ , 1.79 - 3.01
f 1302 4.80
| | 4.81 -6.77
[ 3R7& 1676
,•• 16.77 -28.50
^B 28.51 - 58.68
^•58.69-131.30
Figure ES- 2. Watersheds associated with 110 R-EMAP Great Lakes coastal riverine wetlands, characterized
by flow responsiveness index. Flow responsiveness index is defined as 2-year peak flood volume and
watershed depressional storage volume.
ES-4
-------
I I US State Boundary
I 1 - Large. Medium Volume. Very High Flow, Shallow, Low Salinity
| | 2 - Large. Large Volume. Medium Flow, Deep, High Salinity
| | 3 - Small EDA/Large % Estuary, Small Volume, Low Flow, Deep. High Salinity
| | 4 - Medium EDA/Small % Estuary. Small Volume, High Flow. Shallow. Low Salinity
I 5 - Medium EDA/Small % Estuary. Small Volume, Low Flow. Shallow. High Salinity
| 6 - Medium. Small Volume. Low Flow, Shallow, High Salinity
| | 7 - Medium EDA/Large % Estuary, Medium Volume. Low Flow, Medium Depth. High Salinity
| | 8 - Large EDA/Medium % Estuary. Medium Volume, High Flow, Shallow. Mixed Salinity
_| 9 - Medium EDA/Small % Estuary. Small Volume, Medium Flow, Shallow. Mixed Salinity
( 10 - Large EDA/Small % Estuary, Small Volume. Medium Flow. Shallow. Low Salinity
I 11 - Small EDA/Medium % Estuary. Small Volume, Low Flow, Shallow. Mixed Salinity
Figure ES- 3. Estuarine areas classified through cluster analysis of physical and hydrological variables.
ES-5
-------
I. INTRODUCTION
Coastal areas are some of the most devel-
oped areas in the nation and are a source
of many valuable resources. These include: 1)
fish and shellfish, which support commercial
and recreational fisheries, 2) extensive water
areas that support boating, swimming, and
other water-related recreational activities, 3)
water supplies for cooling, and 4) a large dilu-
tion and flushing capacity, which reduces con-
centrations of wastes from many municipal
and industrial points of discharge. Because a
large proportion of the nation's population
lives in coastal areas, environmental pressures
threaten the resources that make the coast so
desirable (USEPA, 1995; 2001a).
Aquatic Stressors Framework
EPA's common management goal is to main-
tain ecological integrity by protecting aquatic
systems against the degradation of habitat,
loss of ecosystem functions and services, and
reduced biodiversity. To contribute to this
common goal through ecological effects re-
search, the EPA National Health and Envi-
ronmental Effects Research Laboratory
(NHEERL) developed the Aquatic Stressors
Research Framework and Implementation
Plan (USEPA, 2002a). EPA designed this
research framework to develop scientifically
valid approaches for protecting and restoring
the ecological integrity of aquatic ecosystems
from the impacts of multiple aquatic Stressors.
Stressors of Concern
Dorward-King et al. (2001) define Stressors as
environmental factors that have the potential
to cause a significant change in an organism,
population, community, or ecological system.
Stressors may act simultaneously or sequen-
tially at an intensity, duration, and frequency
of exposure that results in a change in eco-
logical condition (USEPA, 2000). In estuarine
systems, Stressors with natural origins include
temperature, salinity, and suspended sediment
load while Stressors from anthropogenic
sources include single chemical classes, such
as pesticides, mixtures of chemicals of differ-
ent classes (e.g., wastewaters), or a combina-
tion of chemicals and habitat degradation.
For the present classification, our research
focused on aquatic ecosystem Stressors de-
rived from anthropogenic sources or activities
with the greatest potential for causing adverse
effects, including habitat alteration, nutrients,
excessive suspended and bedded sediments,
toxic chemicals, and interactions among the
four. The research will attempt to isolate
natural stressor effects from those derived
from anthropogenic sources (USEPA, 2002a).
Furthermore, we will assume that sufficient
information on the ecological effects of many
toxic chemicals exists although information
on effects of chemical mixtures is not avail-
able.
Diagnostics Objectives
To support implementation of Section 303(d)
of the Clean Water Act and other regulatory
programs, NHEERL's Diagnostics research
focuses on the need to diagnose causes of bio-
logical impairment within an integrated
framework linking watersheds with receiving
water. The starting point for diagnostics re-
search is the need to respond to reports of
biological impairment, non-attainment of
aquatic life use, and other indications of ad-
verse effects (e.g., toxicity). Initial assess-
ments can also record evidence of multiple
potential causes of impairment and conflicting
lines of evidence that might complicate a di-
-------
agnosis. Thus, the endpoint for the diagnostic
process includes both the definition of the
primary causes of impairment as well as the
allocation of observed effects among multiple
potential stressors, and the assessment of po-
tential interactive effects among stressors.
In developing a research plan, we considered
the States' implementation stages (i.e., moni-
toring, diagnosis, restoration). We then linked
these implementation stages to the critical
path for research. The following are four
primary objectives for diagnostics research:
• Provide a framework for interpreting
cause-and-effect relationships, includ-
ing
- Conceptual ecosystem models
based on appropriate mechanisms
of action to improve stressor re-
lated impairment decisions
- Conceptual models to define the
natural conditions of ecosystems
and watersheds and their driving
factors for quantifying degree of
impairment and for setting resto-
ration goals
- Classification frameworks that ex-
plain variation in the response of
individuals, populations, commu-
nities, and ecosystems to individ-
ual stressors and to combinations
of stressors at regional, watershed,
water body, and habitat scales.
• Develop single-stressor diagnostic
methods and models to determine the
primary source for biological impair-
ment of aquatic ecosystems.
• Develop methods and models to allo-
cate causality among multiple stressors
and to diagnose interactions among
them.
• Develop methods and models capable
of forecasting the ecological benefits
of source reductions, to investigate
stressor interactions, and to assess the
gains and losses realized by various al-
ternatives for restoration and remedia-
tion.
Need for a Coastal Classification System
Risk to Coastal Systems
Coastal systems at the interface between land
and sea, or between land and the Great Lakes,
are highly productive and diverse. These ar-
eas of dynamic inter-change for both fresh
and oceanic waters, sediments, and organic
and inorganic chemicals create both opportu-
nities and risks for organisms, populations
and communities (Hobbie, 2000). Increases
in coastal human population density and the
concomitant changes in land use affect rates
of interchange, which in turn affect harvests
of fish and shellfish and recreational opportu-
nities for human populations.
The U.S. National Research Council (NRC;
1994) has outlined some important inter-
changes and associated risks to coastal ecosys-
tems that may result from increases in human
uses and population density. Habitats such as
seagrass meadows and emergent marshes are
vital nursery areas for important commercial
fishery species, yet are affected by dredging,
filling, and increased sediment loads. Land
use alterations and water diversions change
the seasonal pattern and amount of freshwater
inflow along with the supply of organic and
inorganic nutrients and the amount of sedi-
ment transported into coastal systems. Nutri-
ents from treated wastewater discharges and
agricultural runoff have increased causing de-
pletion or decreases in oxygen in bottom wa-
ters and shifts in species dominance resulting
in harmful algal blooms. Industrial activities
release toxic contaminants such as polycyclic
aromatic hydrocarbons (PAHs) and heavy
metals in some areas. Formerly, abundant
populations of finfish and shellfish have been
overexploited, reducing stocks and altering
natural cycles. Loss or decline of natural
populations and the introduction of nonindi-
genous species has resulted in loss of biodi-
-------
versity and degradation of resource popula-
tions. Changes in climate and weather pat-
terns affect precipitation and salinity,
circulation patterns and transport of nutrients,
which can also affect biological production
and diversity. In short, coastal resources will
be subject to increasingly complex interac-
tions as human population pressure increases.
Understanding these complex interactions
poses a significant challenge to researchers
and managers.
Classification by Sensitivity to Common Stressors
Although coastal areas are diverse, complex
and heavily utilized, they may show similar
patterns that can be useful for classifying
coastal systems. Specifically, coastal systems
may show commonalities in their sensitivity to
stressors that can be used as classification pa-
rameters to aid diagnosis. Estuaries most sus-
ceptible to pollution have a poor ability to
dilute or flush sediments, toxins or dissolved
substances (NOAA, 1989). Thus, physical
and hydrologic characteristics can be used to
predict the susceptibility of coastal systems.
The NRG (2000) considered the role of both
biological and physical factors in influencing
estuarine susceptibility to nutrients. NOAA
had previously identified the following physi-
cal factors as predictors of sensitivity: physi-
ography, dilution due to area-volume
relationships and mixing processes, water
residence time and flushing rate, and stratifi-
cation, to which NRC added hypsography, or
the relative areal extent of land surface eleva-
tion and depth, and loading (Bricker et al.,
1999). Loading includes nutrient load derived
from both watershed and atmospheric inputs,
suspended material load, which reduces light
penetration through the water column, dis-
solved and particulate organic matter load,
and toxin loads (PAHs, metals, pesticides and
other classes of chemicals as well as mixtures
of classes). We suggest that the loads derived
from sediment, including biotic and abiotic
particles, nutrients, and toxins, are also influ-
ential in these systems.
Biological factors that determine estuarine
response to stressors like nutrient over-
enrichment include primary production, graz-
ing rates, and denitrification (NRC, 2000).
Major types of primary producer communities
include emergent marshes, seagrasses, benthic
macroalgae, periphyton, and phytoplankton in
marine coastal systems, and macrophyte, pe-
riphyton, and phytoplankton in Great Lakes
coastal wetlands. Communities dominated by
marshes are likely to be shallow with short
residence times, while plankton-dominated
systems may be deeper with longer residence
times. Changes in grazing pressure at differ-
ent trophic levels, may result in changes in
food webs, system function and sensitivity to
stressors. Processes like denitrification, sul-
fate reduction and methane generation are all
biogeochemical processes in coastal systems,
and are likely to determine system sensitivity
to the affected pollutants. Biological factors
are less well characterized, for the most part,
compared to physical factors, and should be a
target for future classification and modeling
efforts.
Client Needs
Managers and researchers need classification
frameworks to understand, protect and man-
age coastal resources. A successful classifica-
tion scheme will accomplish several key tasks
needed by EPA clients: (1) describe and in-
ventory near coastal communities and habitat
types, (2) identify and help set priorities for
conservation efforts, (3) aid in the manage-
ment of ecosystem resources, and (4) help
target future research needs. Classification
frameworks are logical approaches to organiz-
ing and grouping information about ecological
systems. Because we compare systems based
on data, we can use classification frameworks
as logical organizing structures and reposito-
ries for data collected from a variety of con-
-------
tributors. Once managers have established a
common terminology and organized catego-
ries of data in a database, they can conduct
inventories to determine the extent and distri-
bution of different ecosystem types and po-
tential stressor effects.
Classification frameworks and the information
base required to develop them can assist envi-
ronmental managers in their efforts to set wa-
ter or sediment criteria, establish reference
conditions, determine the cause of impair-
ment, and predict changes in environmental
condition. Water quality criteria for chemicals
are concentrations that, when not exceeded,
protect aquatic life and human health accord-
ing to available scientific information. Bio-
logical criteria are narrative or numerical
descriptions of the desired biological condi-
tion of aquatic communities inhabiting par-
ticular types of water bodies (Detenbeck,
2001). Reference conditions describe charac-
teristics of water body segments least impaired
by human activities, i.e., those habitat condi-
tions that may exist in pristine areas or those
conditions attainable through management
actions. Establishment of water quality crite-
ria can be aided not only by describing ex-
pected reference conditions for naturally-
occurring substances or communities, but also
by identifying classes of aquatic systems with
differential sensitivity to pollutants.
Determining impairment and diagnosing its
cause(s) are requirements of sections 305(b)
and 303(d) of the Clean Water Act. Under
section 305(b) states and tribes are required to
assess the status of water bodies and to iden-
tify suspected causes of impairment. Section
303(d) requires preparation and submission of
listings of impaired water bodies that violate
water quality standards or exceed water quality
criteria or biocriteria. Grouping of systems by
class can simplify the problem of determining
the cause of observed ecological effects.
Classes behave differently under the influence
of the stressor of concern. Once we define
classes and categorize responses, we can pre-
dict changes in environmental condition re-
sulting from restoration actions, habitat
alterations, or increased contaminant loading
with greater confidence.
Classification can assist environmental man-
agers in meeting water quality standards by
supporting implementation of the Total
Maximum Daily Load (TMDL) program initi-
ated as part of the Clean Water Act. A
TMDL is the projected load of a pollutant
that will result in compliance with a water
quality standard. Of the 40,000 water bodies
currently identified in the nation, 21,000 river
segments, lakes, and estuaries have been iden-
tified as being in violation of one or more
standards (NRC, 2001). States must develop
plans for TMDLs that will result in attainment
of water quality standards under an ambitious
time schedule. In addition, most plans will
require controlling nonpoint source pollution,
which is more difficult to quantify and man-
age than point sources.
For each water body, managers must diagnose
the cause(s) of impairment prior to specifying
a TMDL. Classification can help to establish
the expected ecological conditions for water
bodies by class, which would help to deter-
mine if impairment exists. We can simplify
determination of the cause(s) of impairment
for thousands of systems by developing ro-
bust classification schemes that identify
groups of coastal ecosystems that behave in a
similar manner in the presence of a stressor.
A useful classification framework will provide
regional, state, and tribal regulatory authorities
a tool to collapse the over 40,000 water bodies
requiring TMDLs into a more manageable
number of water body classes, each class
composed of individual water bodies with
common, stressor-sensitive characteristics.
For example, estuaries with slower turnover
times are more susceptible to the effects of
nutrient loading and may form one logical
class. For defined water body classes, manag-
ers could create a TMDL template or plan for
remediating the impairment, which they could
-------
apply to all of the water bodies within the
class with minor adjustments on a case-by-
case basis. This method would eliminate the
need for 40,000 unique TMDLs and remedia-
tion plans.
Finally, classification systems can serve to tar-
get current and future research needs.
In building a database for classification, re-
search gaps readily become apparent. These
gaps point to opportunities for empirical stud-
ies to complete data sets and may help to de-
termine which missing data are most
important to obtain. Analysis of classification
databases may reveal important couplings be-
tween physical, biogeochemical, and ecologi-
cal processes. Numerical modeling
simulations may be a useful approach for bet-
ter understanding these couplings and interac-
tions, and may address important issues like
spatial and temporal variation (Geyer et al.,
2000). Establishing meaningful and measur-
able response indicators and causal links be-
tween loads and response to stressors within
ecosystems are also important areas for re-
search. Combining food web models and
chemical fate models could allow better pre-
diction of exposure and response within
aquatic ecosystems (Baird et al., 2001). Stud-
ies comparing observed loads and responses
among classes of coastal ecosystems com-
bined with modeling approaches will advance
our abilities to make responsible decisions
that will protect coastal ecosystems.
Properties and Limitations of Existing
Classification Systems
Efforts to develop classification systems have
focused principally on terrestrial and freshwa-
ter systems, and on specific regions and habi-
tat types, as well as on entire nations.
Researchers have studied a few coastal sys-
tems and their watersheds intensively, but
have not yet expanded broad-scale classifica-
tion efforts to coastal and estuarine ecosys-
tems (Edgar et al., 2000). Researchers have
conducted even fewer studies to compare sus-
ceptibility or responses to stressors among or
across coastal ecosystems.
Although none of the 25 classification sys-
tems we reviewed specifically met our needs
for a coastal classification based on suscepti-
bility to stressors (Appendix E), each pro-
vided approaches and information to build
upon. Three geographic mapping efforts di-
vided the U.S. into regions with common fea-
tures based on overlays of existing landscape
and climatic data (Bailey, 1976; Keys et al.,
1995; Omermk, 1987; USGS, 1999). These
geographic efforts served to define and com-
partmentalize the entire country into similar
climatic units, aiding environmental manage-
ment and conservation efforts to inventory
and define natural resources. Although com-
prehensive, climatic units are more relevant to
terrestrial systems and result in an impractical
number of classes for our purposes.
Scientists have designed another group of
classification frameworks for inventory and
management of wetlands (Chow-Fraser, Al-
bert, 1998; Cowardm et al., 1979; Day et al.,
1988; Detenbeck, 2001; Keough et al., 1999;
Shaw, Fredine, 1956). The most widely used
of these, developed by Cowardin et al. (1979),
divides environments into groups in a manner
similar to a taxonomic key. Broad categories
of habitat types are successively divided in
hierarchical fashion into groups with more
aspects in common, cascading down to nu-
merous, well-defined classes with many com-
mon features. These systems add
considerations of biological diversity, hydrol-
ogy and retention time, but lack a qualitative
framework making susceptibility to stressors
difficult to predict or measure (Jay et al.,
2000). At the level of refinement necessary
for considerations of susceptibility to stress-
ors, these systems still result in a large number
of classes.
Fluvial systems and watersheds are the focus
of five additional frameworks that use hydrol-
-------
ogy, geomorphology, and sediment transport
as classification parameters (Montgomery,
Buffington, 1993; Rosgen, 1994; Poff, Ward,
1989; USGS, 2003g; Detenbeck et al, 2000).
These systems result in smaller numbers of
classes, and are relevant to classifying aquatic
habitats for the coastal Great Lakes. They are
not designed for direct application to estuar-
ies, nor are they directly applicable to stressor
susceptibility determinations.
Several classification frameworks, published
since the 1950s and 1960s, have addressed
estuaries. Two are simplified to elements of
stratification and circulation (Hansen, Rattray,
1966; Strommel, Farmer, 1952). Two more
recent systems add forcing processes like wind
and waves (Jay et al., 2000), or employ hierar-
chical clustering methods with both quantita-
tive and subjective data for Australia (Digby et
al., 1998). Alice et al. (2000), developed a sys-
tem including biological criteria that was
modeled after the hierarchical framework de-
veloped for wetlands by Cowardin et al.
(1979). Two additional estuarine efforts in-
clude biological aspects. Briggs (1974) out-
lined zoogeographic regions based on the
distribution of indigenous marine organisms,
and the Nature Conservancy has adapted their
freshwater classification framework to estua-
rine habitats for the conservation of biodiver-
sity and specific species of interest (Beck,
Odaya, 2001).
Four classification systems address suscepti-
bility to stressors in a more direct way. Sklar
and Browder (1998) identify the potential im-
pacts of alterations to freshwater flow, com-
paring effects from individual stressors to
multiple stressor effects. This study provided
an in-depth examination of Gulf of Mexico
systems and additional research may demon-
strate broader applicability. Stefan et al.
(1996) used a modeling approach to predict
habitat susceptibility to global climate change.
They considered trophic status and interacting
stressors, although only in closed, northern
lake systems. Ferreira (2000) developed an
estuarine quality index. This decision support
system addressed vulnerability, but required
fish and benthic community diversity meas-
ures and sediment quality indicators that may
not be widely available. Finally, NOAA con-
sidered estuarine susceptibility to nutrient
over-enrichment (NOAA, 1989; Bricker et al.,
1999). Using these calculations and metrics,
managers can classify estuaries effectively
based on nutrient susceptibility, but must rely
partially on subjective measures, and are un-
able to consider other stressors to aquatic sys-
tems using these methods.
When maintaining habitat inventories and
prioritizing habitats for conservation efforts
or examining reference conditions are the
only goals considered, the state of science for
classification of Great Lakes coastal habitats is
similar to the state of science for marine estu-
aries. Existing habitat classification frame-
works for wetlands and deepwater (Cowardin
et al., 1979) and ecoregion or ecological unit
classification frameworks (Keys et al., 1995;
Maxwell et al., 1995; Omermk, 1987) are gen-
erally applicable to the Great Lakes as well as
other regions of the U.S. For Great Lakes
coastal wetlands, McKee et al. (1992) suggest
a modification of Cowardin's system, incorpo-
rating landscape position (system), depth zone
(littoral vs. limnetic subsystems), vegetative or
substrate cover (class and subclass), and modi-
fiers of ecoregions, water level regimes, fish
community structure, geomorphic structure,
and human modification. More detailed habi-
tat type classifications for both coastal and
inland aquatic systems in the Great Lakes ba-
sin are in progress through the USGS Gap
Analysis Program (USGS, GAP, 2003a). In
this approach, managers and researchers use
physical, chemical, and hydrological character-
istics of streams or coastal habitats to classify
habitat types and relate them to organism
presence or absence and biological commu-
nity types. For streams and rivers of the up-
per Midwest, Robertson et al. (2001) have
demonstrated that they can improve predic-
tions of reference condition for nutrient con-
-------
centrations by applying the ecozone concept.
Robertson et al. (2001) identified both natural
and combined natural and anthropogenic wa-
tershed characteristics that discriminate
among classes of watersheds with different
ranges of nutrient concentrations. Previously,
managers and researchers had neither tested
the ecoregion nor the
ecozone classification approaches specifically
for Great Lakes coastal habitats. However,
results from a recent EPA R-EMAP project
(Simon et al., 2003) will provide an opportu-
nity to test ecoregion and ecozone definitions
of reference condition for both nutrients and
indices of biotic integrity.
In contrast to the state of the science for clas-
sification of aquatic habitat types and
reference condition, the science for classifica-
tion of Great Lakes coastal habitats based on
relative susceptibility to stressors is in its in-
fancy. One exception is the construction of
an Environmental Sensitivity Index for both
marine and Great Lakes shorelines by NOAA;
however, NOAA constructed this index spe-
cifically to predict sensitivity to spills of oil
and other hazardous substances (USEPA,
2001 c; NOAA, 2003c). Researchers have
proposed numerous Great Lakes coastal wet-
land classification schemes, but have focused
either on vegetation type or on some combi-
nation of geomorphology and geologic origin,
without establishing any conceptual or em-
pirical relationships between wetland type and
susceptibility to stressors (Chow-Fraser, Al-
bert, 1998; Keough et al., 1999; Maynard, Wil-
cox, 1997; Mine, Albert, 1998; Great Lakes
Commission, 2001).
Conceptual Models
The conceptual framework for evaluating
causes of biological impairment within aquatic
ecosystems of the U.S. is comprised of a
hierarchical, modular set of diagnostic meth-
ods and models and an ecosystem classifica-
tion scheme (Figure 1). Together, these tools
should simplify and improve the accuracy of
water body evaluations, which the federal
government, states, and tribes carry out under
Sections 305(b), and 303(d) of the Clean Wa-
ter Act. The conceptual overview described
here provides the basis for developing and
testing our classification system.
Context for Model Development and Classification
One of our goals is to simplify the process of
diagnosing the causes of impairment by classi-
fying ecosystems based on differences in their
response to stressors. A second goal is that
our models and classification system simplify
and facilitate management at the watershed
scale. To accomplish these goals we need to
understand and predict the actions of stress-
ors on aquatic ecosystems within their water-
sheds. Thus, the nature of watersheds,
aquatic ecosystems and stressors set the con-
text for model development. We define
fundamental watersheds here as networks of eco-
systems that are linked by flow of water and
have a terminal connection to the open sea or
to one of the Great Lakes (Figure 2). They
serve as the largest watershed scale system
within which we must manage wetlands,
stream segments, lakes and estuaries to ensure
that limits established for pollutants and habi-
tat alteration will be effective.
Our classification system uses the properties
of fundamental watersheds as input data and
organizes the information in this way. We can
define a stressor as an injury or impairment to
an ecosystem that results from the overuse of
one or more ecosystem components or proc-
esses. In general, this condition is the result
of a change or perturbation of the normal
(long-term or natural) suite of energy inputs
to a place (the energy signature, Figure 3) that
results in a change in the normal or expected
functioning of the ecosystem under the old
signature.
-------
Watershed Diagnostics
Conceptual Models
Water Body Diagnostics
Classification
Secondary
Data
Case
Studies
if>
PIE Methods
Screening
Phase I
Phase II
*-»
s'
%
^
Secondary
Data
Case
Studies
Integrated Diagnostic Method
A. Causes of Impairment
B. Scale of Control
Development of a TMDL or Other Regulations
Figure 1. Generalized critical path for Diagnostics Research showing role of classification and conceptual
models. PIE = Pollutant Identification Evaluation.
A conceptual representation of the impact
pathway that results in stress in an ecosystem
is a simple chain of cause and effect:
/ B
A
Land
SEA
Figure 2. Network of ecosystem units arranged on the
landscape within fundamental watersheds A, B, C, and
D.
Human activities => pollutant
sources => presence of the
stressor in the environment,
e.g., the concentration of a
pollutant => observed effect,
e.g., a biological impact.
For a classification system to be useful in
simplifying diagnosis of the causes of impair-
ment to aquatic ecosystems, it must be able to
discriminate groups of systems based on dif-
ferences in their response to stress and to a
particular stressor. The four classes of stress-
-------
ors we are considering, nutrients, suspended
and bedded sediments, toxics, and altered
habitat, each have different mechanisms of
action within ecosystems. Therefore, we ex-
pect that the system groupings found by a
classification scheme will differ based on the
factors that control the behavior of aquatic
ecosystems under the influence of one or
more of these stressors. Development of a
TMDL is determined on a pollutant-by-
pollutant basis. Thus, there is a need to un-
derstand the mechanisms of stressor action
and the behavior of systems under stress on a
pollutant-by-pollutant basis. We developed a
second conceptual model for the network of
system interactions controlling the behavior
of stressors within fundamental watersheds
but did not use these watershed models in our
initial classification research.
The Conceptual Model Used as a Basis for Classifica-
tion
We developed a hierarchical, modular frame-
work for constructing the conceptual models
needed to diagnose the causes of impairment
in aquatic ecosystems. At the highest level of
aggregation, we used a set of simple standard-
ized models (Figure 4) to describe the links
between watershed units and their aquatic
ecosystems on the landscape. We illustrate the
common and distinguishing properties of the
four canonical models used to represent
aquatic ecosystems with different hydrological
properties in Figure 4. The base unit for each
of these models is a water body (river reach,
lake, or estuary) and its watershed. We must
specify the characteristics of the water body,
including its geometric and geomorphic prop-
erties along with the loading of the material
stressor from the surrounding watershed and
the quantity of the material that is stored
within the system. We hypothesize that three
primary factors control the stressful actions of
pollutants in aquatic ecosystems. They are (1)
the residence time of water and pollutant in
the system, (2) the natural processing capacity
of the system for the pollutant including the
pathways that decompose, bind, take-up, or
sequester the material, and (3) ancillary factors
that modify the form of a pollutant, the rate
of processing, or the kind of action the pol-
lutant exerts within the ecosystem. We can
evaluate these three factors in a manner that
quantitatively determines the effective dose of
a pollutant experienced in ecosystems of dif-
ferent kinds. We hypothesize that different
ecosystems will have characteristic properties
related to residence time, processing capacity,
and modifying factors that together can be
used to differentiate classes of ecosystems that
develop different biologically effective con-
centrations of a material when loaded with a
given quantity of that pollutant. We can fur-
ther simplify the problem by grouping pollut-
ants according to their mode of action such
that an ecosystem processes all members of a
class in a similar manner. In this case, we can
express the bioeffective concentration in ag-
gregate units, i.e., standard toxicity units.
The factors that we have used to construct the
conceptual models for diagnosis when quanti-
tatively evaluated give an expression for the
exposure of the ecosystem to biologically ac-
tive concentrations of a particular stressor.
Residence *
Time
(days)
Bioeffective - Exposure
Concentration (g nr3 - days)
(gm-3)
We base our classification system on the
premise that effective exposure will differ for
particular pollutants or classes of pollutants
across estuaries and Great Lakes coastal wet-
lands of different kinds. The first term of the
expression given above depends on the physi-
cal forces and flows that control the residence
time of pollutants in the system, whereas the
second term depends on the biological and
chemical factors that determine processing
capacity for the material. Modifying factors
are forcing functions or materials that alter the
effect of stressor action. When the effects of
modifying factors are applied to exposure cal-
-------
culations given above, the effective exposure
results.
System Properties Controlling the Affects of Stressors
Residence Time
This generic property measures the average
time period a molecule of water derived from
riverine sources resides in the estuary. The
longer the residence time, the longer freshwa-
ter and dissolved constituents will remain in
the estuary. Shallow systems exhibit proper-
ties that magnify stress, e.g., increased concen-
trations of riverine inputs, internal waste
products and lower salinities. However, shal-
low systems also decrease stress by removing
some of the dissolved constituents through
decreased freshwater residence times. Large
estuarine systems by virtue of their lower sur-
face area to volume ratio should have lower
concentrations of waste products and, there-
fore, be less stressful. However, because of
longer residence times deeper areas of the es-
tuary can accumulate these organic degrada-
tion products resulting in anoxia with
concomitant changes in other critical popula-
tions, communities and ecosystem functions.
As a result, residence time of both freshwater
and seawater systems are important determi-
nants of the ecological state of estuarine sys-
tems.
Ecosystem processing capacity
Biological, chemical, and physical processes
alter the bioavailable concentrations and resi-
dence time of materials entering the ecosys-
tem. These processes consist of absorption
and desorption onto particles, chemical and
biological uptake, transformation and degra-
dation. As a result of these processes, an-
thropogenic materials and pollutants can be
either removed from the system with no
adverse affects or accumulate in various com-
partments (e.g., sediments) of the aquatic sys-
tem where adverse ecological effects can
occur. Any foreign material that reduces or
interferes with the rate at which these proc-
esses occur can result in further system degra-
dation due to the changes in removal rates of
other compounds normally found at low con-
centrations. The rate of processing of major
compounds (e.g., nutrients) could be ecosys-
tem specific and, therefore, be a determinant
in the classification of an ecological system.
Modifying factors
Modifying factors alter the relationship be-
tween exposure and effect. For example,
aluminum is toxic at low pH but not at a
higher pH. Therefore, a calculated exposure
to aluminum would result in greater mortality
at pH 5 than at pH 7. Another example of a
modifying factor is water column turbidity,
which alters the amount of carbon fixation
realized from a given concentration of phos-
phorus and/or nitrogen. Processes that alter
the residence time of the system such as
changes in the flow regime can also change
the effective exposure of the ecosystem to a
pollutant by altering the time it has to react
with the biota. For example, the amount of
watershed storage in fresh water systems can
influence the amount of biological and chemi-
cal processing, flow regime effects on habitat,
and biological condition of fish communities
in fresh water aquatic ecosystems (Detenbeck
et al., 2000). We should also consider factors
that control the relationship between an ob-
served biological attribute and exposure to a
pollutant as modifying factors, e.g., storage in
fresh water systems, salinity in salt water sys-
tems, and pH (Table 1)
10
-------
l.OOE+21
l.OOE+20
l.OOE+19
l.OOE+18
l.OOE+17
l.OOE+16
^
l.OOE+211
l.OOE+20
l.OOE+19
l.OOE+18
l.OOE+17
l.OOE+16
l.OOE+21
l.OOE+20 j
l.OOE+19
l.OOE+18 \
l.OOE+17
l.OOE+16 •<
*///£
'r &r «& cP
Figure 3. Emergy signatures for a) top: a micro-tidal estuary (Cobscook
Bay, ME), b) middle: a fluvial estuary (York River, VA), and c) bottom: a
lagoon (Mosquito Lagoon, Florida), from Campbell 2000).
11
-------
Figure 4. Conceptual canonical energy system models of the factors controlling the action of stressors in a system with a) top left: unidirectional water
flow, b) top right: unidirectional flows and two different processing capacities, e.g., stratified system (water column or water plus sediment column); c)
bottom left: bi-directional flows, or d) bottom right: bi-directional flows and two different processing capacities.
Residence Time
M = Modifying Factors
P= Pollutant
Residence Time 1
P = Pollutant
M — Modifying Factors
Residence Time 2
Residence Time
M — Modifying Factors
P = Pollutant
Residence Time 1
M = Modifying Factors
P = Pollutant
12
-------
Table 1. Modifying factors and processing rates relevant for major aquatic stressors.
Modifying factors
Processing rates
Toxics
Total and dissolved organic carbon
Acid volatile sulfide
PH
Ionic strength (salinity, hardness)
Temperature
Redox potential (oxic vs. anoxic)
Total suspended solids
Photic depth
Nutrients
Biotic, abiotic degradation (photolysis, hydrolysis)
Bioaccumulation, uptake
Denitrification and nitrification, primary production
Total suspended solids
Other nutrients, Redfield ratio effects
Temperature
Dissolved oxygen
Hardness
Hypsography (depth distribution)
Organic matter loading, shift processes
Primary production and respiration
Remineralization (bacteria)
Grazing rates, food web and chain changes
Nitrification and denitrification
Shear force
Grain size distribution
Large scale structure
Flow magnitude and duration, flashiness
Filter feeding
Bioturbation
Physical resuspension (storm events)
Large-scale hydrologic
Normal channel and basin evolution
13
-------
II. APPROACH
Methods Applied for Stage I Classification
Database
Database structure
We compiled separate databases for clas-
sification of estuarine and Great Lakes
coastal watersheds and wetlands. Both data-
bases will eventually contain the same com-
ponents, i.e., information on system
morphometry and hydrology needed to esti-
mate retention time, watershed characteristics
determining hydrologic regime and/or load-
ings for stressors and modifying factors, and
ecological exposure and effects data necessary
to test the classification systems. In the Stage
I database, some components of the estuarine
database are more complete than those for
coastal Great Lake wetlands, while data for
Great Lakes coastal watersheds are available at
a finer scale of spatial resolution. The delinea-
tion of finer scale watersheds for Great Lakes
coastal wetlands allowed us to develop indica-
tors of hydrologic regime for individual sys-
tems, while for estuarine systems, only
regional thresholds were derived for the Stage
I database. This situation reflects the current
availability of data, and will be rectified during
development of the Stage II database. The
components of the databases and sources of
data are described below.
Estuaries
Any geographical classification requires stan-
dard spatial units. The base unit for this clas-
sification framework was a unique
combination of USGS 8-digit Hydrologic Unit
Codes (HUCs) and Estuarine Drainage
Area/Coastal Drainage Area (EDA/CDA)
defined by NOAA. Watersheds are delineated
by USGS using a nationwide hierarchical sys-
tem based on surface hydrologic features
(Seaber et al., 1987). The hydrologic unit sys-
tem divides the U.S. sequentially into finer
and finer drainage basin subdivisions, with
regions as the largest unit (2-digit code) and
cataloging units as the fourth level sub-
division (8-digit code).
The Coastal Assessment Framework (CAP)
(NOAA, 2003d) is a consistently derived, wa-
tershed-based, national digital spatial frame-
work that is similar to the 8-digit HUC
system. The main difference between CAP
and HUC lies in the method by which NOAA
sub-divides 8-digit HUCs into ED As and
CDAs where the limits of tidal influence
within an estuary or coastal drainage area are
incorporated. An EDA is that component of
an estuary's entire watershed that empties di-
rectly into waters affected by the tides. ED As
may be composed of a portion of a single hy-
drologic unit, an entire hydrologic unit, more
than one hydrologic unit, or several complete
hydrologic units and portions of several adja-
cent hydrologic units. Every EDA has both a
land and water component. A CDA is gener-
ally defined as that component of an entire
watershed that meets the following three cri-
teria: 1) it is not part of any EDA or a corre-
sponding FDA (fluvial drainage area), 2) it
drains directly into an ocean, an estuary, or
the Great Lakes, and 3) it is composed only of
the HUC that is closest to the ocean or shore-
line.
Our unique spatial referencing unit for coastal
watershed classification is based primarily on
the USGS 8-digit HUCs, modified in some
areas along the coast by sub-dividing or com-
bining the HUCs using the EDA/CDA
boundaries. We identified 8-digit HUCs that
drained into an ocean or estuary. For the
conterminous U.S. we selected 277 HUCs that
were directly associated with the coast. We
also identified ED As and CDAs located along
14
-------
the U.S. coast. Of the 348 ED As and CD As
listed in the CAP, we selected a total of 203
(145 ED As and 58 CDAs) for our classifica-
tion database.
The HUCs and ED A/CD As were overlaid on
a map of the U.S. using GIS. HUCs were
geographically referenced to EDA/CDAs.
There were 240 one-to-one matches of HUCs
to EDA/CDAs. Differences in the methods
of delineating EDA/CDAs from HUCS re-
sulted in 64 EDA/CDAs that overlapped the
boundaries of more than one HUC and 44
HUCs that overlapped the boundaries of
more than one EDA/CDA (Figure 5).
We developed unique identification codes for
348 classification units that enabled us to ref-
erence either a HUC or EDA/CDA depend-
ing on our needs and whether data was
aggregated by HUC or EDA/CDA. Five fi-
nal databases were generated for use in classi-
fication and future stressor-response
modeling: 1) physical and hydrologic charac-
teristics, 2) land cover statistics, 3) stressor
loads, 4) in situ stressor concentrations, and 5)
modifying factors. Measurements or indica-
tors of exposure (stressors, loadings, land-
cover) are not intended to be used to classify
systems but to test for differences in response
among classes to stressor gradients. An addi-
tional standard spatial unit database contained
both EDA and HUC identifiers so that data
could be merged from sources referenced by
either EDA or HUC. The physical and hy-
drologic characteristics of ED As were used to
classify estuaries into groups whose members
had similar characteristics. This database in-
cluded area, volume, flow, tides, depth, and
salinity for each EDA (Appendix A-1.2).
Most of the variables were derived directly
from CA&DS. CA&DS is a national- and
regional-level database and mapping analysis
tool designed to access, synthesize, assess, and
apply nationwide data sets to priority coastal
issues such as estuarine eutrophication, essen-
tial fish habitat, coastal monitoring, and sus-
tainable development (NOAA, 2003a).
Where CA&DS data were incomplete, data
from the Estuarine Eutrophication Survey
(NOAA, 1996; NOAA, 1997a-c; NOAA,
1998) were used. Average salinity and depth
were calculated from EMAP data, where
point locations sampled from 1990-2000 were
geo-referenced to ED As.
Great Lakes
Spatial units for Great Lakes systems were the
coastal riverine wetlands and their associated
watersheds. Unlike the case for estuaries, re-
ceiving water bodies could be defined at a
finer scale, and watershed boundaries scaled
appropriately. A recently completed inven-
tory of coastal riverine wetlands in the Great
Lakes identified a total of 283 sites (Simon et
al., 2003). The classification database for the
Great Lakes includes data for those 150
coastal riverine wetlands and associated water-
sheds for which response variables have been
measured through the EPA Region 5 Great
Lakes Coastal Wetlands R-EMAP (Simon et
al., 2003). Although 8-digit HUC boundaries
for coastal drainage areas in the Great Lakes
are defined in CA&DS, the spatial resolution
of these hydrologic unit boundaries is too
coarse to provide the base coverage for a clas-
sification system for Great Lakes coastal wet-
lands. To develop the National Watershed
Boundary Database, delineation of 10- and
12-digit HUCs has begun for the Great Lakes
states, but is not yet complete (Legleiter,
2001). For the current classification database,
watershed boundaries for Great Lakes coastal
riverine wetlands were either imported from
existing state watershed coverages or 12-digit
HUCs, delineated through an automated
process using digital elevation models
(Franken et al., 2001) or digitized onscreen in
an Arclnfo Geographic Information. System
(GIS) using 1:24,000 topographic maps (digi-
tal raster graphics files) and 1:100,000 Na-
tional Hydrography Database (NHD) stream
coverages as a backdrop (USGS, 2003b).
15
-------
BOBt>101_G220*
Figure 5. a) HUC 8090302 (West Central Louisiana Coastal) overlapped two EDAs (G210x-
Terrebonne/Timbalier Bays and G220x-Atchafalaya/Vermillion Bays) resulting in two unique spatial units for
classification (8090302_G210x and 8090302_G220x), b) EDA G310x (Corpus Christi Bay) overlapped four HUCs
(12110111-Lower Nueces, 12110201-North Corpus Christi Bay, 12110202-South Corpus Christi Bay, and
12100405-Aransas Bay) resulting in four unique spatial units for classification (12110111_G310x,
12110201_G310x, 12110202_G310x, and 12100405_G310x). Yellow and green areas = EDA and red outline =
HUC.
The full population of Great Lakes coastal
riverine wetlands was defined operationally
for the Region 5 R-EMAP according to the
following criteria:
*»* wetland area
*»* surface connection to the lake
*»* proximity to the Lake (must be within 10-
20' elevation above lake level)
subset of all coastal wetlands (drowned-
river mouth and riverine wetlands) ex-
tracted based on
^ association with 2nd order streams or
larger (1:24,000) OR
^ association with 1st order streams if
tributary is outflow of a lake or pond.
16
-------
A list-based sample frame of all coastal Great
Lakes wetlands was developed based on an
inventory generated by USFWS (1981a-f) and
supplemented by identification of wetlands
using marsh symbols present on 1:24000 to-
pographic maps, presence of emergent or
floating vegetation on digital orthophoto
quadrangle aerial photos, and experience of
local wetlands experts (Simon et al., 2003). A
subset of 22 coastal riverine wetlands in the
Lake Michigan basin was selected for sam-
pling in 2000 using a probabilistic survey de-
sign with unequal probability weighting for
ecoregion classes and three wetland areal size
classes (<100 acres, 100-1000 acres, and
>1000 acres). In 2001, this set of coastal riv-
erine wetlands was supplemented with sites
selected through a second probabilistic survey
design for the four Great Lakes within EPA
Region 5 (Lakes Superior, Michigan, Huron,
and Erie) and associated connecting channels,
for a sample total of 155 coastal riverine wet-
lands. To distribute sampled sites more
evenly among chosen categories, the second
survey design included unequal probability
weighting by Great Lake or connecting chan-
nel class and by wetland size class (Simon et
al.,2003).
Five separate databases were generated for
Great Lakes coastal systems. These include:
1) a database of watershed characteristics and
hydrologic variables for selected USGS gaug-
ing stations in coastal Great Lakes states
(Glatfelter, 1984; Holtschlag and Croskey,
1984; Jacques and Lorenz, 1988; Krug et al.,
1992; Lumia, 1991; Sauer et al., 1983; Stedfast,
1986), 2) derived hydrologic thresholds by
state climatological region that define bounda-
ries among hydrologic regime classes based on
watershed characteristics, 3) characteristics
and hydrologic regime class for watersheds
associated with a subset of 155 coastal riverine
wetlands, 4) coastal riverine wetland attributes
for the subset of 155 coastal riverine wetlands,
and 5) biological condition estimates for the
subset of 155 coastal riverine wetlands. Hy-
drologic thresholds in the second database
were derived for all coastal marine states as
well.
Databases to characterize hydrologic regime
class of Great Lakes coastal systems were de-
rived by compiling data from state-level
USGS reports on watershed characteristics
related to peak and base flows (see Jennings et
al., 1993 for summary). USGS and state part-
ners first divide each state into homogeneous
hydroclimatological regions. Gaging stations
are identified with long-term time series ade-
quate to define peak flows for recurrence in-
tervals ranging from 2 to 100 years. For each
region, equations have been developed to
predict peak flows based on watershed char-
acteristics. Watershed characteristics to be
included in each analysis are chosen on a
state-by-state basis but typically include con-
tributing drainage area, main channel slope,
and a subset of other variables such as water-
shed storage (lake + wetland and watershed
area), minimum soil permeability, and precipi-
tation (e.g., 2-year, 24-hour rain event magni-
tude, annual snowfall, annual average
precipitation) (Jennings et al., 1993).
For each hydroclimatological region within
each state, a combination of visual graphical
analysis and CART techniques were applied to
identify hydrologic thresholds (Wilkinson,
1999). We define a hydrologic threshold as a
region of rapid change in a hydrologic metric
such as peak discharge for a 2-year recurrence
interval normalized by watershed area as a
function of other watershed characteristics.
For example, hydrologic thresholds that have
been derived and tested for some regions of
the Great Lakes include a watershed storage
threshold of 5-10% and a mature forest
threshold of 50% (Detenbeck et al., 2000).
Below these thresholds, area-normalized peak
flow increases exponentially, such that thresh-
old values can be used to predict which water-
sheds will have stable versus flashy hydrologic
regimes.
17
-------
Watershed characteristics were derived for
each of the 155 Great Lakes coastal riverine
wetlands sampled under the EPA Region 5 R-
EMAP. For Stage I of the database, water-
shed characteristics included land cover as an
indicator of stressors such as potential nutri-
ent, sediment, and toxics inputs, along with
two derived watershed flashiness indices.
Coastal riverine wetland characteristics in-
cluded wetland location and wetland area in
forested, emergent, and submerged wetland
classes, as defined in National Wetlands In-
ventory and state wetland inventory cover-
ages. Initial estimates of coastal wetland areas
were taken from (USFWS 1981a-f).
Data Sources for Stressor Exposure(s)
In the initial version of the classification data-
base, data for stressor exposures were com-
piled only for coastal watersheds along the
marine coast, using the EDA/CDA and HUG
units described above. For the Great Lakes
watersheds, only land cover for coastal river-
ine wetland watersheds has been compiled
because most other information sources in-
clude data that have already been aggregated
to the level of 8-digit HUCs.
Land Cover
USGS and EPA created the National Land
Cover Data (NLCD), (Vogelmann et al.,
2001), for the conterminous U.S. based on
early to mid-1990s, 30-meter Landsat The-
matic Mapper (TM) satellite imagery. The
NLCD consists of 21 level II land cover
classes (Figure 6) (Vogelmann et al., 2001).
To derive acreage statistics for each spatial
referencing unit or EDA, we performed a ma-
trix overlay of our spatial referencing unit
dataset with the NLCD.
For Great Lakes coastal watersheds, land
cover classes were aggregated to the following
categories: agricultural, nonagricultural grass-
lands, commercial and residential, forested,
wetlands, barren, mining, and open water
(Table 2). For estuaries, NLCD level II land
cover classes were aggregated to level I classes
to produce area estimates for water, shrub-
land, grassland, non-natural woody, devel-
oped, barren, forested, agricultural, and
wetland land cover classes (Appendix A-2.1).
18
-------
Emergent Herbaceous Wetlands
Woody Wetlands
Urban/Recreational Grasses
Fallow
Small Grains
Row Crops
Pasture/Hay
Grasslands/Herbaceous
OrchardsA/ineyards/Other
Shrubland
Mixed Forest
Evergreen Forest
Deciduous Forest
Transitional
Quarries/Strip Mines/Gravel Rts
Bare Rock/Sand/day
Commercial/lndustrial/Transportation
High Intensity Residential
Low Intensity Residential
Perennial Ice/Snow
Open Water
11.22
I 2. Ib I
D 0.22
I 1 0.89
I 3.31 |
1 8.98 I
I 1b.4 |
J0.21
\1t.bi I
1 4.24 |
I12.4/ I
|11. bl \
^\ 0.62
0.08
| 1.37
^ 0.56
D 0.26
1 1.02
0.02
I b-02 |
0 5 10 15 20
Figure 6. Breakdown of NLCD land cover classes for coastal EDAs and CDAs.
Table 2. Aggregation of land-cover categories for Great Lakes and estuarine watershed databases.
Original NLCD Class Aggregated Class for Great Lakes Aggregated Class for Estuaries
Open Water
Low Intensity Residential
High Intensity Residential
Commercial/Industrial/Transportation
Bare Rock/ Sand/ Clay
Transitional
Quarries/Strip Mines/Gravel Pits
Deciduous Forest
Evergreen Forest
Mixed Forest
Shrubland
Orchards/Vineyards/Other
Pasture/Hay
Row Crops
Small Grains
Fallow
Urban/Recreational Grasses
Grasslands/Herbaceous
Woody Wetlands
Emergent Herbaceous Wetlands
Open Water
Commercial/Residential
Barren
Mining
Forested
Shrubland
Agricultural
Non-agricultural grasslands
Wetlands
Water
Developed
Barren
Forested
Shrubland
Non-Natural Woody
Agricultural
Herbaceous Upland Natural/ Semi-
Natural Vegetation
Wetlands
19
-------
Risk from Groundwater Nitrate Inputs
Groundwater nitrate risk was obtained from
USGS (2003c). Nitrate risk is based upon in-
put factors (population density and nitrogen
loading) and aquifer vulnerability factors (soil
drainage and woodland and cropland ratio in
agricultural areas). The USGS calculated the
soil drainage class from the State Soil Geo-
graphic (STATSGO) database (USDA, 2001).
Categorical values were converted to numbers
and threshold values determined. Soils with a
hydrologic group value of <2.5 were consid-
ered well-drained; soils >2.5 were considered
poorly drained. The woodland and cropland
ratio was calculated from the 1992 Census of
Agriculture. The data were divided into two
groups: those with a woodland and cropland
value <0.3 and >0.3. Population density was
calculated from 1990 census data by dividing
number of people in a block group by the to-
tal area of the block group. These data were
divided into two groups: regions having <386
people/km2 and regions with >386 peo-
ple/km2. Nitrogen loading was calculated by
adding input from fertilizer, manure and at-
mospheric deposition (Battaglin and Goolsby,
1994). Nitrogen loading was divided into two
groups: <2100 kg/km2 and >2100 kg/km2.
These four data sets were added together to
create four risk categories (Table 3).
Groundwater nitrate collected from wells less
than 100 ft deep generally verified the nitrate
risk patterns. This national coverage was
matched to the HUC and EDA/CDA units
described earlier
Table 3. Categories of risk for groundwater nitrate inputs.
Risk Category
High Risk
Moderately High Risk
Moderately Low Risk
Low Risk
N Loading
>2100 kg/km2
>2100 kg/km2
<2100 kg/km2
<2100 kg/km2
Population Density
>386 people/km2
>386 people/km2
< 386 people/km2
< 386 people/km2
Hydrologic
Group
<2.5
>2.5
<2.5
>2.5
Woodland/
Cropland Ratio
<0.3
>0.3
<0.3
> 0.3
Risk from Pesticides
Risk from pesticides from agricultural sources
was derived from the USDA NRCS National
Pesticide Loss Database (Goss et al., 1998,
NPLD). Goss et al. (1998) applied the pesti-
cide fate and transport model from Ground-
water Loading Effects from Agricultural
Management Systems (GLEAMS) to scenar-
ios for 243 pesticides applied to!20 generic
soils for 20 years of daily weather from each
of 55 climate stations. They then used data
from the NPLD, along with data from the
1992 National Resources Inventory (NRI), on
crop type and 1990-1993 pesticide use data
(application rate by crop) from Gianessi and
Anderson (1995). Estimates of pesticide loss
from the NPLD were imputed onto the
170,000 field sample points in the NRI data-
base according to soil type, geographic loca-
tion, and pesticide. Predicted concentrations
were compared to Maximum Acceptable
Toxicant Concentrations (MATCs) to derive
threshold exceedance units at each point, then
20
-------
multiplied by the number of acres treated and
summed over points in each watershed to cre-
ate aggregate measures of risk. Two risk indi-
ces available from Natural Resources
Conservation Service (NRCS) at the 8-digit
HUG scale were included in the EDA data-
base: 1) Acres (1,000) by watershed where the
potential leaching concentration at the bottom
of the root zone exceeds a multiple of one or
more water quality thresholds for fish, and 2)
Acres (1,000) by watershed where the poten-
tial runoff concentration at the edge of the
field exceeds a multiple of one or more water
quality thresholds for fish.
Loadings
The classification database for stressor loads
includes total and point source nitrogen and
phosphorus, total suspended solids, and indi-
cators of toxic contaminant and sediment
loads (Appendix A-3.2). Nitrogen and phos-
phorus loads were derived from the USGS
Spatially Referenced Regression on Watershed
Attributes (SPARROW) model for point and
non-point sources of nutrients (Smith and
Alexander, 2000; Smith et al., 1997). The
loadings encompass both point and non-point
sources estimated empirically, based on data
collected from approximately 400 long-term
stream monitoring sites, nutrient sources and
physical characteristics of the watershed
(Battaglin and Goolsby, 1994). Datasets gen-
erated using the SPARROW model include
watershed total nitrogen and total phosphorus
export and yield by source, as well as percent
contribution to total load by source (fertilizer,
livestock waste, atmosphere, non-agriculture;
Smith and Alexander, 2000). Additional point
source loads for nitrogen, phosphorus, and
total suspended solids were compiled from
Permit Compliance System (PCS) data housed
in the Better Assessment Science Integrating
point and Nonpoint Sources (USEPA, 2003a,
BASINS). Total nitrogen and phosphorus
loads were calculated by averaging loads over
time for each National Pollutant Discharge
Elimination System (NPDES) identification
code and then summing loads across all
NPDES codes by HUC.
Twenty-nine chemical loads were also com-
piled from the PCS. A principal component
analysis (PCA) was conducted to reduce this
data set to three principal components repre-
senting PAHs, metals, and pesticides.
Sediment load data are not currently in the
classification database; however, a hydrologic
unit model for the U.S. (HUMUS; TAES,
2000) has been developed to estimate runoff,
and sediment, phosphorus and nitrogen loads
to coastal areas. HUMUS utilizes four input
data types (land use, soil survey, digital eleva-
tion, and climate). For the current version of
the database, we have included a relative rank-
ing for the potential for sediment delivery to
HUCs, based on data from the HUMUS da-
tabase, was obtained from the Index of Wa-
tershed Indicators (IWI) database (USEPA,
2003i). The relative rank of watersheds based
on this parameter estimates the potential for
possible water quality problems from in-
stream sediment loads.
In Situ Stressor Exposure Levels
Nutrients
In situ nutrient concentrations were obtained
from EMAP NCA and BASINS (USEPA,
2003a; 2003b) databases (Appendix A-4.1).
Water quality data collected in 2000 for NCA
included dissolved inorganic nitrogen (nitrate,
nitrite and ammonia), which species were
summed for the stressor exposure database.
Where data gaps existed, nitrogen data from
the BASINS database were utilized to com-
plement the EMAP nutrient database. Total
Kjeldahl nitrogen and total phosphorus con-
centrations were extracted from the BASINS
database and averaged by HUC. All data were
geo-referenced to ED As.
21
-------
Suspended Sediments
As a surrogate for suspended sediment con-
centrations, total suspended solids (TSS) data
were obtained from EPA's Storage and Re-
trieval (STORET) data summaries provided in
the BASINS database. TSS concentrations
were averaged by HUG and geo-referenced to
EDA. TSS concentrations are also included
in the stressor exposure database as a modify-
ing factor.
Toxics
Sediment contaminant concentrations
were obtained from EMAP and were con-
verted to toxic units for metals, pesticides,
polychlorinated biphenyls (PCBs), and PAHs.
Sediment toxicity information was obtained
from three sources: National Sediment Inven-
tory (NSI) (USEPA, 1996), STORET and
EMAP. NSI data and STORET data were
sorted first by freshwater or marine location.
STORET data were sorted into freshwater or
marine using the website options; NSI data
were sorted based upon the descriptors in-
cluded in the state-by-state database. NSI
data and STORET data were then sorted by
the presence or absence of 32 toxic sub-
stances and 5 modifying factors (Table 4).
Using best professional judgment and avail-
able data in the literature, these toxics were
chosen as the most commonly occurring in
sediments; choices were consistent with
summaries in the NSI report (USEPA, 1998).
Five modifying factors were chosen for their
ability to assist in interpretation of the bio-
availability of the toxics (Table 1). Entries
containing any of the listed toxics were cho-
sen for further consideration if they contained
either HUG location or latitude- longitude
coordinates that could be associated with an
8-digit HUG.
22
-------
Table 4. Effects level for diagnostic screening.
Toxic Chemical
Name
Total Ammonia5
Water-only Toxicity
Value (ng/L) A
Freshwater
4.68 (pH 6.5)
4.15 (pH 7)
1.71 (pH 8)
0.342 (pH 9)
Marine
ll(pH7)
1.1 (pH 8)
0.14 (pH 9)
Sediment Toxicity Value
(See footnotes for units)
Freshwater
Use water-
only values
Marine
Use water-
only values
Sediment Con-
centration
Units1'2
Metals
Cadmium
Copper
Chromium
Mercury
Nickel
Zinc
2.2
9.0
11D
0.77
52
120
9.3
3.1
50D
0.94
8.2
81
0.99C
31. 6C
43.4C
0.1 8C
22. 7C
120C
1.2C
34C
81C
0.1 5C
21C
150C
Hg/g dwt
Hg/g dwt
Hg/g dwt
Hg/g dwt
l^g/g dwt
Hg/g dwt
Organics
Pesticides
Chlordane
Chlorpyrifos
Total DDTs
Diazinon1
Dieldrin
Total Endosulfans
Endrin
Pyrethroids (Permethrin)1
Total PCBs
0.0043
0.041
0.001
0.05
0.056
0.056
0.036
0.024
0.014
0.004
0.0056
0.001
0.0019
0.0087
0.0023
0.03
3.24E
5.28E
12000F
5400F
35E
1.58E
28000F
990F
60E
ng/g dwt
ng/g dwt
ng/g dwt
ng/g dwt
ng/g OC
ng/g dwt
ng/gOC
ng/g dwt
ng/g dwt
Poly cyclic Aromatic Hydrocarbons (PAHs)G
Acenaphthene
Acenaphthylene
Anthracene
Benzo (a) anthracene
Benzo (b) fluoranthene
Benzo (k) fluoranthene
55.9
307
20.7
2.23
0.68
0.64
491000F
452000F
594000F
841000F
979000F
981000F
ng/g OC
ng/g OC
ng/g OC
ng/g OC
ng/gOC
ng/gOC
23
-------
Toxic Chemical
Name
Benzo (ghi)perylene
Benzo(a)pyrene
Chrysene
Dibenzo (a,h) anthracene
Fluoranthene
Fluorene
Indeno (1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
Total Dioxins
Water-only Toxicity
Value (|ag/L) A
Freshwater Marine
0.44
0.96
2.04
0.28
7.11
39.3
0.28
194
19.1
10.1
1.4 * 10-8
Sediment Toxicity Value
(See footnotes for units)
Freshwater Marine
1100000F
965000F
844000F
1120000F
80H
538000F
1120000F
385000F
596000F
697000F
-
Sediment Con-
centration
Units1-2
ng/g OC
ng/g OC
ng/g OC
ng/gOC
ng/g dwt
ng/gOC
ng/gOC
ng/g OC
ng/g OC
ng/g OC
Dwt = dry weight OC = organic carbon
A Water Quality Criteria (WQC) or Final Chronic Values (FCV) from USEPA (1989; 1999b,c; 2003e,f,g)
B mg/L, freshwater values assume 20°C and marine values assume 20°C and 30%o.
c From MacDonald et al. (2000a) and Long et al. (1995) in |-ig/g dry weight.
D Chromium VI.
E From MacDonald et al. (2000a,b) and Long et al. (1995) in ng/g dry weight.
F From USEPA (2003e,f,g) in ng/g organic carbon.
G Freshwater and marine values are identical.
H From USEPA (1999a).
1 Diazinon and pyrethroid values are from Illinois water quality standards.
(http://www.epa.state.il.us/water/water-quality-standards/water-quality-criteria-list.pdf)
Note 1: Using the units above, sediment dry weight concentrations can be converted to sediment organic carbon
normalized concentrations using the following:
Sediment Concentrationoc = Sediment Concentration dwt -=- (Sediment organic carbon (in %) * 0.01)
Note 2: Using the units above, toxic units are calculated as follows:
Toxic Units dry weight = Sediment Concentrationary weight "=" Sediment Toxicity Valueary weight
Toxic UnitS organic carbon = Sediment Concentrationorgamc carbon -=- Sediment Toxicity Valueorgamc carbon
Note 3: Diazinon and pyrethroid values are from Illinois water quality standards
(http://www.epa.state.il.us/water/water-quality-standards/water-quality-criteria-list.pdf).
These values are lower than those observed in surveys of peer- reviewed literature with one exception
Schulz and Liess, (2000), indicate that fenvalerate, a synethetic pyrethroid, has an effect level lower
than the Illinois chronic water quality criterion (0.024 |^g/L). In their paper, fenvalarate with a one hour
exposure, at a concentration of 0.001 |^g/L has an effect on the temporal emergence of the caddis fly. In
addition the same paper states that 0.01 |^g/L affects the dry weight of adults. The WQC number was
used because the fenvalerate number is an outlier (order of magnitude lower than the rest), and while
fenvalerate is a synthetic pyrethroid, it is not one often measured in waters and sediments, and
we have based the pyrethroid number on permethrin concentrations
24
-------
Data Sources for System Properties Affecting Retention
Time
Most morphometry data were only available for
estuarine systems. For Great Lakes coastal riv-
erine wetlands, only preliminary estimates of
wetland area are included in Stage I of the
classification database.
Estuaries
Morphometry of estuaries includes measures of
area, depth and volume. There are five meas-
ures of area within the database derived from
CA&DS which were converted from square
miles to square kilometers. The total area of
the EDA represents land and water area for the
watershed. Estuary area represents water area
for the watershed. Mixing zone area represents
the area in the estuary where salinity ranges
from 0.5 to 25 ppt. The sea zone area repre-
sents the area in the estuary where salinity is
>25 ppt. The tidal freshwater zone represents
the area in the estuary where salinity is <0.5
ppt. Depth of the estuary in meters was ob-
tained from EMAP, averaged over time and
space. If depth or volume was not available for
an EDA from EMAP data, the average depth
or estuary volume was extracted from NOAA's
Estuarine Eutrophication Survey Regional Re-
ports (NOAA, 1996; NOAA, 1997a-c; NOAA,
1998). All depth values were converted from
feet to meters. Volume was converted from
cubic feet to cubic meters. Where estuarine
volume was missing, it was estimated by multi-
plying estuary area by average depth.
Tidal Range
Tidal height and tidal prism volume were ob-
tained from CA&DS. Average tidal height was
calculated as the means of the height differ-
ences or ratios measured from NOAA National
Ocean Service (NOS) tide gauge stations and
converted from feet to meters. Tidal prism
volume was calculated from the salinity zone
mean range value or the salinity mean tide value
multiplied by two. The salinity zone tidal value
(depth) multiplied by salinity zone area (i.e.,
tidal freshwater zone, mixing zone, seawater
zone) provided volume for each salinity zone.
The tidal prism volume was calculated as the
sum of all salinity zone volumes. If tide infor-
mation was not available for all three salinity
zones, the estuary mean range was used or the
estuary mean tide value multiplied by two. This
value was converted from cubic feet to cubic
meters.
Riverine discharge
Average monthly river flow was obtained from
CA&DS. These values were obtained from the
annual long-term flow average of gauged rivers
from USGS. Where such data were missing,
average daily inflow values were converted to
monthly and were substituted from NOAA's
Estuarine Eutrophication Survey Regional Re-
ports (NOAA, 1996; NOAA, 1997a-c; NOAA,
1998).
Salinity-based Indicators of Retention Time
We estimated the dissolved concentration po-
tential (DCP) of a pollutant as a function of
pollutant load, the volume of freshwater in the
estuary, freshwater inflow, and total estuarine
volume. The volume of freshwater in the estu-
ary was calculated using the freshwater fraction
method:
F£w = (SO-S)/SO) where,
w = Freshwater fraction,
SO = Boundary Salinity, and
S = Average salinity.
Average salinity data were obtained from the
EMAP database for surface and bottom waters.
Boundary salinities were designated as 35 ppt
25
-------
unless average salinities in the estuary were hy-
persaline (>35 ppt). In those cases with salini-
ties >35 ppt, boundary salinities were based on
averaged salinity and rounded up to the nearest
whole number. For example, if the average sa-
linity was 38.5 ppt, then the boundary salinity
was set at 39 ppt. This procedure eliminated
negative values for the volume of freshwater in
the estuary.
The volume of freshwater was calculated using:
Vfw - Ffw x Vtot where,
/fa = volume of freshwater in the estuary,
Tfw = Freshwater fraction, and
Vtot = Estuarine volume.
Estuarine volume data were obtained, by EDA,
from the Estuarine Eutrophication Survey
(NOAA, 1996). Where no data were available
for ED As or CD As, volume was estimated by
multiplying average depth from EMAP data
and estuarine area from CA&DS.
DCP was calculated using the following equa-
tion (NOAA, 1989):
DCP = L(Vfw/Ifw)(l/Vtot) where,
DCP = Dissolved concentration potential,
L = Pollutant Load,
fw =Volume of freshwater in the estuary,
= Average freshwater inflow (daily average river
flow), and
Vtot = Estuarine volume.
In order to compare relative DCP values
among ED As, a constant pollutant load (L) was
assigned to each EDA. Relative DCP values
can be used to estimate the concentration of a
pollutant expected in an estuary assuming that
its concentration is entirely controlled by physi-
cal processes.
Particle retention efficiency (PRE) estimates the
ability of an estuary to trap suspended particles
(i.e., the time a particle remains in an
estuary). PRE is calculated using the formula
(NOAA, 1989):
PRE = C/I where,
C = Volume of the estuary, and
I = freshwater inflow.
Because both nutrients and toxic substances
can bind to particles entering the estuary, it is
necessary to include PRE when evaluating in
situ concentrations and loads. In addition, PRE
can be applied to sediment loads.
Modifying factors are variables that change the
equilibrium or rate of material processing, e.g.,
pH changes the bioeffective concentration of
many toxic metals and turbidity changes the
maximum productivity rate associated with a
given nutrient supply rate. The following vari-
ables have been identified as "modifying fac-
tors", factors that affect the equilibrium or rates
of nutrient processing and are included in the
classification database. Associations between
primary stressors and potential modifying fac-
tors were described in Table 1.
Sediment total organic carbon and acid volatile
sulfide concentrations were obtained from
EMAP. Average dissolved oxygen concentra-
tions, water temperature, salinity, and pH were
calculated from surface and bottom measure-
ments from EMAP NCA databases. Where
gaps existed, values were extracted from the
BASINs database. Values were geo-referenced
to ED As and HUCs.
Total suspended solids, total chloride, sulfate
concentrations, hardness and alkalinity as cal-
cium carbonate in water, and specific conduc-
tance were derived from the BASINS database.
Average values by HUC were geo-referenced to
EDA.
26
-------
Great Lakes
Morphological data for Great Lakes wetlands
are less available and less well developed than
those for estuaries; thus in Stage I, emphasis
was placed on creating a geospatial inventory of
coastal wetlands based on hydrogeomorphic
class. A GIS database was constructed to sup-
port the identification and classification of
Great Lakes coastal wetlands along the U.S.
shoreline for each of the four Great Lakes in-
cluded in the Great Lakes R-EMAP project:
Lakes Superior, Erie, Huron, and Michigan.
Coastal wetlands associated with connecting
channels and Lake St. Clair were also included.
The GIS database consists of a fine-resolution
shoreline coverage for each of the lakes and
connecting channels, the most accurate hydro-
graphy coverages available for each Great Lakes
basin, a compilation of the most up-to-date
digital wetland inventory coverages for the
Great Lakes basin, a point coverage of Great
Lakes coastal wetlands developed from Her-
dendorf s records (USFWS, 1981a-f) which
were supplemented by locations derived from
local experts, and an elevation contour denoting
the upper extent of lake level influence. The
shoreline coverage was constructed from
1:24,000 state hydrography where available,
supplemented by the NOAA medium resolu-
tion shoreline vector for other states (Illinois,
Indiana). The hydrography coverage was con-
structed from 1:24,000 state hydrography where
available, supplemented by 1:24,000 digital line
graph (DLG) coverages for the state of Ohio.
Digital wetland inventory coverages were com-
piled from the National Wetlands Inventory
(USFWS, 2003), Wisconsin Wetlands Inventory
(WIDNR, 2003), Ohio Wetlands Inventory
(OHDNR, 2003), and New York wetlands in-
ventories developed by the state and by the
Adirondack Park Agency. Wetland inventory
coverages were queried to extract the wetland
polygons expected to be associated with coastal
wetland areas. A coastal zone was defined by
intersecting the Lake Michigan Basin with a
buffered version of the Great Lakes high-water
elevation contour. In future versions of the
Great Lakes coastal wetlands database, coastal
wetlands will be identified by intersecting wet-
lands inventory coverages with buffered shore-
line and stream and river hydrography seg-
ments, and then classifying these according to
proximity to lake (mainland and island) shore-
lines and streams or rivers.
Hydrologic Regimes
Two sources of data were used to define hydro-
logic regime classes of Great Lakes and marine
coastal watersheds. These included 1) the deri-
vation of empirical hydrologic thresholds for
state hydro-climatological regions based on
USGS datasets (Ries and Grouse, 2002), and 2)
the derivation of watershed flashiness indices
(Great Lakes coastal watersheds only). In fu-
ture versions of the classification database, all
coastal watersheds will be assigned to hydro-
logic regime types based on regions defined by
Saco and Kumar (2000).
Two different approaches for the development
of watershed indicators of hydrologic regime
were tested. The first approach relied on visual
analysis of empirical relationships derived by
USGS between flood magnitude of given return
intervals and associated watershed attributes
(Detenbeck et al., 2000; Jennings et al., 1993).
Flood prediction equations created by the
USGS are of the form:
Qn = Aa Bb O
where Qn = peak flow with recur-
rence interval of n years
A, B, C = watershed
attributes
Typically, watershed area is one predictive vari-
able; other variables commonly included are
main channel slope (S) and watershed storage
(ST, fraction of watershed area covered by lakes
and wetlands). In some regions, soil permeabil-
ity (SP) or texture, precipitation (snowpack
27
-------
(SN) or 2-year, 24-hour rainfall event (I24_2)),
and land cover (% forest, % urban or impervi-
ous surface area) are also included. The expo-
nent for watershed area (A) is typically close to
1, so it is reasonable to normalize peak flows
for watershed area (Qn/A). Graphical analysis
can be used to examine plots of single predic-
tors (e.g., Qn/A vs ST) or combined predictor
variables (e.g., Qn/A vs ST *SP) to determine
thresholds of response. Where exponents have
a common sign, it is appropriate to examine
multiplicative terms; where exponents of equa-
tion variables have different signs, it is appro-
priate to examine ratios of variables. We also
supplemented graphical analysis with CART
analysis in SYSTAT to determine if thresholds
could be identified in a more quantitative fash-
ion (Wilkinson, 1999).
We developed a second, model-based approach
for development of indicators of hydrologic
regime to combine effects of changing land
cover and variation in watershed storage. Run-
off volume from peak snowpack or from a de-
sign storm (2-year, 24-hour event) was
compared with estimated watershed storage
volume. Runoff volume was calculated as
maximum potential snowmelt, the water
equivalent of total snowfall for a watershed, or
as runoff expected from a design rainfall event,
using the Soil Conservation Service (SCS) curve
number approach (Neitsch et al., 2002). Design
storm magnitudes were derived by scanning
isopleth maps (Huff and Angel, 1992) and con-
verting these to grid coverages in Arclnfo.
Mean snowfall was estimated based upon geo-
spatial coverages derived from the Parameter-
elevation Regressions on Independent Slopes
Model (Daly et al., 1994, PRISM). We esti-
mated watershed storage volume using different
weighting factors for lake and wetland class
polygons from wetland inventory coverages.
For the latter, we used digital NWI coverages
where available, supplemented by Wisconsin
Wetland Inventory coverages. We estimated
soil hydrologic group (Types A, B, C, and D)
coverage and average slope using STATSGO
layers (USDA, 1994), and derived land cover
class coverage from the NLCD (Vogelmann et
al., 2001).
Methods for Developing and Testing Clas-
sification System
Data Reduction Methods
If too many correlated variables are included in
the final analysis, the results may not be robust.
The purpose of an initial data reduction step is
to identify those sets of variables that are
strongly correlated so that redundant informa-
tion is not included in the final analysis. Three
methods were used to reduce the number of
variables analyzed: PCA of stressor indicators
and exposure data, non-metric dimensional
scaling (NMDS) to reduce dimensionality of
biological community data, and calculation of
toxicity equivalent units (TEUs).
PCA was used to reduce the number of vari-
ables for contaminant load data. Annual aver-
age loads (Ib/yr) for 29 contaminants were
compiled from PCS available through BASINS
(Table 5). The first three principal components
accounted for 76% of the variation With some
individual exceptions the eigenvectors were
weighted on PAHs for the first principal com-
ponent (PCI), metals for the second principal
component (PC2), and pesticides for the third
principal component (PCS). We substituted
these three principal components for the actual
load data in our database.
The magnitude of anthropogenic contam-
ination in estuaries is an important criterion for
testing any classification system. The number
and magnitude of the concentrations of these
contaminants, however, is highly variable
among estuaries, which makes spatial compari-
son complex and unwieldy. For simplification, a
toxic units approach was used to reduce the
number of variables related to contaminant ex-
posure. The measured concentration of each
contaminant in surface water and sediment was
28
-------
Table 5. Contaminant loads available from Permit Compliance System.
Metals
Pesticides
PAHs
Cadmium
Copper
Chromium
Mercury
Nickel
Zinc
Chlordane
Chlorpyrifos
Total DDTs
Diazinon
Dieldrin
Total Endosulfans
Endrin
Acenaphthene
Acenaphthylene
Anthracene
Benzo(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(ghi)perylene
Benzo(a)pyrene
Chrysene
Dibenzo(ah)anthracene
Fluoranthene
Fluorene
Indeno(123-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
divided by the corresponding toxicity values
(Table 4), which are based upon proposed ma-
rine sediment quality guidelines. The resulting
fractions for all contaminants were then
summed for each estuary and presented as a
whole number for sediment. These two values
can then be compared among estuaries for
similar media where higher numbers indicate a
greater likelihood of toxicity.
Methods for A Priori Development of Classes
Classification approaches can be applied either
a priori or posteriori. A. priori classification is
based on a conceptual model or hypothesis
concerning expected differences in behavior of
ecological response along stressor gradients as a
function of watershed or water body character-
istics. In contrast, a posteriori classification is
driven by analysis and interpretation of avail-
able data. We developed and tested a priori clas-
sification strategies based on conceptual models
of watershed hydrology by determining dis-
criminating factors for classification based on
hydrological endpoints as integrators of ex
pected ecological effects (Detenbeck et al.,
2000).
Model-based and hydrologic integration
The hydrologic regime can be used as an inte-
grating factor to indicate the sensitivity of
aquatic systems to stressors (Clausen and Biggs,
2000; Detenbeck et al., 2000; Poff and Ward,
1989). Detenbeck et al. (2000), have derived
hydrologic thresholds based on watershed char-
acteristics that control magnitude and frequency
of peak flows (Breiman et al., 1984; Kass, 1980;
Jennings et al., 1993). Thresholds are deter-
mined as the level of some watershed character-
istic below or above which 2-year flood
discharge and watershed area increases sharply
and exponentially. Peak flows with 2-year re-
currence interval (Q2) values were used as the
response variable because floods with a 2 - 2.5
year return interval are known to have the
greatest influence on channel morphology,
transporting the greatest amount of sediment
and associated pollutants (Rosgen, 1996). A
29
-------
database of peak flow statistics and relevant
watershed characteristics was compiled from
records for USGS gauging stations with long-
term discharge records, using data reported as
part of the USGS National Flood Frequency
program (USGS, 2003e, f). The database in-
cludes all Great Lakes states and marine coastal
states with estuarine systems, as identified in
CA&DS. Hydrologic thresholds were derived
from the database by using a combination of
CART analysis and piecewise linear regression
(Breiman et al, 1984; Wilkinson, 1999). CART
was applied to determine the identity and mag-
nitude of variables associated with significant
shifts in area-normalized peak flows. Piecewise
linear regression analysis was used to determine
breakpoints in the slope of cumulative fre-
quency plots of Q2/area values > regional me-
dian values as a function of classification
variables such as watershed storage. Threshold
parameters were applied to Great Lakes coastal
riverine wetland watersheds for which long-
term flow records were unavailable to classify
them into stable versus flashy hydrologic re-
gime categories. Thresholds will ultimately be
applied to coastal estuarine watersheds for
which flow records are unavailable, as predic-
tors of the type of hydrologic regime present.
Approaches for ^4 Posteriori Development of Classes
Physical and hydrologic characteristics were
compiled for each of the estuarine classification
units from several sources. Parameters in-
cluded area, volume, flow, tides, depth, and sa-
linity. Because the frequency of missing values
was high and often data were unavailable at the
HUC level, we aggregated the original matrix
up to 203 EDA/CD As. This involved taking
the sum or average across classification units
(or HUCs) associated with each EDA/CDA as
appropriate. This procedure reduced the fre-
quency of missing values to 15%. The matrix
used for classification of physical and hydro-
logic characteristics was 203 rows
(ED A/CD As) by 15 columns (Appendix A-
1.2). All numeric values were log-transformed.
Remaining missing values were imputed using
the multiple imputation procedure, PROC MI,
available in SAS/STAT (SAS Institute, 2001).
Multiple imputation provides a useful strategy
for dealing with data sets with missing values.
Instead of filling in a single value for each miss-
ing value, Rubin's (1987) multiple imputation
procedure replaces each missing value with a
set of plausible values that represent the uncer-
tainty about the right value to impute. PROC
MI yielded five completed data sets differing
only in the plausible imputed missing data val-
ues. These five data sets were found to be not
significantly different from each other, so one
data set was chosen at random to be analyzed
by the cluster routine available in PRIMER
(Plymouth Marine Lab). Average-linkage clus-
ter analysis was performed on normalized
Euclidean distances using PRIMER. Eleven
clusters or groups of EDA/CDAs were identi-
fied from examining the dendogram result from
the cluster analysis. Box-plots for each variable
used in the cluster analysis were examined to
compare the mean, median, 25th and 75th per-
centiles, minimum and maximum values across
clusters. This resulted in the derivation of la-
bels to describe the properties of each cluster of
EDA/CDAs.
Tests of Classification Approaches for
Coastal Watersheds
An initial test of a priori hydrologic classes de-
veloped for Lake Michigan watersheds was
done through analysis of covariance
(ANCOVA) of Great Lakes coastal wetland R-
EMAP data using hydrologic regime classes,
loading or land cover gradients vs. stressor and
exposure metrics and biological response vs.
exposure or land cover gradients (Detenbeck et
al., 2003b)
30
-------
III. RESULTS
Estuaries
The final geographic coverage used in cluster
analysis for classification of estuarine
coastal systems contained 203 unique
classification units, representing EDA/CDAs
and HUCs as described in the methods section.
Each unit was identified by an EDA/CDA
code and an estuary name.
The final database used in cluster analysis for
classification of estuarine coastal systems
contained 15 physical and hydrologic
parameters for each of 203 classification units
(Appendix A-1.2). The physical and hydrologic
database was not complete. Every parameter
had at least one missing value except for the
area of the estuarine drainage area (km2). The
frequency of missing values is indicated in
Table 6. Imputation procedures were used so
that the final database contained no missing
values.
Cluster Analysis of Estuarine Systems
In agglomerative hierarchical cluster analysis
using Euclidean distance, individual units with
the lowest dissimilarity are joined first. Size of
the estuarine drainage area and estuary area
were the primary variables contributing to the
separation of clusters, although volume, flow,
depth and salinity contributed as well. Com-
parisons of the means between clusters were
used to describe the properties of the clusters.
We grouped the means of primary variables
into large, medium and small according to the
values shown in Table 7. If the EDA area and
percent of the EDA that is defined as estuary
fell in the same size group, the label simply in-
dicates large, medium or
small as the first descriptor. If, however, those
two variables fell in different groups, the label
indicates both (e.g., Large EDA / Small % Es-
tuary). The labels are intended to be generally
descriptive of the properties of the estuaries
within each cluster. The values in Table 7 rep-
resent the average values for the physical and
hydrological variables for the estuaries within
each cluster. This does not mean that every
estuary in a cluster falls within the values in Ta-
ble 7. Eleven clusters and the number of
EDA/CDAs in each cluster are listed in Table
8. Finally, the clusters were mapped on the
geographic coverage of ED As (Figure 7) and
the estuaries within each cluster were listed in
Appendix Dl.
Some of the classes showed clearly distinct
characteristics. The last class to be identified by
the cluster analysis included the Chesapeake Bay
Mainstem, Potomac River, the tidal portion of
the Mississippi River, and the Columbia River.
This large, river-dominated class had the highest
average river flow, lowest average salinity, and
largest watershed areas when compared with all
other classes. The second to last class included
Long Island Sound, Cape Cod Bay, Puget
Sound and San Pedro Channel Islands. On av-
erage, this class had the largest estuarine area,
largest volume, and was the deepest of all other
classes. The class containing the smallest estu-
aries with the shallowest depths included pri-
marily estuaries and sub-estuaries located in
California and the Mid-Atlantic (e.g., Chester
River, Elk/Sassafras Rivers, Tijuana Estuary,
Waquoit Bay, Mission Bay, Morro Bay). The
first two estuaries to join in the cluster analysis
(i.e., most similar) were in this class: Morro Bay
and Anaheim Bay. It is important to reiterate
that not every estuary fits the general character-
istics of the class. Cluster analysis identifies pat-
-------
terns and similarities, the interpretation of
which is often subjective. In Stage II of this
classification effort, we will enhance this classi-
fication by using an improved database and ap-
plying the conceptual models described earlier.
Table 6. Frequency of missing values for physical and hydrologic parameters.
Parameters
Area of Estuary (km2)
Area of Estuarine Drainage Area (km2)
Mixing Zone Surface Area (km2)
Seawater Zone Surface Area (km2)
Tidal Freshwater Zone Surface Area (km2)
Average Tide Height (m)
Average Monthly River Flow (m3/day)
Maximum Monthly River Flow (m3/day)
Estuarine Volume (109 m3)
Tidal Prism Volume (109 m3)
Average Bottom Salinity (ppt)
Average Surface Salinity (ppt)
Average Depth (m)
Dissolved Concentration Potential of Pol-
lutant (DCP in mg/L)
Time for Freshwater to Displace Entire
Volume of Estuary (PRE)
Missing Value
Frequency
6
0
57
57
57
24
15
27
20
12
36
36
19
47
31
s
Percent
3%
0%
28%
28%
28%
12%
7%
13%
10%
6%
18%
18%
9%
23%
15%
32
-------
Table 7. Ranges of values for classification variables, used to describe clusters.
Variable Large/High/Deep Medium Small/Low/Shallow
„„. , , 9X >6000 2000-6000 < 2000
EDA Area (km2)
% Estuary > 30 10-30 < 10
Estuary Volume (109 m3) > 20 2-20 < 2
River Flow (m3/day) > 100 25-100 10-25 < 10
Depth (m) > 10 5-10 < 5
Salinity > 25 10-25 < 10
33
-------
Great Lakes
The final geographic coverage for Great Lakes
coastal riverine wetlands and associated water-
sheds included 155 (55%) of the 283 total sys-
tems chosen through a probability-weighted
survey design process, and distributed among
all of the Great Lakes and connecting channels
except for Lake Ontario. The database of wa-
tershed characteristics, hydrologic variables, and
associated CART-derived thresholds for hydro-
logic regimes in coastal watersheds includes
data from all Great Lakes states and from all
coastal marine states in the conterminous U.S.,
with the exception of Virginia and Alabama.
The Great Lakes coastal wetlands and water-
sheds database is complete for all base data:
watershed area, wetland area, land cover, soils,
and climatic variables, and for derived flashi-
ness indicators. To date, hydrologic thresholds
have only been derived and tested for Lake
Michigan coastal watersheds (n = 55).
The database of hydrologic variables and drain-
age basin characteristics for gauged watersheds
in coastal and Great Lakes states is currently
inconsistent with respect to variables included,
because data were derived from reports pro-
duced by different state USGS offices and state
agencies (USGS, 2003f) which derived variables
using a variety of manual versus digital tech-
niques. All equations predicting peak discharge
contained a term for watershed area or contrib-
uting drainage area. Of the remaining catego-
ries of variables representing natural features,
channel or basin slope (32 % of cases) and an-
nual precipitation, snowfall or precipitation in-
tensity (29% of cases) were most often
significantly related to peak flows. Variables
related to depressional storage or soil and un-
derground storage were also frequently in-
cluded (32 % of cases total), with inclusion of
depressional storage being more common.
Given the historic loss of wetlands throughout
the U.S., variation in surficial storage has both
natural and anthropogenic components. Land
cover was rarely found to be a significant factor
affecting peak runoff in rural watershed (10%
of cases), but some measure of urbanization or
impervious surface area was found to be a sig-
nificant predictor of peak flows for all urban
studies and for one statewide study (Table 9).
Derivation of Hydrologic Regime Classes for Water-
sheds in Coastal and Great Lake States
CART analysis successfully discriminated
among classes of watersheds by state or by state
and hydrologic region combinations based on
flow responsiveness. Mean 2-year peak flow
per unit watershed area differed among classes
by 3-4 orders of magnitude within each region
of the country (Great Lakes, Atlantic coast,
Gulf coast, Pacific coast), although variation
was somewhat less for the Atlantic and Pacific
coastal states (Figures 9 a-d). Categories of
variables identified through CART analysis as
the best discriminators among watershed peak
flow classes normalized for drainage basin area
(Q2/A) are summarized for all coastal and
Great Lakes states in Table 10, and are listed in
complete form in Appendix C-l.l. Average
percent reduction in error produced by CART
analyses was 55% (range = 10 — 83%) for
analyses conducted on all rural watersheds
within individual states, 58% (range = 16 -
94%) for analyses conducted on rural water-
sheds within hydrologic regions by state, and
58% (37 - 76%) for analyses conducted on ur-
ban watersheds. For both rural and urban wa-
tershed analyses combined, the most frequent
predictor variables identified were slope (54%),
basin shape or lag time (40%) and depressional
or soil storage variables (30%). For analyses
conducted with data from entire states, ignoring
hydrologic regions, storage was less commonly
identified as an explanatory variable (reduction
from 30% to 21%). However, for rural water-
sheds analyzed by hydrologic region within
coastal and Great Lake states, depressional and
soil storage variables were retained as signifi-
cant predictors in 50% of cases.
34
-------
Table 8. Estuarine classes resulting from cluster analysis of physical and hydrologic variables.
Estuarine Class
Number of
ED A/CD As in
Class
Large Area, Very High Flow, Shallow, Low Salinity
Large Area, High Volume, Deep, High Salinity
Small EDA/Large % Estuary, Low Volume, Low Flow, High Salinity
Medium EDA/Small % Estuary, Low Volume, High Flow, Low Salinity
Medium EDA/Small % Estuary, Low Volume, Low Flow, High Salinity
Medium Area, Low Volume, Shallow, Mixed Salinity
Medium Area and Volume, High Salinity
Large Area, High Flow, Shallow, Mixed Salinity
Medium EDA/Small % Estuary, Low Volume, High Flow, Mixed Salinity
Large EDA/Small % Estuary, Low Volume
Small Area, Low Volume, Low Flow, Shallow, Mixed Salinity
9
16
2
6
2
37
37
23
24
23
24
Table 9. Frequency of variable inclusion by state and hydrologic region in equations predicting peak discharge
for coastal and Great Lakes states.
Water- Slope Storage - Soil Imper- Forest Precipi- Runoff Temp- Eleva- Shape
shed surficial storage vious tation erature tion
area surface
RURAL 28 17
CASES
(n=28)
URBAN 12 2
CASES
(n=13)
TOTAL 40 19
CASES
(n=41)
area
11 614 10 2054
20 12 020000
13 6 13 4 12 2 0 5 4
35
-------
Derivation of model-based flow classes for
Great Lakes coastal wetland watersheds
demonstrated a wide gradient of flow regimes,
from those predicted to be stable (peak 2-year
flood volume: watershed depressional storage
volume < 1) to those predicted to be extremely
flashy (ratio > 10; Figure 10).
Initial Testing of Classification Frame-
works
To date, watershed hydrologic regime classes
have only been derived and tested for Lake
Michigan coastal riverine wetlands. Overall,
hydrologic regime classes were more successful
than either nutrient ecozones (Robertson et al.,
2001) or Omernik's nutrient ecoregions
(Omernik et al., 2002) in explaining differences
in sensitivity of water quality and biological re-
sponse to land cover gradients and nutrient
concentrations. With minor exceptions, classi-
fication of Lake Michigan coastal riverine wet-
lands by either Robertson's nutrient ecozones
or Omernik's nutrient ecoregions showed no
significant differences in reference condition (y-
intercept) or response (slope) in regressions of
surface water nutrient concentration versus
fraction watershed developed.
Hydrologic Thresholds for Subwatersheds in Lake
Michigan Basin
Hydrologic thresholds for peak flows were
similar conceptually for Michigan and Wiscon-
sin, based on the product of indicators for soil
permeability and watershed storage, although
the exact variables included in analyses differed
among states (Detenbeck et al., 2003b). Hydro-
logic thresholds for Michigan Hydrologic Re-
gions 1-3 were all based on the product of
fraction coarse substrate (outwash + coarse gla-
cial till) with fraction channel storage (fraction
main channel in wetlands), although the posi-
tion of the threshold was higher for Region 2
than for Regions 1 and 3.
The hydrologic threshold for Wisconsin Hydro-
logic Region 4 was similar in construct to those
developed for Michigan, but was based on a
product of soil permeability and watershed
storage (percent watershed area covered by
lakes and wetlands).
Relationship between WatershedFlashiness Indicators
Derived for Lake Michigan Basin
The significance of relationships between em-
pirical hydrologic thresholds and model-based
flashiness indicators was not guaranteed, be-
cause model-based indicators explicitly included
the interaction of land cover with precipitation,
soils and storage attributes in modifying peak
flows, whereas empirical hydrological thresh-
olds did not. Model-based flashiness indicators
for rain- versus snowmelt-based events are
strongly correlated for Lake Michigan coastal
wetland watersheds (r2 = 0.90, p < 0.0001). In
addition, the magnitude of the model-based
flashiness indicator for rain events was signifi-
cantly higher for flashy watersheds as compared
to stable watershed classes, as assigned by the
empirical models for peak flows (Kruskal-
Wallis test, p = 0.001; (Detenbeck et al., 2003).
Hydrologic Thresholds and Flashiness Indices as Predic-
tors of Water Quality andBiotic Condition in Lake
Michigan Coastal Riverine Wetlands
With the exception of nitrate, log-transformed
surface water concentrations of nutrients, sus-
pended solids, and turbidity in Lake Michigan
coastal riverine wetlands were significantly cor-
related with the model-based index of water-
shed flashiness for rainfall events. For total
phosphorus, total nitrogen, turbidity, and total
suspended solids in coastal wetlands, the rela-
tionship between fraction of watershed devel-
oped and surface water concentrations differed
significantly between stable and flashy water-
shed classes, with the slope of the relationship
being lower for flashy watersheds (p < 0.05).
For soluble reactive phosphorus, ammonium,
and nitrate + nitrite-nitrogen, surface water
concentrations increased significantly with frac-
tion watershed developed, but neither slope nor
36
-------
y-intercept varied among stable versus flashy
watershed classes (p > 0.05; Detenbeck et al.,
2003).
Midsummer temperature of Lake Michigan wet-
land tributaries and of coastal riverine wetlands
differed significantly among hydrologic classes,
and no additional variability could be explained
by including wetland latitude in ANCOVAs.
Midsummer temperatures were 4-5 degrees
(EC) lower in tributaries and wetlands associ-
ated with stable hydrologic regimes as com-
pared to flashy hydrologic regimes (Detenbeck
etal.,2003b).
Fish IBI scores in Lake Michigan coastal river-
ine wetlands differed significantly among stable
versus flashy watershed classes, but did not re-
spond linearly to fraction watershed developed,
either within watershed classes or in the com-
bined data set (p > 0.05). Fish IBI scores also
were significantly related to the model-based
index of flashiness for rainfall events (p =
0.006). In contrast, plant IBI scores for 2001
sites decreased significantly with fraction water-
shed developed, but neither the slope nor y-
intercept for the relationship differed among
watershed classes (Detenbeck et al., 2003b).
Both phytoplankton chlorophyll a and periphy-
ton chlorophyll (mg chl a' cnr2) increased ex-
ponentially as a function of surface water total
P in Lake Michigan coastal riverine wetlands.
For phytoplankton, responses differed by hy-
drologic class, with higher chlorophyll a levels
at low total phosphorus for wetlands with sta-
ble hydrologic regimes as compared to those
with flashy hydrologic regimes. Periphyton
composition also changed along a watershed
development gradient, with increases in green
and blue-green algae, and concomitant de-
creases in diatoms, occurring at lower levels of
development for coastal wetlands with stable
hydrologic regimes (Detenbeck et al., 2003b).
37
-------
I I US State Boundary
| | 1 - Large. Medium Volume. Very High Flow. Shallow. Low Salinity
| | 2 - Large. Large Volume. Medium Flow, Deep. High Salinity
| | 3 - Small EDA/Large % Estuary, Small Volume. Low Flow. Deep, High Salinity
| | 4 - Medium EDA/Small % Estuary. Small Volume. High Flow. Shallow. Low Salinity
] 5 - Medium EDA/Small % Estuary, Small Volume, Low Flow. Shallow. High Salinity
| 6 - Medium. Small Volume, Low Flow, Shallow. High Salinity
| | 7 - Medium EDA/Large % Estuary, Medium Volume. Low Flow. Medium Depth. High Salinity
3 6- Large EDA/Medium % Estuary, Medium Volume, High Flow. Shallow. Mixed Salinity
| 9 - Medium EDA/Small % Estuary. Small Volume. Medium Flow. Shallow. Mixed Salinity
| 10 - Large EDA/Small % Estuary, Small Volume, Medium Flow. Shallow. Low Salinity
| 11 - Small EDA/Medium % Estuary. Small Volume. Low Flow. Shallow. Mixed Salinity
Figure 7. Estuarine classes resulting from cluster analysis of physical and hydrological variables.
-------
Table 10. Categories of variables identified through CART analysis by state, region, and urban area that best discriminate among area-normalized
peak flow classes for coastal and Great Lake states.
Analysis for full state, Analysis by region,
rural watersheds rural watersheds
Analysis by urban area
Total
Watershed area
Slope
Storage — surficial
Soil storage
Impervious surface area
Forest
Precipitation
Runoff/ Evap
Temperature
Elevation
Shape/ Lag time
Region
Potential total
Cases
4
14
4
2
1
2
5
2
1
2
9
3
29
Percent of
total
14%
48%
14%
7%
3%
7%
17%
7%
3%
7%
31%
10%
Cases
4
15
8
2
1
6
3
3
2
4
10
0
20
Percent of
total
20%
75%
40%
10%
5%
30%
15%
15%
10%
20%
50%
0%
Cases
0
2
0
1
4
0
1
0
0
0
4
0
8
Percent of total
0%
25%
0%
13%
50%
0%
13%
0%
0%
0%
50%
0%
Cases
8
31
12
5
6
8
9
5
3
6
23
o
J
57
Percent
of total
14%
54%
21%
9%
11%
14%
16%
9%
5%
11%
40%
5%
39
-------
(* 2.33-year for PA, 5-year for Ml)
i T i T t
ilii jiiiiniiis
160
120
80
40
~0
D)
T T
30
20
10
Q X O O
4 • » t T +
State-Class
100
80
60
40
20
i !
20
j i i
State-Class
10
E>
o
.«
Figure 8. Peak 2-year flow classes identified through CART analysis of data from USGS gauging stations. Top left: Great Lakes states, top right: At-
lantic coastal states, bottom left: Gulf states, bottom right: Pacific coast states.
40
-------
\
Map Legend
Great Lakes Shoreline
| 'j State Boundaries
REMAP Wateisheds
Flow Responsiveness Index
^B 0.04-0.50
^BM.SI -1 .06
| |1.07-1 .78
[ 11.79-3.01
I I 3.02-4.80
| | 4.31 -6.77
F l6.7S-16.76
^B 16.77 -28.50
^B 28.51 -98.68
^•58.69-131.30
Figure 9. Flow responsiveness watershed index (peak 2-year flood volume and watershed depressional storage volume) for watersheds associated with
Great Lakes coastal riverine wetlands.
41
-------
Toxic Units - Metals
Gulf Coast
Figure 10. Average total toxic equivalent units for PAHs (top) and metals (bottom) within
estuarine sediments of Gulf of Mexico, color-coded for associated EDAs and CD As.
42
-------
IV. STAGE II PLANS
We will build upon the Stage I classifica-
tion framework by improving geo-
graphic coverage, reassessing scale issues, fill-
ing in missing values and missing parameters
in the database, adding data sources on coastal
condition, and evaluating different approaches
for testing the coastal classification frame-
work. Specific examples of these improve-
ments are included in the following text. In
addition, we will improve consistency between
estuarine and Great Lakes databases by in-
cluding more and better estimates of physical
and hydrological variables for Great Lakes
coastal wetlands. We will explore the implica-
tions of using average retention time in estuar-
ies versus examining temporal patterns in
retention time. Finally we will supplement the
Stage I database with additional modifying
factors and indicators of key ecosystem proc-
esses.
Geographic Coverage
Extent
Currently, the coastal estuarine classifica-
tion database covers the full set of estuar-
ies in the conterminous United States as de-
fined in CA&DS. In the future, we will ex-
pand the geographic extent of the database to
include systems in Alaska and Hawaii. In the
first version of the database, watershed char-
acteristics were included only for ED As and
CDAs, i.e., those 8-digit HUCs immediately
upstream of ED As. In the future, we will ex-
pand the database to include upstream fluvial
drainage areas that are hydrologically con-
nected to estuarine watersheds.
The Great Lakes coastal riverine wetland and
watershed database currently includes only
those coastal riverine wetlands sampled as
part of the 2000-2001 EPA Region 5 R-
EMAP project on coastal Great Lakes wet-
lands (Simon et al., 2003), and specifically ex-
cludes Lake Ontario as it is outside of EPA
Region 5. Future versions of the classification
database will include all of the Great Lakes
and connecting channels. Contingent upon
GIS resource support, we will delineate water-
sheds for remaining coastal riverine wetlands
using state watershed boundaries and the Na-
tional Watershed Boundary Database as a
starting point, and refined using watershed
delineation tools developed through an inter-
agency agreement with USGS (2002). An al-
ternative, but less desirable option, would be
to characterize the full set of "reachsheds"
defined by the Natural Resources Research
Institute under EPA Science to Achieve Re-
sults (STAR) grant to the Great Lakes Envi-
ronmental Indicators (GLEI) project
(University of Minnesota, 2003). Investiga-
tors have defined the watersheds feeding
coastal reaches (reachsheds) for the entire
Great Lakes shoreline using endpoints defin-
ing lengths of the shoreline, i.e., reaches iden-
tified in NOAA's medium resolution
shoreline vector database (NOAA, 2003b).
Scale of Units
The most recent guidance from EPA Office
of Water indicated that the proper scale for
TMDLs was an important issue that should be
addressed (USEPA, 2003c). The level of spa-
tial resolution for marine ED As is adequate in
some regions for addressing TMDL issues but
not in others. Using a combination of local
knowledge, coastal states' definition of report-
ing units for 305(b) reports and 303(d) listing
(shapefile coverages available from EPA,
43
-------
20021) and shoreline reaches defined in
NOAA's medium resolution shoreline vector
coverage (NOAA, 2003b), we will examine
the boundaries of ED As to determine areas
needing improved spatial resolution. Al-
though the level of spatial resolution in cur-
rent ED As might be adequate for TMDL
purposes, different stressors might be better
detected at different scales (Edgar and Barrett,
2002; Momsey et al., 1992). Finer-scale stud-
ies should also allow investigation of variabil-
ity within estuaries.
Parameter Improvements
Loadings
The Stage I classification database lacks quan-
titative measures of suspended solids loadings,
nutrient loading data of appropriate spatial
resolution, and complete geographic coverage
for toxics exposure data for all of the Great
Lakes and marine coastal states. Refinements
to the Stage II database will address these is-
sues. We will obtain improved estimates of
suspended sediment loadings by calculating
actual loads from USGS databases (USGS,
2003d). For ED As with complete data, we
will compare results with the EPA IWI index
of potential sediment loading from the USDA
HUMUS database to determine if we can use
these data to calibrate the IWI relative ranking
index (USEPA, 2003i; TAES, 2000). An addi-
tional data source for sediment yield data may
include the Soil and Water Assessment Tool
(USDA, 2002; SWAT). SWAT is a river basin
scale model developed to quantify the impact
of land management practices in large, com-
plex watersheds and can now be readily pa-
rameterized in a GIS using AGWA (Miller et
al. 2002).
We will obtain improved estimates of nutrient
loadings from the next version of the
SPARROW model through direct collabora-
tion with the USGS. Version 2 of the na-
tionwide SPARROW database has been
improved by developing continuous estimates
of nutrient loading by stream reach rather
than just at the outlets of 8-digit HUCs, and
will include inputs from upstream HUCs (R.
Alexander, U.S. Geological Survey, personal
communication).
The current database only includes toxicity
loadings estimates from EPA's PCS database
and sediment toxics data from the NSI and
STORET for marine estuaries. In future ver-
sions of the database, we will expand nutrient,
sediment, and toxics loading and exposure
data to include Great Lakes coastal riverine
wetlands. In cooperation with EPA Office of
Water, we will screen the PCS database for
outliers to improve the quality of toxics load-
ing estimates.
Retention Time Estimates
The current version of the marine EDA clas-
sification database includes estimates of reten-
tion time and the related parameters, particle
retention efficiency (PRE) and dissolved con-
centration potential (DCP), which were calcu-
lated using readily available salinity data from
EMAP and average values for freshwater dis-
charge. We will improve the spatial and tem-
poral resolution of the salinity portion of the
database by requesting data from coastal state
monitoring programs. Using data from well-
characterized systems, we will assess the de-
gree of error inherent in calculating retention
time from standard comprehensive data
sources like EMAP.
The current database includes values
for average tidal volume derived from
CA&DS. In future work, we will examine the
effects of temporal variation in tidal volumes
on residence time. We will also assess tempo-
ral variability in retention time indirectly
through classification of systems by hydrocli-
matic region (Saco and Kumar, 2000) and hy-
drologic regime (naturally stable or regulated
vs. flashy; (Detenbeck et al., 2000). We will
improve the USGS database supporting deri-
vation of hydrologic regime classes by stan-
44
-------
dardizing the parameters included and region-
alization schemes across state boundaries.
We will improve hydrological and physical
databases for ED As by filling in missing data
where possible from NOAA's coastal mor-
phometry databases. We can estimate values
for missing hydrologic data using regional
equations for prediction of flow (Koltun and
Whitehead, 2001). We will improve hydro-
logical and physical databases for Great Lakes
coastal wetlands by estimating average dis-
charge for ungauged systems, using regional
equations (e.g., (Koltun and Whitehead, 2001)
and estimates included in the EPA Reach File
1 (RF1) database (USEPA, 2003h). We will
refine area estimates for Great Lakes coastal
riverine wetlands to distinguish among differ-
ent wetland cover classes (e.g., open water vs.
emergent vs. forested). We will examine the
feasibility for assessing coastal wetland vol-
umes by digitizing changes in wetland open
water boundaries for wet versus dry years
(with associated known high and low lake lev-
els). We will assess improvements in coastal
bathymetry based on data from Light Detec-
tion and Ranging (LIDAR) initiatives coordi-
nated by the Federal Emergency Management
Agency (FEMA) as another source of fine
resolution digital elevation data for direct cal-
culation of wetland volume using GIS.
We will record hydrologic regime types for all
coastal tributaries as defined by Saco and
Kumar (2000) through spectral analysis of
discharge time series. Saco and Kumar (2000)
defined three distinct hydrologic regime types:
1) a long seasonal (LS) mode, 2) a short sea-
sonal (SS) mode, and 3) a high small-scale
(HSS) mode. The LS mode is characterized
by a seasonal cycle of streamflow associated
with either sustained or frequent above aver-
age flow conditions across several months.
The SS mode is similar to the LS mode, but
above average flow conditions occur over a
period of only 2-3 months, with higher peaks
of short duration overriding them. The HSS
mode is associated with very high variability at
timescales of 6 days to 1 month.
Modifying Factors
The Stage I database contains data on modify-
ing factors that were readily available from
EMAP, STORET, and the National Sediment
Inventory.
Data on some modifying factors such as acid
volatile sulfide (AVS) were incomplete. Data
on other modifying factors were available but
require further analysis before they can be
used. We will expand modifying factors in-
cluded in the coastal classification database to
include:
improved estimates of suspended
sediment concentrations. The USGS
has compiled data on mineral and or-
ganic suspended sediments from their
monitoring programs into a single
database (USGS, 2003d).
values of dissolved organic carbon.
DOC in the water column interacts
with suspended sediments by influenc-
ing light penetration, and influences
partitioning of organic toxins and the
effect they exert.
estimates of photic depth. The
Gallegos model allows derivation of
predictive relationships for extinction
coefficients based on dissolved or-
ganic carbon, chlorophyll a, and total
suspended solids data from EMAP
(Gallegos, 2001).
Morphometric interactions with the
photic zone. We can calculate change
in % estuarine bottom within the
photic zone as a function of increased
suspended solids or chlorophyll a.
AVS predictions. We will assess the
feasibility of predicting AVS from
readily available data (e.g., organic
45
-------
carbon, redox potential, Fe, particle
size) based on systems with complete
data sets.
=> aluminum: heavy metal ratios. These
ratios can be used to correct for natu-
ral background in metals content.
=> energy regime. NOAA's Environ-
mental Sensitivity Index for coastal
systems contains indicators of the en-
ergy regime (NOAA, 2003c).
System Processing Capacity
Processing capacity is determined by the nor-
mal cycle of interactions processing materials
in natural systems; generally, rate functions
that are driven or limited by internal or exter-
nal modifying factors, e.g., denitrification,
carbon and nitrogen fixation, primary produc-
tion, and grazing. The current version of the
coastal classification database does not include
any direct measurements or indicators of sys-
tem processing capacity, although most of the
data required to estimate denitrification po-
tential is available. In future iterations of the
database, we will include indicators of:
=> N processing potential. Dissolved in-
organic N [DIN] in the water column
of an aquatic system as the result of
the integration of total system proc-
esses (nitrification, denitrification,
sediment remineralization, etc.) = the
difference between the measured
[DIN] in the water column and the
conservative [DIN] expected in the
water column from riverine inputs.
The determination of these variables
will be system dependent and incorpo-
rates flushing time, volume, river flow,
riverine nutrient inputs, etc.
=> biological filtering capacity. Shellfish
bed area will be used as an indicator of
biological filtering capacity, based on
information in CA&DS.
=> coastal wetland extent. We will com-
bine NWI and state wetland inventory
data as necessary.
=> primary productivity potential. We
will explore whether productivity var-
ies systematically as a function of cli-
matic factors such as mean annual
temperature and seasonally (Phytoso-
ciological Research Center, 1995).
The processing capacity of estuaries for nutri-
ents is dependent upon a combination of
physical and biological factors. In situ concen-
trations of nutrients and toxics are indicators
of an estuary's ability to process contaminants
based on pollutant load, flushing factors, mix-
ing, and biogeochemical cycling. Initial classi-
fication resulting from cluster analysis
incorporated physical and hydrological pa-
rameters and DCP of pollutants based on a
standardized load. By comparing in situ nutri-
ent concentrations within an estuary to the
DCP calculated based on actual or estimated
load (SPARROW), we can evaluate the proc-
essing capacity of the estuary. If nutrient
concentrations measured in the water column
are below the DCP calculated for the estuary,
it can be assumed that the rate of removal due
to internal processes, whether biological or
physical, exceeds the rate of regeneration due
to internal processes. Conversely, if nutrient
concentrations exceed the DCP, then internal
regeneration exceeds removal by internal
processes. While the DCP calculation in-
cludes flushing, it does not account for inter-
nal or recycled nutrients. In order to compare
the processing capacity within and among es-
tuarine classes, we need to compare the in situ
concentrations to the DCP. Based on physi-
cal and hydrological data, we can assign three
subclasses to systems based on system re-
sponse to nutrient load: below capacity, at ca-
pacity, or above capacity. We will use case
studies to validate class assignments within
this scheme and to provide further informa-
46
-------
tion on the processes driving the system re-
sponse. These processing rates include, but
are not limited to, the following:
Nitrification
Denitrifi cation
Sediment Phosphorus Regeneration
Primary Production
Bacterial Production
Sediment Nutrient Flux Rates
Oxygen Metabolism (sediment and
water column oxygen demand)
Data for these processes are not available for
every unit; however, investigators have quanti-
fied these rates in several well-studied estua-
rine systems. By comparing process rates in
these systems between and among classes, we
could validate class designations based on the
DCP. Intensively studied systems for which
processing rate data are most likely available
include:
Class I Chesapeake Bay Mainstem,
Albermarle Sound
Class II Puget Sound, Long Island Sound
Class III Damariscotta River
Class IV Connecticut River, Klamath River
Class VI Florida Bay, Corpus Christi Bay
Class VII Buzzards Bay, San Francisco Bay
Class VIII Tampa Bay, Galveston Bay,
Pamlico Sound
Class IX Great Bay, Charleston Harbor
Class X Pensacola Bay, Neuse River
Class XI Waquoit Bay
Data associated with processes for many of
the estuaries can be obtained from the Na-
tional Estuaries Program (USEPA, 2003d;
NEP) or those associated with coastal Long-
Term Ecological Research (LTER) sites (Flor-
ida Coastal Everglades LTER, Georgia
Coastal Ecosystems LTER, and Plum Island
LTER).
Data Sources on Coastal System Condi-
tion
National Coastal .Assessment
EMAP has monitored and assessed the condi-
tion of coastal estuarine systems in the U.S.
since 1990. In addition to indicators of
stressor exposure and habitat condition, ben-
thic macroinvertebrate and fish community
data have been collected to determine biotic
integrity. Several benthic indices of condition
have been developed through EMAP NCA
for different biogeographic regions of the U.S:
Virginian Province (Paul et al., 2001; Weisberg
et al., 1993), Chesapeake Bay (Weisberg et al.,
1997), Carolinian Province (Van Dolah et al.,
1999), Gulf of Mexico (Engle and Summers,
1999). These multimetric indices combine
measures of abundance, species richness and
diversity, and relative abundance of sensitive
species to distinguish between reference and
degraded benthic communities. In the re-
gions of the U.S. for which a benthic condi-
tion index has not yet been developed (i.e., the
Pacific West Coast and the Northeast), meas-
ures of diversity were used to assess benthic
condition. The original benthic community
data and calculated indicators were available
from EMAP (USEPA, 2003b).
Great Lakes Regional Environmental Monitoring
and Assessment Program
Data on condition of 155 Great Lakes coastal
riverine wetlands were obtained from a EPA
Region V R-EMAP project (Simon et al.,
2003). For testing of Phase I of the classifica-
tion database, indices of biotic integrity (IBIs)
and associated metrics for vegetation, macro-
invertebrates, and fish communities were
available only for a subset of Lake Michigan
sites. Subsequent phases of the classification
database can be tested using the full dataset,
which will include fish IBIs and metrics for all
155 sites.
47
-------
Testing of Estuarine Classification System
Classification approaches can be applied ei-
ther a priori or posteriori, as discussed previ-
ously. We will explore additional approaches,
both empirical- and model-based.
Improvements in ^4 Priori Testing
The significance and robustness of classes
identified in Stage I through cluster analysis
will be tested using the nonparametric multi-
response permutation procedures (MRPP)
available in PC-ORD software (Mielke, 1984).
Discriminant function analysis will then be
applied to determine which watershed and
estuarine characteristics can be used to dis-
criminate among hydromorphological types.
After applying stepwise discriminant analysis
to narrow down the range of explanatory flow
or velocity metrics, we will use the selected
subset of metrics to define linear discriminant
functions, using PROG DISCRIM (SAS Insti-
tute, 1990). Classification error rates will be
estimated using the CROSSVALIDATE op-
tion.
Development of Model-leased Classes for Testing
Simple mechanistic models can be used in
conjunction with physically-based empirical
classes of estuaries or coastal riverine wetlands
to determine critical differences in behavior
among systems, based on predictions of
stressor levels or ecological assessment end-
points. For example, Stefan et al. (1996) has
used this approach to predict loss of habitat
volume in different physically-based classes of
lakes in response to climate change. In Stefan
et al.'s (1996) work, habitat volume was de-
scribed as a function of temperature and dis-
solved oxygen requirements for different
thermal guilds of fish.
In the next year, we will use our conceptual
models and the database for estuaries pre-
sented in this paper to develop and test
stressor-based classification systems. We will
apply simple canonical models of stressor ef-
fects and interactions to determine disconti-
nuities in stressor-response surfaces for
estuaries as a function of water-body retention
time, modifying factors, and processing capac-
ity (Campbell et al., 2003; Stefan et al., 1995;
Stefan et al., 1996). We illustrate our planned
approach in Figurell. The aquatic systems to
be classified include water bodies and their
watersheds. Any system defined in this way
can be classified, e.g., a stream reach and its
watershed, an estuary and its drainage area, a
lake and its watershed. The classes will be
based primarily on water body characteristics
and are stressor specific. Basic information
for the pollutant (stressor) will be determined
along with the loading rate from the adjacent
watershed, watersheds upstream, the atmos-
phere, and the ocean for estuaries. In addi-
tion, we will determine the stored quantity of
the pollutant presently residing in the aquatic
system. Implicit in the discussion that follows
is the assumption that all the information
needed for each step in the model-based clas-
sification will be present in the database. In
the next year, we may have to augment the
existing database with needed information.
We will first apply the classification process
for a unit load of pollutant and predict the
expected biologically effective concentrations
for different classes of aquatic systems. The
first step in classification is to place the system
to be classified into one of the four canonical
models controlling residence time (Figure 11).
Once this is accomplished, we will divide the
systems into ranges of average temperature
and into one of two classes (Continuous or
discontinuous) based on the way materials are
processed. Thus, we will distinguish between
temperate and tropical systems at this step.
Temperature determines the rate of metabolic
processing of the pollutant processing above,
we will determine the average residence time
for the system and the relevant range of tem-
poral variation. If residence time varies mark-
edly over the area of the system under study,
48
-------
we will divide the system into subsystems and
analyze each subsystem separately. We will
separate systems into residence time classes,
which will be determined based on the
chronic dose-response characteristics of the
particular stressor. Knowing the variation of
turnover with time will allow us to partition a
system into more than one residence time
class, if necessary.
Once we have divided systems into classes
based upon residence time, we will split
classes again using factors that control proc-
essing capacity, e.g., the ratio of wetland to
water body area, dissolved organic carbon, or
AVS. Once again, we could use two or more
classes based on the processing capacity of the
wetland for the various pollutants. We hy-
pothesize that the presence or absence of wet-
lands will be the factor of greatest importance
after temperature in processing pollutants.
Wetlands are both a response variable an indi-
cator of processing capacity. Initially, the dis-
tribution of wetlands is determined by natural
factors. However, wetlands can be lost
through direct physical stressors (dredging,
fill) as well as through indirect stressors (eu-
trophication) which hamper the growth of
submerged aquatic vegetation. The loss of
vegetation is expected to create a feedback
effect, further limiting the retention of sedi-
ments and nutrients within an estuarine sys-
tem.
We will consider other processing factors at
this stage based on the particular pollutant
being evaluated. At this point, we will esti-
mate the bioeffective concentration expected
in the class and multiply it by the residence
time to determine an exposure. We will con-
struct the expected exposure-effect relation-
ship for the pollutant from past studies in the
literature and predict the effects on biological
output variables from the exposures deter-
mined for each class.
Next, we will consider the effects of modify-
ing factors to determine the alterations in the
biological impacts expected in particular sys-
tems. We will group modifying factors ac-
cording to their effects on the pollutant. We
will combine those that have a positive (de-
creased response effect) and those that have a
negative (increased response effect) to esti-
mate the net effect on the biological response.
Once we have determined a positive or nega-
tive effect then we will apply it to the expo-
sure calculated above to determine an
effective exposure in the system containing
modifying factors. If not modifying factors
are present, the exposure value determined
above is the effective exposure and it passes
directly to the bottom line (Figure 11). We
hypothesize that effective exposure will char-
acterize sets of aquatic systems where similar
biological effects will be observed.
Next we will apply the actual loads entering
the aquatic systems and determine the effec-
tive exposures. We will plot the observed
values for the biological output variables from
the aquatic systems against the effective expo-
sures to construct an exposure-effect curve
for the pollutant. We will compare this rela-
tionship to the one expected from past stud-
ies. We can expect system classes to plot as a
family of curves on the exposure-effect plane
or as a single curve on the effective exposure-
effects plane. Managers could allow greater
loading in a class of aquatic systems that is
less sensitive to the pollutant to attain a given
level of effect deemed acceptable.
Approaches for A. Posteriori Development of Classes
We will derive water-body classes empirically
both through indirect and factor-based meth-
ods, using cluster analysis of water-body and
watershed characteristics, and through direct
and response-based approaches, using Bayes-
ian approaches to determine natural break-
points in assessment endpoints as a function
of stressor gradients and classification factors
(Breiman et al., 1984; Kass, 1980).
Indirect classification procedures such as clus-
ter analysis use information on the variation
49
-------
of potential classification variables among
coastal watersheds and wetlands. In contrast,
response-based classification procedures use
information on both independent variables
(classification factors, stressor indicators) and
dependent variables (ecological assessment
endpoints such as indices of biotic integrity).
Procedures that can be used to empirically
discriminate differences among classes in re-
sponse of ecological endpoints along stressor
gradients include Bayesian techniques such as
CART (Breiman et al., 1984) and Chi-square
Automatic Interaction Detector (CHAID, see
(Kass, 1980). CART produces a binary tree in
which a response variable is sequentially sepa-
rated into two classes, using either categorical
variables or breakpoints for continuous vari-
ables. CHAID is analogous to CART, but
extends the procedure to multiple classes at
each level of the tree. Unlike parametric pro-
cedures such as canonical correlation analysis,
CART and CHAID do not require assump-
tions of normality, homoscedasticity, or addi-
tivity of effects. The techniques are ideally
suited for teasing out interactions among fac-
tors, e.g., the interaction between stressor or
exposure gradients and watershed and water
body classes. Interaction terms between cate-
gorical and continuous variables also can be
explicitly included in a model to tease out dif-
ferences between main effects and interaction
terms, analogous to what is done in an analy-
sis of variance (Statsoft, Inc., 2003).
Spatio-temporal classification
Classification approaches can be applied to
distinguish among system behaviors either
based on spatial differences among systems at
one point in time, or among system behaviors
over climatic cycles. We will explore a spatio-
temporal classification approach, which de-
fines spatial aggregations of watershed units
based on similarities in system hydrology
across climatic cycles (Saco and Kumar, 2000).
Climate change can determine the magnitude
and timing of freshwater flow and nutrient
delivery to coastal systems from rivers, as well
as directly affecting marine and freshwater
organisms through alterations in salinity and
temperature (Chang et al., 2001; Drinkwater et
al., 2003; Staile et al., 2003). Coastal systems,
at the interface of fresh and salt water, can be
expected to be especially vulnerable. Climatic
cycles driven by El Nino, El Nino-Southern
Oscillation, and the North Atlantic Oscillation
(NAO) can affect different regions of the
coast differently during the same time period
(Cayan et al., 1998; Dettmger et al., 1998;
Walker et al., 2002). Hydroclimatic temporal
regimes defined by Saco and Kumar (2000)
will be tested as part of an a priori classifica-
tion scheme. Rather than focusing on differ-
ences in hydrologic response within a region
of homogeneous climate, hydroclimatic re-
gimes take into account differences among
regions in hydrologic response over time re-
lated to atmospheric circulation patterns.
Based on spectral analysis of long-term dis-
charge records from U.S. coastal segments,
Saco and Kumar (2000) identified three
classes of temporal regimes based on strength
of seasonality and frequency of high flows.
Once watersheds have been separated into
hydroclimatic regimes, and the atmospheric
forcing functions identified for each region,
response data can also be categorized by posi-
tion along climatic cycles, using indicators
such as the NAO index, or Palmer's drought
index applied at a regional scale (Cayan et al.,
1998; Dettmger et al., 1998; Walker et al.,
2002).
50
-------
A Classification Scheme Starting with the Canonical Models
Freshwater well-mixed Freshwater stratified Estuary, well-mixed
Estuary, stratified
Temperature
Residence Time
] Processing
Exp 0 S UT e (Bioeffectnr* cone.
Modifying Factors
EEX EE7
EE Effective Exposure
Figure 11. A classification tree to group estuaries by effective exposure regimes based on our conceptual model of the controlling factors.
time)
51
-------
V. ACKNOWLEDGMENTS
We gratefully acknowledge the following indi-
viduals for their open discussions of ongoing
work on the principles and applications of
classification within different organizations:
Suzanne Bricker (NOAA) — for discussion of
NOAA CA&DS database and ongoing classi-
fication efforts, Mary Lammert (The Nature
Conservancy) — for discussion of TNC classi-
fication efforts; Dave Flemer (EPA Office of
Water) — for discussion of classification issues
relative to nutrient criteria for estuaries; and
Mary Moffett (MED) for discussion of classi-
fication of Great Lakes coastal wetlands. Su-
perb CIS (GED, AED, MED) and data
entry and manipulation support (MED) were
provided by the USGS Gulf Breeze Project
Office at GED (GIS: Pete Bourgeois; graph-
ics: Renee Conner), and by Computer Sci-
ences Corporation under the FAIR II
Contract at AED (GIS: Doug McGovern,
Jane Copeland; database extraction: Melissa
Hughes) and MED-Duluth (Matthew Starry,
Roger Meyers, Benjamin Bertsch, and James
Quinn, GIS; Dianne Spehar and Susan Mat-
tis-Turner, data scanning and entry). We
thank Peggy Rogers (NCBA) at GED for her
patient and diligent efforts in bibliographic
data entry and text formatting. Cover photo
by Renee Connor.
52
-------
VI. REFERENCES CITED
Alice, R. J., Dethier, M., Brown, D., Deegan, L., Ford, R. G., Hourigan, T. F., Maragos, J., Schoch,
C., Sealey, K., Twilley, R., Weinstein, M. P., and Yoklavich, M. (2000). Marine and Estuarine Eco-
system and Habitat Classification (Technical Memorandum NMFS-F/SPO-43). Silver Spring, MD:
National Oceanic and Atmospheric Administration, National Marine Fisheries Service.
Bailey, R. G. (1976). Ecoregions of the United States. Ogden, UT: U.S. Department of Agriculture,
Forest Service.
Baird, D. J., Brock, T. C. M., De Ruiter, P. C., Boxall, A. B. A., Gulp, J. M., Eldndge, P., Hommen,
U., Jak, R. G., Kidd, K. A.,, DeWitt, T. (2001). The Food Web Approach in the Environmental
Management of Toxic Substances. (Technical Memorandum NMFS-F/SPO-43). Pensacola, FL:
Society of Environmental Toxicology and Chemistry (SETAC).
Bartley, J. D., Buddemeier, R. W.,, Bennett, D. A. (2001). Coastline Complexity: A Parameter for
Functional Classification of Coastal Environments. Journal of SEA Research, 46, 87-89.
Battaglin, W. A.,, Goolsby, D. A. (1994). Spatial data in GIS format on agricultural chemical use,
land use, and cropping practices in the United States. U.S. Geological Survey Water-Resources In-
vestigations Report 94-4176. Retrieved on October 3, 2003 from
http://water.usgs.gov/pubs/bat/batOOO.html
Beck, M. W.,, Odaya, M. (2001). Ecoregional Planning in Marine Environments: Identifying Priority
Sites for Conservation in the Northern Gulf of Mexico. Aquatic Conservation: Marine and Fresh-
water Ecosystems, 11, 235-242.
Breiman, L., Friedman, J. FL, Olshen, R. A.,, Stone, C. J. (1984). Classification and Regression
Trees. Monterey, CA: Wadsworth, Brooks/Cole Advanced Books, Software.
Bricker, S. G., Clement, C. G., Pirhalla, S. P., Orlando, S. P.,, Farrow, D. R. G. (1999). National Es-
tuarine Eutrophication Assessment: Effects of Nutrient Enrichment in the Nation's Estuaries. Sil-
ver Spring, MD: National Oceanic and Atmospheric Administration.
Briggs, J. C. (1974). Marine Zoogeography. New York, NY: McGraw Hill Co. 475 pp.
Campbell, D. E. (2000). Using energy systems theory to define, measure, and interpret ecological
integrity and ecosystem health. Ecosystem Health, 6, 181-204.
Campbell, D. E., Hill, B., Detenbeck, N. E., Engle, V. D., Ho, K., Kurtz,]., Pelletier, M. C., Bur-
gess, R. M., Perez, K. T.,, Snarski, V. M. (2003). Conceptual Models and Methods for Diagnosing
the Causes of Impairment to Aquatic Ecosystems (EPA Report): National Health and Environ-
53
-------
mental Effects Research Laboratory, Gulf Ecology Division, Mid-Continent Ecology Division and
Atlantic Ecology Division.
Cayan, D. R., Dettmger, M. D., Diaz, H. F.,, Graham, N. E. (1998). North-South Precipitation Pat-
terns in Western North America on Interannual-to-Decadal Timescales. Journal of Climatology,
11(12), 3148-3166.
Chang, H., Evans, B. M.,, Easterling, D. R. (2001). The effects of climate change on stream flow
and nutrient loading. Journal of the American Water Resources Association, 37(4), 973-985.
Chow-Fraser, P.,, Albert, D. A. (1998). Identification of "Eco-reaches" of Great Lakes Coastal Wet-
lands That Have Biodiversity: A Discussion Paper for the State of the Lakes Ecosystem Conference.
Paper presented at the State of the Lakes Ecosystem Conference, Chicago, IL and Burlington, On-
tario, Canada. Retrieved on October 3, 2003 from
http://www.epa.gov/glnpo/solec/98/present/wetlandsbia/wetiandsbia.html.
Clausen, B.,, Biggs, B. J. F. (2000). Flow variables for ecological studies in temperate streams:
groupings based on covariance. Journal of Hydrology, 237, 184-197.
Cowardin, L. M., Carter, V., Golet, F. C.,, LaRoe, E. T. (1979). Classification of Wetlands and
Deepwater Habitats of the United States (FWS.OBS-79/31). Washington, DC: U.S. Fish and Wild-
life Service.
Curtis, G. W. (1987). Technique for estimating flood-peak discharges and frequencies on rural
streams in Illinois (Water-Resources Investigations Report 87-4207). Urbana, IL: U.S. Geological
Survey.
Daly, C., Neilson, R. P.,, Phillips, D. L. (1994). A statistical-topographic model for mapping clima-
tological precipitation over mountainous terrain. Journal of Applied Meteorology, 33, 140-158.
Day, R. T., Keddy, P. A., McNeill, J.,, Carleton, T. (1988). Fertility and Disturbance Gradients: A
Summary Model for Riverine Marsh Vegetation. Ecology, 69(4), 1044-1054.
Detenbeck, N. E., Batterman, S. L., Brady, V. J., Brazner, J. C., Snarski, V. M., Taylor, D. L.,
Thompson, J. A.,, Arthur, J. W. (2000). A Test of Watershed Classification Systems for Ecological
Risk Assessment. Environmental and Toxicological Chemistry, 19, 1174-1181.
Detenbeck, N. E. (2001). Methods for Evaluating Wetland Condition: Wetlands Classification
(EPA 822-R-01-007g). Washington, DCU.S. EPA, Office of Water.
Detenbeck, N. E., Elonen, C. M., Taylor, D. L., Anderson, L. E., Jicha, T. M.,, Batterman, S. L.
(2003a). Effects of hydrogeomorphic region, watershed storage and mature forest on baseflow and
snowmelt stream water quality in second-order Lake Superior Basin tributaries. Freshwater Biology,
48(5), 911-927.
Detenbeck, N. E., Moffett, M. F., Pearson, M., Simon, T. P.,, Albert, D. (2003b). Flow- and nutri-
ent-based classification of Lake Michigan coastal riverine wetlands.
54
-------
Dethier, M. N. (1992). Classifying Marine and Estuarine Natural Communities: An Alternative to
the Cowardin System. Natural Areas Journal, 12(2), 90-100.
Dettmger, M. D., Cayan, D. R., Diaz, H. F.,, Meko, D. M. (1998). Decadal Variability of Precipita-
tion over Western North America. Journal of Climatology, 11(12), 3095-3111.
Dettmann, E. T. (2001). Effect of Water Residence Time on Annual Export and Denitrification of
Nitrogen in Estuaries: A model Analysis. Estuaries, 24(4), 481-490.
Digby, M. J., Saenger, P., Whelan, M. B., McCdonchie, D., Eyre, B., N., H.,, Bucher, D. (1998). A
Physical Classification of Australian Estuaries (Report prepared for the Urban Water Research Asso-
ciation of Australia 16/99, Land and Water Resource Research and Development Corporation).
Lismore, NSW: Centre for Coastal Management, Southern Cross University.
Dorward-King, E. J., Suter, G. W., Kapustka, L. A., Mount, D. R., Reed-Judkins, D. K., Cormier, S.
M., Dyer, S. D., Luxon, M. G., Parrish, R.,, Burton, G. A. (2001). Distinguishing Among Factors
That Influence Ecosystems. In: D. J. Baird, G. A. Burton (Eds.), Ecological Variability: Separating
Natural from Anthropogenic Causes of Ecosystem Impairment (pp.1-26). Pensacola, FL: Society of
Environmental Toxicology and Chemistry.
Drinkwater, K. F., Belgram, A., Borja, A., Conversi, A., Edwards, M., Greene, C. H., Ottersen, G.,
Pershing, A. J.,, Walker, H. (2003). The Response of Marine Ecosystems to Climate Variability As-
sociated with the North Atlantic Oscillation. In: Y. H. J. W. Hurrell, Y. Kushnir, G. Ottersen and
M. Visbeck (Ed.), The North Atlantic Oscillation: Climate Significance and Environmental Impact,
Geophysical Monograph 134 (pp. 211-234). Washington, DC.
Dyer, K. (1973). Estuaries: A Physical Introduction. Wiley-Interscience, 140 pp.
Edgar, G. J., Barrett, N. S., Graddon, D. J.,, Last, P. R. (2000). The Conservation Significance of
Estuaries: A Classification of Tasmanian Estuaries Using Ecological, Physical and Demographic At-
tributes As a Case Study. Biological Conservation, 92, 383-397.
Edgar, G. J.,, Barrett, N. S. (2002). Benthic macro fauna in Tasmanian estuaries: scales of distribu-
tion and relationships with environmental variables. Journal of Experimental Marine Biology and
Ecology, 270, 1-24.
Engle, V. D.,, Summers, J. K. (1999). Refinement, validation, and application of a benthic condition
index for northern Gulf of Mexico estuaries. Estuaries, 22(3A), 624-635.
Ferreira, J.G. (2000). Development of an estuarine quality index based on key physical and biogeo-
chemical features. Ocean, Coastal Management, 43(1), 99-122.
Fischer, H. (1976). Mixing and Dispersion of Estuaries. Annual Review of Fluid Mechanics, 8(1),
107-133.
55
-------
Franken, S. K., Tyler, D. J.,, Verdin, K. L. (2001). Development of a National Seamless Database of
Topography and Hydrologic Derivatives. Paper presented at the 2001 ESRI International User's
Conference, San Diego, CA.
Frederichs, C. T.,, Madsen, M., O. S. (1992). Nonlinear Diffusion of the Digal Gignal in Frictionally
Dominated Embayments. Journal of Geophysical Research, 97, 5637-5650.
Gall egos, C. L. (2001). Calculating Optical Water Quality Targets to Restore and Protect Submersed
Aquatic Vegetation: Overcoming Problems in Partitioning the Diffuse Attenuation Coefficient for
Photosynthetically Active Radiation. Estuaries, 24(3), 381-397.
Geyer, W. R., Morris, J. T., Prahl, F. G.,, A., J. D. (2000). Interaction Between Physical Processes
and Ecosystem Structure: A Comparative Approach In J. E. Hobbie (Ed.), Estuarine Science (pp.
177-210). Washington, DC: Island Press.
Gianessi, L. P.,, Anderson,}. E. (1995). Pesticide Use in U. S. Crop Production; National Data Re-
port. Washington, DC: National Center for Food and Agricultural Policy.
Gibbs, J. P. (1993). Importance of Small Wetlands for the Persistence of Local Populations of Wet-
land-Associated Animals. Wetlands, 13, 25-31.
Glatfelter, D. R. (1984). Techniques for estimating magnitude and frequency of floods on streams
in Indiana (Water-Resources Investigations Report 84-4134). Indianapolis, IN: U. S. Geological
Survey (USGS).
Goss, G. W., Kellog, R. L., Sanabna, J., Wallace, S.,, Kniesel, W. (1998). The National Pesticide
Loss Database: A Tool for Management of Large Watersheds. Poster presented at the 53rd Annual
SWCS Conference, San Diego, CA. Retrieved on October 3, 2003 from
http://www.nrcs.usda.gov/technical/land/pubs/gosstext.html
Great Lakes Commission. (2001) Summary of Great Lakes Coastal Wetlands Classification, No-
vember 21, 2001. http://www.glc.org/wetlands/pdf/wetlands-class-scheme.pdf (October 3, 2003).
Groves, C. R., Jensen, D. B., Valutis, L. L., Redford, K. H., Shaffer, M. L., Scott, J. M., Baumgartner,
J. V., Higgins, J. V., Beck, M. W.,, Anderson, M. G. (2002). Planning for Biodiversity Conservation:
Putting Conservation Into Practice. BioScience, 52(6), 499-512.
Hansen, D. V.,, Rattray, M. J. (1966). New Dimensions in Estuary Classification. Limnology and
Oceanography, 11, 319-326.
Hawkins, C. P., Kershner, J. L., Bisson, P. A., Bryant, M. D.,, Decker, L. M. (1993). A Hierarchical
Approach to Classifying Stream Habitat Features. Fisheries, 18, 3-12.
Hearn. (1998). Application of the Strommel Model to Shallow Mediterranean Estuaries and their
Characterization. Journal of Geophysical Research, 103(10), 391-405.
Hobbie, J. E. (2000). Estuarine Science: The Key to Progress in Coastal Ecological Research: Ex-
56
-------
ecutive Summary. Washington, DC: Island Press.
Holtschlag, D. J.,, Croskey, H. M. (1984). Statistical models for estimating flow characteristics of
Michigan streams (Water-Resources Investigations 84-4207). Lansing, MI: U. S. Geological Survey
(USGS).
Huff, F. A.,, Angel,]. R. (1992). Rainfall Frequency Atlas of the Midwest (Bulletin 71). Champaign,
IL: Illinois State Water Survey.
Ippen, A.,, Harlemann, D. (1961). Committee for Tidal Hydraulics (Technical Bulletin 51): U. S.
Army Corps of Engineers.
Jacques,}. E.,, Lorenz, D. L. (1988). Techniques for estimating the magnitude and frequency of
floods of unguaged streams in Minnesota (Water Resources Investigations Report 87-4170). Wash-
ington, DC: U. S. Geological Survey (USGS).
Jay, D. A.,, Smith, J. D. (1988). Circulation in and Classification of Shallow, Stratified Estuaries. In
J. Dronkers, W. van Leussen (Eds.), Physical Processes in Estuaries (pp. 21-41) New York, NY:
Springer-Verlag.
Jay, D. A., Geyer, W. R.,, Montgomery, D. R. (2000). An Ecological Perspective on Estuarine Clas-
sification. In J. E. Hobbie (Ed.), Estuarine Science: A Synthetic Approach to Research and Practice
(pp. 149-176). Washington, DC: Island Press.
Jennings, M. E., Thomas, W. O. J.,, Riggs, H. C. (1993). Nationwide Summary of U. S. Geological
Survey Regional Regression Equations for Estimating Magnitude and Frequency of Floods for Un-
gauged Sites (94-4002). Reston, VA: U. S. Geological Survey (USGS).
Jensen, M. E., Goodman, I. A., Bourgeron, P. S., Poff, N. L.,, Brewer, C. K. (2001). Effectiveness of
Biophysical Criteria in the Hierarchical Classification of Drainage Basins Water. Water Resources
Bulletin, 37(5), 1155-1167.
Kass, G. V. (1980). An exploratory technique for investigating large quantities of categorical data.
Applied Statistics, 29, 119-127.
Keough, J. R., Thompson, T. A., Guntenspergen, G. R.,, Wilcox, D. A. (1999). Hydrogeomorphic
Factors and Ecosystem Responses In Coastal Wetlands of the Great Lakes. Wetlands, 19, 821-834.
Keys, J. E., Jr., Carpenter, C. A., Hooks, S. L., Koenig, F., McNab, W. H., Russell, W. E.,, Smith, M.
L. (1995). Ecological Units of the Eastern United States: First Approximation: U. S. Department of
Agriculture (USDA), Forest Service.
Koltun, G.F.,, Whitehead, M.T. (2001) Techniques for Estimating Selected Streamflow Characteris-
tics of Rural, Unregulated Streams in Ohio. (Water Resources Investigations Report 02—4068) Co-
lumbus, OH: U.S. Geological Survey.
Krug, W. R., Conger, D. H.,, Gebert, W. A. (1992). Flood-frequency characteristics of Wisconsin
57
-------
streams (Water Resources Investigation Report 91-4128). Madison, WI: U. S. Geological Survey.
Legleiter, K. J. (2001). Interagency Development of National Watershed and Subwatershed Hydro-
logic Units. Paper presented at the 2001 ESRI International User's Conference, San Diego, CA.
Long, E. R., MacDonald, D. D., Smith, S. L.,, Calder, D. (1995). Incidence of Adverse Biological
Effects Within Ranges of Chemical Concentrations in Marine and Estuarine Sediments. Environ-
mental Management, 19, 81-97.
Lumia, R. (1991). Regionalization of flood discharges for rural, unregulated streams in New York,
excluding Long Island (Water-Resources Investigations Report 90-4197). Albany, NY: U. S. Geo-
logical Survey.
MacDonald, D. D., Ingersoll, C. G.,, Berger, T. A. (2000a). Development and Evaluation of Con-
sensus-Based Sediment Quality Guidelines for Freshwater Ecosystems. Arch. Environ. Contam.
Toxicol., 39, 20-31.
MacDonald, D. D., Dipmto, L. M., Fields,]., Ingersoll, C. G., Long, E. R.,, Swartz, R. C. (2000b).
Development and evaluation of consensus-based sediment effect concentrations for polychlorinated
biphenyls. Environ. Toxicol., 19, 1403-1413.
Maxwell, J. R., Edwards, C. J., Jensen, M. E., Paustian, S. J., Parott, H.,, Hill, D. M. (1995). A Hierar-
chical Framework of Aquatic Ecological Units in North America (Neararctic Zone) (Technical Re-
port NC-176): U. S. Department of Agriculture (USDA), Forest Service.
Maynard, L.,, Wilcox, D. (1997,). Coastal Wetlands of the Great Lakes. Paper presented at the
State of the Lakes Ecosystem Conference 1966, Chicago, IL.
McKee, P. M., Batterson, T. R., Dahl, T. E., Glooschenko, V., Jaworski, E., Pearce, J. B., Raphael,
C. M., Whillans, T. H.,, LaRoe, E. T. (1992). Great Lakes Aquatic Habitat Classification Based on
Wetland Classification Systems. In W. Dieter, M. Busch,, P. G. Sly (Eds.), The Development of an
Aquatic Habitat Classification System for Lakes . Ann Arbor, MI: C. R. C. Press.
McKune, B.,, Mefford, M. J. (1999). Multivariate Analysis of Ecological Data (Version 4.14) [Soft-
ware]. Gleneden Beach, OR: MjM Software.
Mielke, P. W. J. (1984). Meteorological applications of permutation techniques based on distance
functions: Elsevier Science Publishers.
Mine, L. D.,, Albert, D. A. (1998). Great Lakes Coastal Wetlands: Abiotic and Floristic Characteri-
zation.
Montgomery, D. R.,, Buffington, J. M. (1993). Channel classification, prediction of channel re-
sponse, and assessment of channel condition (TFW-SH10-93-002). Olympia, WA 84 p: Washing-
ton Department of Natural Resources.
Morrisey, D. J., Howitt, L., Underwood, A. J.,, Stark, J. S. (1992). Spatial variation in soft sediment
58
-------
benthos. Marine Ecology Progress Series, 81, 197-204.
Neitsch, S. I., Arnold,]. G., Kiniry, J. R., Williams,]. R.,, King, K. W. (2002). Soil and Water As-
sessment Tool: Theoretical Documentation (TWRI Report TR-191). College Station, TX: Texas
Water Resources Institute.
NOAA. (1989). Susceptibility and Status of Gulf of Mexico Estuaries to Nutrient Discharges. Silver
Spring, MD: National Oceanic and Atmospheric Administration, Office of Oceanography and Ma-
rine Assessment.
NOAA. (1996). Estuarine Eutrophication Survey, v-1 1: South-Atlantic region. Silver Spring, MD:
National Oceanic and Atmospheric Administration, Office of Ocean Resources Conservation and
Assessment.
NOAA. (1997a). NOAA's Estuarine Eutrophication Survey, Vol. 2: Mid-Atlantic region . Silver
Spring, MD: National Oceanic and Atmospheric Administration, Office of Ocean Resources Con-
servation and Assessment.
NOAA. (1997b). NOAA's Estuarine Eutrophication Survey, Vol. 3: North-Atlantic region. Silver
Spring, MD: National Oceanic and Atmospheric Administration, Office of Ocean Resources Con-
servation and Assessment.
NOAA. (1997c). NOAA's Estuarine Eutrophication Survey, Vol. 4: Gulf of Mexico region. Silver
Spring, MD: National Oceanic and Atmospheric Administration, Office of Ocean Resources Con-
servation and Assessment.
NOAA. (1998). NOAA's Estuarine Eutrophication Survey, Vol. 5: Pacific Coast region. Silver
Spring, MD: National Oceanic and Atmospheric Administration, Office of Ocean Resources Con-
servation and Assessment.
NOAA. (2000). Marine and Estuarine Ecosystem and Habitat Classification (NOAA Technical
Memorandum NMFS-F/SPO-43). Silver Spring, MD: U. S. Department of Commerce, National
Oceanic and Atmospheric Administration (NOAA).
NOAA. (2003a). Coastal Assessment, Data Synthesis System (CA&DS)
(http://spo.nos.noaa.gov/projects/cads/ftp_data_download.html): National Oceanic and Atmos-
pheric Administration.
NOAA. (2003b). NOAA's Medium Resolution Digital Vector Shoreline.
http://spo.nos.noaa.gov/proiects/shoreline/shoreline.html (October 3, 2003).
NOAA. (2003c). Office of Response and Restoration, National Ocean Service, National Oceanic
and Atmospheric Administration. ESI Maps. January 6, 2003.
http://response.restoration.noaa.gov/esi/esiintro.html (October, 3, 2003).
NOAA. (2003d). ORCA's Coastal Assessment Framework
http://spo.nos.noaa.gov/projects/caf/caf.html (October 3, 2003).
59
-------
NRC. (1994). Priorities For Coastal Ecosystem Science. Washington, DC: U.S. National Research
Council, National Academy Press.
NRC. (2000). Clean Coastal Waters. Washington, DC: U.S. National Research Council, National
Academy Press.
NRC. (2001). Assessing the TMDL Approach to Water Quality Management. Washington, DC:
U.S. National Research Council, National Academy Press.
Nunes Vas, R. A.,, Lennon, C. W. (1991). Modulation of Estuarine Stratification and Mass Trans-
port at Tidal Frequencies. In B. B. Parker (Ed.), Progress in Tidal Hydrodynamics (pp. 502-520).
New York: John Wiley and Sons.
Odum, H. T. (1962) Man and Ecosystem. Proceedings, Lockwood Conference on the Suburban
Forest and Ecology. Bulletin Connecticut Agricultural Station
Odum, H. T. (1994). Ecological and General Systems. Niwot, CO: University Press
of Colorado.
Oey. (1984). On the Steady Salinity Distribution and Circulation in Partially Mixed and Well Mixed
Estuaries. Journal Physical Oceanography, 14, 629-645.
Officer, C. B. (1976). Physical Oceanography of Estuaries and Associate Coastal Waters, Wiley In-
terscience. New York.
OHDNR. (2003). Ohio Department of Natural Resources. Wetland Maps for Ohio.
http: / /www.dnr.state.oh.us /wetlands /mapping.htm (October 3, 2003).
Omernik, J. M. (1987). Ecoregions of the Conterminous United States. Ann Assoc Am Geogr, 77,
118-125.
Omernik, J. M., 1995. Ecoregions: A Spatial Framework for Environmental Management. In: Bio-
logical Assessment and Criteria: Tools for Water Resource Planning and Decision Making, W. Davis
and T. Simon (eds). Boca Raton, FL: Lewis Publishers.
Omernik, J. M., Rohm, C. M., Woods, A. J.,, Stoddard, J. L. (2002). Regional Characteristics of Nu-
trient Concentrations in Streams and their Application to Nutrient Criteria Development. Journal of
the American Water Resources Association, 38(1), 213-240.
Orlando, P.,, Klein, J. (1999). A Classification of Physical Transport Processes and the Susceptibility
to Nutrients-Related Water Quality Impacts In 138 U.S. Estuaries: National Oceanic and Atmos-
pheric Administration (NOAA).
Paul J. F., Scott, K. J., Campbell, D. E., Gentile, J. H., Strobel, C. J., Valente, R. M., Weisberg, S. B.,
Holland, A. F., Ranasinghe, J. A. (2001). Developing and applying a benthic index of estuarine con-
60
-------
dition for the Virginian Biogeographic Province. Ecological Indicators l(2):83-99.
Phytosociological Research Center. (1995). Global Bioclimatic's Org.
http://www.globalbioclimatics.org (October 3, 2003).
Poff, N. L.,, Ward, J. V. (1989). Implications of streamflow variability and predictability for lotic
community structure: A regional analysis of streamflow patterns. Ca. J. Fish Aquat. Sci., 46, 1805-
1818.
Poff, N. L.,, Allan, J. D. (1995). Functional Organization of Stream fish assemblages in relation to
Hydrologic Variability. Ecology, 76, 606-627.
Prandle, D. (1986). Generalized theory of estuarine dynamics. In J. van deKreek (Ed.), Physical of
Shallow Estuaries and Bays (pp. 42-57). Berlin: Springer-Verlag.
Ries, K. G., Ill,, Grouse, M. Y. (2002). The National Flood Frequency Program, Version 3 (Com-
puter Program for Estimating Magnitude and Frequency of Floods for Ungaged Sites 02-4168, 42
p.): U. S. Geological Survey.
Robertson, D. M., Saad, D. A.,, Wieben, A. M. (2001). An alternative rationalization scheme for de-
fining nutrient criteria for rivers and streams (Water-Resources Investigations Report 01-4073).
Middleton, WI: U.S. Geological Survey.
Rosgen, D. L. (1994). A Classification of Natural Rivers. Catena, 22, 169-199.
Rosgen, D. (1996). Applied River Morphology. Pagosa Springs, CO.
Rubin, D. B. (1987). Multiple Imputation for Nonresponse in Surveys, New York:
John Wiley, Sons, Inc.
Saco, P.,, Kumar, P. (2000). Coherent modes in multiscale variability of streamflow over the United
States. Water Resources Research 4th ed, 36, 1049-1067.
SAS Institute, Inc. (1990). SAS/STAT® User's Guide, Version 6, 4th Ed., Vols. 1-2.
Gary, NC: SAS Institute, Inc.
SAS Institute, Inc. (2001). The MI Procedure.
http://support.sas.com/rnd/app/da/new/802ce/stat/chap9/sect4.htm (October 10, 2003).
Sauer, V. B., Thomas, W. O. J., Stacker, V. A.,, Wilson, K. V. (1983). Flood characteristics of urban
watersheds in the United States (Water-Supply Paper 2007): U.S. Geological Survey (USGS).
Seaber, P. R., Kapinos, F. P.,, Knapp, G. L. (1987). Hydrologic Unit Maps (Water-Supply Paper
2294, 63 p.): U.S. Geological Survey (USGS).
Shaw, S. P.,, Fredme, C. G. (1956). Wetlands of the United States (Circular 39): U.S. Fish and Wild-
life Service (FWS).
61
-------
Simon, T. P., Dufour, R. L., Stewart, P. M.,, Moffett, M. F. (2003). Distribution, Classification, and
Selection Rationale of Coastal Wetlands in Lake Michigan. In T. Edsall (Ed.), The Lake Michigan
Ecosystem: Ecology, Health, and Management. London, England: Elsevier Press.
Simpson,].,, Hunter,]. R. (1974). Fronts in the Irish Sea. Nature, 250, 404-406.
Sklar, F. H.,, Browder, J. A. (1998). Coastal environmental impacts brought about by alterations to
freshwater flow in the Gulf of Mexico. Environmental Management, 22, 547-562.
Smith, R. A., Swartz, G. E.,, Alexander, R. B. (1997). Regional interpretation of water-quality moni-
toring data. Water Resources Research, 33, 2781-2798.
Smith, R. A.,, Alexander, R. B. (2000). Sources of Nutrients in the Nation's Watersheds
(http://water.usgs.gov/nawqa/sparrow/nut_sources/nut_sources.htm). Reston, VA: U.S. Geologi-
cal Survey (USGS).
Staile, D., Livingstone, D. M., Weyhenmeyer, G. A.,, George, D. G. (2003). The Response of
Freshwater Ecosystems to Climate Variability Associated with the North Atlantic Oscillation. In J.
W. Hurrell (Ed.) (pp. 265-28?).
Statsoft, Inc. (2003). Electronic Textbook Statsoft, 1984-2003.
http://www.statsoft.com/textbook/stathome.html (October 3, 2003).
Stedfast, D. A. (1986). Evaluation of six months for estimating magnitude and frequency of peak
discharges on urban streams in New York (Water-Resources Investigations Report
84-4350). Albany, NY: U.S. Geological Survey (USGS).
Stefan, H. G., Hondzo, M., Eaton, J. G.,, McCormick, J. H. (1995). Validation of a Fish Habitat for
Lakes. Ecological Modeling, 82, 211-224.
Stefan, H. G., Hondzo, M., Fang, X., Eaton,]. G.,, McCormick, J. H. (1996). Simulated Long-Term
Temperature and Dissolved Oxygen Characteristics in the North Central United States and Associ-
ated Fish Habitat Limits. Limnology and Oceanography, 41(5), 1124-1135.
Strommel, H.,, Farmer, H. (1952). On the Nature of Estuarine Circulation (Reference Notes 52-51,
52-63, 52-88): Woods Hole Oceanographic Institute.
TAES (2000). Texas Agricultural Experiment Station. Hydrologic Unit Modeling for the United
States (HUMUS), http://srph.brc.tamus.edu/humus/ (October 3, 2003).
University of Minnesota. (2003). Welcome to the Great Lakes Environmental Indicators Project,
2001-2003. http://glei.nm.umn.edu/default/default.htm (October 3, 2003).
USDA (1994). State Soil Geographic (STATSGO) Database Data Use Information (Miscellaneous
Publication 1492). Fort Worth, TX: U.S. Department of Agriculture, National Resources Conserva-
tion Service, National Soil Survey Center.
62
-------
USDA (2001). Natural Resources Conservation Service, Soil Survey Division. National STATSGO
Database, July 30, 2001. http://www.ftw.nrcs.usda.gov/stat data.html. (October 3, 2003).
USDA (2002). USDA Agricultural Research Service, Grassland, Soil and Water Research Labora-
tory. Soil and Water Assessment Tool (SWAT), November 12, 2002.
http://www.brc.tamus.edu/swat/ (October 3, 2003).
USEPA (1989). Ambient Water Quality Criteria for Ammonia (Saltwater) (EPA-440/5-88-004).
Washington, DC: U.S. Environmental Protection Agency, Office of Water.
USEPA (1995). The Great Lakes: An Environmental Atlas and Resource Book, 3rd edition. Chi-
cago, IL and Toronto, Ontario, Canada: U.S. Environmental Protection Agency, Great Lakes Na-
tional Program Office and Government of Canada.
USEPA (1996). The National Sediment Quality Survey (EPA-823-D-96-002). Washington, DC:
U.S. Environmental Protection Agency, Office of Science and Technology.
USEPA (1998). National Sediment Inventory (NSI) Version 1.2 for the Conterminous U.S., No-
vember 18, 1998. http://www.epa.gov/waterscience/basins/metadata/nsi.htm
(October3, 2003).
USEPA (1999a). Comparative toxicity testing of selected benthic and epibenthic organisms for the
development of sediment quality test protocols (EPA-600-R-99-085).
U.S. Environmental Protection Agency.
USEPA (1999b). National Recommended Water Quality Criteria - Correction (EPA-822-Z-99-001).
Washington, DC: U.S. Environmental Protection Agency, Office of Water.
USEPA (1999c). Update of Ambient Water Quality Criteria for Ammonia (EPA-822-R-R-99-014).
Washington, DC: U.S. Environmental Protection Agency, Office of Water.
USEPA (2000). Stressor Identification Guidance Document (EPA/822/B-00/025). Washington,
DC: U.S. Environmental Protection Agency, Office of Water and Office of Research and Develop-
ment.
USEPA (2001a). National Coastal Condition Report (EPA 620/R-01/005). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development and Office of Water.
USEPA (200Ib). Nutrient Criteria Technical Guidance Manual: Estuarine and Coastal Marine Wa-
ters. (EPA-822-B-01-003). Washington, DC: U.S. EPA, Office of Water.
http://www.epa.gov/waterscience/standards/nutrients/marine/index.html
USEPA (200Ic). Great Lakes National Program Office. Sensitivity of Coastal Environments and
Wildlife to Spilled Oil. July 18, 2001. http://www.epa.gov/glnpo/lakemich/esi/ (October 3,
2003).
63
-------
USEPA (2002a). Aquatic Stressors: Framework and Implementation Plan for Effects Research
(EPA 600/R-02/074). Research Triangle Park, NC: U. S. Environmental Protection Agency, Office
of Research and Development.
USEPA (2002b). Watershed Assessment, Tracking, Environmental Results: Data Downloads, Sep-
tember 3, 2002. http: / 7www.epa.gov/waters /data/downloads.html
(Octobers, 2003).
USEPA (2003a). BASINS: Better Assessment Science Integrating Point and Non-point Sources,
October 3, 2003. http://www.epa.gov/waterscience/basins/ (October 3. 2003).
USEPA (2003b). Environmental Monitoring and Assessment Program, June 16, 2003.
http: / 7www.epa.gov/emap (October 3, 2003).
USEPA (2003c). Guidance for 2004 Assessment, Listing and Reporting Requirements Pursuant to
Sections 303(d) and 305(B) of the Clean Water Act (TMDL-01-03). Washington, DC: U.S. Envi-
ronmental Protection Agency, Office of Water.
USEPA (2003d). National Estuary Program: Links, October 3, 2003.
http://www.epa.gov/owow/estuaries/links.htm#neps (October 3, 2003).
USEPA (2003e). Procedures for Deriving Equilibrium Partioning Sediment Benchmarks (ESBs) for
the Protection of Benthic Organisms: Dieldrm (EPA-600-R-02-010). Washington, DC: U.S. Envi-
ronmental Protection Agency, Office of Research and Development.
USEPA (2003f). Procedures for Deriving Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: Endnn. (EPA-600-R-02-009). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development.
USEPA (2003g). Procedures for Deriving Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: PAH Mixtures (EPA-600-R-02-013). Washington, DC:
U.S. Environmental Protection Agency, Office of Research and Development.
USEPA (2003h). U.S. EPA Reach File 1 (RF1) for the Conterminous United States in BASINS,
October 2, 2003. http: / /www.epa.gov/waterscience/basins /metadata/rfl .htm (October 3, 2003).
USEPA (2003i). Watershed Information Network. Overall Watershed Characterization - National
Maps, October 3, 2003. http://www.epa.gov/iwi/1999sept/catalog.html
(October3, 2003).
USFWS (1981a). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the
United States. (Overview: FWS/OBS-81/02-vl ed.). (Vol. 1): U.S. Fish and Wildlife Service.
USFWS (1981b). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the
United States. (Vol. 2): U.S. Fish and Wildlife Service.
USFWS (1981c). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the
64
-------
United States. (Vol. 3): U.S. Fish and Wildlife Service.
USFWS (1981d). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the
United States. (Vol. 4): U.S. Fish and Wildlife Service.
USFWS (1981e). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the Un-
tied States. (Vol. 5): U.S. Fish and Wildlife Service.
USFWS (1981f). Fish and Wildlife Resources of the Great Lakes Coastal Wetlands within the
United States. (Vol. 6): U.S. Fish and Wildlife Service.
USFWS (1998-92). National Wetlands Inventory (NWI) Michigan and Illinois
(www.dnr.state.of.us/wetlands/mapping.shtml). St. Petersburg, FL.
USFWS (2003). National Wetlands Inventory, July 8, 2003. http: / /wetlands.fws.gov/ (October 3,
2003).
USGS (1999). Aquatic ecoregions of the conterminous United States, June 22, 1999.
http://water.usgs.gov/GIS/metadata/usgswrd/ecoregion.html (October 3, 2003).
USGS (2002). EROS Data Center, Elevation Derivatives for National Applications, December,
2002. http://edna.usgs.gov/Edna/ (October 3, 2003).
USGS (2003a). Great Lakes Aquatic Gap Analysis Project, September 10, 2003.
http: / /www.glsc.usgs.gov/GLGAP.htm (October 3, 2003).
USGS 2003b). National Hydrography Dataset, September 18, 2003. http://nhd.usgs.gov/
(October3, 2003).
USGS (2003c). Retrieval for spatial dataset gwnsk, October 3, 2003.
http://water.usgs.gov/lookup/getspatialPgwrisk (October 3, 2003).
USGS (2003d). SUSPENDED-SEDIMENT DATABASE: Daily Values of Suspended Sediment
and Ancillary Data, January 29, 2003. http://webserver.cr.usgs.gov/sediment/seddatabase.cfm
(October3, 2003).
USGS (2003e). Table of updated equations for the NFF Program, March 20, 2003.
http://water.usgs.gov/osw/programs/nffpubs.html (October 3, 2003).
USGS (2003f). Water Resources Applications Software, NFF, Version 3.2, National Flood Fre-
quency program, September 5, 2003. http://water.usgs.gov/software/nff.html
(Octobers, 2003).
USGS (2003g). Water Resources - Hydrologic Unit Maps, (http://water.usgs.gov/GIS/huc.html):
U.S. Geological Survey.
65
-------
Van Dolah, R. F., Hyland, J. L., Holland, A. F., Rosen, J. S.,, Snoots, T. R. (1999). A benthic index
of biological integrity for assessing habitat quality in estuaries of the southeastern USA. Marine En-
vironmental Research, 48(4), 269-283.
Vogelmann, J. E., Howard, S. M., Yang, L., Larson, C. R., Wylie, B. K.,, Van Driel, N. (2001).
Completion of the 1990s National Land Cover Data Set for the Conterminous United States from
Landsat Thematic Mapper Data and Ancillary Data Sources. Photogramm. Eng. Remote Sensing,
67, 650-662.
Walker, H. A., Yarnal, B.,, Najjar, R. (2002). Associations between NAO variability and US Mid-
Atlantic Region Hydroclimatology. Paper presented at the Proceedings, Chapman Conference,
Spain.
Weisberg, S. L., Frithsen, J. B., Holland, A. F., Paul, J. F., Scott, K. J., Summers, J. K., Wilson, H. T.,
Valente, R., Heimbuch, D., Gemtsen, J., Schimmel, S. C.,, Latimer, R. W. (1993). EMAP-Estuaries
Virginian Province 1990 Demonstration Project Report (EPA/620/R-93/006): Office of Research
and Development.
Weisberg, S. B., Ranasmghe, J. A., Dauer, D. M., Schaffner, L. C., Diaz, R. J.,, Frithsen, J. B. (1997).
An estuarine benthic index of biotic integrity (B-IBI) for Chesapeake Bay. Estuaries, 20(1), 149-158.
Wilkinson, L. (1999). Systat 9.0 Statistics I. Chicago, IL: SPSS, Inc.
WIDNR (2003). Wisconsin Department of Natural Resources. Wisconsin Wetlands: Wetland
Mapping, April 16, 2003. http: / /www.dnr.state.wi.us /org/water/flip/wetlands /mapping.shtml
(October3, 2003).
66
-------
-------
SEPA
United States
Environmental Protection
Agency
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HEREfJ ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA 600/R-04/XXX
May 2004
www.epa.gov
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
-------
EPA 600/R-04/061
May 2004
Appendices
Classification Framework
for Coastal Systems
U.S. Environmental Protection Agency
Office of Research and Development
National Health and Environmental Effects Research Laboratory
Research Triangle Park, NC 27711
-------
List of Appendices
A-l.l Estuarine Drainage Area physical and hydrological characteristics: metadata
A-1.2 Estuarine Drainage Area physical and hydrological characteristics: data1
A-2.1 EDA/CDA land-use and land-cover: metadata
A-2.2 EDA/CDA land-use and land-cover: data1
A-3.1 EDA stressor loadings: metadata
A-3.2 EDA stressor loadings: data1
A-4.1 EDA stressor exposure: metadata
A-4.2 EDA stressor exposure: data1
A-5.1 EDA modifying factors: metadata
A-5.2 EDA modifying factors: data1
B-l.l Great Lakes coastal riverine wetland watersheds: metadata
B-1.2 Great Lakes coastal riverine wetland watersheds: data1
C-l.l Marine and Great Lakes coastal watersheds: equations for peak flow predictions: metadata
C-1.2 Marine and Great Lakes coastal watersheds: equations for peak flow predictions: data1
C-1.3 Marine and Great Lakes coastal watersheds: equations for peak flow predictions: references
C-2.1 Marine and Great Lakes coastal watersheds: peak flow classes identified by CART analysis:
metadata
C-2.2 Marine and Great Lakes coastal watersheds: peak flow classes identified by CART analysis: data1
C-3 Hydrologic regions for marine and Great Lakes coastal states
D Classification of ED As by cluster analysis
E Matrix of properties of existing classification schemes
F Regional maps of sediment toxic units by chemical class (metals, pesticides, PAHs) for estuaries
1 These appendices are data files (i.e., Excel files); available in a zipped downloadable file with the main document:
Appendices are located in the zipped file: A-1.2, A-2.2, A-3.2, A-4.2, A-5.2, B-1.2, C-1.2, C-2.2 at the same weblink as this
document.
-------
A-l.l Estuarine Drainage Area physical and hydrological characteristics: metadata
Database: EDAPHYSHYDRO.XLS
Variable: EDA
Label: Estuarine Drainage Area Code
Units:
Format: uppercase alpha-numeric $5.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system
Metadata: EDA is derived from the variable, EDASUBEDA, which is in the
CA&DS dataset, Reference EDA H Data, available from the above
download site. 203 ED As were chosen to be used in classification
Variable: EDANAME
Label: Estuarine Drainage Area Name
Units:
Format: uppercase alpha $43.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office, Coastal Assessment & Data
Synthesis (CA&DS) system
Metadata: EDANAME is derived from the variable, EDA_NAME, which is in
the CA&DS dataset, Reference EDA H Data, available from the
above download site.
Variable: EDATASQKM
Label: Total Area of EDA
Units: sq km
Format: numeric 12.
Source: http://spo.nos.noaa.gov/projects/cads/ftp data download.html
NOAA/NOS/Special Projects Office, Coastal Assessment & Data
Synthesis (CA&DS) system, SAS Dataset: cads_surfarea, physhydro,
Excel File: cads_surfarea, physhydro
Metadata: EDATASQKM is derived from the variable, EDATSAMI2, which is
in the CA&DS dataset, PandH_EDA_h Data, available from the
above download site. It represents the Total Area (Land + Water)
for the Coastal Watershed (ED A/CD A). This was converted from
square miles to square kilometers.
Variable: ESTUARYAREA
Label: Area of estuary in EDA
Units: sq km
Format: numeric 12.
Source: NOAA's Estuarine Eutrophication Survey — Volumes 1-5, NOAA,
Office of Ocean Resources Conservation Assessment 1996; SAS
A-l.l - 1
-------
Dataset: cads_pandh, physhydro, Excel File: cads_pandh,
physhydro
Metadata: ESTUARYAREA is derived from the variable, WATRE_AREA,
which is in the CA&DS dataset, PandH_EDA_h Data, available
from the above download site. It represents the Water Area for the
Coastal Watershed (EDA/CDA). This was converted from square
miles to square kilometers
Variable: MIXZONEAREA
Label: Mixing Zone Surface Area
Units: sq km
Format: numeric 12.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; SAS Dataset: cads_pandh, physhydro,
Excel File: cads_pandh, physhydro
Metadata: MIXZONEAREA is derived from the variable, MIXZSAMI2, which
is in the CA&DS dataset, PandH_EDA_h Data, available from the
above download site. It represents the Mixing Zone (0.5 - 25.0 ppt)
Surface Area. This was converted from square miles to square
kilometers.
Variable: SEAZONEAREA
Label: Seawater Zone Surface Area
Units: sq km
Format: numeric 12.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; SAS Dataset: cads_pandh, physhydro,
Excel File: cads_pandh, physhydro
Metadata: SEAZONEAREA is derived from the variable, SEAZSAMI2, which
is in the CA&DS dataset, PandH_EDA_h Data, available from the
above download site. It represents the Seawater Zone (>25.0 ppt)
Surface Area. This was converted from square miles to square
kilometers.
Variable: TFZONEAREA
Label: Tidal Freshwater Zone Surface Area
Units: sq km
Format: numeric 12.
Source: http://spo.nos.noaa.gov/projects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; SAS Dataset: cads_pandh, physhydro;
Excel File: cads_pandh, physhydro
Metadata: TFZONEAREA is derived from the variable, TFZSAMI2, which is
in the CA&DS dataset, PandH_EDA_h Data, available from the
above download site. It represents the Tidal Freshwater Zone (<0.5
A-l.l -2
-------
ppt) Surface Area. This was converted from square miles to square
kilometers.
Variable: TIDEHT
Label: Height of tide
Units: m
Format: numeric 12.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; SAS Dataset: cads_pandh, physhydro;
Excel File: cads_pandh, physhydro
Metadata: TIDEHT is derived from the variable, AESTMTDFT, which is in
the CA&DS dataset, PandH_EDA_h Data, available from the above
download site. It represents the Average Tidal Height calculated as
means of the height differences or ratios measured at NOS tide gauge
stations. This was converted from feet to meters.
Variable: RIVERFLOW
Label: Average Monthly River Flow
Units: cu m/day
Format: numeric 12.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html.
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; NOAA's Estuarine Eutrophication
Survey — Volumes 1-5, NOAA, Office of Ocean Resources
Conservation Assessment 1996; SAS Dataset: cads_pandh,
physhydro, neesdata; Excel File: cads_pandh, physhydro, neesdata
Metadata: RIVERFLOW is derived from the variable, ANNLTFLWAV, which
is in the CA&DS dataset, PandH_EDA_h Data, available from the
above download site. It represents the Annual Long-Term Flow
Average of Gauged Rivers obtained from USGS Gage stations data.
If values were missing for RIVERFLOW, the average daily inflow
values from NOAA's Estuarine Eutrophication Survey were
substituted. Both sets of values were converted from cubic feet per
second to cubic meters per day.
Variable: ESTUARYVOL
Label: Estuary Volume
Units: billion cu m
Format: numeric 12.
Source: http://spo.nos.noaa.gov/projects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; NOAA's Estuarine Eutrophication
Survey — Volumes 1-5, NOAA, Office of Ocean Resources
Conservation Assessment 1996; SAS Dataset: physhydro, neesdata;
Excel File: physhydro, neesdata
Metadata: ESTUARYVOL was typed in from hardcopies of the 5 regional
reports of the Estuarine Eutrophication Survey. It comes directly
A-l.l -3
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
from the bottom left cell, labeled "Volume (billion cu ft)" in the
table, titled "Physical and Hydrologic Characteristics" for each EDA.
This was converted from billion cubic feet to billion cubic meters.
This value represents the volume of the estuary only (water only). If
ESTUARYVOL was missing then estimates of estuary volume were
calculated as estuaryarea(m2)*depth_m/1000000000.
TIDALPRISMVOL
Tidal Prism Volume
cu m
numeric 12.
http://spo.nos.noaa.gov/projects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system; SAS Dataset: cads_pandh, physhydro;
Excel File: cads_pandh, physhydro
TIDALPRISMVOL is derived from the variable, TPVOLBCF,
which is in the CA&DS dataset, PandH_EDA_h Data, available
from the above download site. It represents the Tidal Prism Volume
calculated using the salinity zone mean-range value when available; if
not, the salinity mean-tide value multiplied by two was used instead.
This salinity zone tide value multiplied by the salinity zone area
provided volume for each salinity zone. The sum of all salinity zone
volumes provided the tidal prism volume representative for the
estuary. If tide information was not available for all three-salinity
zones, the estuary mean-range was used when available, if not, the
estuary mean-tide value multiplied by two was used instead. This
estuary tide-value times the estuary water area provided the tidal
prism volume representative for the estuary. This value was
converted from billion cubic feet to billion cubic meters.
Variable:
Label:
Units:
Format:
Source:
BTM_SAL
Salinity at bottom depth
ppt
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP) 1990-
1997; National Coastal Assessment (NCA) 2000; SAS datasets:
emapwq, physhydro; Excel files: emapwq, physhydro
Metadata: Salinity was measured at surface and bottom depths of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). All EMAP stations
were geo-referenced to ED As and HUCs. BTM_SAL represents the
salinity measured at bottom depths averaged across space and time
for each EDA.
A-l.l -4
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
SRF_SAL
Salinity at surface depth
ppt
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP) 1990-
1997; National Coastal Assessment (NCA) 2000; SAS datasets:
emapwq, physhydro; Excel files: emapwq, physhydro
Salinity was measured at surface and bottom depths of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). All EMAP stations
were geo-referenced to ED As and HUCs. SRF_SAL represents the
salinity measured at surface depths averaged across space and time
for each EDA.
Variable:
Label:
Units:
Format:
Source:
DEPTH_M
Depth at the bottom
m
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP) 1990-
1997; National Coastal Assessment (NCA) 2000; SAS datasets:
emapdepth, physhydro; Excel files: emapdepth, physhydro
Metadata: Depth was measured at the bottom of stations sampled during the
summer through EMAP. This includes stations in the Virginian
Province (1990-1993), Carolinian Province (1994-1997), West Indian
Province (1995), and Gulf of Mexico (1991-1994). This also includes
stations sampled through the National Coastal Assessment (Western
Pilot, 1999; NCA, 2000). All EMAP stations were geo-referenced to
ED As and HUCs. DEPTH_M represents the bottom depths
averaged across space and time for each EDA. If depth was missing
for an EDA, the average depth from NOAA's Estuarine
Eutrophication Survey Regional reports was used instead. This is
found in the middle left cell labeled, Average Depth (ft) Estuary, in
the table titled, Physical and Hydrologic Characteristics, for each
EDA. This depth was converted from feet to meters.
Variable: DCP
Label:
Units:
Format:
Source:
Metadata:
Dissolved Concentration Potential
mg/L
numeric 10.4
National Oceanic and Atmospheric Administration (NOAA). 1989.
Susceptibility and Status of Gulf of Mexico Estuaries to Nutrient
Discharges. Silver Spring, MD.Office of Oceanography and Marine
Assessment.
The variable DCP is a calculated variable estimating the dissolved
concentration potential of a pollutant as a function of pollutant load,
A-l.l -5
-------
Variable: PRE
Label:
Units:
Format:
Source:
Metadata:
the volume of freshwater in the estuary, freshwater inflow, and total
estuarine volume. The volume of freshwater in the estuary was
calculated using the freshwater fraction method, where,
Ffw = (SO-S)/SO), Ffw- Freshwater fraction,
SO= Boundary Salinity and S= Average salinity
The volume of freshwater was calculated using:
Vfw = Ffw*Vtot, where,
Vfw= volume of freshwater in the estuary,
Ffw= Freshwater Fraction, and
Vtot= Estuarine volume
Dissolved concentration potential (DCP) was calculated using the
following equation: DCP= L(Vfw/Ifw)(l/Vtot), where,
DCP= Dissolved concentration potential,
L= Pollutant Load,
Vfw=Volume of freshwater in the estuary,
Ifw= Average freshwater inflow (daily average river flow)
Vtot= Estuarine volume.
In order to compare DCP values among ED As, an estimated
pollutant load (L) of 25,000 kg/d was assigned to each EDA and
substituted in the DCP equation. Based on the standard pollutant
load, DCP values can be used to estimate the concentration of a
pollutant expected in an estuary.
Particle Retention Efficiency
years
numeric 10.4
National Oceanic and Atmospheric Administration (NOAA). 1989.
Susceptibility and Status of Gulf of Mexico Estuaries to Nutrient
Discharges. Silver Spring, MD.Office of Oceanography and Marine
Assessment.
Particle retention efficiency (PRE) is estimates an estuary's ability to
trap suspended particles, i.e., the time a particle remains in an estuary.
PRE is calculated using the formula: PRE= C/I, where,
C= Volume of the estuary
1= freshwater inflow
A-l.l -6
-------
Appendix A-2.1 EDA/CDA land-use and land-cover: metadata
Database: EDALANDCOVER.XLS
Variable: EDA
Label: Estuarine Drainage Area Code
Units:
Format: uppercase alpha-numeric $5.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system
Metadata: EDA is derived from the variable, EDASUBEDA, which is in the
CA&DS dataset, Reference EDA H Data, available from the above
download site. 203 ED As were chosen to be used in classification
Variable: WATER
Label: Area with Land Cover Type = Water
Units: sq km
Format: numeric 8.
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc7nlcd.html
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset.
Variable: URBANCOMM
Label: Area with Land Cover Type = Urban/Commercial
Units: sq km
Format: numeric 8.
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc7nlcd.html
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset. Urban/Commercial was created by summing the area for the
A-2.1 - 1
-------
following land cover types: High & Low Intensity Residential and
Commercial, Industrial, Transportation.
Variable: BARREN
Label: Area with Land Cover Type = Barren
Units: sq km
Format: numeric 8.
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc7nlcd.html
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset. Barren was created by summing the area for the following
land cover types: Bare rock, sand, clay and Quarry, strip mine, gravel
pit and Transitional from barren.
Variable: FORESTED
Label: Area with Land Cover Type = Forested
Units: sq km
Format: numeric 8.
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc/nlcd.html
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset. Forested was created by summing the area for the following
land cover types: Deciduous, Evergreen, Mixed Forest and
Shrubland.
Variable: AGRICULTURE
Label: Area with Land Cover Type = Agriculture
Units: sq km
Format: numeric 8.
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc/nlcd.html
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
A-2.1 -2
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset. Agriculture was created by summing the area for the
following land cover types: orchard, vineyard, other and grassland,
herbaceous, and pasture, hay and row crops and small grains and
fallow and urban/recreational grass.
WETLAND
Area with Land Cover Type = Wetland
sq km
numeric 8.
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992, http: / /www.epa.gov/mrlc7nlcd.html
The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive acreage statistics for each spatial referencing unit or
EDA, USGS (NWRC - Gulf Breeze Project Office) performed a
matrix overlay of our spatial referencing unit dataset with the NLCD
dataset. Wetland was created by summing the area for the following
land cover types: woody wetland and emergent, herbaceous wetland
A-2.1 -3
-------
Appendix A-3.1 EDA stressor loadings: metadata
Database: EDALOADS.XLS
Variable: EDA
Label: Estuarine Drainage Area Code
Units:
Format: uppercase alpha-numeric $5.
Source: http://spo.nos.noaa.gov/proiects/cads/ftp data download.html.
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system
Metadata: EDA is derived from the variable, EDASUBEDA, which is in the
CA&DS dataset, Reference EDA H Data, available from the above
download site. 203 ED As were chosen to be used in classification
Variable: TOTALN
Label: Total Nitrogen Load from Point and Non-point Sources
Units: kg/day
Format: numeric 10.4
Source: SPARROW Surface Water Quality Modeling Nutrients in
Watersheds of the Conterminous U.S.,
http://water.usgs.gov/nawqa/sparrow/wrr97/results.html; SAS
datasets: nexport, npexport, hucedanpexport; Excel files: nexport,
npexport
Metadata: Total N Load was modeled from point and nonpoint source water
quality data. The models empirically estimate the delivery of
nutrients to streams and the outlets of watersheds from point and
nonpoint sources. Estimates of stream transport (dependent variable
in the SPARROW models) are adjusted to reflect 1987 nutrient
inputs and long-term mean flow conditions (1970-1988), based on
records of the concentration and flow for the period 1974 to 1989.
Nitrogen nonpoint source data are for 1987. Point source data are for
the period 1977-81.
Variable: POINTN
Label: Total Nitrogen Load from Point Sources
Units: kg/day
Format: numeric 10.4
Source: SPARROW Surface Water Quality Modeling Nutrients in
Watersheds of the Conterminous U.S.,
http://water.usgs.gov/nawqa/sparrow/wrr97/results.htmh SAS
datasets: nexport, npexport, hucedanpexport; Excel files: nexport,
npexport
Metadata: Point source N load was modeled from point source water quality
data. The models empirically estimate the delivery of nutrients to
streams and the outlets of watersheds from point and nonpoint
sources. Estimates of stream transport (dependent variable in the
A-3.1 - 1
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
SPARROW models) are adjusted to reflect 1987 nutrient inputs and
long-term mean flow conditions (1970-1988), based on records of the
concentration and flow for the period 1974 to 1989. Point source
data are for the period 1977-81.
TOTALP
Total Phosphorus Load from Point and Non-point Sources
kg/day
numeric 10.4
SPARROW Surface Water Quality Modeling; Nutrients in
Watersheds of the Conterminous U.S.;
http://water.usgs.gov/nawqa/sparrow/wrr97/results.htmh SAS
datasets: pexport, npexport, hucedanpexport; Excel files: pexport,
npexport
Total P Load was modeled from point and nonpoint source water
quality data. The models empirically estimate the delivery of
nutrients to streams and the outlets of watersheds from point and
nonpoint sources. Estimates of stream transport (dependent variable
in the SPARROW models) are adjusted to reflect 1987 nutrient
inputs and long-term mean flow conditions (1970-1988), based on
records of the concentration and flow for the period 1974 to 1989.
Point source data are for the period 1977-81.
POINTP
Total Phosphorus Load from Point Sources Only
kg/day
numeric 10.4
SPARROW Surface Water Quality Modeling Nutrients in
Watersheds of the Conterminous U.S.;
http://water.usgs.gov/nawqa/sparrow/wrr97/results.htmh SAS
datasets: pexport, npexport, hucedanpexport; Excel files: pexport,
npexport
Point source P load was modeled from point source water quality
data. The models empirically estimate the delivery of nutrients to
streams and the outlets of watersheds from point and nonpoint
sources. Estimates of stream transport (dependent variable in the
SPARROW models) are adjusted to reflect 1987 nutrient inputs and
long-term mean flow conditions (1970-1988), based on records of the
concentration and flow for the period 1974 to 1989. Point source
data are for the period 1977-81.
A-3.1 -2
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
PAHPC
Principal Component - PAHs
numeric 10.4
EPA/OW BASINS Water Quality Data by HUG Permit Compliance
System Data;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: pcsload, loadpca; Excel files: pcsload
Loads of individual chemicals were derived from BASINS data.
BASINS data represents average concentrations by HUG. HUCs
were geo-referenced to ED As. Individual loads were calculated for
each EDA by averaging over time for each NPDES ID and then
summing across all NPDES IDs within a HUG and then summing by
EDA. All missing loads were assigned a 0. More information on how
pollutant loads are calculated from the Permit Compliance System
can be found at http://www.epa.gov/owmitnet/pcsguide.htm.
Principal Component Analysis was conducted on the full data set of
individual chemical loads. All loads were In-transformed prior to
analysis. Three principal components accounted for 75% of the
variance. The first principal component was weighted on PAHs.
Variable:
Label:
Units:
Format:
Source:
METALPC
Principal Component - Metals
numeric 10.4
EPA/OW BASINS Water Quality Data by HUC Permit Compliance
System Data;
http: / /www.epa.gov/waterscience/ftp /basins /gis data/hue /SAS
datasets: pcsload, loadpca; Excel files: pcsload
Metadata: Loads of individual chemicals were derived from BASINS data.
BASINS data represents average concentrations by HUC. HUCs
were geo-referenced to ED As. Individual loads were calculated for
each EDA by averaging over time for each NPDES ID and then
summing across all NPDES IDs within a HUC and then summing by
EDA. All missing loads were assigned a 0. More information on
how pollutant loads are calculated from the Permit Compliance
System can be found at
http: / /www.epa.gov/owmitnet/pcsguide.htm. Principal Component
Analysis was conducted on the full data set of individual chemical
loads. All loads were In-transformed prior to analysis. Three
principal components accounted for 75% of the variance. The first
principal component was weighted on metals.
A-3.1 -3
-------
Variable:
Label:
Units:
Format:
Source:
PESTPC
Principal Component - PAHs
numeric 10.4
EPA/OW BASINS Water Quality Data by HUG Permit Compliance
System Data;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: pcsload, loadpca; Excel files: pcsload
Metadata: Loads of individual chemicals were derived from BASINS data.
BASINS data represents average concentrations by HUG. HUCs
were geo-referenced to ED As. Individual loads were calculated for
each EDA by averaging over time for each NPDES ID and then
summing across all NPDES IDs within a HUG and then summing by
EDA. All missing loads were assigned a 0. More information on how
pollutant loads are calculated from the Permit Compliance System
can be found at http://www.epa.gov/owmitnet/pcsguide.htm.
Principal Component Analysis was conducted on the full data set of
individual chemical loads. All loads were In-transformed prior to
analysis. Three principal components accounted for 75% of the
variance. The first principal component was weighted on pesticides.
Variable:
Label:
Units:
Format:
Source:
SED_RANK
Relative ranking for the potential for sediment delivery
numeric 10.4
EPA/OW/OWOW Watershed Information Network Index of
Watershed Indicators;
http://www.epa.gov/wateratlas/geo/maplisthtml: SAS datasets:
sedranks, hucedasedranks; Excel files: sedranks
Metadata: Full metadata is located at http://www.epa.gov/eims/index.html and
found by going to EIMS Search > > Advanced Search > > Entry ID
= 1757. SEDIMENT DELIVERY TO RIVERS AND STREAMS
FROM CROPLAND AND PASTURELAND 1990-1995 was
estimated from two simulation model outputs: Hydrologic Unit
Modeling of the United States (HUMUS) and Soil and Water
Assessment Tool (SWAT). Soils characteristics for each subarea are
taken from the STATSGO soils database. A 30-year weather database
is available for each watershed. A process model incorporating
hydrology, weather, sedimentation, crop growth, and agricultural
management (SWAT—Soil and Water Assessment Tool) is applied to
each subarea to simulate the relationships among rainfall, runoff,
leaching, groundwater return flow, farm management practices, eros!
ion, and surface flow in rivers and streams. One of the outputs of
the model is average annual sediment delivery to rivers and streams
from sheet and rill erosion from cropland and pastureland, as shown
on this map.
A-3.1 -4
-------
Variable: TSSLOADPCS
Label: Total Suspended Solids Load
Units: kg/7r
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUG Permit Compliance
System Data;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: pcsload; Excel files: pcsload
Metadata: Total Suspended Solids Load (Storet Code=00530) was derived from
BASINS data. BASINS data represents average concentrations by HUG.
HUCs were geo-referenced to ED As. The total phosphorus load was
calculated for each EDA by averaging over time for each NPDES ID and
then summing across all NPDES IDs within a HUG and then summing by
EDA. All missing loads were assigned a 0. More information on how
pollutant loads are calculated from the Permit Compliance System can be
found at http://www.epa.gov/owmitnet/pcsguide.htm
Variable: TPLOADPCS
Label: Total Phosphorus Load
Units: kg/7r
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUG Permit Compliance
System Data;
http: / /www.epa.gov/waterscience/ftp /basins /gis data/hue /SAS
datasets: pcsload; Excel files: pcsload
Metadata: Total Phosphorus Load (Storet Code=00665) was derived from
BASINS data. BASINS data represents average concentrations by
HUG. HUCs were geo-referenced to ED As. The total phosphorus
load was calculated for each EDA by averaging over time for each
NPDES ID and then summing across all NPDES IDs within a HUG
and then summing by EDA. All missing loads were assigned a 0.
More information on how pollutant loads are calculated from the
Permit Compliance System can be found at
http: / /www.epa.gov/owmitnet/pcsguide.htm
Variable: TNLOADPCS
Label: Total Nitrogen Load
Units: kg/7r
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUG Permit Compliance
System Data;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: pcsload; Excel files: pcsload
Metadata: Total Nitrogen Load (Storet Code = 00600) was derived from
BASINS data. BASINS data represents average concentrations by
HUC. HUCs were geo-referenced to ED As. The total nitrogen load
was calculated for each EDA by averaging over time for each
NPDES ID and then summing across all NPDES IDs within a HUC
A-3.1 -5
-------
and then summing by EDA. All missing loads were assigned a 0.
More information on how pollutant loads are calculated from the
Permit Compliance System can be found at
http: / /www.epa.gov/owmitnet/pcsguide.htm
A-3.1 -6
-------
Appendix A-4.1 EDA modifying factors: metadata
Database:
EDAEXPOSURE.XLS
Variable: EDA
Label:
Units:
Format:
Source:
Metadata:
Estuarine Drainage Area Code
uppercase alpha-numeric $5.
http://spo.nos.noaa.gov/proiects/cads/ftp data download.html;
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system
EDA is derived from the variable, EDASUBEDA, which is in the
CA&DS dataset, Reference EDA H Data, available from the above
download site. 203 ED As were chosen to be used in classification
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
DIN_MGL
Dissolved Inorganic Nitrogen Concentration
mg/L
numeric 10.4
Environmental Monitoring and Assessment Program
EMAP)National Coastal Assessment (NCA) 2000 EPA/OW
BASINS Water Quality Data by HUC;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: emapnuts, basinwq; Excel files: emapnuts, basinwq,
"edanutrients calculations"
Nitrate+Nitrite (NO3+NO2), and Ammonia (NH4) were measured
at 888 coastal stations nationwide in the summer of 2000.
EMAP/NCA stations were geo-referenced to ED As and HUCs by
USGS/NWRC Gulf Breeze Project Office. DIN was calculated as
the sum of NH4 and NO2NO3. The average DIN concentration
was calculated for each EDA. When there was no EMAP data for an
EDA, BASINS data was used if available. From BASINS, DIN was
calculated as the sum of NH4_MGL and NO2NO3_MGL. BASINS
data represents average concentrations by HUC. HUCs were geo-
referenced to ED As. The average DIN concentration was calculated
for each EDA.
TKN_MGL
Total Kjeldahl Nitrogen Concentration
mg/L
numeric 10.4
EPA/OW BASINS Water Quality Data by HUC;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: basinwq; Excel files: basinwq,
TKN_MGL was derived from BASINS data. BASINS data
represents average concentrations by HUC. HUCs were geo-
A-4.1 - 1
-------
referenced to ED As. The average TKN concentration was calculated
for each EDA.
Variable: TP_MGL
Label: Total Phosphorus Concentration
Units: mg/L
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUG;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq, "edanutrients calculations"
Metadata: TP_MGL was derived from BASINS data. BASINS data represents
average concentrations by HUG. HUCs were geo-referenced to
ED As. The average TP concentration was calculated for each EDA.
Variable: TSS
Label:
Units:
Format:
Source:
Metadata:
Total Suspended Solids Concentration
mg/L
numeric 10.4
EPA/OW BASINS Water Quality Data by HUG;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS 7
TSS was derived from BASINS data. BASINS data represents
average concentrations by HUC. HUCs were geo-referenced to
ED As. The average TSS concentration was calculated for each
EDA.
Variable:
Label:
Units:
Format:
Source:
Metadata:
METALTUSUM
Sediment Metals Toxic Unit Sum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http:/7www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment metals were measured from bottom sediments of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). Marine sediment
toxicity values were derived from McDonald et al. (2000) and Long et
al. (1995) [see table below]. Toxic units for Cd, Cu, Cr, Hg, Ni, and
Zn at each EMAP station were calculated by dividing the measured
concentration by the appropriate toxicity value (e.g. Cd T.U. = [Cd] /
1.2). Toxic units for all metals were summed for each station. All
EMAP stations were geo-referenced to ED As and HUCs. The
average metal toxic unit sum forEDAs and HUCs were calculated by
averaging the toxic unit sums for all stations located within the EDA
or HUC.
A-4.1 -2
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
PESTTUSUM
Sediment Pesticides and Polychlorinated Biphenyls Toxic Unit Sum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http://www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment pesticides and total PCBs were measured from bottom
sediments of stations sampled during the summer through EMAP.
This includes stations in the Virginian Province (1990-1993),
Carolinian Province (1994-1997), West Indian Province (1995), and
Gulf of Mexico (1991-1994). This also includes stations sampled
through the National Coastal Assessment (Western Pilot, 1999; NCA,
2000). Marine sediment toxicity values were derived from McDonald
et al. (2000) and Long et al. (1995) [see table below]. Because the
toxicity values for dieldrin and endrin were in units of organic
carbon, concentrations of these pesticides were converted from
"ng/g dry weight" to "ug/g OC". Toxic units for total DDTs,
Dieldrin, Endrin, and total PCBs at each EMAP station were then
calculated by dividing the measured concentration by the appropriate
toxicity value (e.g. Dieldrin T.U. = [Cd] / 28). Toxic units for the
four contaminants were summed for each station. All EMAP
stations were geo-referenced to ED As and HUCs. The average
pesticide/PCB toxic unit sum for ED As and HUCs was calculated by
averaging the toxic unit sums for all stations located within the EDA
or HUC.
Variable:
Label:
Units:
Format:
Source:
Metadata:
PAHTUSUM
Sediment Polycyclic Aromatic Hydrocarbons Toxic Unit Sum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http://www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment PAHs were measured from bottom sediments of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). Marine sediment
toxicity values were derived from McDonald et al. (2000) and Long et
al. (1995) [see table below]. Because the toxicity values for all PAHs
were in units of organic carbon, PAH concentrations were converted
from "ng/g dry weight" to "ug/g OC". Toxic units for 16 PAHs at
each EMAP station were calculated by dividing the measured
concentration by the appropriate toxicity value (e.g. Acenaphthene
T.U. = [Acenaphthene] / 491). Toxic units for all PAHs were
summed for each station. All EMAP stations were geo-referenced to
A-4.1 -3
-------
ED As and HUCs. The average PAH toxic unit sum for ED As and
HUCs were calculated by averaging the toxic unit sums for all
stations located within the EDA or HUG.
Variable:
Label:
Units:
Format:
Source:
Metadata:
METALTUMAX
Sediment Metals Toxic Unit Maximum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http:/7www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment metals were measured from bottom sediments of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). Marine sediment
toxicity values were derived from McDonald et al. (2000) and Long et
al. (1995) [see table below]. Toxic units for Cd, Cu, Cr, Hg, Ni, and
Zn at each EMAP station were calculated by dividing the measured
concentration by the appropriate toxicity value (e.g. Cd T.U. = [Cd] /
1.2). The maximum toxic unit for all metals was calculated for each
station. All EMAP stations were geo-referenced to ED As and
HUCs. The average metal toxic unit maximum for ED As and HUCs
were calculated by averaging the toxic unit maxima for all stations
located within the EDA or HUC.
Variable:
Label:
Units:
Format:
Source:
Metadata:
PESTTUMAX
Sediment Pesticides and Polychlorinated Biphenyls Toxic Unit
Maximum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http:/7www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment pesticides and total PCBs were measured from bottom
sediments of stations sampled during the summer through EMAP.
This includes stations in the Virginian Province (1990-1993),
Carolinian Province (1994-1997), West Indian Province (1995), and
Gulf of Mexico (1991-1994). This also includes stations sampled
through the National Coastal Assessment (Western Pilot, 1999; NCA,
2000). Marine sediment toxicity values were derived from McDonald
et al. (2000) and Long et al. (1995) [see table below]. Because the
toxicity values for dieldrin and endrin were in units of organic
carbon, concentrations of these pesticides were converted from
"ng/g dry weight" to "ug/g OC". Toxic units for total DDTs,
Dieldrin, Endrin, and total PCBs at each EMAP station were then
A-4.1 -4
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
calculated by dividing the measured concentration by the appropriate
toxicity value (e.g. Dieldrin T.U. = [Cd] / 28). The maximum toxic
unit for the four contaminants was calculated for each station. All
EMAP stations were geo-referenced to ED As and HUCs. The
average pesticide/PCB toxic unit maximum for ED As and HUCs
was calculated by averaging the toxic unit maxima for all stations
located within the EDA or HUG.
PAHTUMAX
Sediment Polycyclic Aromatic Hydrocarbons Toxic Unit Maximum
no units
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
http:/7www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Sediment PAHs were measured from bottom sediments of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). Marine sediment
toxicity values were derived from McDonald et al. (2000) and Long et
al. (1995) [see table below]. Because the toxicity values for all PAHs
were in units of organic carbon, PAH concentrations were converted
from "ng/g dry weight" to "ug/g OC". Toxic units for 16 PAHs at
each EMAP station were calculated by dividing the measured
concentration by the appropriate toxicity value (e.g. Acenaphthene
T.U. = [Acenaphthene] / 491). The maximum toxic unit for all
PAHs was calculated for each station. All EMAP stations were geo-
referenced to ED As and HUCs. The average PAH toxic unit
maximum for ED As and HUCs were calculated by averaging the
toxic unit maxima for all stations located within the EDA or HUC.
Variable: PRSKF_LEA
Label: Potential Leaching Concentration at the Bottom of the Root Zone
Exceeds at Least One Water Quality Threshold for Fish
Units: % Acres
Format: numeric 10.4
Source: USDA NRCS National Pesticide Loss Database;
http://www.nrcs.usda.gov/technical/land/pubs/gosstext.htmh SAS
datasets: pestlriskfish; Excel files: riskfjea, pestlriskfish
Metadata: A National Pesticide Loss Database was created for use as a look-up
table for estimates of pesticide losses from farm fields in leachate and
runoff. Pesticide leaching and runoff losses were estimated using the
pesticide fate and transport model GLEAMS 1. Pesticide leaching
was movement beyond the bottom of the root-zone. Final pesticide
loss results are reported as 1) the percentage of total mass of
pesticide applied, and 2) the annual concentration of pesticide leaving
A-4.1 -5
-------
the field, expressed as the percentage of total mass of pesticide
applied per million parts of water or sediment. Mass loss and annual
concentration were calculated for each pesticide at each sample point.
Mass loss estimates were then aggregated over acres treated in each
watershed to produce national maps. Concentrations were compared
to water quality thresholds to derive a measure of environmental risk
at each NRI sample point. Maximum Acceptable Toxicant
Concentrations (MATCs) were used as "safe" thresholds for fish,
which were calculated using toxicity data published by EPA. The
extent to which the concentration exceeded the threshold was used as
a measure of risk for each pesticide. PRSKF_LEA is an index of the
percent of the land in the watershed (nonfederal rural land) where the
potential leaching concentration at the bottom of the root zone
exceeds at least one water quality threshold for fish.
Variable: PRSKF_RUN
Label: Potential Runoff Concentration at the Edge of the Field Exceeds at
Least One Water Quality Threshold for Fish
Units: % Acres
Format: numeric 10.4
Source: USDA NRCS National Pesticide Loss Database;
http://www.nrcs.usda.gov/technical/land/pubs/gosstext.htmh SAS
datasets: pestrriskfish; Excel files: riskf_run, pestrriskfish
Metadata: A National Pesticide Loss Database was created for use as a look-up
table for estimates of pesticide losses from farm fields in leachate and
runoff. Pesticide leaching and runoff losses were estimated using the
pesticide fate and transport model GLEAMS 1. Pesticide leaching
was movement beyond the bottom of the root-zone. Final pesticide
loss results are reported as 1) the percentage of total mass of
pesticide applied, and 2) the annual concentration of pesticide leaving
the field, expressed as the percentage of total mass of pesticide
applied per million parts of water or sediment. Mass loss and annual
concentration were calculated for each pesticide at each sample point.
Mass loss estimates were then aggregated over acres treated in each
watershed to produce national maps. Concentrations were compared
to water quality thresholds to derive a measure of environmental risk
at each NRI sample point. Maximum Acceptable Toxicant
Concentrations (MATCs) were used as "safe" thresholds for fish,
which were calculated using toxicity data published by EPA. The
extent to which the concentration exceeded the threshold was used as
a measure of risk for each pesticide. PRSKF_RUN is an index of the
percent of the land in the watershed (nonfederal rural land) where the
potential runoff concentration at the edge of the field exceeds at least
one water quality threshold for fish.
Variable: PRSKH_LEA
Label: Potential Leaching Concentration at the Bottom of the Root Zone
Exceeds at Least One Water Quality Threshold for Humans
A-4.1 -6
-------
Units: % Acres
Format: numeric 10.4
Source: USDA NRCS National Pesticide Loss Database;
http://www.nrcs.usda.gov/technical/land/pubs/gosstext.htmh SAS
datasets: pestlriskhuman; Excel files: riskh_lea, pestlriskhuman
Metadata: A National Pesticide Loss Database was created for use as a look-up
table for estimates of pesticide losses from farm fields in leachate and
runoff. Pesticide leaching and runoff losses were estimated using the
pesticide fate and transport model GLEAMS 1. Pesticide leaching
was movement beyond the bottom of the root-zone. Final pesticide
loss results are reported as 1) the percentage of total mass of
pesticide applied, and 2) the annual concentration of pesticide leaving
the field, expressed as the percentage of total mass of pesticide
applied per million parts of water or sediment. Mass loss and annual
concentration were calculated for each pesticide at each sample point.
Mass loss estimates were then aggregated over acres treated in each
watershed to produce national maps. Concentrations were compared
to water quality thresholds to derive a measure of environmental risk
at each NRI sample point. Health Advisories (HAs) and Maximum
Contaminant Levels (MCLs) were used for humans for pesticides
that have been assigned drinking water standards by EPA. For other
pesticides, "safe" thresholds were estimated from EPA Reference
Dose values and cancer slope data. The extent to which the
concentration exceeded the threshold was used as a measure of risk
for each pesticide. PRSKH_LEA is an index of the percent of the
land in the watershed (nonfederal rural land) where the potential
leaching concentration at the bottom of the root zone exceeds at
least one water quality threshold for humans.
Variable: PRSKH_RUN
Label: Potential Runoff Concentration at the Edge of the Field Exceeds at
Least One Water Quality Threshold for Fish
Units: % Acres
Format: numeric 10.4
Source: USDA NRCS National Pesticide Loss Database;
http://www.nrcs.usda.gov/technical/land/pubs/gosstext.htmh SAS
datasets: pestrriskhuman; Excel files: riskh_run, pestrriskhuman
Metadata: A National Pesticide Loss Database was created for use as a look-up
table for estimates of pesticide losses from farm fields in leachate and
runoff. Pesticide leaching and runoff losses were estimated using the
pesticide fate and transport model GLEAMS 1. Pesticide leaching
was movement beyond the bottom of the root-zone. Final pesticide
loss results are reported as 1) the percentage of total mass of
pesticide applied, and 2) the annual concentration of pesticide leaving
the field, expressed as the percentage of total mass of pesticide
applied per million parts of water or sediment. Mass loss and annual
concentration were calculated for each pesticide at each sample point.
A-4.1 -7
-------
Mass loss estimates were then aggregated over acres treated in each
watershed to produce national maps. Concentrations were compared
to water quality thresholds to derive a measure of environmental risk
at each NRI sample point. Health \Advisories (HAs) and Maximum
Contaminant Levels (MCLs) were used for humans for pesticides
that have been assigned drinking water standards by EPA. For other
pesticides, "safe" thresholds were estimated from EPA Reference
Dose values and cancer slope data. The extent to which the
concentration exceeded the threshold was used as a measure of risk
for each pesticide. PRSKH_RUN is an index of the percent of the
land in the watershed (nonfederal rural land) where the potential
runoff concentration at the edge of the field exceeds at least one
water quality threshold for humans.
A-4.1 -8
-------
Appendix A-5.1 EDA modifying factors: metadata
Database:
EDAMODIFIERS.XLS
Variable: EDA
Label:
Units:
Format:
Source:
Metadata:
Estuarine Drainage Area Code
uppercase alpha-numeric $5.
http://spo.nos.noaa.gov/proiects/cads/ftp data download.html
NOAA/NOS/Special Projects Office Coastal Assessment & Data
Synthesis (CA&DS) system
EDA is derived from the variable, EDASUBEDA, which is in the
CA&DS dataset, Reference EDA H Data, available from the above
download site. 203 ED As were chosen to be used in classification
Variable: TOC
Label:
Units:
Format:
Source:
Metadata:
Variable: AVS
Label:
Units:
Format:
Source:
Metadata:
Total Organic Carbon Concentration in Sediment
%
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP)
http: / /www.epa.gov/emap; SAS datasets: emapsedchem; Excel
files: emapsedchem
Total Organic Carbon (TOC) was measured from bottom sediments
of stations sampled during the summer through EMAP. This
includes stations in the Virginian Province (1990-1993), Carolinian
Province (1994-1997), West Indian Province (1995), and Gulf of
Mexico (1991-1994). This also includes stations sampled through the
National Coastal Assessment (Western Pilot, 1999; NCA, 2000). All
EMAP stations were geo-referenced to ED As and HUCs. TOC
represents the average DO across depth, space and time for each
EDA.
Acid- Volatile Sulfide Concentration in Sediment
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP),
http: / /www.epa.gov/emap; SAS datasets: emapsedchem; Excel files:
emapsedchem
Acid- Volatile Sulfide (AVS) was measured from bottom sediments of
stations sampled during the summer through EMAP. This includes
stations in the Virginian Province (1990-1993), Carolinian Province
(1994-1997), West Indian Province (1995), and Gulf of Mexico
(1991-1994). This also includes stations sampled through the
National Coastal Assessment (Western Pilot, 1999; NCA, 2000). All
EMAP stations were geo-referenced to ED As and HUCs. AVS
A-5.1 - 1
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
represents the average AVS across depth, space and time for each
EDA.
AV_DO
Average Dissolved Oxygen Concentration in Water
mg/L
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP);
National Coastal Assessment (NCA) 2000; EPA/OW BASINS
Water Quality Data by HUC;
http:/7www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: emapwq, basinwq; Excel files: emapwq, basinwq, "WQ
CALCS"
Dissolved Oxygen (DO) was measured at surface and bottom depths
of stations sampled during the summer through EMAP. This
includes stations in the Virginian Province (1990-1993), Carolinian
Province (1994-1997), West Indian Province (1995), and Gulf of
Mexico (1991-1994). This also includes stations sampled through the
National Coastal Assessment (Western Pilot, 1999; NCA, 2000).
EMAP/NCA stations were geo-referenced to ED As and HUCs by
USGS/NWRC Gulf Breeze Project Office. The average DO
concentration was calculated for each EDA. When there was no
EMAP data for an EDA, BASINS data was used if available. From
BASINS, DO_MGLwas used. BASINS data represents average
concentrations by HUC. HUCs were geo-referenced to ED As. The
average DO concentration was calculated for each EDA.
AV_SAL
Average Salinity in Water
ppt
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP),
http:/7www.epa.gov/emap; SAS datasets: emapwq; Excel files:
emapwq
Salinity was measured at surface and bottom depths of stations
sampled during the summer through EMAP. This includes stations in
the Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). All EMAP stations
were geo-referenced to ED As and HUCs. AV_SAL represents the
average salinity across depth, space and time for each EDA.
Variable: AV_PH
Label: Average pH in Water
Units:
Format: numeric 10.4
A-5.1 -2
-------
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
TSS
Environmental Monitoring and Assessment Program (EMAP)
National Coastal Assessment (NCA) 2000; EPA/OW BASINS
Water Quality Data by HUG;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: emapwq, basinwq; Excel files: emapwq, basinwq, "WQ
CALCS"
PH was measured at surface and bottom depths of stations sampled
during the summer through EMAP. This includes stations in the
Virginian Province (1990-1993), Carolinian Province (1994-1997),
West Indian Province (1995), and Gulf of Mexico (1991-1994). This
also includes stations sampled through the National Coastal
Assessment (Western Pilot, 1999; NCA, 2000). EMAP/NCA
stations were geo-referenced to ED As and HUCs by USGS/NWRC
Gulf Breeze Project Office. The average pH was calculated for each
EDA. When there was no EMAP data for an EDA, BASINS data
was used if available. From BASINS, PH was used. BASINS data
represents average concentrations by HUC. HUCs were geo-
referenced to ED As. The average pH was calculated for each EDA.
AV_TEMP
Average Water Temperature
degrees C
numeric 10.4
Environmental Monitoring and Assessment Program (EMAP)
National Coastal Assessment (NCA) 2000; EPA/OW BASINS
Water Quality Data by HUC;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/; SAS
datasets: emapwq, basinwq; Excel files: emapwq, basinwq, "WQ
CALCS"
Water Temperature was measured at surface and bottom depths
ofstations sampled during the summer through EMAP. This includes
stations in the Virginian Province (1990-1993), Carolinian Province
(1994-1997), West Indian Province (1995), and Gulf of Mexico
(1991-1994). This also includes stations sampled through the
National Coastal Assessment (Western Pilot, 1999; NCA, 2000).
EMAP/NCA stations were geo-referenced to ED As and HUCs by
USGS/NWRC Gulf Breeze Project Office. The average temperature
was calculated for each EDA. When there was no EMAP data for an
EDA, BASINS data was used if available. From BASINS,
WTRTEMP_C was used. BASINS data represents average
concentrations by HUC. HUCs were geo-referenced to ED As. The
average water temperature was calculated for each EDA.
Total Suspended Solids Concentration in Water
mg/L
numeric 10.4
A-5.1 -3
-------
Source: EPA/OW BASINS Water Quality Data by HUG;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
Metadata: TSS was derived from BASINS data. BASINS data represents
average concentrations by HUG. HUCs were geo-referenced to
ED As. The average TSS concentration was calculated for each EDA.
Variable: HARDNESS
Label: Hardness as CACO3 in Water
Units: mg/L
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUG;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
Metadata: HARDNESS was derived from BASINS data. BASINS data
represents average concentrations by HUC. HUCs were geo-
referenced to ED As. The average HARDNESS was calculated for
each EDA.
Variable: ALKALINITY
Label: Total Alkalinity as CACO3 in Water
Units: mg/L
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUC;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
Metadata: ALKALINITY was derived from BASINS data. BASINS data
represents average concentrations by HUC. HUCs were geo-
referenced to ED As. The average ALKALINITY was calculated for
each EDA.
Variable: CHLORIDE
Label: Total Chloride Concentration in Water
Units: mg/L
Format: numeric 10.4
Source: EPA/OW BASINS Water Quality Data by HUC;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
Metadata: CHLORIDE was derived from BASINS data. BASINS data
represents average concentrations by HUC. HUCs were geo-
referenced to ED As. The average CHLORIDE concentration was
calculated for each EDA.
Variable: COND
Label: Specific Conductance in Water
Units: |J,mhos/cm
Format: numeric 10.4
A-5.1 -4
-------
Source: EPA/OW BASINS Water Quality Data by HUG;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
Metadata: COND was derived from BASINS data. BASINS data represents
average concentrations by HUG. HUCs were geo-referenced to
ED As. The average COND concentration was calculated for each
EDA.
Variable: SO4
Label:
Units:
Format:
Source:
Metadata:
Sulfate Concentration in Water
mg/L
numeric 10.4
EPA/OW BASINS Water Quality Data by HUC;
http://www.epa.gov/waterscience/ftp/basins/gis data/hue/: SAS
datasets: basinwq; Excel files: basinwq,
SO4 was derived from BASINS data. BASINS data represents
average concentrations by HUC. HUCs were geo-referenced to
ED As. The average SO4 concentration was calculated for each
EDA.
A-5.1 -5
-------
Appendix B-l.l. Metadata for Great Lakes R-EMAP Coastal Riverine Wetland Watershed
Classification Database
Database: APPENDIXB1 l.XLS
Variable:
Label:
Units:
Format:
Source:
Metadata:
WSHDAREA_KM2
Watershed area
kilometer2
numeric, 8.2
U.S. Environmental Protection Agency, Mid-Continent Ecology
Division, Duluth, MN, REMAP03WSHDS,
detenbeck.naomi@epa.gov
Watershed boundaries for 155 Great Lakes coastal riverine wetlands
sampled for a EPA Region V Regional Assessment and Monitoring
Program (R-EMAP) project were digitized in ArcMap using digital
raster graphics (DRGs, 1:24,000) as backdrops. Existing watershed
boundaries (National Watershed Boundary Database, state watershed
boundary databases, and watershed boundaries derived by USGS
EROS Data Center through an automated process) were used as a
base coverage when available, and modified so that the watershed
outlet was consistent with R-EMAP sampling points. Watershed
areas were calculated in Arclnfo in meter2 and converted to square
kilometers by dividing by 106.
Variable:
Label:
Units:
Format:
Source:
FWATER
Fraction open water in watershed
Fraction, unitless
numeric 6.5
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Open water class
consists of the sum of areas with grid codes: 10-11
Variable: FURBAN
Label: Fraction urban land in watershed
Units: Fraction, unitless
Format: numeric 6.5
B-l.l - 1
-------
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Urban land-use class
consists of the sum of areas with grid codes: 21-23, 84-85
Variable: FBARREN
Label: Fraction barren land in watershed
Units: Fraction, unitless
Format: numeric 6.5
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Barren cover class
consists of the sum of areas with grid codes: 31-33
Variable: FFOREST
Label: Fraction forested land in watershed
Units: Fraction, unitless
Format: numeric 6.5
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: 7 7www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
B-l.l -2
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Forested land-cover
class consists of the sum of areas with grid codes: 41-43
FSHRUB
Fraction shrubland in watershed
Fraction, unitless
numeric 6.5
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / 7www.epa.gov/mrlc/nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Shrub land-cover class
consists of the sum of areas with grid codes: 51
FGRASS
Fraction nonagricultural grassland in watershed
Fraction, unitless
numeric 6.5
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc/nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Non-agricultural
grassland-cover class consists of the sum of areas with grid codes: 71
Variable: FAGRIC
Label: Fraction agricultural land in watershed
Units: Fraction, unitless
Format: numeric 6.5
Variable:
Label:
Units:
Format:
Source:
Metadata:
B-l.l -3
-------
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Agricultural land-use
class consisted of the sum of areas with grid codes: 81- 85
Variable: FNLCDWTLD
Label: Fraction wetland area in watershed, NLCD-based
Units: Fraction, unitless
Format: numeric 6.5
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Wetland land-cover
class consists of the sum of areas with grid codes: 91-92
Variable: FLWINTRES
Label: Fraction low intensity residential area in watershed
Units: Fraction, unitless
Format: numeric 6.5
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: 7 7www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
B-l.l -4
-------
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Low intensity
residential land-use class consists of the sum of areas with grid codes:
21
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
FHINTRES
Fraction high intensity residential area in watershed
Fraction, unitless
numeric 6.5
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. High intensity
residential land-use class consists of the sum of areas with grid codes:
22
FCOMINDTR
Fraction commercial, industrial, and transportation area in watershed
Fraction, unitless
numeric 6.5
Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages.
Commercial/industrial/transportation land-use class consists of the
sum of areas with grid codes: 23
B-l.l -5
-------
Variable: FMINING
Label: Fraction mined area in watershed
Units: Fraction, unitless
Format: numeric 6.5
Source: Multi-Resolution Land Characteristics Consortium National Land
Cover Data — 1992 http: / /www.epa.gov/mrlc7nlcd.html and U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN, derived from US EPA MED-Duluth watershed
boundaries
Metadata: The USGS and the USEPA created a nationwide land cover dataset
(National Land Cover Data - NLCD) for the conterminous U.S.
based on early to mid-1990s 30-meter Landsat Thematic Mapper
(TM) satellite imagery. The NLCD consists of 21 land cover
categories classified in a consistent manner across the conterminous
U.S. To derive area of different land-cover/land-use classes within
each wetland watershed, US EPA MED-Duluth intersected
watershed boundaries with NLCD coverages. Mined land-use class
consists of the sum of areas with grid codes: 32
Variable: FSTORAGE
Label: Fraction watershed storage area (lakes and wetland area)
Units: Fraction, unitless
Format: numeric 6.5
Source: Derived from following digital wetland inventory databases: National
Wetlands Inventory (NWI, http://wetiands.fws.gov/ ). Wisconsin
Wetlands Inventory (WWI,
http: / /www.dnr.state.wi.us /org/water/flip /wetlands /mapping.shtml.
http: / /wisclinc.state.wi.us /datadisc/wimeta browser.html see Wetlands of
Wisconsin), and Ohio Wetlands Inventory
(http: / /www.dnr.state.oh.us /wetlands 7mapping.htm)
Metadata: Calculated from digital wetlands inventory coverages as fraction of
area occupied by lacustrine deepwater plus palustrine wetland classes
Variable: FIMPERV
Label: Estimated fraction impervious surface area in watershed
Units: Fraction, unitless
Format: numeric 6.5
Source: Derived from Multi-Resolution Land Characteristics Consortium
National Land Cover Data - 1992
http://www.epa.gov/mrlc/nlcd.html and U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN,
derived from US EPA MED-Duluth watershed boundaries
Metadata: Estimated from NLCD database, using estimates of impervious land
in each class for weighting factors. Fraction impervious = (0.55 *
fraction low intensity residential) + (0.9 * fraction high intensity
residential) + fraction commercial/industrial/transportation. (
B-l.l -6
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
FHYDGA
Fraction soil in hydrologic group A (high infiltration rate) in
watershed
Fraction, unitless
numeric, 6.5
Derived from U.S. Department of Agriculture State Soil Geographic
Database (STATSGO, http://www.ftw.nrcs.usda.gov/stat data.html
) and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
Fraction soils in hydrologic group A was estimated from STATSGO
by averaging percent soil components in hydrologic soil group A for
each Map Unit (MUID) with percent soil components (PCTCOMP)
from the COMPLAYER.DBF files as a weighting factor, then
averaging percent hydrologic group A across the watershed using
MUID area as a weighting factor. Soil components with a hydrologic
group of A/D were assumed to be drained (group A) at a frequency
proportional to the co-occurrence of agricultural land-use by MUID.
FHYDGAB
Fraction soil in hydrologic group A (high infiltration rate) or B
(moderate infiltration rate) in watershed
Fraction, unitless
numeric, 6.5
Derived from U.S. Department of Agriculture State Soil Geographic
Database (STATSGO, http://www.ftw.nrcs.usda.gov/stat data.html
) and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
Fraction soils in hydrologic groups A and B were estimated from
STATSGO by averaging percent soil components in hydrologic soil
groups A and B for each Map Unit (MUID) with percent soil
components (PCTCOMP) from the COMPLAYER.DBF files as a
weighting factor, then averaging percent hydrologic groups A and B
across the watershed using MUID area as a weighting factor. Soil
components with a hydrologic group of A/D or B/D were assumed
to be drained (groups A or B) at a frequency proportional to the co-
occurrence of agricultural land-use by MUID.
AVMNPERM
Average minimum soil permeability in watershed
inches/hour
numeric, 6.2
Derived from U.S. Department of Agriculture State Soil Geographic
Database (STATSGO, http://www.ftw.nrcs.usda.gov/stat data.html
) and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
B-l.l -7
-------
Metadata: Minimum soil permeability was calculated for each soil component
by selecting the minimum soil permeability across soil layers, then
averaging by map unit (MUID) weighting by percent soil component
(PCTCOMP), and finally averaging across the watershed weighting
by map unit area.
AVSLOPE
Average watershed slope
percent
numeric, 6.2
Derived from U.S. Department of Agriculture State Soil Geographic
Database (STATSGO, http://www.ftw.nrcs.usda.gov/stat data.html
) and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
Metadata: Average watershed slope was derived from STATSGO by extracting
slope by map unit (MUID), then averaging across watersheds using
map unit area as a weighting factor.
Variable:
Label:
Units:
Format:
Source:
Variable:
Label:
Units:
Format:
Source:
WLVOLUME
Total wetland plus lake storage volume per watershed
meters3
numeric, 12.
Derived from following digital wetland inventory databases: National
Wetlands Inventory (NWI, http://wetlands.fws.gov/ ). Wisconsin
Wetlands Inventory (WWI,
http: / /www.dnr.state.wi.us /org/water/flip /wetlands /mapping.shtml.
http: / /wisclinc.state.wi.us /datadisc/wimeta browser.html see Wetlands
of Wisconsin), and Ohio Wetlands Inventory
(http: / /www.dnr.state.oh.us /wetlands 7mapping.htm)
Metadata: Wetland plus lake storage volume was derived by multiplying the area
of each wetland type by an appropriate depth, based on descriptions
found in wetland inventory metadata.
Variable:
Label:
Units:
Format:
Source:
124 2
Metadata:
Average rainfall intensity for 2-year, 24-hour event in watershed
water depth, inches/24 hours
numeric, 4.2
U.S. Northeast 2-Year 24-Hour Rain Event (neus2y24hcnt), derived
by U.S. Environmental Protection Agency from Wilks and Cember
(1993) and U.S. Midwest 2-Year 24-Hour Ram Event
(mwus2y24hcnt) derived by U.S. Environmental Protection Agency
from Huff and Angel (1992).
NEUS2Y24HCNT was georeferenced & vectorized from scanned
image of "Map 1, 2-yr return period, 1-day ppt accumulation" from
Wilks, D.S. & R.P. Cember, Atlas of Ppt Extremes for the NE U.S.
& SE, Northeast Regional Climate Center Research Publ. RR93-5, 40
pp. MWUS2Y24HCNT was georeferenced and vectorized from
B-l.l -8
-------
scanned image of [Figure 6] Spatial distribution of 2-year 24-hour
rainfall events (inches). Huff, Floyd A., and James R. Angel. Rainfall
Frequency Atlas of the Midwest. Illinois State Water Survey,
Champaign, Bulletin 71, 1992.
Variable: I24_2MM
Label: Average rainfall intensity for 2-year, 24hour event in watershed,
metric
Units: water depth, millimeters/24 hours
Format: numeric, 4.1
Source: U.S. Northeast 2-Year 24-Hour Ram Event (NEUS2Y24HCNT),
derived by U.S. Environmental Protection Agency from Wilks and
Cember (1993) and U.S. Midwest 2-Year 24-Hour Ram Event
(MWUS2Y24HCNT) derived by U.S. Environmental Protection
Agency from Huff and Angel (1992).
Metadata: NEUS2Y24HCNT was georeferenced & vectorized from scanned
image of "Map 1, 2-yr return period, 1-day ppt accumulation" from
Wilks, D.S. & R.P. Cember, Atlas of Ppt Extremes for the NE U.S.
& SE, Northeast Regional Climate Center Research Publ. RR93-5, 40
pp. MWUS2Y24HCNT was georeferenced and vectorized from
scanned image of [Figure 6] Spatial distribution of 2-year 24-hour
rainfall events (inches). Huff, Floyd A., and James R. Angel. Rainfall
Frequency Atlas of the Midwest. Illinois State Water Survey,
Champaign, Bulletin 71, 1992. Inches were converted to millimeters
using a conversion factor of 25.4.
SNWTOTL
Average estimated snowfall per year in watershed, water equivalents
depth in mm, water equivalents
numeric, 8.1
Parameter-elevation Regressions on Independent Slopes Model
database (PRISM, Climate Source, Corvallis, OR,
http://www.climatesource.com/us/fact sheets/meta snowfall us.ht
ml)
Metadata: Average total snowfall was estimated by intersecting PRISM coverage
for annual snowfall with coastal wetland watershed boundaries which
were derived by US EPA MED-Duluth.
Variable:
Label:
Units:
Format:
Source:
Variable:
Label:
Units:
Format:
Source:
CN2
Runoff curve number 2 for watershed
Unities s
numeric, 5.1
Derived from Derived from Multi-Resolution Land Characteristics
Consortium National Land Cover Data — 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html) and from coastal
B-l.l -9
-------
wetland watershed boundaries derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
Metadata: Calculated based on USDA Soil Conservation Service curve number
method, using curve numbers for combinations of soil hydrologic
groups and major land-use classes based on tables in Soil and Water
Assessment Tool (SWAT) model
(http://www.brc.tamus.edu/swat/swatdoc.html):
CNA = ((77 * FBARREN) + (61.5 * FURBAN) + (50.8 * FAGRIC) +
(25 * FFOREST))/TOTCLASS
CNB = ((86 * FBARREN) + (76.5 * FURBAN) + (68 * FAGRIC) +
(55 * FFOREST))/TOTCLASS;
CNC = ((91 * FBARREN) + (84.5 * FURBAN) + (78.5 * FAGRIC) +
(70 * FFOREST))/TOTCLASS;
CND = ((94 * FBARREN) + (88 * FURBAN) + (83.5 * FAGRIC)
+ (77 * FFOREST))/TOTCLASS;
then averaging across watershed using map unit area (MUID) as a
weighting factor
Variable:
Label:
Units:
Format:
Source:
CN3
Metadata:
Variable:
Label:
Units:
Format:
Source:
CN2S
Runoff curve number 3 for watershed
Unities s
numeric, 5.1
Derived from Derived from Multi-Resolution Land Characteristics
Consortium National Land Cover Data — 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html) and from coastal
wetland watershed boundaries derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
Calculated based on USDA Soil Conservation Service curve number
method, based on documentation in Soil and Water Assessment Tool
(SWAT) model (http://www.brc.tamus.edu/swat/swatdoc.html):
CN3 = CN2 * exp (0.00673 * (100 _ CN2)
Slope-corrected runoff curve number 2 for watershed
Unities s
numeric, 5.1
Derived from Derived from Multi-Resolution Land Characteristics
Consortium National Land Cover Data — 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html) and from coastal
wetland watershed boundaries derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
B-l.l - 10
-------
Metadata:
Variable:
Label:
Units:
Format:
Source:
CN3S
Metadata:
Calculated based on USDA Soil Conservation Service curve number
method, based on documentation in Soil and Water Assessment Tool
(SWAT) model (http://www.brc.tamus.edu/swat/swatdoc.html):
CN2S = ((1/3) * (CN3-CN2) * (1 - (2 * e
(_13.86 * avslope/100)
'))) + CN2
Slope-corrected runoff curve number 3 for watershed
Unities s
numeric, 5.1
Derived from Derived from Multi-Resolution Land Characteristics
Consortium National Land Cover Data — 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html) and from coastal
wetland watershed boundaries derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
Calculated based on USDA Soil Conservation Service curve number
method, based on documentation in Soil and Water Assessment Tool
(SWAT) model (http://www.brc.tamus.edu/swat/swatdoc.html):
CN3S = CN2S * e1
Variable: S
Label:
Units:
Format:
Source:
Estimated soil storage compartment associated with 2-year, 24-hour
rainfall event
depth in millimeters
4.1
Derived from Multi-Resolution Land Characteristics Consortium
National Land Cover Data - 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html), 2-year 24-hour
rainfall intensity (Wilks and Cember, 1993; Huff and Angel, 1992)
and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
Calculated based on USDA Soil Conservation Service curve number
method, based on documentation in Soil and Water Assessment Tool
(SWAT) model (http://www.brc.tamus.edu/swat/swatdoc.html):
S = 254* ((100/CN3S) - 1),
where S = soil storage component
CN3S = curve number for average soil moisture
conditions, corrected for watershed slope
Variable: Q2_24
Label: Estimated runoff associated with 2-year, 24hour rainfall event per
watershed
Units: depth in millimeters
Format: 4.1
Metadata:
B-l.l - 11
-------
Source: Derived from Multi-Resolution Land Characteristics Consortium
National Land Cover Data - 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department of
Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html), 2-year 24-hour
rainfall intensity (Wilks and Cember, 1993; Huff and Angel, 1992)
and from coastal wetland watershed boundaries derived by U.S.
Environmental Protection Agency, Mid-Continent Ecology Division,
Duluth, MN.
Metadata: Calculated based on USDA Soil Conservation Service curve number
method, based on documentation in Soil and Water Assessment Tool
(SWAT) model (http://www.brc.tamus.edu/swat/swatdoc.html):
S = 254* ((100/CN3S) - 1)
if Rmm > (0.2 * S) then Q = ((Rmm - (0.2*S))**2)/(Rmm + (0.8*S))
if Rmm le (0.2 * S) then Q = 0,
where Rmm = rainfall from 2-year, 24-hour event (mm)
S = soil storage component (mm), and
Q = runoff (mm) associated with 2-year, 24-hour
event
Variable: RVcum
Label: Estimated runoff volume associated with 2-year, 24hour rainfall
event per watershed
Units: depth in millimeters
format: 12.
Source: Derived from Multi-Resolution Land Characteristics
Consortium National Land Cover Data — 1992
http://www.epa.gov/mrlc/nlcd.html. from U.S. Department
of Agriculture State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html), 2-year 24-hour
rainfall intensity (Wilks and Cember, 1993; Huff and Angel,
1992) and from coastal wetland watershed boundaries derived
by U.S. Environmental Protection Agency, Mid-Continent
Ecology Division, Duluth, MN.
Metadata: Calculated based on USDA Soil Conservation Service curve
number method, based on documentation in Soil and Water
Assessment Tool (SWAT) model
(http://www.brc.tamus.edu/swat/swatdoc.html):
RVcum = (Q/1000) * WSHDAREA
where Q = runoff depth associated with 2-year, 24-
hour precipitation even
RVcum = cumulative runoff volume
Variable: RDFLINDX
Label: Watershed index of flow responsiveness, rain events
Units: Unities s
format: numeric, 6.2
B-l.l - 12
-------
Source:
Metadata:
Derived from Multi-Resolution Land Characteristics Consortium
National Land Cover Data - 1992
http://www.epa.gov/mrlc/nlcd.html. U.S. Department of Agriculture
State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html), 2-year 24-hour rainfall
intensity (Wilks and Cember, 1993; Huff and Angel, 1992) and from
wetland volumes based on digital wetland inventory databases:
National Wetlands Inventory (NWI, http://wetlands.fws.gov/ ).
Wisconsin Wetlands Inventory (WWI,
http: / /www.dnr.state.wi.us /org/water/flip /wetlands /mapping.shtml.
http: / /wisclinc.state.wi.us /datadisc/wimeta browser.html see
Wetlands of Wisconsin), and Ohio Wetlands Inventory
(http://www.dnr.state.oh.us/wetlands/mapping.htm) and coastal
wetland watershed boundaries, derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
The watershed index of flow responsiveness for rainfall events is
calculated as the ratio of potential runoff volume from a 2-year, 24-
hour event to watershed depressional storage volume.
Variable:
Label:
Units:
Format:
Source:
Metadata:
SNFLINDX
Watershed index of flow responsiveness, snowmelt
Unities s
6.2
Derived from Multi-Resolution Land Characteristics Consortium
National Land Cover Data - 1992
http://www.epa.gov/mrlc/nlcd.html. U.S. Department of Agriculture
State Soil Geographic Database (STATSGO,
http://www.ftw.nrcs.usda.gov/stat data.html), estimated annual
snowfall (Parameter-elevation Regressions on Independent Slopes
Model, PRISM, Climate Source, Corvallis, OR,
http://www.climatesource.com/us/fact sheets/meta snowfall us.htm
1) from wetland volumes estimated from digital wetland inventory
databases: National Wetlands Inventory (NWI,
http: / /wetlands.fws.gov/ ), Wisconsin Wetlands Inventory (WWI,
http: / /www.dnr.state.wi.us /org/water/flip /wetlands /mapping.shtml.
http: / /wisclinc.state.wi.us /datadisc/wimeta browser.html see
Wetlands of Wisconsin), and Ohio Wetlands Inventory
(http://www.dnr.state.oh.us/wetlands/mapping.htm), and from
coastal wetland watershed boundaries, derived by U.S. Environmental
Protection Agency, Mid-Continent Ecology Division, Duluth, MN.
The watershed index of flow responsiveness for snowmelt events is
calculated as the ratio of potential maximum runoff volume from
snowmelt to watershed depressional storage volume.
B-l.l - 13
-------
Appendix C-l.l. Marine and Great Lakes coastal watersheds: equations for peak flow
predictions: Metadata for summary of state regression equations to predict peak flows
Database:
NFF COASTAL.XLS
Variable: State
Label:
Units:
Format:
Source:
Metadata:
State-City
AAAnn-X, where AA = two digit state abbreviation or URB (all urban
areas combined), nn = year of report if more than two are included for a
given state, X = U (urban), W (west), P (Portland), H (Houston)
alphanumeric, uppercase $ 7.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state agencies,
have produced a series of reports containing flood frequency data and
predictive equations derived using watershed characteristics. Typically,
analyses are performed separately for urban versus rural areas, as
urbanized watersheds often have artificial flow regulation and impervious
surface areas, which greatly influence peak flows.
Variable: Region
Label:
Units:
Format:
Source:
Metadata:
Hydrologic region within state
N/A
alphanumeric, uppercase $ 5.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state agencies,
have produced a series of reports containing flood frequency data and
predictive equations derived using watershed characteristics. Typically,
regression analyses are performed separately for different hydrologic
regions of each state, based on examination of spatial distribution of
regression residuals, as well as for urban areas.
Variable:
Label:
Units:
Format:
Source:
Metadata:
Transformations
Description of transformations applied to variables in USGS peak flow
prediction equations
N/A
alphanumeric description, $34.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state agencies,
have produced a series of reports containing flood frequency data and
predictive equations derived using watershed characteristics.
Variable: Mult_factor
Label:
Units:
Format:
Multiplication factor in nonlinear regression equation
Unities s
numeric, 10.4.
C-l.l - 1
-------
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Label:
Units:
Format:
Source:
Metadata:
DAREA
Exponent for drainage area term in USGS equation
Unities s
numeric, 5.3.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2 for all states except ME, km2), B =
other watershed characteristic(s), and a and b are exponents derived
through nonlinear regression analysis.
Variable: CDA
Label:
Units:
Format:
Source:
Metadata:
Variable:
Label:
Units:
Format:
Source:
Exponent for contributing drainage area term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for main channel slope term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
C-l.l -2
-------
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: BR
Label:
Units:
Format:
Source:
Metadata:
Variable: ST
Label:
Units:
Format:
Source:
Metadata:
Exponent for basin relief term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for watershed storage term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Isabel:
Units:
Format:
Source:
Metadata:
LAKES
Exponent for percent lakes term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
C-l.l -3
-------
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: WETLANDS
Label: Exponent for percent wetlands term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: CHSWAMP
Label: Exponent for percent channel swamp term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: CONTRA
Label: Exponent for regulated contributing drainage area term in USGS
equation
Units: Unities s
Format: numeric, 6.4
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
C-l.l -4
-------
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: CORSD
'Label: Exponent for percent coarse glacial drift term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: HYD_A
Label: Exponent for percent hydrologic soil group A term in USGS
equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: HYD_D
Label: Exponent for percent hydrologic soil group D term in USGS
equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: 7 /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
C-l.l -5
-------
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: GARB
Label: Exponent for percent area with carbonate bedrock term in USGS
equation
Units: Unities s
Format: numeric, 6.4
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: OUTWASH
Label: Exponent for percent outwash surficial deposits term in USGS
equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: FINEM
Label: Exponent for percent fine-grained glacial surface deposits term in
USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: 7 /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
C-l.l -6
-------
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: MEDTILL
Label: Exponent for percent medium-grained glacial till term in USGS
equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: MUCK
Label: Exponent for percent muck surficial deposits term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: CLAY
Label: Exponent for percent clay surficial deposits term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: 7 /water.usgs .gov/software 7nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
C-l.l -7
-------
Variable:
Label:
Units:
Format:
Source:
Metadata:
TILROCK
Exponent for percent bare rock/thin till term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Label:
Units:
Format:
Source:
CORGT
Exponent for coarse-grained glacial till term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: SP
Label:
Units:
Format:
Source:
Metadata:
Variable: IA
Label:
Exponent for minimum soil permeability term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for impervious surface area term in USGS equation
C-l.l -8
-------
Units:
Format:
Source:
Metadata:
Variable: BDF
Label:
Units:
Format:
Source:
Metadata:
Variable: UI
Label:
Units:
Format:
Source:
Metadata:
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for basin development factor term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for urbanization intensity term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Label:
Units:
Format:
URBAN
N
Exponent for percent urban area term in USGS equation
Unities s
numeric, 6.4
C-l.l -9
-------
Source:
Metadata:
Variable: LU12
Label:
Units:
Format:
Source:
Metadata:
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for percent land-use 12 area term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: GUTR
Label:
Units:
Format:
Source:
Metadata:
Exponent for percent area with gutters/storm drainage term in
USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
FOREST
Isabel: Exponent for percent forested area term in USGS equation
Units: Unities s
Format: numeric, 6.4
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
C-l.l - 10
-------
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: PREC
Label:
Units:
Format:
Source:
Metadata:
Exponent for annual precipitation term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Label:
Units:
Format:
Source:
Metadata:
Variable:
SNOFALL
Exponent for snowfall term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
12 24
Isabel: Exponent for rainfall intensity, 2-yr, 24-hour storm term in USGS
equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
C-l.l - 11
-------
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable:
Label:
Units:
Format:
Source:
Metadata:
124 100
Variable: I2_2
Label:
Units:
Format:
Source:
Metadata:
Variable: RC
Label:
Units:
Format:
Source:
Metadata:
Exponent for rainfall intensity, 100 yr, 24-hour storm term in USGS
equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for rainfall intensity, 2-yr, 2-hour storm term in USGS
equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for runoff coefficient term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
C-l.l - 12
-------
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: RO
Label: Exponent for annual runoff term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: JANMIN
Label Exponent for minimum January temperature term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Variable: ELEV
Label: Exponent for basin elevation term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
C-l.l - 13
-------
Variable: H
Label:
Units:
Format:
Source:
Metadata:
Variable: BSF
Label:
Units:
Format:
Source:
Metadata:
Variable: L
Label:
Units:
Format:
Source:
Metadata:
characteristic(s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for average main channel elevation, 15%ile and 85%ile
term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for basin shape factor term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for main channel length term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
C-l.l - 14
-------
Variable: SLENRAT
Label: Exponent for basin slenderness ratio term in USGS equation
Units: Unities s
Format: numeric, 5.3
Source:
Metadata:
Variable: RQ2
Label:
Units:
Format:
Source:
Metadata:
Variable: SEE
Label:
Units:
Format:
Source:
Metadata:
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Exponent for rural 2-year peak flow term in USGS equation
Unities s
numeric, 5.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
Standard error of estimate for USGS equation
Percent
numeric, 3.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Equations are typically in the form: Q2 = MF *
AaBb... where Q2 = 2-year peak discharge (cfs), MF = multiplication
factor, A = watershed area (mil2), B = other watershed
characteristic (s), and a and b are exponents derived through nonlinear
regression analysis.
C-l.l - 15
-------
Appendix C1.3. Flood frequency equation references (Study abbreviation).
Asquith, W.H., and Slade, Raymond, Jr., 1997, Regional equations for estimation of peak-stream
flow frequency for natural basins in Texas: U.S. Geological Survey Water-Resources
Investigations Report 96-4307, 68p. (TX)
Asquith, W.H. and R.M. Slade, Jr. 1999. Site-specific estimation of peak-streamflow frequency using
generalized least-squares regression for natural basins in Texas. : U.S. Geological Survey
Water-Resources Investigations Report 99-4172. (TX)
Atkins, J.B., 1996, Magnitude and frequency of floods in Alabama: U.S. Geological Survey Water-
Resources Investigations Report 95—4199, 234 p. (AL)
Bisese, J.A., 1995, Methods for estimating the magnitude and frequency of peak discharges of rural,
unregulated streams in Virginia: U.S. Geological Survey Water-Resources Investigations
Report 94^148, 70 p. (VA)
Bohman, L.R., 1992, Determination of flood hydrographs for streams in South Carolina: Volume 2.
Estimation of peak-discharge frequency, runoff volumes, and flood hydrographs for urban
watersheds: U.S. Geological Survey Water-Resources Investigations Report 92-4040, 79 p.
(SC-U)
Dillow, J.J.A., 1996, Technique for estimating magnitude and frequency of peak flows in Delaware:
U.S. Geological Survey Water-Resources Investigations Report 95—4153, 26 p. (DE)
Dillow, J.J.A., 1996, Technique for estimating magnitude and frequency of peak flows in Maryland:
U.S. Geological Survey Water-Resources Investigations Report 95—4154, 55 p. (MD)
Ensminger, P.A., 1998, Floods in Louisiana, magnitude and frequency, Fifth Edition: Louisiana
Department of Transportation and Development Water Resources Technical Report No. 60,
353 p. (LA)
Feaster, T.D. and G.D. Tasker. 1999. Techniques for estimating the magnitude and frequency of
floods in rural basins of South Carolina, 1999. U.S. Geological Survey Water-Resources
Investigations Report 02-4140. (SC)
Guimaraes, W.B., and Bohman, L.R., 1992, Techniques for estimating magnitude and frequency of
floods in South Carolina, 1988: U.S. Geological Survey Water-Resources Investigations
Report 91-4157, 174 p. (SC)
Gunter, H.C., Mason, R.R., and Stamey, T.C., 1987, Magnitude and frequency of floods in rural and
urban basins of North Carolina: U.S. Geological Survey Water-Resources Investigations
Report 87^096, 52 p. (NC-U,R)
Hodgkins, G. 1999. Estimating the magnitude of peak flows for streams in Maine for selected
recurrence intervals. U.S. Geological Survey Water-Resources Investigations Report 99-4008.
(ME)
C-1.3- 1
-------
Inman, E.J., 1995, Flood-frequency relations for urban streams in Georgia—1994 update: U.S.
Geological Survey Water-Resources Investigations Report 95-4017, 27 p. (GA-U)
Jones, S.H. and C.B. Fahl. 1993. Magnitude and frequency of floods in Alaska and conterminous
basins of Canada. Geological Survey Water-Resources Investigations Report 83-4179. (AK)
Koltun, G.F. and M.T. Whitehead. 2001. Techniques for Estimating Selected Streamflow
Characteristics of Rural, Unregulated Streams in Ohio. US Geological Survey Water
Resources Investigations Report WRIR 02^4068. (OH)
Lorenz, D.L., G.H. Carlson, and C.A. Sanocki. 1997. Techniques for estimating peak flow on small
streams in Minnesota. U.S. Geological Survey Water-Resources Investigations Report 97-
4249. (MM)
Olin, D.A., and Bingham, R.H., 1982, Synthesized flood-frequency of urban streams in Alabama:
U.S. Geological Survey Water-Resources Investigations Report 82—683, 23 p. (AL-U)
Pope, B.F., and Tasker, G.D., 2001, Estimating the magnitude and frequency of floods in rural
basins of North Carolina — revised: U.S. Geological Survey Water-Resources Investigations
Report 01^207, 44 p. (NC)
Robbins, J.C., and Pope, B.F., 1996, Estimation of flood-frequency characteristics of small urban
streams in North Carolina: U.S. Geological Survey Water-Resources Investigations Report
96^084, 21 p. (NC-U)
Stamey, T.C., and Hess, G.W., 1993, Techniques for estimating magnitude and frequency of floods
in rural basins of Georgia: U.S. Geological Survey Water-Resources Investigations Report
93-4016, 75 p. (GA)
Stuckey, M.H. and L.A. Reed. 2000. Techniques for estimating magnitude and frequency of peak
flows for Pennsylvania streams. U.S. Geological Survey Water-Resources Investigations
Report 00-4189. (PA)
Sumioka, S.S., Kresch, D.L., and Kasnick, K.D., 1998, Magnitude and frequency of floods in
Washington: U.S. Geological Survey Water-Resources Investigations Report 97-4277, 91 p.
(WA)
Weiss, L.A., 1975, Flood flow formulas for urbanized and nonurbanized areas of Connecticut:
Watershed and Management Symposium, Logan, Utah, Irrigation and Drainage Division,
American Society of Civil Engineers, p. 658-675. (CT-U,R)
1983, Evaluation and design of a streamflow-data network for Connecticut: Connecticut
Water Resources Bulletin No. 36, 30 p. (CT83)
C-l.3-2
-------
Older references (from Jennings et al. 1983):
Bridges, W.C. 1982. Technique for estimating magnitude and frequency of floods on natural-flow
streams in Florida. U.S. Geological Survey Water-Resources Investigations Report 82-4012.
(FL)
Carpenter, D.H. 1980. Technique for estimating magnitude and frequency of floods in Maryland.
U.S. Geological Survey Water-Resources Investigations Open-File Report 80-1016. (MD)
Curtis, G.W. 1987. Technique for estimating flood-peak discharges and frequencies on rural streams
in Illinois. U.S. Geological Survey Water-Resources Investigations Report 87-4207. (IL)
Flippo, H.N. 1977. Floods in Pennsylvania. Commonwealth of Pennsylvania Department of
Environmental Resources and U.S. Geological Survey, Harrisburg, PA. Water Resources
Bulletin No. 13. (PA)
Glatfelter, D.R. 1984. Techniques for estimating magnitude and frequency of floods on streams in
Indiana. U.S. Geological Survey Water-Resources Investigations Report 84-4134. (IN)
Guimaraes, W.B. and L.R. Bohman. 1988. Techniques for estimating magnitude and frequency of
floods in South Carolina, 1988. U.S. Geological Survey Water-Resources Investigations
Report 91-4157. (SC)
Gunter, H.C., R.R. Mason, and T.C. Stamey. 1987. Magnitude and frequency of floods in rural and
urban basins of North Carolina. U.S. Geological Survey Water-Resources Investigations
Report 87-4096. (NC-U,R)
Harris, D.D., Hubbard, L.L., and Hubbard, L.E., 1979, Magnitude and frequency of floods in
western Oregon: U.S. Geological Survey Open-File Report 79-553, 35 p (OR-W)
Harris, D.D., and Hubbard, L.E., 1983, Magnitude and frequency of floods in eastern Oregon: U.S.
Geological Survey Water-Resources Investigations Report 82-4078, 39 p. (OR)
Holtschlag, D.J. and H.M. Croskey. 1984. Statistical models for estimating flow characteristics of
Michigan streams. U.S. Geological Survey Water-Resources Investigations Report 84-4207.
(MI)
Inman, E.J. 1988. Flood-frequency relations for urban streams in Georgia. U.S. Geological Survey
Water-Resources Investigations Report 88-4085. (GA-U)
Jacques, J.E. and D.L. Lorenz. 1987. Techniques for estimating the magnitude and frequency of
floods in Minnesota. U.S. Geological Survey Water- Resources Investigations Report 87-
4170. (MN87)
Johnson, C.G. and G.A. Laraway. 1975. Flood magnitude and frequency of small Rhode Island
streams: Preliminary estimating relations. U.S. Geological Survey Water-Resources Division
Administrative Report (Jennings for full reference) (RI)
C-l.3-3
-------
Koltun, G.F., and Roberts, J.W., 1990, Techniques for estimating flood-peak discharges of rural,
unregulated streams in Ohio: U.S. Geological Survey Water- Resources Investigations
Report 89-1126, 68 p. (mining effects) (OH)
Krug, W.R., D.H. Conger, and W.A. Gebert. 1992. Flood-frequency characteristics of Wisconsin
streams. U.S. Geological Survey Water-Resources Investigations Report 91-4128. (WI)
Laenen, A. 1980. Storm runoff as related to urbanization in the Portland, Oregon-Vancouver,
Washington area. U.S. Geological Survey Water-Resources Investigations Open-File Report
80-689. (OR-U-P)
Landers, M.N. 1985. Floodflow frequency of streams in the alluvial plain of the Lower Mississippi
River in Mississippi, Arkansas, and Louisiana. U.S. Geological Survey Water-Resources
Investigations Report 85-4150. (MS)
Landers, M.N. and K.Van Wilson, Jr. 1991. Flood characteristics of Mississippi streams. U.S.
Geological Survey Water-Resources Investigations Report 91-4037. (MS)
LeBlanc, D.R., 1978, Progress report on hydrologic investigations of small drainage areas in New
Hampshire-Preliminary relations for estimating peak discharges on rural, unregulated
streams: U.S. Geological Survey Water-Resources Investigations Report 78-47, 10 p. (NH)
Liscum, F. and B.C. Massey. 1980. Technique for estimating the magnitude and frequency of floods
in the Houston, Texas, metropolitan area. U.S. Geological Survey Water-Resources
Investigations Report 80-17. (TX-U-H)
Lopez, M.A. and W.M. Woodham. 1983. Magnitude and frequency of flooding on small urban
watersheds in the Tampa Bay area, West-central Florida. U.S. Geological Survey Water-
Resources Investigations Report 82-42. (FL-U-T)
Lumia, R. 1991. Regionalization of flood discharges for rural, unregulated streams in New York,
excluding Long Island. U.S. Geological Survey Water-Resources Investigations Report 90-
4197. (NY)
Morrill, R.A. 1975. A technique for estimating the magnitude and frequency of floods in Maine. U.S.
Geological Survey Open File Report Number 75-292. (ME)
Sauer, V.B., W.O. Thomas, Jr., V.A. Strieker, and K.V. Wilson. 1983. Flood characteristics of urban
watersheds in the United States. U.S. Geological Survey Water-Supply Paper 2207. (URB)
Sherwood, J.M., 1993a, Estimation of flood volumes and simulation of flood hydrographs for
ungaged small rural streams in Ohio: U.S. Geological Survey Water-Resources Investigations
Report 93-4080, 52 p. (OH)
Sherwood, J.M., 1993b, Estimation of peak-frequency relations, flood hydrographs, and volume-
duration-frequency relations of ungaged small urban streams in Ohio: U.S. Geological
Survey Open-File Report 93-135, 53 p. (OH-U)
C-l.3-4
-------
Simmons, R.H. and D.H. Carpenter. 1978. Technique for estimating magnitude and frequency of
floods in Delaware. U.S. Geological Survey Water-Resources Investigations Report 78-93.
Stedfast, D.A. 1986. Evaluation of six methods for estimating magnitude and frequency of peak
discharges on urban streams in New York. U.S. Geological Survey Water-Resources
Investigations Report 84-4350. (NY-U)
Veenhuis, J.E. and D.G. Gannett. 1986. Effects of urbanization on floods in the Austin
metropolitan area, Texas. U.S. Geological Survey Water-Resources Investigations Report 86-
4069. (TX-U-A)
Wandle, S.W., Jr. 1983. Estimating peak discharges of small, rural streams in Massachusetts. U.S.
Geological Survey Water-Supply Paper 2214. (MA)
Weiss, L.A. 1983. Evaluation and design of a stream-flow data network for Connecticut. U.S.
Geological Survey and Connecticut Dept of Environmental Protection, Connecticut Water
Resources Bulletin No. 36. (CT83)
C-l.3-5
-------
Appendix C-2.1 Marine and Great Lakes coastal watersheds: peak flow classes identified by
CART analysis: metadata
Database:
CART RESULTS.XLS
Variable: U/R
Label:
Units:
Format:
Source:
Metadata:
Urbanized or rural watersheds
U = urban, R = rural
alphanumeric, uppercase $ 1.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software 7nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, analyses are performed separately for urban
versus rural areas, as urbanized watersheds often have artificial flow
regulation and impervious surface areas, which greatly influence peak
flows.
Variable: State
Label:
Units:
Format:
Source:
Metadata:
State-City
AAAnn-X, where AA = two digit state abbreviation or URB (all
urban areas combined), nn = year of report if more than two are
included for a given state, X = U (urban), W (west), P (Portland), H
(Houston)
alphanumeric, uppercase $ 7.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, analyses are performed separately for urban
versus rural areas, as urbanized watersheds often have artificial flow
regulation and impervious surface areas, which greatly influence peak
flows.
Variable: Region
Label:
Units:
Format:
Source:
Metadata:
Hydrologic region within state
Unitless, coding varies by state
alphanumeric, uppercase $ 6.
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
C-2.1 - 1
-------
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals.
Variable: Group
Label:
Units:
Format:
Source:
CART classification group
Unities s
numeric, 2.
U.S. Geological Survey, National Flood Frequency Program Reports
(http://water.usgs.gov/software/nff.html) provided the raw data on
2-year peak flows and watershed characteristics for the Classification
and Regression Tree analyses
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables.
Variable:
Label: Number of USGS gaging station watersheds
Units: Number of watersheds
Format: numeric, 3.
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
Metadata: U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables. The variable n represents the number of
observations included in each CART analysis.
Variable:
Criteria
Label: Criteria for separation of peak flow classes
Units: N/A, description
Format: alphanumeric, $18.
Source: U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
C-2.1 -2
-------
Metadata:
Variable: Mean
Label:
Units:
Format:
Source:
Metadata:
Variable: SD
Label:
Units:
Format:
Source:
Metadata:
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables. This column reports the identity and cutoff
points associated with divisions among flow classes (2-year peak flow
normalized to watershed area). Variable names are defined in
Appendix X of US EPA (2003).
Class average 2-year peak discharge normalized to watershed area
cfs/mil2
numeric, 4.1
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables. This column reports the mean value for
each flow class identified by CART analysis.
Class standard deviation 2-year peak discharge normalized to
watershed area
cfs/mil2
numeric, 5.1
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
C-2.1 -3
-------
Variable: PRE
Label:
Units:
Format:
Source:
Metadata:
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables. This column reports the standard deviation
for each flow class identified by CART analysis.
Percent reduction in error
Fraction, unitless
numeric, 4.3
U.S. Geological Survey, National Flood Frequency Program Reports
(http: / /water.usgs .gov/software /nff.html)
U.S. Geological Survey state offices, in cooperation with state
agencies, have produced a series of reports containing flood
frequency data and predictive equations derived using watershed
characteristics. Typically, regression analyses are performed
separately for different hydrologic regions of each state, based on
examination of spatial distribution of regression residuals. US EPA
performed an analysis of each of the state or urban area data sets
from these reports, using a Classification and Regression Tree
approach with 2-year peak flows normalized to watershed area as the
dependent variable and watershed variables included in state reports
as independent variables. This column contains the total percent
reduction in error associated with each CART analysis, roughly
analogous to the r2 value for a regression analysis.
C-2.1 -4
-------
Appendix C.3.
Hydrologic regions by state
Figures No. Title of Figure
Figure C-3.1 Hydrologic regions for Alabama.
Figure C-3.2 Hydrologic regions for California.
Figure C-3.3 Hydrologic regions for Delaware.
Figure C-3.4 Hydrologic regions for Florida.
Figure C-3.5 Hydrologic regions for Georgia.
Figure C-3.6 Hydrologic regions for Illinois.
Figure C-3.7 Hydrologic regions for Indiana.
Figure C-3.8 Hydrologic regions for Louisiana.
Figure C-3.9 Hydrologic regions in Massachusetts.
Figure C-3.10 Hydrologic regions for Maryland.
Figure C-3.11 Hydrologic regions for Michigan.
Figure C-3.12 Hydrologic regions for Mississippi.
Figure C-3.13 Hydrologic regions for Minnesota.
Figure C-3.14 Hydrologic regions for New York.
Figure C-3.15 Hydrologic regions for Ohio.
Figure C-3.16 Hydrologic regions for Oregon.
Figure C-3.17 Hydrologic regions in Pennsylvania.
Figure C-3.18 Hydrologic regions for South Carolina.
Figure C-3.19 Hydrologic regions for Texas.
Figure C-3.20 Hydrologic regions for Virginia.
Figure C-3.21 Hydrologic regions for Washington.
Figure C-3.22 Hydrologic regions for Wisconsin.
C-3 - 1
-------
Digital bate from U.S. Geological Survey 1:2.000.000.1970
flvf rsa I Transverse Mercaror pro|ec lion. Zone 16
EXPLANATION
1 Regbn ixtrKbry
LAj Rcglcn Idenlfllcalon
Figure C-3.1. Hydrologic regions for Alabama.
C-3 -2
-------
U.S. Geological
Nation;*] Flood Frequency
Water-Rwources
Report 94-4002
' '* " % "~~^- -
i r~~j'+K~~'--i ii°°
M&r %jr42-
, Northeast :
1 + y
/
South Lahontan- \,
Colorado Desert
+ -t ^
"^ + "-
34*+ * \; >•
w* ^^-S-:;-^
. South
^ -Ti v Coast
10ft 150 MILES
I |
i r
n .5C imi 1.W KIIOMFTFRS
DlflHal tusto tram U& e«clocilod Surwsv
1:4000,000, 1S70
North
Coast
EXPLANATION
l hnundaiy
cqu^-ama projection based on
standard uarjfels 29.5 and 4-5.5
Figure C-3.2. Hydrologic regions for California.
C-3 -3
-------
76*
40"
39*
7S€
EXPLANATION
Streams
Fallline
— County boundry
— State boundry
10 20 MILES
MARYLAND
FROM U.S. GEOLOGICAL SURVEY 1:600.0001 19T9
Figure C-3.3 Hydrologic regions for Delaware.
C-3 -4
-------
U.S. Geological Surrey
National Flood Fifsqnwity
TjtresUgatiadS Report 94-4002
fift"
5U 1UO MILES
J |
n sn Km KIlOMFTFIb,
Digital b««» from UL&. Q««kalod Survey
1:4000,000, 1S70
Airera
sfc»idan1
projection based on
29.5 and 4-5.5 degn
EXPLANATION
Figure C-3.4. Hydrologic regions for Florida.
C-3 -5
-------
94"
91'
33
30"
Mississippi
A.1.JC
. ..,
0 ED KILOMETERS
PHYSKX3HAPHIC
DIVISION BOUNDARY
Figure C-3.5. Hydrologic regions for Georgia.
C-3 -6
-------
U.S. Getfogical
NatkiTiaJ Flood Frequency
Wdter-R
-------
EXPLANATION
cir,1 bcaibn
Figure C-3.7. Hydrologic regions for Indiana.
C-3 -8
-------
U.S. Geological
National Flood Fnequenty
Water-R«ouDrccs
Report JM-4MK2
•J
-1
BO
I
MILES
I
n .511 1110 K1IOMFTFI&
Digital bate liom LLS. taobcjItfaJ Sunny
EXPLANATION
Area htHjnfiary
+ 41'
eqi4!*-arca projection based on
29.5 and 4-5.5 degroec
Figure C-3.8. Hydrologic regions for Louisiana.
C-3 -9
-------
U.S. Geological
National Flood
Lori-aigaidniis Report JM-4001
70°
50 MILE,
_J
0 25 .511 KllfJMFTFRS
Dloltal butt tram UL&. fisufcualod Survey
1:2,000,000, 1670
pnoicction based on
oarJlels 29.5 and 45.5 degrees
EXPLANATION
Areas In whlrh fl«irf-esHVnatlni!
i* dn "Of apply
l boundary
Figure C-3.9. Hydrologic regions in Massachusetts.
C-3 - 10
-------
79='
76°
40°
38=
\
PENNSYLVANIA
WEST VIRGINIA
_L
EASE FROM U.S. GEOLOGICAL SURVEY, 1:500,000
Figure C-3.10. Hydrologic regions for Maryland.
C-3 - 11
-------
U.S. Geological ffarrey
NatUmal Flood FVsqDeitc
Wdter-Retjources InrestigatiaBS Report 94-4002
+
33°
+ 47°
+
*>'W/*T
via * - -*
c/ ^^7'W,K +
f» a ^,
42" +
8?
50
1(10 MILES
I
I I
n sn Kin KIIOMFTFI&
Digital bata trum Ll& Q««bcilad Survey
1:3,000,000, 1&7U
Atxars equal-area projection based on
l uarrffcls 29.5 and 45.5 degrees
EXPLANATION
l hnundaiy
Figure C-3.11. Hydrologic regions for Michigan.
C-3 - 12
-------
U.S. Geological Surrey
National Flood Frequency
li]>t*l«j(.iiiiLs Report 94-4003
31'
92s
» V--' «
W v-
JO
I
100 MILES
I
0 50 100 KIIOMFTFRS
Digital bao fram ULS. Geekmkml Survey
1:iOQO.OW), 1S70
AiTcrs equal-area projection based on
i 29.5 and 45.5
EXPLANATION
l hnunrisry
Figure C-3.12. Hydrologic regions for Mississippi.
C-3 - 13
-------
s «d 2)
Figure C-3.13. Hydrologic regions for Minnesota.
C-3 - 14
-------
U.S. Geological
National Flood Frequency Prugram
W;rter-Re*our™s InreatjgatiaiK Report 94-4MB
SO'
(1 SO 1ti[) KIIOMFTFRSi
Digital tests from US. (jeekdlod Survoy
1:3,000,000, 1670
ud-dna projection based on
F>ar^lelc 29.5 and 45.5 degrees
EXPLANATION
Figure C-3.14. Hydrologic regions for New York.
C-3 - 15
-------
U.S. Geological
Flood FVeqnency Program
Report #4-4002
i
sn
SO
I
inn K1IOMFTFK.
100 MILES
I
l l>**s from US. Q««k>alod Sunny
1:2,000,000, 1S70
projection based on
standard uaraHc-ls 29.5 and 1-5.5 degreci
EXPLANATION!
I bounds ry
Figure C-3-15. Hydrologic regions for Ohio.
C-3 - 16
-------
U.S. Geological
National FIoo4 Frequency Program
Imrestlgilttoiis Report 94-4W2
46
124°
Eastern Cascades
High Cascades
42°
117"
Sfl
toil MILES
u SO 100 KILOUE1HJ5
Digital beea fnam US. Guolcglcal Smny
1:2,Q«.MD, 197O
aqual-araa projaetlon bond on
BtandanJ psmllate 29.5 and 46 G dsgrtwe
Coast
EXPIANATION
tAtetem Oregon
Eastern Oregon
Regional boundary
Figure C-3.16. Hydrologic regions for Oregon.
C-3 - 17
-------
o at
Figure C-3.17. Hydrologic regions in Pennsylvania.
C-3 - 18
-------
Figure C-3.18. Hydrologic regions for South Carolina.
C-3 - 19
-------
104°
10CP
96°
34°
30°
26°
100
200 MILES
2
EXPLANATION
Hydrobgic region boundary
(from Asquith and Slade, 1997)
Hydrobgic region number
Abers Equal-Area Conic Projection baaed on
standard parallels at 45.5" and 29.5™
&ale 1:7.250,000
Figure C-3.19. Hydrologic regions for Texas.
C-3 - 20
-------
930
760
360
35°
34°
EXPLANATION
_ Hytirologlc region boundary
Hydrologlc region
Digral Case ffotn U.S. Gedotjcal Surrey
O, 197D
Figure C-3.20. Hydrologic regions for Virginia.
C-3-21
-------
Figure C-3.21. Hydrologic regions for Washington.
C-3 - 22
-------
U.S. Geological Sumy
National Flood Frequency
W
-------
Appendix D. Classification of EDAs by cluster analysis: Classes of Estuarine/Coastal Drainage
Areas based on Physical and Hydrologic Characteristics.
Large. Very High Flow. Shallow. Low Salinity
Lake Pontchartrain
Mississippi River
Atchafalaya-Vermilion Bays
Potomac River
Chesapeake Bay Mainstem
Mattole
Columbia River
Queets-Quinault
Albemarle Sound
Large, High Volume, Deep, High Salinity
Long Island Sound
Southern Long Island
Eastern Lower Delmarva
Maine Coastal
Penobscot Bay
Cape Cod Bay
Cape Cod
San Pedro Channel Islands
Santa Barbara Channel
Central Coastal
Hoh- Quillayute
Hood Canal
Skagit Bay - Whidbey Bn.
Puget Sound
Northeast Cape Fear
Daytona-St. Augustine
Small EDA/Large % Estuary, Low Volume, Low Flow, High Salinity
Damariscotta River
New
Medium EDA/Small % Estuary. Low Volume. High Flow. Low Salinity
Mermentau River
Connecticut River
Susquehanna River
Russian
Klamath River
Nooksack
Medium EDA/Small % Estuary. Low Volume. Low Flow. High Salinity
Passamaquoddy Bay
South Puget Sound
D- 1
-------
Medium. Low Volume. Shallow. Mixed Salinity
Florida Bay
Big Cypress Swamp
Sarasota Bay
Waccasassa River
Econfina-Steinhatchee River
St. Andrew Bay
Austin-Oyster
Aransas Bay
Corpus Christi Bay
Baffin Bay
Upper Laguna Madre
Lower Laguna Madre
Great South Bay
Barnegat Bay
New Jersey Inland Bays
Delaware Inland Bays
Chincoteague Bay
Saco Bay
Plum Island Sound
San Diego Bay
San Diego
Santa Margarita
Santa Ana
San Pedro Bay
Calleguas
Santa Clara
Santa Maria River
Tomales Bay
Humboldt Bay
Coquille River
Coos Bay
Yaquina Bay
Fraser
Bogue Sound
New River
St Marys River-Cumberland Sound
Indian River
Medium Area & Volume, High Salinity
Buzzards Bay
Narragansett Bay
Pawcatuck-Wood
Gardiners Bay
Englishman-Machias Bay
Narraguagus Bay
Blue Hill Bay
Muscongus Bay
Sheepscot Bay
D-2
-------
Casco Bay
Boston Harbor
Massachusetts Bay
San Louis Rey-Escondido
Aliso-San Onofre
Santa Monica Bay
Santa Ynez
Central Coastal
Carmel
Monterey Bay
San Francisco Coastal South
San Francisco Bay
Tomales-Drakes Bay
Bodega Bay
Gualala-Salmon
Chetco
Siltcoos
Necanicum
Crescent-Hoko
Dungeness-Elwha
Port Orchard Sound
San Juan Islands
Strait of Georgia
South Carolina Coast
Broad River
Ogeechee Coastal
St. Catherines-Sapelo Sounds
Nassau
Large, High Flow, Shallow, Mixed Salinity
Charlotte Harbor
Tampa Bay
Crystal-Pithlachascotee
Apalachee Bay
Apalachicola Bay
Mobile Bay
East Mississippi Sound
West Mississippi Sound
Breton-Chandeleur Sound
Galveston Bay
San Antonio Bay
Hudson River-Raritan Bay
Delaware Bay
James River
Central San Francisco-San Pablo-Suisun Bays
Big Navaro-Garcia
Mad-Redwood
Smith
Wilson-Trusk-Nestuccu
Pamlico Sound
D-3
-------
St. Johns River
Cape Canaveral
Biscayne Bay
Medium EDA/Small % Estuary, Low Volume, High Flow, Mixed Salinity
Suwannee River
Brazos River
Kennebec-Androscoggin
Great Bay
Merrimack River
Eel River
Rogue River
Umpqua River
Siuslaw River
Alsea River
Siletz Bay
Tillamook Bay
Nehalem River
Willapa Bay
Grays Harbor
Cape Fear River
North-South San tee Rivers
Charleston Harbor
Stono-North Edisto Rivers
St. Helena Sound
Savannah River
Ossabaw Sound
Altamaha River
St. Andrew-St. Simons Sounds
Large EDA/Small % Estuary, Low Volume
South Ten Thousand Islands
North Ten Thousand Islands
Caloosahatchee River
Charlotte Harbor
Withlacoochee
Choctawhatchee Bay
Pensacola Bay
Perdido Bay
Lake Borgne
Barataria Bay
Calcasieu Lake
Sabine Lake
Matagorda Bay
Patuxent River
Rappahannock River
York River
Choptank River
Tangier-Pocomoke Sound
Piankatank River-Mobjack Bay
D-4
-------
Patapsco-Gunpowder Rivers
Pamlico-Pungo Rivers
Neuse River
Winyah Bay
Small, Low Volume, Low Flow, Shallow, Mixed Salinity
Rookery Bay
Terrebonne-Timbalier Bays
Rio Grande
Maryland Inland Bays
Chester River
Lynnhaven River
Poquoson-Back Rivers
Ingram-Fleets Bays
Elk-Sassafras Rivers
Eastern Bay
Wells Bay
Hampton Harbor
Waquoit Bay
Tijuana Estuary
Mission Bay
Newport Bay
Anaheim Bay
Alamitos Bay
Ventura
San Antonio
Morro Bay
Elkhorn Slough
Drakes Estero
Netarts Bay
D-5
-------
Appendix E. Matrix of properties of existing classification schemes.
Classification
system
Ecoregions of the
US (Bailey, 1976)
Ecological units of
the Eastern US
(Keys et al., 1995;
Maxwell et al.,
1995)
Ecoregions of the
conterminous US
(Omernik, 1987)
Objective
Map / develop
hierarchical
framework of
terrestrial habitats
Map ecological
units: based on
physical and
biological
components that
influence ecological
relationships,
processes and
potential
Map ecosystem
regionalities: define
regions and explore
the processes and
effects of human
activities
Classification
Factors
Considered
Land use
Soil types
Landform
Climax Vegetation
Temperature
Geomorphology
Geology
Human use
Soil types
Climate
Surface water
characteristics
Growing season
Vegetation
potential
Temperature
Land use
Geology
Climate
Physiography
Soil types
Hydrology
Vegetation
Stressor
Pertinence
Habitat
alteration
Habitat
alteration
Nutrients
Suspended
sediments
Thermal
regime
Habitat
alteration
Nutrients
Suspended
sediments
Flow regime
Extent,
Spatial
Temporal
Variability
US
Spatial
US
Spatial
US
Spatial
Data
Availability,
Gaps
Nationwide
Eastern US, digital
Gaps include areas
other than eastern
US and some
portions of west
Maps, local
databases are
available to
support some
reference
condition
locations
Limitations/
Status of
Testing/
Modifications
Terrestrial focus
does not include
hydrology.
Used by the Nature
Conservancy, but
not tested for
wetlands
Terrestrial focus
does not include
hydrology. Case
studies have
demonstrated
watersheds to group
by these ecological
units (Jensen, 2001,
Detenbeck, 2000)
Terrestrial focus.
Utilized by a number
of states to develop
biological criteria, set
water quality
standards and lake
management goals
but not tested for
most wetlands
E- 1
-------
Classification
system
Circular 39
(Shaw and Fredine,
1956)
Classification of
wetlands and
deepwater habitats
of the US
(Cowardin et al,
1979)
Riverine Marsh
Disturbance
Gradients
(Day et al., 1988)
Objective
Wetlands:
Inventory the
distribution, extent
and quality of
wetlands along
with their value as
wildlife
Habitat
Wetlands:
Inventory status
and trends in
coverage and
deepwater habitat
types
Wetlands: Identify
vegetation
community
response types
Classification
Factors
Considered
Water depth
Flooding regime
Salinity
Vegetation type
Ecosystem typea
Relative elevation
Substrate type
pH and soil type
Flooding regime
Water regime
Water chemistry
Vegetation
Hydrologic regime
Disturbance regime
Vegetative
composition
Stressor
Pertinence
Habitat
alteration
Hydrologic
regime
Habitat
alteration
Nutrients
Habitat
alteration
Extent,
Spatial
Temporal
Variability
US
Spatial
US
Spatial
Coastal and
riverine wetlands
Data
Availability,
Gaps
Nationwide
Gaps in mapping
Nationwide
Gaps in
digitization
NWI, state
inventories,
extension to other
regions necessary
based on species
composition/traits
Limitations/
Status of
Testing/
Modifications
Served as a simple
but effective basis
for later efforts
Widely used,
extensive number of
classes may be
impractical. Gibbs,
1993 case study, but
stressor sensitivity
aspects (density,
spatial configuration,
temporal variability)
need to be tested.
Modified by McKee,
1992, and for marine
systems (Detier,
1992).
Untested for stressor
sensitivity
applications
E-2
-------
Classification
system
Coastal Wetland
Ecosystems (Chow-
Fraser and Albert,
1998)
Coastal Wetlands of
the Great Lakes
(Keough et al, 1999)
Great Lakes
Wetlands
Consortium (Great
Lakes Commission,
2003)
Fluvial
Classification
(Montgomery and
Buffington, 1993)
Objective
Wetlands: identify
habitats of
biodiversity for
conservation
Wetlands: classify
by functional
groups
Wetland habitats:
refine of
inventories for
tracking a real
status and trends,
stratify types for
monitoring
programs
Fluvial: apply
hierarchical
geomorphic
classification to
predict risk of
sediment input,
transport
Classification
Factors
Considered
Geomorphology
Vegetation types
Hydrogeomorphic
types: open coast,
drowned-river
mouth, flooded
delta and protected
Hydrology
Geomorphology
Shoreline processes
Resuspension
Residence time
Geology
Climate
Hydrology
Sediment transport
Stressor
Pertinence
Not directly
linked with
stressor
susceptibility
Not directly
linked with
stressor
susceptibility,
but hydrogeo-
morphic types
may differ in
retention time,
settling
efficiency
Nutrients
Suspended
sediments
Toxics
Hydrologic
regime
Not linked
with stressor
susceptibility
per se but
influenced by
sediment
loading
Extent,
Spatial
Temporal
Variability
Great Lakes
Great Lakes
Great Lakes
Fluvial systems
Data
Availability,
Gaps
NWI inventoried
wetlands only
Existing NWI
state inventories,
GLEI
NWI, state
wetlands
inventories under
assessment, point
and areal
coverages need
matching
State inventories
Limitations/
Status of
Testing/
Modifications
Extensive number of
classes may be
impractical
Some difficulty with
separation of classes
in practice
Untested for stressor
sensitivity
applications
Modified by Jay et al,
1999 for estuaries
E-3
-------
Classification
system
Channel Types,
(Rosgen, 1994)
Flow Regimes
(Poff and Allan,
1995)
Hydrologic
landscape Regions
oftheUS(USGS,
2003)
Objective
Streams: predict
direction and
magnitude of
changes due to
natural and human
disturbances
Fluvial: classify
hydrologic regime
by
flooding/drought
magnitude,
frequency, and
predictability and
relate to biological
community types
Watersheds: group
according to
similarities n
landscape and
climate
characteristics to
assist with water
quality assessments
Classification
Factors
Considered
Geomorphology
Channel slope
Substrate type
Instream Sediment -
sources/sinks
Climate
Hydrology
Land-surface form
Geologic texture
(Soil and bedrock
permeability)
Climate variables
Stressor
Pertinence
Suspended
sediments,
sediment-
associated
pollutants
Nutrients
Suspended
sediments
Toxics
Hydrologic
regime
Not linked
with stressor
susceptibility
Extent,
Spatial
Temporal
Variability
Fluvial systems
Temporal
Spatial
US
Temporal
43,931 small
watersheds (200
sq km) in the US
Data
Availability,
Gaps
Not mapped
Nationwide
STATSGO
USGS National
Atlas, data bases
Limitations/
Status of
Testing/
Modifications
Riverine only,
instream channel
form focus.
See also, Hawkins et
al, 1993
Untested for stressor
sensitivity
applications
Parameters are
assumed to affect
hydrologic processes
and so may relate to
retention time
predictions
E-4
-------
Classification
system
Comparative
Watershed
Framework
(Detenbeck et al,
2000)
Hydrodynamic,
single parameter,
(Strommel and
Farmer, 1952)
Objective
Freshwater Lotic
Systems: predict
susceptibility of
biota, habitat and
water quality to
nonpoint stressors
mediated by
changes in
hydrology
Estuarine: describe
types based on
stratification
Classification
Factors
Considered
Hydrogeomorphic
region
Watershed storage
(retention time)
Land use
Thresholds relative
to hydrologic
regime
Hydrology
River flow
Stratification
Tidal currents
Stressor
Pertinence
Habitat
alteration
Nutrients
Suspended and
bedded
sediments
Toxics
Thermal
regime
Hydrologic
regime
Nutrients
Toxics
Suspended
sediments
Extent,
Spatial
Temporal
Variability
US: Humid
regions, runoff-
dominated urban
regions
Temporal
Narrow estuaries
Data
Availability,
Gaps
MRLC, NWI,
NWSD,
USFS-EU, USGS
Gaps in NWI-
digital coverage
(storage
calculations) and
NWBD, flow
thresholds
NOAA CA&DS
Limitations/
Status of
Testing/
Modifications
Arid regions,
groundwater-
dominated systems
not covered.
Tested in Lake
Superior and
Michigan Basins
Does not consider
estuarine types,
focus on narrow
estuaries. Modified
by Ippen and
Harlemann, 1961;
Prandle, 1986;
Fischer, 1976;
Simpson and
Hunter, 1974; Nunes
Vax and Lennon,
1991
E-5
-------
Classification
system
Hydrodynamic, two
parameter (Hansen
and Rattray, 1966)
Ecological
Perspective on
Estuarine
Classification
(Jay et al., 1999)
Objective
Estuarine: describe
types based on
stratification and
circulation
Estuarine: identify
environments
found in different
estuaries and
describe their
sediment transport
processes
Classification
Factors
Considered
Hydrology
Geomorphology
Freshwater,
Tidal influences
Stratification
River Flow
Tidal Flow
Residence Time
Forcing Processes:
wind, waves,
sea ice
Stressor
Pertinence
Nutrients
Toxics
Suspended
sediments
Suspended and
bedded
sediments
Hydrologic
regime
Extent,
Spatial
Temporal
Variability
Narrow
estuaries, fjords,
and river
dominated
estuaries;
Worldwide
Estuaries
Spatial
Temporal
Data
Availability,
Gaps
NOAA CA&DS
LMER
Limitations/
Status of
Testing/
Modifications
Unsuitable for broad
shallow embayments
and systems subject
to wind forcing or
temporal variability.
Modified by Fischer,
1976; Officer, 1976;
Oey, 1984; Jay and
Smith, 1988;
Friederichs and
Madsen, 1992;
Hearn, 1998; Geyer
et al, 1999
Effectively oriented
towards
susceptibility to
suspended and
bedded sediments as
a stressor, but
untested for other
aquatic stressors, i.e.
does not address
eutrophication
E-6
-------
Classification
system
Physical
Classification of
Australian estuaries
(Digby et al., 1999)
NOAA Estuarine
Classification (
Alice et al., 2000)
Coastal Provinces
(Briggs, 1974)
Objective
Estuarine: Develop
a framework for
780 Australian
estuaries based on
quantifiable,
biologically
important physical
characteristics and
transfer knowledge
between estuaries
with similar
characteristics
Marine and
Estuarine: describe
the spatial
heterogeneity of
marine and
estuarine
landscapes and link
to underlying
mechanisms
structuring the
ecosystem and
biotic communities
Near Coastal and
Marine: Outline
zoogeographic
regions
Classification
Factors
Considered
Geomorphology
Climactic zones
Tidal range
Shoreline
Intertidal
proportion
Hydrology
Geomorphology
Topology
Ecosystem type
Substratum
Climate Zones
Wave/wind Energy
Temperature
Salinity, alkalinity
Extreme events
Biological
interactions
Coastal ocean
currents
Distribution of
marine organisms,
indigenous species
Stressor
Pertinence
Nutrients
Toxics
Suspended and
bedded
sediments
Habitat
alteration
Nutrients
Not directly
linked with
stressor
susceptibility
Extent,
Spatial
Temporal
Variability
Australia
Spatial
US: Marine and
Estuarine
Spatial
Near coastal
Data
Availability,
Gaps
Data available for
623 of 780
estuaries
Gaps in seagrass
coverage, and in
accounting for
temporal
variability of this
and other
parameters
NOAA CADS
Bathymetry
Topography
Limitations/
Status of
Testing/
Modifications
Temperate estuaries
become a large
category that may be
amenable to
subdivision based on
a biological
parameter, ie.
seagrass coverage or
diversity offish or
macrobenthic
communities
Leads to a large
number of classes,
reduction strategies
may include system
response factors
Freshwater systems
require modification
Untested for stressor
sensitivity
assessment
E-7
-------
Classification
system
The Nature
Conservancy,
(e.g. Beck and
Odaya, 2001)
Coastal Impacts
from Freshwater
Flow Alterations
(Sklar and Browder,
1998)
Objective
Aquatic Systems:
map and inventory
community and
ecosystem habitats
for conservation of
biodiversity and
target species
Near Coastal:
identify potential
impacts of
alterations to
freshwater flow to
the Gulf of Mexico
Classification
Factors
Considered
Physicochemical -
Geologic factors
System attributes
Target species
Habitat type
Land use
Road, Dam density
Point source
density
Freshwater Flow
Salinity, isohaline
zone
Dissolved Oxygen
System Geometry
Discharge
component
Vegetative habitat
Stressor
Pertinence
Habitat
alteration
Point sources:
Toxins,
Nutrients
Suspended and
bedded
sediments
Hydrologic
regime
Habitat
alteration
Nutrients
Suspended and
bedded
sediments
Toxics
Hydrologic
regime
Compares
individual
stressor effects
with multiple
stressor effects
Extent,
Spatial
Temporal
Variability
US, Central and
South America
Spatial
Gulf of Mexico
Spatial
Temporal
Data
Availability,
Gaps
Species
Aquatic system
targets
Conservation
areas
Gaps in aquatic
insect species
coverage, snails,
crayfish, fish and
mussels, and in
tidal marsh
habitats
Data for
individual systems
Limitations/
Status of
Testing/
Modifications
Used to select
priority areas for
conservation action,
not tested for
stressor susceptibility
applications
Currently used for
in-depth
examination of
individual systems,
not tested for
extrapolating across
systems.
E-8
-------
Classification
system
Geomorphic
Modeling Approach
(Stefan et al., 1995,
1996)
Estuarine Qualilty
Index
(Ferreira, J.G.,
2000)
Estuarine
Susceptibility,
(NOAA 1989;
Bricker et al., 1999)
Objective
Lakes: Predict
susceptibility of
fish habitat to
global climate
change
Estuarine; design a
decision support
system to provide
an index or score
based on estuarine
condition to
facilitate
classification
Estuaries: classify
by susceptibility to
nutrient over-
enrichment
Classification
Factors
Considered
Stratification (as a
function of mean
depth, area)
Trophic status
Latitudinal gradient
Thermal regime
Vulnerability
Water Quality
Sediment Quality
Trophodynamic s
Nutrient Load
Dilution
Flushing
Dissolved
concentration
potential (DCP)
Particle retention
efficiency (PRE)
Estuarine Eport
potential (EXP)
Stressor
Pertinence
Habitat
alteration
Nutrients
Suspended and
bedded
sediments
Thermal
regime
Interacting
stressors
Nutrients
Toxics
Suspended and
bedded
sediments
Nutrients
Extent,
Spatial
Temporal
Variability
Potential for US
applicability
Temporal
Spatial
US and Europe
138 US estuaries
or Estuarine
drainage units
(EDUs)
Data
Availability,
Gaps
Lake
morphometry and
trophic status
NOAA CADS
BASINS
NOAA CADS
Gaps in
considering
temperature, wind
mixing, inlet
configuration,
estuarine plume
exchange with
nearshore oceanic
water, ratio of
shoreline length to
estuarine surface
area
Limitations/
Status of
Testing/
Modifications
Predictions have
applicability to
multiple stressors.
Currently limited
regional (MN lakes)
Benthic community,
sediment quality, and
fish diversity data
may not be widely
available.
Effectively oriented
towards nutrient
susceptibility, but
untested for other
aquatic stressors.
Prediction less useful
for estuaries in
Maine, small
estuaries in southern
California, and Puget
Sound estuaries
E-9
-------
BASINS- Better Assessment Science Integrating Point and Nonpoint Sources
GLEI = Great Lakes Environmental Indicators Project
LMER - Land Margin Ecosystem Research Program
MRLC - National Land-use database
NGDC - www.ngdc.noaa.gov.mgg/mggd.html
NWSD-National Watershed Boundary Database
NOAA CADS - National Oceanic and Atmospheric Administration, Coastal Assessment and Data Synthesis System
NWI- National Wetlands Inventory
STATSGO database (U.S. Department of Agriculture, 1994)
USFS-EU - USFS- Ecological Units
USGS- Regional USGS peak flow prediction equations; water.usgs.gov/GIS/metadata/usgswrd/hlrus.htm
Ecosystem type: Marine, Estuarine, Riverine, Lacustrine, and Palustrine
E- 10
-------
Appendix F-l: Regional maps of sediment toxic units by chemical class (metals,
pesticides, PAHs), for estuaries.
Figure F-l Toxic Units - Metals in the Northeast
Figure F-2 Toxic Units - Metals in the Gulf Coast
Figure F-3 Toxic Units - Metals in the Southeast
Figure F-4 Toxic Units - Metals in the Northwest
Figure F-5 Toxic Units - Metals in the Southwest
Figure F-6 Toxic Units - PAH's in the Northeast
Figure F-7 Toxic Units - PAH's in the Gulf Coast
Figure F-8 Toxic Units - PAH's in the Southeast
Figure F-9 Toxic Units - PAH's in the Northwest
Figure F-10 Toxic Units - PAH's in the Southwest
Figure F-ll Toxic Units - Pesticide's in the Northeast
Figure F-7 Toxic Units - Pesticide's in the Gulf Coast
Figure F-8 Toxic Units - Pesticide's in the Southeast
Figure F-9 Toxic Units - Pesticide's in the Northw est
Figure F-10 Toxic Units - Pesticide's in the Southwest
-------
Toxic Units - Metals
Northeast
-------
Legend
*
I I SB.
Toxic Units
^H
05-10
1.0-3.0
dy Unit Boundary
Toxic Units - Metals
Gulf Coast
M
-f
-------
Toxic Units - Metals
Southeast
Legend
States
Study Unit Boundary
Toxic Units
^H
-------
Toxic Units - Metals
Northwest
Legend
Study Unit Boundary
Toxic Units
0.5- 1.0
1.0-3.0
| 3.0 -10
N
+
-------
Legend
States
Study Unit Boundary
Toxic Units
^H
-------
Toxic Units - PAHs
Northeast
-------
Toxic Units - PAHs
Gulf Coast
Legend
_JSMt.
Study Unit Boundary
Toxic Units
0 5 -1 C
1,0-3.0
^| 3.0 -10
-------
Toxic Units - PAHs
Southeast
Legend
States
Study Unit Boundary
Toxic Units
^H
-------
Toxic Units - PAHs
Northwest
Legend
UK States
Study Unit Boundary
Toxic Units
HI <0,5
H 0.5-1.0
1.0-3.0
^H 3.0-10
-------
Toxic Units - PAHs
Southwest
Legend
States
Study Unit Boundary
Toxic Units
^H
-------
Toxic Units - Pesticides
Northeast
Legend
States
Study Unit Boundary
Toxic Units
^H <0.5
- 0.5-1.0
1.0-3.0
| 3.0 -10
^H > 10
N
4-
-------
Legend
Study Unit Boundary
Toxic Units
^H ' '
05-10
1 0 - 3.0
30- 10
Toxic Units - Pesticides
Gulf Coast
S
+
-------
.
Toxic Units - Pesticides
Southeast
Legend
States
Study Unit Boundary
Toxic Units
^H <05
I 0.5-1.0
10-3.0
• 3.0-10
IS
-f
-------
Toxic Units - Pesticides
Northwest
Legend
Study Unit Boundary
Toxic Units
0.5- 1.0
1.0-3.0
3.0- 10
N
+
-------
Toxic Units - Pesticides
Southwest
States
Study Unit Boundary
Toxic Units
------- |