&EPA
United Sfc*i6
Enviraimerilfll PiutmBmi
Agamy
Policy Assessment for the Review of the
Secondary National Ambient Air Quality
Standards for NOx and SOx
Second External Review Draft
-------
EPA 452/P-10-008
September, 2010
Policy Assessment for the Review of the
Secondary National Ambient Air Quality Standards
for NOx and SOx:
Second External Review Draft
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, North Carolina
-------
DISCLAIMER
This document has been reviewed by the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency (EPA), and approved for publication. This draft
document has been prepared by staff from the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. Any opinions, findings, conclusions, or
recommendations are those of the authors and do not necessarily reflect the views of the EPA
Mention of trade names or commercial products is not intended to constitute endorsement or
recommendation for use. This document is being provided to the Clean Air Scientific Advisory
Committee for their review, and made available to the public for comment. Any questions or
comments concerning this document should be addressed to Bryan Hubbell, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, C504-06, Research Triangle
Park, North Carolina 27711 (email: hubbell.bryan@epa.gov ).
Acknowledgements (if any)
-------
Policy Assessment for the Review of the Secondary National Ambient Air
Quality Standards for NOX and SOX: Second External Review Draft
EXECUTIVE SUMMARY
Introduction
This second draft Policy Assessment is an evaluation of the policy implications of the key
scientific information contained in the Integrated Science Assessment (ISA) for Oxides of Nitrogen and
Sulfur-Ecological Criteria, prepared by EPA's National Center for Environmental Assessment (NCEA),
and the results from the analyses contained in the Risk and Exposure Assessment (REA) for Review of
the Secondary National Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur.
This second draft also presents preliminary EPA staff conclusions regarding the adequacy of the current
standards and various policy options that we believe are appropriate to consider as part of the current
review of the secondary (welfare-based, e.g. focused on non-health effects including impacts on soils,
water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility, and climate) NOX and
SOXNAAQS.
This policy assessment is intended to help "bridge the gap" between the scientific assessment
contained in the ISA and the judgments required of the EPA Administrator in determining whether, and if
so, how, it is appropriate to revise the secondary NAAQS for NOX and SOX. This policy assessment
considers the available scientific evidence and quantitative risk-based analyses, together with related
limitations and uncertainties, and focuses on the basic elements of air quality standards: indicators,
averaging times, forms, and levels. These elements, which serve to define each standard, must be
considered collectively in evaluating the welfare protection afforded by the secondary NOX and SOX
NAAQS.
In conducting this periodic review of the NOX and SOX secondary NAAQS, EPA has decided to
jointly assess the scientific information, associated risks, and standards because ambient NOX and SOX,
and their associated transformation products, such as deposited N and S, are linked from an atmospheric
chemistry perspective, as well as jointly contributing to environmental effects.
Scope
This assessment primarily focuses on the effects of the deposition of ambient NOX and SOX on
multiple ecological receptors. Highlighted effects include those associated with acidification and nitrogen
nutrient enrichment. Based on these highlighted effects, EPA's objective is to develop a framework for
setting standards that are ecologically relevant and that reflect the common impacts of these two
pollutants as they deposit to sensitive ecosystems.
For this second draft policy assessment, we have chosen to focus much of our discussion on
effects in sensitive aquatic ecosystems caused by acidifying deposition of nitrogen and sulfur, which is a
transformation product of ambient NOX and SOX. We have a high degree of confidence in the link to
aquatic acidification effects as well as more information available with which to develop an ecologically
meaningful structure for the standards. We recognize in doing so that the resulting standards will not
likely provide full protection against terrestrial acidification effects or against adverse nutrient enrichment
effects in sensitive terrestrial and aquatic ecosystems. It is however likely that some additional protection
for those ecosystems will be provided as overall NOX and SOX levels in the environment are decreased in
response to the aquatic acidification based standards.
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
In the atmospheric science community, NOX typically refers to the sum of nitrogen dioxide (NO2),
and nitric oxide (NO). In contrast, the Clean Air Act uses "NOX" to refer to any gaseous mixture of
species composed solely of nitrogen and oxygen (e.g., NO2, NO, nitrous oxide [N2O], nitrogen trioxide
[N2O3], nitrogen tetroxide [N2O4], and dinitrogen pentoxide [N2O5]). The term used by the scientific
community to represent the complete set of oxidized nitrogen compounds, including those listed in CAA
Section 108(c), is total oxidized nitrogen (NOy). NOy includes all nitrogen oxides, including NO, NO2,
HNO3, peroxyacetyl nitrate (PAN), 2N2O5, HONO, NO3, organic nitrates, and particulate NO3. In the
policy assessment, unless otherwise indicated, we use the term "NOy" to refer to the complete set of
oxidized nitrogen compounds.
For this assessment, the full definition of SOX includes all oxides of sulfur, including both
gaseous substances (e.g., SO2, sulfur monoxide [SO], sulfur trioxide [SO3], thiosulfate [S2O3], and
heptoxide [S2O7], as well as particulate species, such as ammonium sulfate [(NH^SOJ). However,
throughout this document we refer to SOX as the sum of SO2 and sulfate to be consistent with standard
monitoring instrumentation. Sulfate is referred to as SO4 and nitrate as NO3, recognizing that they refer to
the ions that have charges of-2 for sulfate and -1 for nitrate.
Conceptual Framework
The figure below depicts the framework by which we are considering the structure of an
ecologically relevant secondary standard. It is a conceptual diagram that illustrates how a level of
protection related to an indicator of ecological effect(s) equates to atmospheric concentrations of NOX and
SOX indicators. This conceptual diagram illustrates the linkages between ambient air concentrations and
resulting deposition metrics, and between the deposition metric and the ecological indicator of concern.
The Deposition Transference Ratios translate between NOX and SOX deposition metrics and ambient
atmospheric concentrations of NOX and SOX, while the Ecological Response to Deposition Function
relates the deposition metric into the ecological indicator.
ocation
specific
Modifying
Factors
Location
specific
Modifying
Factors
Atmospheri
Deposition
Transference
Ratio
Ecological
Response
to
Deposition
Function
Indicator(s)
Our policy assessment is structured around this conceptual model, and includes an evaluation of the
effects associated with deposition of NOX and SOX to ecosystems, as well as an assessment of the
ES-2
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
adequacy of the existing NOX and SOX standards in protecting against these effects. This policy
assessment also develops a more complete understanding of the conceptual structure needed to address
the variable ecosystem and atmospheric factors which modify the impacts of deposited NOX and SOX on
ecosystems. Development of the form for the standard and options for ambient atmospheric indicators for
NOX and SOX, averaging times, and levels of the standard are also discussed.
Ecological Effects from NOX and SOX Deposition
Effects are broadly categorized into those related to acidification and nutrient-enrichment. Acidification
occurs in both aquatic and terrestrial ecosystems, with most aquatic effects occurring in freshwater lakes
and streams. Nutrient enrichment also occurs in both aquatic and terrestrial ecosystems; however, the
types and prevalence of nutrient enrichment effects vary between freshwater and estuarine aquatic
ecosystems.
In the process of acidification, chemical components of terrestrial and freshwater aquatic ecosystems are
altered in a way that leads to effects on biological organisms. Because NOX and SOX deposited to
terrestrial ecosystems often move through the soil and eventually leach into adjacent water bodies,
deposition to terrestrial ecosystems is also a cause of acidification in aquatic ecosystems.
The scientific evidence is sufficient to infer a strong causal relationship between acidifying deposition and
effects on biogeochemical processes and biota in aquatic ecosystems, and between acidifying deposition
and changes in biogeochemistry in terrestrial ecosystems. Acidic deposition is observed to alter sulfate
and nitrate concentrations in surface waters, balance of base cations, acid neutralizing capacity (ANC),
inorganic aluminum, calcium, and surface water pH. These changes can result in the loss of acid-
sensitive biological species such as salmonid fish species and disrupt food web dynamics causing
alteration to the diet, breeding distribution and reproduction of certain species of bird, such as goldeneye
ducks and loons. Acidification in terrestrial ecosystems has been shown to cause decreased growth and
increased susceptibility to disease and injury in sensitive tree species, including red spruce and sugar
maple.
Principal factors governing the sensitivity of terrestrial and aquatic ecosystems to acidification from
sulfur and nitrogen deposition include geology, biological uptake of nitrogen, soil depth, and elevation.
Geologic formations having low base cation supply generally underlie the watersheds of acid-sensitive
lakes and streams. Other factors that contribute to the sensitivity of soils and surface waters to acidifying
deposition include topography, soil chemistry, land use, and hydrology. Episodic and chronic
acidification tends to occur primarily (but not exclusively) at relatively high elevations in areas that have
base-poor bedrock, high relief, and shallow soils.
Based on published analyses of surface water data from freshwater ecosystem surveys and monitoring,
the most sensitive lakes and streams are located in New England, the Adirondack Mountains, the
Appalachian Mountains (northern Appalachian Plateau and Ridge/Blue Ridge region), the Upper Midwest
and high elevation Western ecosystems.
ANC is the most widely used indicator of acid sensitivity and has been found in various studies to be the
best single indicator of the biological response and health of aquatic communities in acid-sensitive
systems. Annual or multiyear average ANC is a good overall indicator of sensitivity, capturing the ability
of an ecosystem to withstand episodic events such as spring melting that can lower ANC over shorter
time spans. Biota are generally not harmed when annual average ANC levels are >100 microequivalents
per liter (ueq/L). At annual average ANC levels between 100 and 50 ueq/L, the fitness of sensitive
species (e.g., brook trout, zooplankton) begins to decline. When annual average ANC is <50 ueq/L,
ES-2
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
negative effects on aquatic biota are observed, including large reductions in diversity offish species, and
declines in health offish populations, affecting reproductive ability and fitness. Annual average ANC
levels below 50 ueq/L are generally associated with death or loss of fitness of biota that are sensitive to
acidification.
Recent studies indicate that acidification of lakes and streams can result in significant loss in economic
value, which is one indicator of adversity associated with loss of ecosystem services. A 2006 study of
New York residents found that they are willing to pay between $300 and $800 million annually for the
equivalent of improving lakes in the Adirondacks region to an ANC of 50. In addition, several states
have set goals for improving the acid status of lakes and streams, generally targeting ANC in the range of
50 to 60 ueq/L, and have engaged in costly activities to decrease acidification.
Forests of the Adirondack Mountains of New York, Green Mountains of Vermont, White Mountains of
New Hampshire, the Allegheny Plateau of Pennsylvania, and high-elevation forest ecosystems in the
southern Appalachians are the regions most sensitive to terrestrial acidification effects from acidifying
deposition. A commonly used indicator of terrestrial acidification is the base cation to aluminum ratio,
Bc/Al. Many locations in sensitive areas of the U.S. have Bc/Al levels below benchmark levels we have
classified as providing low to intermediate levels of protection to tree health. At a Bc/Al ratio of 1.2
(intermediate level of protection), red spruce growth can be reduced by 20 percent. At a Bc/Al ratio of 0.6
(low level of protection), sugar maple growth can be reduced by 20 percent. While not defining whether
a 20 percent reduction in growth can be considered significant, existing economic studies suggest that
avoiding significant declines in the health of spruce and sugar maple forests may be worth billions of
dollars to residents of the Eastern U.S.
The numerous ecosystem types that occur across the U.S. have a broad range of sensitivity to N
deposition. Organisms in their natural environment are commonly adapted to a specific regime of nutrient
availability. Change in the availability of one important nutrient, such as N, may result in imbalance in
ecosystems, with effects on ecosystem processes, structure and function. In certain N-limited ecosystems,
including many ecosystems managed for commercial production, N deposition can result in beneficial
increases in productivity. Nutrient enrichment effects from NOX deposition are difficult to disentangle
from overall effects of nitrogen enrichment. This is caused by two factors: the inputs of reduced nitrogen
from deposition and, in estuarine ecosystems, a large fraction of nitrogen inputs from non-atmospheric
sources.
Adequacy of the Existing NOX and SOX Standards to Protect Against
Acidification and Nutrient Enrichment Effects
Current NOX and SOX secondary standards are designed to protect against direct exposure of vegetation to
ambient concentrations of NOX and SOX. Almost all areas of the U.S. are in attainment of the current NOX
and SOX secondary standards. The NOX secondary standard is 0.053 parts per million (ppm), annual
arithmetic average, calculated as the arithmetic mean of the 1-hour NO2 concentrations. The SOX
secondary standard, which uses SO2 as the atmospheric indicator, is a 3-hour average of 0.5 ppm, not to
be exceeded more than once per year.
Recent acidification status of aquatic ecosystems indicate that in the Adirondacks and Shenandoah areas,
rates of acidifying deposition of NOX and SOX are still well above pre-acidification (1860) conditions.
Forty-four percent of Adirondack lakes evaluated exceed the critical load for an ANC of 50 ueq/L, and in
these lakes recreationally important fish species such as trout are missing due to acidification. In the
Shenandoah area, 85 percent of streams evaluated exceed the critical load for an ANC of 50 ueq/L,
resulting in losses in fitness in species such as the Blacknose Dace.
ES-4
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
The REA only evaluated a small number of sensitive areas as case studies. However, in the sugar maple
case study area (Kane Experimental Forest, Pennsylvania), recent (2002) deposition levels are associated
with a Bc/Al ratio below 1.2, indicating the potential for a greater than 20 percent reduction in growth. In
the red spruce case study area (Hubbard Brook Experimental Forest, New Hampshire), recent deposition
levels are associated with a Bc/Al ratio slightly above 1.2, indicating slightly less potential for significant
reductions in growth.
Available ecological indicators for estuarine nutrient enrichment are not sufficiently sensitive to changes
in atmospheric NOX to be of use in assessing the adequacy of existing NOX standards. Atmospheric NOX
can be an important contributor of N to estuarine nutrient enrichment, but additional analysis is required
to develop an appropriate indicator for assessing levels of protection from nutrient enrichment effects in
estuaries related to deposition of NOX.
Nitrogen deposition can alter species composition and cause eutrophication in freshwater systems. In the
Rocky Mountains, for example, deposition loads of 1.5 to 2 kg/ha/yr, which are within the range
associated with ambient NOX levels meeting the current standards, are known to cause changes in species
composition in diatom communities indicating impaired water quality. From this we initially conclude
that the existing secondary standard for NOX does not protect such ecosystems and their resulting services
from impairment.
Most terrestrial ecosystems in the US are N-limited, and therefore they are sensitive to perturbation
caused by N additions. Under recent conditions, nearly all of the known sensitive mixed conifer forest
ecosystems receive total N deposition levels above 3.1 N kg/ha/yr, which is the ecological benchmark for
changes in lichen species. Lichens are sentinels for broader ecosystem change in terrestrial systems.
Some portions of the Sierra Nevadas receive total N deposition levels above 5.2 N kg/ha/yr, which is the
ecological benchmark for shifts in the dominant species of lichen from acidophytic to tolerant species. In
addition, in Coastal Scrub Sage ecosystems in California, N deposition exceeds the 3.3 N kg/ha/yr
benchmark above which nitrogen is no longer a limiting nutrient, leading to potential alterations in
ecosystem composition. Because excessive N deposition and effects are observed in areas where, under
recent conditions, NOX ambient concentrations are at or below the current NOX secondary standards, we
initially conclude those standards are not adequate to protect against anticipated adverse impacts from N
nutrient enrichment in sensitive ecosystems (systems where N is limiting) that are not managed for
commercial agricultural and forest production.
Sulfur deposition is also linked with the formation of methylmercury. The production of methylmercury
in aquatic ecosystems requires sulfate as well as mercury. The evidence is sufficient to infer a causal
relationship between sulfur deposition and increased mercury methylation in wetlands and aquatic
environments. However, while the production of methylmercury requires the presence of sulfate and
mercury, the amount of methylmercury produced varies with oxygen content, temperature, pH, and
supply of labile organic carbon. Due to limits in data, we are unable to assess the adequacy of the
existing standards in protecting against effects associated with increased mercury methylation.
Conceptual Design of an Ecologically Relevant Standard
The overall concept for ecologically relevant standards recognizes that the fundamental welfare effects
associated with ambient NOX and SOX occur through the process of deposition to sensitive ecosystems.
There are four main components to the conceptual design of the standard: atmospheric and ecological
indicators, deposition metrics, functions that relate indicators to deposition metrics and factors that
modify the functions. In this policy assessment, the focus is on developing the conceptual design for a
ES-5
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
standard that protects against effects associated with acidifying deposition of NOX and SOX in aquatic
ecosystems, but this general conceptual framework is intended to apply to a broader set of potential
endpoints.
For the conceptual design of an aquatic acidification standard, ANC is suggested as the ecological
indicator. ANC is suggested as the ecological indicator because it is the most widely used chemical
indicator of acid sensitivity in aquatic ecosystems and has been found through numerous studies to be the
best single indicator of the biological response and health of aquatic communities in acid-sensitive
systems. Furthermore, ANC can be directly linked to both underlying water chemistry, e.g. pH and
aluminum, and to biological impairment, specifically the number offish species in a water body.
Acidification models represent the ecological response relationship between ANC and deposition of N
and S. Acidification models are designed for the catchment scale. However, for consideration of a
national standard, aggregation to a broader spatial scale is desirable; therefore a method to evaluate
critical loads (the levels of deposition of N and S below which defined levels of harmful effects on
specified sensitive elements of the environment do not occur) across the national landscape is presented.
The atmospheric transformation functions then convert deposition to ambient concentrations of NOX and
SOX.
Acidification models relate ANC to deposition of N and S at the catchment scale. We suggest using an
acidification model that incorporates environmental variables that modify the ecological response
relationship. This includes a variable to account for nitrogen uptake by ecosystems. The acidification
model can be used to calculate critical loads for individual catchments based on a selected level of ANC.
The load of deposition that causes a selected level of ANC varies across the nation depending on the
characteristics of the catchment, such as base cation weathering rates, nitrogen retention and the level of
naturally occurring organic acids.
Although critical loads for a selected level of ANC will vary catchment by catchment it is not practical for
a national standard to evaluate every catchment in the U.S. Therefore, we propose two general approaches
to establish critical loads. One approach is to develop a national distribution of critical loads over all
levels of sensitivity, recognizing that there is a high degree of heterogeneity in acid sensitivity even at
relatively small spatial scales. The second approach is to subdived the landscape of the U.S.into acid-
sensitivity categories, such that within a category there are generally similar acid sensitivity
characteristics. Each national acid-sensitivity category is represented by a population of catchments for
which critical loads at a specified ANC limit are calculated. This second draft policy assessment explores
a number of methods for developing the acid sensitivity categories.
National acid-sensitivity categories should be based on features that govern ecological sensitivity to
acidification. Areas that have similar underlying geology, mineral weathering rates, and hydrology
should show similar sensitivity to NOX and SOX deposition. Ecoregions are useful geographic defiitions
that holistically incorporate a number of important factors related to acid sensitivity, including geology,
physiography, vegetation, climate, soils, land use, and hydrology. As such, the determination of acid-
sensitivity categories begins with aggregation of lakes and streams by ecoregion. From there, acid
sensitivity is further characterized using measured ANC. As noted earlier, ANC is a good indicator of the
overall sensitivity of a water body to acidification. ANC measurement is also widely available in the
U.S., making it a useful metric for further classifying ecoregions. The policy assessment explores several
methods for using ANC to classify ecoregions using both simple and more complex methods.
Once acid-sensitivity categories are defined, a sample of catchments will be selected to represent each
category. The acidification models will be used to evaluate the critical load for each catchment in the
ES-6
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
population at the selected level of ANC. There is a distribution of critical load values within each acid-
sensitivity category reflecting the range of sensitivity of catchments within each category. The goal of
aggregating critical loads from multiple catchments is to develop an appropriately representative
deposition value based on the distribution of critical loads, called the deposition metric, which protects a
percentage of the population of water bodies within a national acid-sensitivity category from exceeding
their critical load for the target value of ANC.
Distributions of catchment level critical loads are based on a combination of previously conducted steady-
state critical load modeling and new critical load modeling conducted as part of this policy assessment.
To ensure the population of water bodies included in the analysis were those sensitive to acidity caused
by atmospheric deposition, several criteria were applied to the critical loads dataset to remove catchments
in which organic acids, acid mine drainage or naturally low base cation weathering caused acidification.
Once a deposition metric is calculated, the value is modified by addition of a term to represent the amount
of N that will be taken up by vegetation, immobilized in soil or degassed from the ecosystem. Next, a
tradeoff curve for N and S deposition is generated for the acid sensitivity category. This function
(illustrated below) is characterized by three nodes: 1) the maximum of amount of N deposition when S
deposition equals zero 2) the amount of N deposition that will be captured by the ecosystem before it
leaches and 3) the maximum amount of S sulfur deposition considering the N captured by the ecosystem.
The function represents all pairs of N and S deposition that will equal the deposition metric for acidifying
deposition for a specific target ANC.
The depositional load function
Reduced forms of nitrogen, such as ammonia, are either taken up by plants and microbes or converted to
nitrate in the environment and use up the assimilative capacity of ANC at the same rate as oxidized forms
of nitrogen deposition; therefore, deposition of reduced nitrogen must be accounted for in the watershed.
The suggested approach is to subtract the loadings of reduced forms of nitrogen derived for a given
spatial area from the deposition metric that represents a selected percentage of critical loads for a given
population, such that the resultant deposition metric is for sulfur and oxidized nitrogen only. The policy
assessment explores several methods for specifying the loadings of reduced nitrogen for specific
geographic areas.
Deposition is related to ambient concentrations of NOX and SOX through deposition velocity, which is the
rate at which an ambient pollutant is deposited. Deposition velocity varies over time and location, and is
affected by land use conditions and meteorology. Our conceptual model requires conversion of
deposition of N and S into ambient concentrations of NOX and SOX. Since the policy objective is to set an
ambient air quality standard for total oxidized sulfur and nitrogen, and this is also the chemical resolution
provided by the ecosystem models, it is convenient to use conversion factors based on available estimates
ES-7
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
of concurrent deposition and ambient concentrations. The ratio of total deposition of NOX and SOX to
ambient concentrations of NOX and SOX, respectively are referred to as deposition transference ratios.
Applying the estimated deposition transference ratios to depositional load tradeoff curves leads to parallel
tradeoff curves for ambient NOX and SOX (see figure below).
The national ambient air quality standard must be able to tie ANC to deposition and deposition to ambient
air concentrations, incorporating ecological conditions and the contribution of reduced nitrogen. To
incorporate all of these aspects, we develop an index that will provide a consistent standard nationally that
is directly expressed in terms of concentrations of NOX and SOX. The form of this standard is referred to
as the Atmospheric Acidification Protection Index (AAPI), which can be applied across the nation to
convey the protection of aquatic ecosystems from acidification due to atmospheric deposition.
The AAPI represents the level of protection against the effects of acidification given local ecological
conditions, the level of reduced nitrogen being deposited, and the levels of NOX and SOX that are limited
by the standard. The AAPI is linked to a target ANC for a chosen percent of lakes and streams, and
determines the combinations of NOX and SOX that will jointly result in the target ANC, taking into
account uncertainties and other factors.
This AAPI can also be expressed as the tradeoff curves which show the combinations of NOX and SOX
that meet the standard, generated for specific values of AAPI, and provide a representation of the
standards in terms of atmospheric concentrations. These tradeoff curves will vary across the U.S. based
on ecosystem sensitivity, reduced nitrogen deposition levels, and other factors. An example tradeoff
curve for two different percentiles of protection at a target ANC of 50 (ieq/L is provided below.
90% and NHx=5
-90% and NHx=54
•75% 8, NHx =5
-75% 8, NHx =54
NOy (meq/m fyr)
D
j
\~\
90% and NHx=5
90% and NHx=54
75% & NHx =5
k 75% & NHx =54
'""•-.. A Adirondack air quality
N (meq/m /yr)
2 3
NOy(ug/m3)
Graphs demonstrate the steps to develop a tradeoff curve for the AAPI including A) N vs.
SO2+SO4from the deposition metric and modified with Neco B) Two loads of NHx indicated C)
after subtracting NHx load the resultant NOy vs SO2+SO4 curve, D) After applying the
transference ratio the resultant air concentrations.
Options for Elements of the Standards
ES-8
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
Ambient air quality indicators other than NO2 and SO2 should be considered as the appropriate pollutant
indicators for protection against the acidification effects associated with deposition of NOX and SOX. This
consideration is based on the recognition that all forms of oxidized nitrogen and sulfur in the atmosphere
contribute to deposition and resulting acidification, and as such concentrations of NO2 and SO2 are
incomplete indicators. Furthermore, concentration of NOy is proposed as an appropriate indicator for
oxides of nitrogen. The sum of concentrations of SO2 and SO4 is proposed as an appropriate indicator for
oxides of sulfur.
Welfare effects associated with acidification result from annual cumulative deposition of nitrogen and
sulfur, reflected in effects on the chronic ANC level (measured as annual or multiyear average ANC).
Short-term (i.e., hours or days) episodic changes in water chemistry can also have significant biological
effects. Episodic chemistry refers to conditions during precipitation or snowmelt events. Surface water
chemistry has lower pH and acid neutralizing capacity (ANC) during these events than under baseflow
conditions. One of the most important effects of acidifying deposition on surface water chemistry is the
short-term change in chemistry that is termed "episodic acidification." While ecosystems are also
affected by episodic increases in acidity due to pulses of acidity during high rainfall periods and
snowmelts, protection against these episodic acidity events can be achieved by establishing a higher
chronic ANC level. Episodic acidification can result from either shorter term deposition episodes, or
from longer term deposition on snowpack. Snowmelt can release stored N deposited throughout the
winter, leading to episodic acidification in the absence of increased deposition during the actual episodic
acidification event. Long term (3 to 5 year average) ambient NOX and SOX concentrations are appropriate
to provide protection against low chronic ANC levels and episodic acidification.
The current forms of the secondary standards for NOX and SOX do not take into account the combined
contributions of NOX and SOX in the causation of effects associated with acidification of aquatic
ecosystems. Based on the causal linkages between NOX and SOX, deposition of N and S, and the indicator
of acidification, ANC, the current forms should be replaced with an atmospheric acidification protection
index (AAPI), which reflects the important roles of underlying ecosystem characteristics, determinants of
deposition, and deposition of reduced nitrogen in determining the potential effects from deposition of
NOX and SOX.
The value of AAPI can be calculated for any observed values of NOX and SOX. However, the level of the
standard for AAPI should reflect a wide number of factors, including desired level of protection indicated
by a target ANC limit, the target percentile of water bodies to achieve the target ANC, and the various
factors and uncertainties involved in specifying all of the other aspects of the standard, such as the acid
sensitivity classification method, the specification of deposition of reduced nitrogen, nitrogen retention,
the deposition transference ratios, and the averaging time. The administrator may choose an AAPI level
reflecting an ANC level higher or lower than the target ANC limit to account for the combined effect of
all of the components of the standard and their related uncertainties. The resulting AAPI, in the context of
the overall standard, will reflect her informed judgment as to a standard that is sufficient but not more
than necessary to protect against adverse public welfare effects.
Within the AAPI form, EPA will specify the parameter values for all elements excepting NOX and SOX,
which will be the measured atmospheric indicators. The values for pre-industrial base cation weathering,
nitrogen retention and uptake, and runoff, are based on the same inputs used to develop the deposition
metrics. The values for reduced nitrogen and the deposition transference ratios will be calculated using
output from the most up-to-date version of EPA's Community Multiscale Air Quality (CMAQ) model.
EPA is considering methods to account for the dynamic nature of deposition of reduced nitrogen in
specifying the NOX and SOX standards.
ES-9
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
As demonstrated by the tradeoff curves, multiple combinations of concentrations of NOX and SOX can
yield the same value of the AAPI. No single combination of NOX and SOX will produce a particular value
of AAPI in all locations. Measured concentrations of annual average NOX and SOX necessary to meet the
standards are thus expressed conditionally by the AAPI and not by fixed quantities.
A target ANC limit based on a desired level of protection is an important input to the decisions of the
level of AAPI and the percent of ecosystems to be protected. Specific levels of ANC are associated with
differing levels of ecosystem impairment, with higher levels of ANC resulting in fewer ecosystem
impacts, and lower levels resulting in both higher intensity of impacts and a broader set of impacts. For
example, the number offish species present in a waterbody has been shown to be positively correlated
with the ANC level in the water, with higher values supporting a greater richness and diversity of fish
species. This relationship is illustrated in the following figure.
Severe Elevated Moderate
14
* • • •^
/ it-'
-200 -100 0 100 200 300 400 500
ANCipeq L)
Number of fish species per lake or stream versus ANC level and
aquatic status category for lakes in the Adirondack Case Study
Area
The target ANC level specified in designing the standard is only one part in determining the overall
protectiveness of the standard. The degree of protectiveness is based on all elements of the standard,
including the target ANC selected by the Administrator, the size of the spatial areas over which the
standard is applied, the percent of aquatic ecosystems targeted within a spatial area that is selected by the
Administrator to achieve the selected ANC level, and the underlying parameters of the AAPI, including
the atmospheric indicator, the critical load models used to determine pre-industrial base cation levels, the
calculated values for the deposition transference ratios, and the calculated value for deposition of reduced
nitrogen. There are widely varying degrees of uncertainty associated with all of these elements, some
being much more certain and others being much less certain. The specified target ANC level is a crucial
part of developing a standard that is requisite to protect public welfare, but it is the overall design and
content of the standard that must be considered in judging the adequacy of protection it provides.
The secondary NAAQS will reflect the public welfare policy judgments of the Administrator, based on
the science, as to the level of air quality which is requisite to protect the public welfare from any known
or anticipated adverse effects associated with the pollutant in the ambient air. In certain naturally acidic
ecosystems, even though the ecological benchmarks are exceeded, e.g. ANC may be quite low; NOX and
ES-10
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
SOX are not contributing to effects because those systems have chronic natural acidity and will not benefit
from reductions in atmospheric deposition. The secondary NAAQS are not intended to provide
protection in these types of naturally acidic systems. Instead, the secondary NAAQS are focused on
providing protection in areas where ambient NOX and SOX are resulting in effects in ecosystems with low
natural levels of acidification that are highly sensitive to additional inputs of acid deposition. The
approaches for specifying populations of critical loads to develop deposition metrics explicitly excludes
lakes and streams that are naturally acidic and those not likely to benefit from decreases in atmospheric
deposition.
ANC levels below 20 (ieq/L are generally associated with high probability of low pH, leading to death or
loss of fitness of biota that are sensitive to acidification. Overall, there is little uncertainty that significant
effects on aquatic biota are occurring at ANC levels below 20 (ieq/L. Based on the field data from the
Adirondacks and Shenendoah case study areas, ANC levels less than 50 (ieq/L are adverse to ecosystem
health, and are likely to lead to reductions in ecosystem services related to recreational fishing. However,
the types of effects, specific species, and prevalence of effects across water bodies in the U.S. is more
uncertain at ANC levels between 20 and 50 (ieq/L. Targeting ANC levels between 50 and 100 (ieq/L
would provide additional protection; however, uncertainties regarding the additional reduction in adverse
welfare effects are much larger for target ANC levels above 50 (ieq/L.
Specifying an appropriate range of levels for an AAPI standard that is designed and specified as discussed
above involves consideration of the degree to which any specific AAPI would lead to achieving the
desired ANC level, and a judgment as to the degree of protection of public welfare that is warranted.
Selection of a range of AAPI and selection of a specific level of AAPI within that range should
incorporate a wide number of considerations, including the percent of water bodies within acid sensitive
areas that the Administrator determines should be protected at the targeted ANC level.
In determining the requisite level of protection for the public welfare from effects on aquatic ecosystems,
the Administrator will need to weigh the importance of the predicted risks of these effects in the overall
context of public welfare protection, along with a determination as to the appropriate weight to place on
the associated uncertainties and limitations of this information.
Co-Protection for Other Effects Provided by an Acidification Standard
To understand the level of protection provided by aNOx/SOx standard based on aquatic acidification to
protect against terrestrial acidification effects, we compared the critical loads for lakes and streams that
would maintain an aquatic ANC of 50 to the critical loads to maintain either a terrestrial Be: Al ratio of 1.2
or 10 averaged across a watershed area.
Results for the Adirondacks showed that critical loads for 29 lakes at an ANC of 50 were lower for 13 of
those lakes than the critical load for the terrestrial watershed areas at a Bc:Al ratio of 10 and for 21 lakes
at a Bc:Al ratio of 1.2. Perhaps more significant was the result that 13 of the 16 lakes in the highly and
moderately sensitive areas had a lower critical load than the Bc:Al 10 areas and 16 of 16 lakes in the
highly and moderately sensitive areas had lower critical loads than the Bc:Al 1.2 areas. The Shenandoah
region reflected similar results.
In general, the aquatic critical acid loads offered greater protection to the watersheds than did the
terrestrial critical loads. Generally in situations where the terrestrial critical loads were more protective,
the lakes or streams in the watershed were rated as having "Low Sensitivity" or "Not Sensitive" to
acidifying nitrogen and sulfur deposition. Conversely, when the water bodies were more sensitive to
ES-11
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
deposition ("Highly Sensitive" or "Moderately Sensitive"), the aquatic critical acid loads generally
provided a greater level of protection against acidifying nitrogen and sulfur deposition in the watershed.
Initial Conclusions
In this current review, important new information has become available since the last reviews (1996 for
NOX, and 1988 for SOX) that supports revising the current NOX and SOX standards. Specifically, the ISA
has concluded that there are causal relationships between NOX and SOX acidifying deposition and effects
on aquatic and terrestrial ecosystems, and the ISA and REA provide substantial quantitative evidence of
effects occurring in locations that meet the current NO2 and SO2 standards. In addition, observational
data and rigorous atmospheric modeling has become available regarding the role of both nitrogen and
sulfur deposition in acidification of sensitive water bodies. This information is sufficient to inform the
development of revised secondary standards for NOX and SOX to protect against the effects of aquatic
acidification in sensitive ecosystems. Additional information is needed to set separate standards to
protect terrestrial ecosystems from acidification effects. While there is also new information available on
the role of nitrogen deposition on nutrient enrichment effects in sensitive terrestrial and aquatic
ecosystems, and the ISA concludes there is a causal relationship between NOX and nutrient enrichment
effects, for this second draft policy assessment, we have focused on acidification effects due to the
substantially greater amount of information available to inform the development of secondary standards.
We highlight the progress made in considering the joint nature of ecosystem responses to acidifying
deposition of NOX and SOX, and note that the ability to consider revisions to the NOX and SOX secondary
standards has been enhanced by our ability to consider a joint standard for NOX and SOX to protect against
acidification effects. The development of an appropriate form of the standard linked to a common
indicator of aquatic acidification, ANC, is also a significant step forward, as it allows for development of
a standard for aquatic acidification designed to provide generally the same degree of protection across the
country, while still reflecting the underlying variability in ecosystem sensitivity to acidifying NOX and
SOX deposition.
We provide the following initial conclusions regarding the NOX and SOX secondary standards:
• The available effects-based evidence for aquatic and terrestrial acidification and nutrient
enrichment suggests consideration of NOX and SOX standards that are at least as protective as the
current standard. Consideration of joint standards for NOX and SOX is appropriate given the
common atmospheric processes governing the deposition of NOX and SOX to sensitive
ecosystems.
• On the basis of the acidification and nutrient enrichment effects that have been observed to still
occur under current ambient conditions and those predicted to occur under the scenario of just
meeting the current secondary NAAQS, we find support for consideration that the current
secondary NAAQS are inadequate to protect the public welfare from known and anticipated
adverse welfare effects from aquatic and terrestrial acidification associated with deposition of
NOX and SOX.
• We find support for consideration that current levels of NOX and SOX are associated with
deposition that leads to ANC values below benchmark values that cause ecological harm and
losses in ecosystem services in sensitive ecosystems, including significant mortality in sensitive
aquatic biota and losses in fish species richness, which is associated with reductions in
recreational fishing services, among others.
• We find support for consideration that current levels of ambient NOX and SOX are associated with
deposition that leads to Be: Al values below benchmark values for terrestrial acidification that
cause ecological harm and losses in ecosystem services in sensitive ecosystems, including losses
ES-12
-------
Draft Policy Assessment for the Review of the Secondary NAAQS for NOX and SOX:
Executive Summary
in tree health and growth, which are associated with reductions in timber production, among other
services.
• We suggest that effects due to aquatic acidification are most suitable for defining secondary
standards for NOX and SOX. We note that in developing a standard designed to protect against the
effects of aquatic acidification due to deposition of NOX and SOX, the resulting standards may not
provide adequate protection against known effects associated with nutrient enrichment in
sensitive aquatic and terrestrial ecosystems or acidification in sensitive terrestrial ecosystems.
• Based on the causal linkages between NOX and SOX, deposition of N and S, and the indicator of
acidification, ANC, consideration should be given to an additional secondary standard with a
form defined by an atmospheric acidification protection index (AAPI), which reflects the
important roles of underlying ecosystem characteristics, determinants of deposition, and
deposition of reduced nitrogen in determining the potential effects from deposition of NOX and
SOX.
• Staff has concluded, based on the evidence and risk based information, and consideration of
information related to definitions of adversity, that:
• a target level of ANC of 20 (ieq/L will protect against significant losses in fish mortality in
many sensitive lakes, but will place less weight on protection against losses in aquatic
biodiversity, and will be less protective against potential acidification episodes,
• a target level of ANC of 50 (ieq/L will protect against significant mortality in aquatic
organisms and loss offish health and biodiversity in sensitive lakes and streams, and will
give weight to considerations of uncertainties in the time to recovery of aquatic ecosystems,
• target levels of ANC above 50 (ieq/L may provide additional protection against declines in
fitness of sensitive species (e.g., brook trout, zooplankton), however, overall health of aquatic
communities may not be impacted.
ES-13
-------
TABLE OF CONTENTS
List of Figures vii
List of Tables xi
List of Acronyms/Abbreviations xiii
List of Key Terms xvi
1 INTRODUCTION 1-1
1.1 DEFINITIONS OF NOx AND SOx FOR THIS ASSESSMENT 1-3
1.2 POLICY OBJECTIVES 1-8
1.3 CRITICAL POLICY ELEMENTS 1-10
1.4 HISTORICAL CONTEXT 1-12
1.4.1 History of NOx and SOx NAAQS Review 1-12
1.4.2 History of Related Assessments and Agency Actions 1-14
1.5 PROPOSED CONCEPTUAL FRAMEWORK FOR COMBINED
NOX SOX STANDARDS 1-17
1.6 POLICY RELEVANT QUESTIONS 1-21
2 KNOWN OR ANTICIPATED ADVERSE ECOLOGICAL EFFECTS 2-1
2.1 Acidification: Evidence of effects on structure and function of terrestrial and
freshwater ecosystems 2-2
2.1.1 What is the nature of acidification related ecosystem responses to
reactive nitrogen and sulfur deposition? 2-3
2.1.2 What types of ecosystems are sensitive to such effects?
In which ways are these responses affected by atmospheric, ecological, and
landscape factors? 2-6
2.1.3 What is the magnitude of ecosystem responses to acidifying deposition? .. 2-6
2.1.4 What are the key uncertainties associated with acidification? 2-17
2.2 Nitrogen-enrichment: Evidence of effects on structure and function of terrestrial and
freshwater ecosystems 2-19
2.2.1 What is the nature of terrestrial and freshwater ecosystem responses to
reactive nitrogen and sulfur deposition? 2-20
-------
2.2.2 What types of ecosystems are sensitive to such effects? How are these
responses affected by atmospheric, ecological, and landscape factors 2-22
2.2.3 What is the magnitude of ecosystem responses to nitrogen deposition? ... 2-23
2.2.4 What are the key uncertainties associated with nutrient enrichment? 2-32
2.3 What ecological effects are associated with gas-phase NOX and SOX? 2-33
2.3.1 What is the nature of ecosystem responses to gas-phase
nitrogen and sulfur? 2-33
2.3.2 What types of ecosystems are sensitive to such effects? How are these
responses affected by atmospheric, ecological, and landscape factors?.... 2-34
2.3.3 What is the magnitude of ecosystem responses to gas phase effects
ofNOxandSOx? 2-34
2.4 SUMMARY 2-35
2.5 REFERENCES 2-37
3 CONSIDERATIONS OF ADVERSITY TO PUBLIC WELFARE 3-1
3.1 How do we characterize adversity? What are the relevant definitions and how are
they addressed in this document? 3-1
3.1.1 What are the benchmarks for adversity from other sources? 3-2
3.1.2 Other EPA Programs and Federal Agencies 3-4
3.2 What are ecosystem services and how does this concept relate to
public welfare? 3-10
3.3 Applying Economic Valuation to Ecosystem Services 3-17
3.3.1 Ecosystem services and links to public welfare 3.17
3.3.2 Economics as a framework to illustrate changes in public welfare 3-17
3.3.3 The role of economics in defining "adversity" 3-19
3.4 Effects of Acidification and Nutrient Enrichment on Ecosystem Services 3-24
3.4.1 Evidence for adversity related to aquatic acidification 3-26
3.4.2 Evidence for Adversity Related to Terrestrial Acidification 3-31
3.4.3 Evidence for Adversity Related to Aquatic Nutrient Enrichment 3-35
3.4.4 Evidence for Adversity Related to Terrestrial Nutrient Enrichment 3-39
3.5 REFERENCES 3-42
4 ADDRESSING THE ADEQUACY OF THE CURRENT STANDARDS 4-1
11
-------
4.1 Are the structures of the current NOx and SOx secondary standards based on
relevant ecological indicators such that they are adequate to determine and protect
public welfare against adverse effects on ecosystems? 4-1
4.2 To what extent are the structures of the current NOx and SOx secondary standards
meaningfully related to relevant ecological indicators of public welfare effects? .. 4-3
4.3 To what extent do current monitoring networks provide a sufficient basis for
determining the adequacy of current secondary NOx and SOx standards? 4-6
4.3.1 To what extent does the NADP monitoring network provide an adequate
characterization of deposition and what are the major limitations? 4-13
4.3.2 What are the relative strengths and important gaps in existing atmospheric
monitoring networks to address a combined NOX/SOX
secondary standard? 4-15
4.3.3 How do we characterize deposition through monitoring and model s? 4-16
4.4 What is our best characterization of atmospheric concentrations of NOy and SOx,
and deposition of N and S? 4-18
4.4.1 What are the current atmospheric concentrations of reactive nitrogen, NOy,
reduced nitrogen, NHX, sulfur dioxide, SO2, and sulfate, SO4? 4-18
4.5 Are adverse effects on the public welfare occurring under current air quality
conditions for NO2 and SO2 and would they occur if the nation met the current
standards? 4-38
4.5.1 To what extent do the current NOx and SOx secondary standards provide
protection from effects associated with deposition of atmospheric NOx and
SOx which results in acidification in sensitive aquatic and terrestrial
ecosystems? 4-40
4.5.2 To what extent does the current NOx secondary standard provide protection
from effects associated with deposition of oxidized nitrogen from
atmospheric NOx, which results in nutrient enrichment effects in sensitive
aquatic and terrestrial ecosystems? 4-45
4.5.3 Aquatic Nutrient Enrichment 4-47
4.5.4 Terrestrial Nutrient Enrichment 4-48
in
-------
4.6 To what extent do the current NOx and/or SOx secondary standards provide
protection from other ecological effects (eg. mercury methylation) associated with
the deposition of atmospheric NOX, and/or SOX? 4-49
4.7 REFERENCES 4-51
5 OPTIONS FOR ELEMENTS OF THE STANDARD 5-1
5.1 What atmospheric indicators of oxidized nitrogen and sulfur are appropriate for use
in a secondary NAAQS that provides protection for public welfare from exposure
related to deposition of NOx and SOx? 5-3
5.2 What is the appropriate averaging time for the air quality indicators NOy and SOx to
provide protection of public welfare from adverse effects from
aquatic acidification? 5-4
5.3 What form(s) of the standard are most appropriate to provide protection of sensitive
ecosystems from the effects of acidifying deposition related to ambient NOx and
SOx concentrations? 5-5
5.3.1 Conceptual Design of the Form: General Overview 5-5
5.3.2 Conceptual design of the form: Linking the ecosystem indicator
to deposition 5-11
5.3.3 Conceptual Design: Linking Deposition to Atmospheric Concentration . 5-53
5.3.4 AAPI 5-61
5.4 Options for specifying the targets for the ecological indicator for aquatic
acidification (ANC) 5-68
5.4.1 What levels of impairment are related to alternative target levels of
ANC? 5-69
5.4.2 Additional information on Target ANC levels 5-84
5.4.3 Adversity of ecological impacts associated with alternative target ANC
levels 5-85
5.5 Options for specifying the targets for the deposition metric 5-87
5.6 REFERENCES 5-90
6 CO-PROTECTION FOR OTHER EFFECTS USING STANDARDS TO
PROTECT AGAINST ACIDIFICATION 6-1
IV
-------
6.1 To what extent would a standard specifically defined to protect against aquatic
acidification likely provide protection from terrestrial acidification? 6-1
6.2 To what extent would a standard specifically defined to protect against aquatic
acidification likely provide protection from terrestrial nutrient enrichment? 6-4
6.3 To what extent would a standard specifically defined to protect against aquatic
acidification likely provide protection from aquatic nutrient enrichment? 6-6
6.4 REFERENCES 6-7
7 Evaluation of Uncertainty and Variability in the Context of an AAPI standard,
including Model Evaluation, Sensitivity Analyses, and Assessment of Information
Gaps 7-1
7.1 INTRODUCTION and PURPPOSE 7-1
7.2 Uncertainty associated with ecosystem effects and dose - response relationships . 7-5
7.3 Uncertainty in benefits estimates 7-7
7.4 CMAQ application and evaluation 7-9
7.4.1 Overview of cmaq model application 7-9
7.4.2 CMAQ Evaluation, Sensitivity and Variability Analyses 7-10
7.4.3 Variability and sensitivity of CMAQ generated component 7-12
7.5 Sensitivity of AAPI to component parameters 7-30
7.6 Uncertainty in Critical Load and ANC modeling 7-30
7.6.1 MAGIC modeling 7-30
7.6.2 SSWC modeling 7-31
7.7 Modeling and Data Gaps 7-34
7.8 Summary and Conclusions 7-36
7.9 REFERENCES 7-40
8 AMBIENT AIR MONITORING 8-1
8.1 What are the appropriate ambient air indicators to consider in developing the
standards? 8-1
8.2 Reactive Oxidized Nitrogen and Sulfur Species 8-5
8.3 What measurements would be used to characterize NOy and SOx ambient air
concentrations for the purposes of the AAPI based standard? 8-7
8.4 What additional complementary measurements are recommended? 8-8
-------
8.5 What sampling frequency would be required? 8-9
8.6 What are the spatial scale issues associated with monitoring for compliance, and
how should these be addressed? 8-10
8.7 What specific monitoring methods would be used? 8-10
8.8 REFERENCES 8-14
9 INITIAL CONCLUSIONS 9-1
9.1 CONSIDERATION OF ALTERNATIVE STANDARDS 9-3
9.1.1 Indicators 9-5
9.1.2 Averaging Times 9-5
9.1.3 Form 9-6
9.1.4 Alternative levels of the AAPI standard 9-9
9.1.5 Additional protections for ecosystems against the effects of terrestrial
acidification and terrestrial and aquatic nutrient enrichment 9-15
9.1.6 Summary of options 9-16
9.2 CONCLUSIONS 9-16
9.3 REFERENCES 9-21
APPENDICES
Appendix A Additional Information for Chapter 5
Appendix B Analysis of Critical Loads, Comparing Aquatic and Terrestrial
Acidification
Appendix C Elasticity and ANOVA AAPI sensitivity analyses
VI
-------
List of Figures
Figure 1-1 Framework of an alternative secondary standard 1-21
Figure 2-1 Conceptual model of direct and indirect acidification effects on aquatic biota... 2-4
Figure 2-2 Average NCV concentrations (orange), SC>42" concentrations (red), and ANC
(blue) across the 44 lakes in the Adirondack Case Study Area modeled using
MAGIC for the period 1850 to 2050 2-11
Figure 2-3 ANC concentrations of preacidification (1860) and current (2006) conditions
based on hindcasts of 44 lakes in the Adirondack Case Study Area modeled using
MAGIC 2-12
Figure 2-4 Critical loads of acidifying deposition that each surface water location can receive
in the Adirondack Case Study Area while maintaining or exceeding an ANC
concentration of 50 ueq/L based on 2002 data 2-13
Figure 2-5 Average N(V concentrations orange), SO42"concentrations (red), and ANC (blue)
levels for the 60 streams in the Shenandoah Case Study Area modeled using
MAGIC for the period 1850 to 2050 2-15
Figure 2-6 ANC levels of 1860 (preacidification) and 2006 (current) conditions based on
hindcasts of 60 streams in the Shenandoah Case Study Area modeled using
MAGIC 2-15
Figure 2-7 Critical loads of surface water acidity for an ANC of 50 ueq/L for Shenandoah
Case Study Area streams. Each dot represents an estimated amount of acidifying
deposition (i.e., critical load) that each stream's watershed can receive and still
maintain a surface water ANC >50 ueq/L 2-16
Figure 2-8 Benchmarks of atmospheric nitrogen deposition for several ecosystem indicators
(REA 5.3.1.2) MCF-Mixed Conifer Forest, CSS-Coastal Sage Scrub 2-25
Figure 2-9 Observed effects from ambient and experimental atmospheric nitrogen deposition
loads in relation to using CMAQ 2002 modeling results and NADP monitoring
data 2-26
Figure 3-1 Common anthropogenic stressors and the essential ecological attributes they
affect 3-3
Figure 3-2 European maps of eutrophication (left) and acidification (right) which protect
95% of natural areas in 50x50 km2 European Monitoring and Evaluation
Programme grid 3-9
Figure 3-3 Representation of the benefits assessment process indicating where some
ecological benefits may remain unrecognized, unquantified, or unmonetized.. 3-12
Figure 3-4 Conceptual model showing the relationships among ambient air quality indicators
and exposure pathways and the resulting impacts on ecosystems, ecological
responses, effects and benefits to characterize known or anticipated adverse
effects to public welfare 3-14
Figure 3-5 Locations of Eastern U.S. National Parks (Class I areas) relative to deposition of
Nitrogen and Sulfur in sensitive aquatic areas 3-15
Figure 3-6 Location of Western U.S. National Parks (Class I areas) relative to deposition of
Nitrogen and Sulfur 3-16
Figure 3-7 Conceptual model linking ecological indicator (ANC) to affected ecosystem
services 3-25
vn
-------
Figure 4-1 Routinely operating surface monitoring stations measuring forms of atmospheric
nitrogen 4-9
Figure 4-2 Routinely operating surface monitoring stations measuring forms of atmospheric
sulfur 4-10
Figure 4-3 Anticipated network of surface based NOy stations based on 2009 network design
plans 4-12
Figure 4-4 Location of approximately 250 National Atmospheric Deposition Monitoring
(NADP) National Trends Network (NTN) sites illustrating annual ammonium
deposition for 2005 4-14
Figure 4-5 2005 CMAQ modeled annual average NOY (ppb) 4-21
Figure 4-6 2005 CMAQ modeled annual average total reduced nitrogen (NHx) (as ug/m3
nitrogen) 4-22
Figure 4-7 2005 CMAQ modeled annual average ammonia, NH3, (as ug/m3 N) 4-23
Figure 4-8 2005 CMAQ modeled annual average ammonium, NH4, (as ug/m3 N) 4-24
Figure 4-9 2005 CMAQ modeled annual average SOX, (as ug/m3 S from SO2 and SO4) .. 4-25
Figure 4-10 2005 CMAQ modeled annual average SO2(asug/m3 S) 4-26
Figure 4-11 2005 CMAQ modeled annual average SO4 (as ug/m3 S) 4-27
Figure 4-12 2005 annual average sulfate concentrations based on CASTNET generated by the
Visibility Information Exchange Web Sysytem (VIEWS) 4-28
Figure 4-13 2005 annual average sulfate concentrations based on CASTNET generated by the
Visibility Information Exchange Web Sysytem (VIEWS) 4-29
Figure 4-14 Annual average 2005 NOy concentrations from reporting stations in AQS 4-30
Figure 4-15 2005 CMAQ modeled oxidized nitrogen deposition (kgN/Ha-yr) 4-31
Figure 4-16 2005 CMAQ modeled oxidized sulfur deposition (kgS/Ha-yr) 4-32
Figure 4-17 2005 CMAQ derived annual average ratio of reduced to total nitrogen
deposition 4-33
Figure 4-18 Three hour average maximum 2005 SO2 concentrations based on the SLAMS
reporting to EPA's Air Quality System (AQS) database 4-34
Figure 4-19 Annual average 2005 NO2 concentrations based on the SLAMS reporting to
EPA's Air Quality System (AQS) data base 4-35
Figure 4-20 2005 CMAQ derived annual average ratio of (NOY - NO2)/NOY 4-36
Figure 4-21 Scatter plots of total oxidized nitrogen deposition with average annual NO2,
HNOs and NOy concentrations on each 12 km2 grid based on 2005 CMAQ results
for the Adirondack region 4-37
Figure 4-22 National map highlighting the 9 case study areas evaluated in the REA 4-41
Figure 5-1 Conceptual design of the form of NOx and SOx secondary standard 5-6
Figure 5-2 Conceptual design of the form of the standard based on aquatic acidification ... 5-7
Figure 5-3 Detailed conceptual design of the form of the standard based on aquatic
acidification 5-10
Figure 5-4 A. The relationship between pH and ANC under equilibrium conditions with
mineral phase gibbsite. Triangles indicate calculated values while circles indicate
measurements (Bi and Liu 2001). B. The relationship between precipitation (wet
and average year), ANC and the risk to exposure of a pH below 5.5 (Gerritsen et
al. 1996) 5-13
Figure 5-5 Illustration of a generalized N + S deposition tradeoff curve that is calculated by
using the FAB approach 5-19
Vlll
-------
Figure 5-6 Relationship between CLs using pre-industrial base cation weathering ([BC]o*
=BC0) calculated by MAGIC versus the F-factor methods 5-24
Figure 5-7 Sites of CLs calculated by SSWC that are used in the Policy Assessment
analysis 5-27
Figure 5-8 Cumulative distribution of the critical loads for ANC =50 ueq/L considered in the
analysis (n =5281) 5-30
Figure 5-9 Map that illustrates Omernick's ecoregions (level 3) 5-33
Figure 5-10 Map of sensitive ecoregions (red) using binary categorization approach (sensitive
vs. less-sensitive) with ANC observations between -50 and 200 ueq/L 5-37
Figure 5-11 Map of sensitive ecoregions based on clusters of log ANC values between -50 and
200 ueq/L 5-39
Figure 5-12 Deposition metrics for 90%,75% and 50% of the population for options, 1 (all
data), 2a (binary classification), and 2b (clusters based on log ANC) 5-44
Figure 5-13 Comparison of sites that are protected if the nation is not subdivided and if the
nation is subdivided into a sensitive and non-sensitive category 5-45
Figure 5-14 Cumulative distribution of the 90% deposition metric for
ecoregions is shown 5-46
Figure 5-15 Examples of N + (SC>2 +S(V2) deposition tradeoff curves for option 1 (A) and
option 2a(B and C) 4-48
Figure 5-16 A U.S. map of NH3 deposition (meq/m2/yr) overlaying ecoregions (level 3)
boundaries 5-50
Figure 5-18 Examples of N + (SC>2 +SO/f2) deposition tradeoff curves for option 5-51
Figure 5-19 Examples of NOy + (SC>2 +SCV2) deposition tradeoff curves for option 1 (A) and
option 2a(B and C). Deposition metrics that correspond to protection 75% or 90%
of the ecosystems considering both the high and low value for NHx
deposition 5-52
Figure 5-20 2005 Tt values for each grid cell in the eastern U.S. domain 5.58
Figure 5-21 Schematic Diagram illustrating the procedure for converting deposition tradeoff
curves of sulfur and nitrogen to atmospheric concentrations of SOx and NOy 5-59
Figure 5-22 Inter-annual coefficients of variation (CV) of a) NOy and b) SOx Tt values,
based on a series of 2002-2005 CMAQ v4.7 simulation 5-60
Figure 5-23 Relationship between ANC and pH levels 5-70
Figure 5-24 Summary of results from Gerritson et al 1996. Blue back herring larval mortality
occurs at pH 6.2. The higher the ANC the less likely it is that the pH will dip
below 6.2 5-71
Figure 5-25 Mean residual number of species per lake for lakes in Ontario, by pH interval.
The residual number of species for a lake is the deviation of the observed number
from the number predicted by lake area 5-76
Figure 5-26 Relationship between ANC and number offish species present in aquatic
freshwater ecosystems in Shenandoah National Park 5-79
Figure 5-27 Relationship between ANC and number offish species present in aquatic
freshwater ecosystems in Shenandoah National Park 5-80
Figure 6-1 Benchmarks of atmospheric nitrogen deposition for several
ecosystem indicators 6-5
Figure 7-1 2002 CMAQv4.6 annual average SO2 predicted concentrations versus
observations at CASTNet sites in the eastern domain 7-17
IX
-------
Figure 7-2 2002 CMAQv4.6 annual average SO2"4 predicted concentrations versus
observations at CASTNet sites in the eastern domain 7-18
Figure 7-3 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the eastern domain 7-19
Figure 7-4 2002 CMAQv4.6 annual average NH4+ predicted concentrations versus
observations at CASTNet sites in the eastern domain 7-20
Figure 7-5 2002-2005 Domain-wide average SC>42" predicted concentrations and
observations by month at CASTNet Sites in the eastern domain 7-21
Figure 7-6 2002-2005 Domain-wide average TNOs predicted concentrations and
observations by month at CASTNet sites in the eastern domain 7-22
Figure 7-7 2002-2005 Domain-wide average NH4+ predicted concentrations and
observations by month at CASTNet sites in the eastern domain 7-23
Figure 7-8 Comparison of CMAQ predictions and measurements for 12-hour (6am-6pm)
average NH3 concentrations, with a monitoring cycle of 4 days on and 4days off,
at a high emission site (Kenansville) and a low emission urban site (Raleigh) in
North Carolina compared to CMAQ for July 2004 (top) and August 2004
(bottom), from Dennis et al., 2010 7-24
Figure 7-9 Unadjusted (left) and PRISM (right) adjusted CMAQ annual wet deposited sulfate
for 2002 7-25
Figure 7-10 Unadjusted (left) and PRISM and bias (right) adjusted CMAQ annual wet
deposition of nitrate (top) and ammonium (bottom) 7-26
Figure 7-11 Spatial and interannual variability of inverse deposition transference ratios, l/TSox
and I/TNOY, for Adirondack (top) and Shenandoah case study areas 7-27
Figure 7-12 Summary of inter-annual and emissions sensitivity variability of sulfur and
nitrogen deposition transference ratios 7-28
Figure 7-13 Cumulative regional NH3 budget of advection, wet- and dry deposition,
calculated for an expanding box starting at the high-emitting Sampson County NC
cell 7-29
Figure 7-14 Simulated versus observed annual average surface water SO42-, NO3-, ANC, and
pH during the model calibration period for each of the 44 lakes in the
Adirondacks Case Study Area 7-33
Figure 7-15 Simulated versus observed annual average surface water SO42-, NO3-, ANC, and
pH during the model calibration period for each of the 60 streams in the
Shenandoah Case Study Area. The black line is the 1:1 line 7-34
Figure 8-1 Annual 2002 - 2004 CMAQ derived annual average fraction of ambient
concentrations (above ) and deposition (below) of individual NOy species
delineated by the Adirondack and Shenandoah case study areas and the remainder
of the Eastern U.S. domain 8-15
Figure 8-2 Examples of the Relative Abundance of Several NOy Species Measured at Two
Rural Southeastern Canadian Sites as a Fraction of the Total Measured NOy
Concentration 8-16
Figure 8-3 Annual average fraction of NOy ambient air contributed by NC>2 based on 2005
CMAQ Eastern U.S. simulation at 12 km grid cell resolution 8-17
Figure 8-4 Annual average fraction of NOy ambient air contributed by HNOs based on 2005
CMAQ Eastern U.S. simulation at 12 km grid cell resolution 8-18
-------
Figure 8-5 Annual average fraction of NOy ambient air contributed by PAN based on 2005
CMAQ Eastern U.S. simulation at 12 km grid cell resolution 8-19
Figure 8-6 Annual 2002 - 2004 CMAQ derived annual average fraction of ambient
concentrations (above ) and deposition (below) of individual SOx species
delineated by the Adirondack and Shenandoah case study areas and the remainder
of the Eastern U.S. domain 8-20
Figure 9-1 Spatially interpolated CMAQ estimates of deposition of reduced nitrogen (2002-
2004 average) 9-8
Figure 9-2 Range of NOx-SOx Tradeoff Curves Across Acid Sensitivity Categories for
Alternative Levels of Reduced Nitrogen Deposition, for a Target ANC of 50
ueq/L and a Target of Protection for 90 Percent of Water Bodies 9-14
List of Tables
Table 1-1 Description of parameters, units and conventions 1-5
Table 2-1 Ecological Effects Associated with Alternative Levels of Acid Neutralizing
Capacity (ANC) 2-8
Table 2-2 Summary several commonly used acidification models 2-9
Table 3-1 Crosswalk between Ecosystem Services and Public Welfare Effects 3-11
Table 3-2 Count of Impacted Lakes 3-28
Table 3-3 Present Value and annualized Benefits, Adirondack Region 3-29
Table 3-4 Aggregate Annual Benefit Estimates for the Zero-Out Scenario 3-30
Table 3-5 Annual participation and value of outdoor (forest related) activity
in the northeast 3-32
Table 3-6 Summary of Studies of Select Terrestrial Ecosystem Services 3-33
Table 3-7 Summary of Values for Current Levels of Services and Changes in
Service Levels in $2007 3-37
Table 3-8 Summary of Annual Damages to Services due to Atmospheric Loading 3-38
Table 3-9 Summary of Current Levels of Ecosystem Services 3-40
Table 4-1 Summary of Monitoring Networks 4-7
Table 5-1 Descriptive statistics of the national CL data set 5-29
Table 5-2 Summary of data sources considered for the evaluation of national ANC 5-34
Table 5-3. National ANC data. Observations from 6318 sites remained after filtering data to
remove sites with an average value of ANC<-50 and >200 ueq/L 5-36
Table 5-4 Descriptive statistics of the CL populations that result when the US is divided into
two categories, sensitive and less-sensitive based on ANC data 5-37
Table 5-5 Descriptive statistics of the CL populations that result when the US is divided into
5 clusters based on log ANC values between -50 and 200 ueq/L 5-39
Table 5-6 Descriptive statistics of the populations of CL ANC=50 ueq/L values that result
for each ecoregion 5-40
Table 5-7 Summary of the number of deposition metrics that would result from the options
for how to categorized the landscape based on acid-sensitivity 5-43
Table 5-9 Summary of Fish Mortality Response to pH 5-73
Table 5-10 Threshold response of increased mortality offish to low pH listed from least
sensitive to most sensitive 5-74
XI
-------
Table 5-11 General summary of biological changes anticipated with surface water
acidification, expressed as a decrease in surface water pH 5-77
Table 5-12 Comparison of percentage protection from ANC values less than 50 and less than
20 using DL that result when the US is considered one population 5-88
Table 5-13 Comparison of percentage protection from ANC values less than 50 and less than
20 using DL that result when the US is divided into two categories, sensitive and
less sensitive based on ANC data 5-89
Table 6-1 Results of the comparison of lake and stream aquatic critical loads (ANC of 50
jieq/L) to terrestrial critical loads (Bc:Al molar ratios of 10.0 in soil solution)
calculated for the full watershed in each of the 16 watersheds in the Adirondacks
and Shenandoah Case Study Areas 6-3
Table 6-2 Results of the comparison of lake and stream aquatic critical loads (ANC of 50
ueq/L) to terrestrial critical loads (Bc:Al molar ratios of 1.2 in soil solution)
calculated for the full watershed in each of the 16 watersheds in the Adirondacks
and Shenandoah Case Study Areas 6-3
Table 7-1 Normalized Mean Bias Statistics for Predicted and Observed Pollutant
Concentration 7-11
Table 7-2 Summary (Incomplete) of Qualitative Uncertainty Analysis of Key Modeling
Elements in the NOx/SOx AAPI 7-38
xn
-------
List of Acronyms and Abbreviations
AAPI
ADR
A13+
ANC
AQCD
AQRV
ASSETS El
Bc/Al
C
Ca/Al
Ca2+
CAA
CASAC
CASTNet
CCS
Chi a
CLE
CMAQ
CSS
CWA
DIN
DO
DOT
EMAP
EPA
FHWAR
FIA
FWS
GIS
GPP
H+
H20
H2SO4
ha
HAB
HFC
Hg+2
Hg°
HNO3
HONO
HUC
IMPROVE
ISA
K+
Atmospheric Acidification Potential Index
Adirondack Mountains of New York
aluminum
acid neutralizing capacity
Air Quality Criteria Document
air quality related values
Assessment of Estuarine Trophic Status eutrophication index
Base cation to aluminum ratio, also Be: Al
carbon
calcium to aluminum ratio
calcium
Clean Air Act
Clean Air Scientific Advisory Committee
Clean Air Status and Trends Network
coastal sage scrub
chlorophyll a
critical load exceedance
Community Multiscale Air Quality model
coastal sage scrub
Clean Water Act
dissolved inorganic nitrogen
dissolved oxygen
U.S. Department of Interior
Environmental Monitoring and Assessment Program
U.S. Environmental Protection Agency
fishing, hunting and wildlife associated recreation survey
Forest Inventory and Analysis National Program
Fish and Wildlife Service
geographic information systems
gross primary productivity
hydrogen ion
water vapor
sulfuric acid
hectare
harmful algal bloom
hydrofluorocarbon
reactive mercury
elemental mercury
nitric acid
nitrous acid
hydrologic unit code
Interagency Monitoring of Protected Visual Environments
Integrated Science Assessment
potassium
Xlll
-------
kg/ha/yr
km
LRMP
LTER
LTM
MAGIC
MCF
MEA
Mg2+
N
N2
N20
N203
N204
N205
Na+
NAAQS
NADP
NAPAP
NAWQA
NEEA
NEP
NH3
NH4+
(NH4)2SO4
NHx
NO
NO2
NO3"
NOAA
NOX
NOy
NPP
NFS
NRC
NSWS
NTN
NTR
03
OAQPS
OW
PAN
PFC
pH
ppb
kilograms per hectare per year
kilometer
Land and Resource Management Plan
Long Term Ecological Monitoring and Research
Long-Term Monitoring
Model of Acidification of Groundwater in Catchments
Mixed Conifer Forest
Millennium Ecosystem Assessment
magnesium
nitrogen
gaseous nitrogen
nitrous oxide
nitrogen tri oxide
nitrogen tetr oxide
dinitrogen pentoxide
sodium
National Ambient Air Quality Standards
National Atmospheric Deposition Program
National Acid Precipitation Assessment Program
National Water Quality Assessment
National Estuarine Eutrophication Assessment
net ecosystem productivity
ammonia gas
ammonium ion
ammonium sulfate
category label for NH3 plus NH4+
nitric oxide
nitrogen dioxide
reduced nitrite
reduced nitrate
National Oceanic and Atmospheric Administration
nitrogen oxides
total oxidized nitrogen
net primary productivity
National Park Service
National Research Council
National Surface Water Survey
National Trends Network
organic nitrate
ozone
Office of Air Quality Planning and Standards
Office of Water
peroxyacyl nitrates
perfluorocarbons
relative acidity
parts per billion
xiv
-------
ppm
ppt
PSD
REA
REMAP
S
S203
S207
SAV
SF6
SMP
SO
SO2
SO3
SO32
SO4
SO42
SOM
SOX
SPARROW
SRB
STORE!
TIME
TMDL
TP
USFS
USGS
ueq/L
Hg/m3
2-
2-
parts per million
parts per trillion
prevention of significant deterioration
Risk and Exposure Assessment
Regional Environmental Monitoring and Assessment Program
sulfur
thiosulfate
heptoxide
submerged aquatic vegetation
sulfur hexafluoride
Simple Mass Balance
sulfur monoxide
sulfur dioxide
sulfur trioxide
sulfite
wet sulfate
sulfate ion
soil organic matter
sulfur oxides
SPAtially Referenced Regressions on Watershed Attributes
sulfate-reducing bacteria
STORage and RETrieval
Temporally Integrated Monitoring of Ecosystems
total maximum daily load
total phosphorus
U.S. Forest Service
U.S. Geological Survey
microequivalents per liter
micrograms per cubic meter
xv
-------
List of Key Terms
Acidification: The process of increasing the acidity of a system (e.g., lake, stream, forest soil).
Atmospheric deposition of acidic or acidifying compounds can acidify lakes, streams,
and forest soils.
Air Quality Indicator: The substance or set of substances (e.g., PM2.5, NO2, SO2) occurring in
the ambient air for which the National Ambient Air Quality Standards set a standard level
and monitoring occurs.
Alpine: The biogeographic zone made up of slopes above the tree line, characterized by the
presence of rosette-forming herbaceous plants and low, shrubby, slow-growing woody
plants.
Acid Neutralizing Capacity: A key indicator of the ability of water to neutralize the acid or
acidifying inputs it receives. This ability depends largely on associated biogeophysical
characteristics, such as underlying geology, base cation concentrations, and weathering
rates.
Arid Region: A land region of low rainfall, where "low" is widely accepted to be less than 250
mm precipitation per year.
Base Cation Saturation: The degree to which soil cation exchange sites are occupied with base
cations (e.g., Ca2+, Mg2+, K+) as opposed to A13+ and H+. Base cation saturation is a
measure of soil acidification, with lower values being more acidic. There is a threshold
whereby soils with base saturations less than 20% (especially between 10%-20%) are
extremely sensitive to change.
Ecologically Relevant Indicator: A physical, chemical, or biological entity/feature that
demonstrates a consistent degree of response to a given level of stressor exposure and
that is easily measured/quantified to make it a useful predictor of ecological risk.
Critical Load: A quantitative estimate of an exposure to one or more pollutants, below which
significant (as defined by the analyst or decision maker) harmful effects on specified
sensitive elements of the environment do not occur, according to present knowledge.
Denitrification: The anaerobic reduction of oxidized nitrogen (e.g., nitrate or nitrite) to gaseous
nitrogen (e.g., N2O or N2) by denitrifying bacteria.
Dry Deposition: The removal of gases and particles from the atmosphere to surfaces in the
absence of precipitation (e.g., rain, snow) or occult deposition (e.g., fog).
Ecological Risk: The likelihood that adverse ecological effects may occur or are occurring as a
result of exposure to one or more stressors (U.S. EPA, 1992).
Ecological Risk Assessment: A process that evaluates the likelihood that adverse ecological
effects may occur or are occurring as a result of exposure to one or more stressors (U.S.
EPA, 1992).
Ecosystem: The interactive system formed from all living organisms and their abiotic (i.e.,
physical and chemical) environment within a given area. Ecosystems cover a hierarchy of
spatial scales and can comprise the entire globe, biomes at the continental scale, or small,
well-circumscribed systems such as a small pond.
Ecosystem Benefit: The value, expressed qualitatively, quantitatively, and/or in economic terms,
where possible, associated with changes in ecosystem services that result either directly
or indirectly in improved human health and/or welfare. Examples of ecosystem benefits
that derive from improved air quality include improvements in habitats for sport fish
species, the quality of drinking water and recreational areas, and visibility.
xvi
-------
Ecosystem Function: The processes and interactions that operate within an ecosystem.
Ecosystem Services: The ecological processes or functions having monetary or non-monetary
value to individuals or society at large. These are (1) supporting services, such as
productivity or biodiversity maintenance; (2) provisioning services, such as food, fiber, or
fish; (3) regulating services, such as climate regulation or carbon sequestration; and (4)
cultural services, such as tourism or spiritual and aesthetic appreciation.
Eutrophication: The process by which nitrogen additions stimulate the growth of autotrophic
biota, usually resulting in the depletion of dissolved oxygen.
Nitrogen Enrichment: The process by which a terrestrial system becomes enhanced by nutrient
additions to a degree that stimulates the growth of plant or other terrestrial biota, usually
resulting in an increase in productivity.
Nitrogen Saturation: The point at which nitrogen inputs from atmospheric deposition and other
sources exceed the biological requirements of the ecosystem; a level beyond nitrogen
enrichment.
Occult Deposition: The removal of gases and particles from the atmosphere to surfaces by fog
or mist.
Semi-arid Regions: Regions of moderately low rainfall, which are not highly productive and are
usually classified as rangelands. "Moderately low" is widely accepted as between 100-
and 250-mm precipitation per year.
Sensitivity: The degree to which a system is affected, either adversely or beneficially, by NOx
and/or SOx pollution (e.g., acidification, nutrient enrichment). The effect may be direct
(e.g., a change in growth in response to a change in the mean, range, or variability of
nitrogen deposition) or indirect (e.g., changes in growth due to the direct effect of
nitrogen consequently altering competitive dynamics between species and decreased
biodiversity).
Total Reactive Nitrogen: This includes all biologically, chemically, and radiatively active
nitrogen compounds in the atmosphere and biosphere, such as NH3, NH4+, NO, NO2,
HNO3, N2O, NO3-, and organic compounds (e.g., urea, amines, nucleic acids).
Valuation: The economic or non-economic process of determining either the value of
maintaining a given ecosystem type, state, or condition, or the value of a change in an
ecosystem, its components, or the services it provides.
Variable Factors: Influences which by themselves or in combination with other factors may
alter the effects on public welfare of an air pollutant (section 108 (a)(2))
(a) Atmospheric Factors: Atmospheric conditions that may influence transformation,
conversion, transport, and deposition, and thereby, the effects of an air pollutant on
public welfare, such as precipitation, relative humidity, oxidation state, and co-pollutants
present in the atmosphere.
(b) Ecological Factors: Ecological conditions that may influence the effects of an air
pollutant on public welfare once it is introduced into an ecosystem, such as soil base
saturation, soil thickness, runoff rate, land use conditions, bedrock geology, and
weathering rates.
Vulnerability: The degree to which a system is susceptible to, and unable to cope with, the
adverse effects of NOx and/or SOx air pollution.
Welfare Effects: The effects on soils, water, crops, vegetation, man-made materials, animals,
wildlife, weather, visibility, and climate; as well as damage to and deterioration of
xvn
-------
property, hazards to transportation, and the effects on economic values and on personal
comfort and well-being, whether caused by transformation, conversion, or combination
with other air pollutants (Clean Air Act Section 302[h]).
Wet Deposition: The removal of gases and particles from the atmosphere to surfaces by rain or
other precipitation.
xvin
-------
1 1 INTRODUCTION
2
3 The U.S. Environmental Protection Agency (EPA) is presently conducting a review of
4 the secondary National Ambient Air Quality Standards (NAAQS) for oxides of nitrogen (NOx)
5 and oxides of sulfur (SOx). The EPA's overall plan and schedule for this review were presented
6 in the Integrated Review Plan for the Secondary National Ambient Air Quality Standards for
7 Nitrogen Dioxide and Sulfur Dioxide (US EPA, 2007). The Integrated Review Plan (IRP)
8 outlined the Clean Air Act (CAA or the Act) requirements related to the establishment and
9 reviews of the NAAQS, the process and schedule for conducting the current review, and the key
10 components in the NAAQS review process: an Integrated Science Assessment (ISA), Risk and
11 Exposure Assessment (REA), and policy assessment/rulemaking. It presented key policy -
12 relevant issues to be addressed in this review as a series of questions that frames our
13 consideration of whether the current secondary (welfare-based) NAAQS for NOx and SOx
14 should be retained or revised.
15 As part of this review, staff in the U.S. Environmental Protection Agency's (EPA) Office
16 of Air Quality Planning and Standards (OAQPS) prepared this second draft Policy Assessment.1
17 The objective of this assessment is to evaluate the policy implications of the key scientific
18 information contained in the document Integrated Science Assessment for Oxides of Nitrogen
19 and Sulfur-Ecological Criteria (USEPA, 2008; henceforth referred to as the ISA), prepared by
20 EPA's National Center for Environmental Assessment (NCEA) and the results from the analyses
21 contained in the Risk and Exposure Assessment for Review of the Secondary National Ambient
22 Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur (U.S. EPA, 2009; henceforth
23 referred to as the REA). This second draft also presents staff conclusions on a range of policy
24 options that we believe are appropriate for the Administrator to consider concerning whether,
25 and if so how, to revise the secondary (welfare-based) NOx and SOx NAAQS.
1 Preparation of a PA by OAQPS staff reflects Administrator Jackson's decision to modify the NAAQS review
process that was presented in the IRP. See http://www.epa.gov/ttn/naaqs/review.html for more information on the
current NAAQS review process.
September, 2010 1-1 Draft - Do Not Quote or Cite
-------
1 This policy assessment is intended to help "bridge the gap" between the scientific
2 assessment contained in the ISA and the judgments required of the EPA Administrator in
3 determining whether it is appropriate to retain or revise the secondary NAAQS for NOx and
4 SOx. This policy assessment considers the available scientific evidence and quantitative risk-
5 based analyses, together with related limitations and uncertainties, and focuses on the basic
6 components of air quality standards: indicators2, averaging times, forms3, and levels. These
7 components, which serve to define each standard, must be considered collectively in evaluating
8 the welfare protection afforded by the secondary NOx and SOx NAAQS. Our development of
9 this policy assessment is based on the assessment and integrative synthesis of information
10 presented in the ISA and on staff analyses and evaluations presented in this document, and is
11 further informed by comments and advice received from an independent scientific review
12 committee, the Clean Air Scientific Advisory Committee (CASAC), in their review of the
13 previous integrated science assessment, risk and exposure assessment, and first draft policy
14 assesment. To view related documents developed as part of the planning, science, and risk
15 assessment phases of this review see
16 http://www.epa.gov/ttn/naaqs/standards/no2so2sec/index.html.
17 This document is organized around a conceptual framework for a combined NOx and SOx
18 secondary NAAQS and is focused on answering key policy questions related to the
19 implementation of that conceptual framework. Chapter 2 provides a summary of ecological
20 effects from the deposition of ambient NOx and SOx to sensitive ecosystems, drawing from the
21 ISA and REA. Chapter 3 places those ecological effects within the context of "public welfare"
22 by linking effects to ecosystem services or other benchmarks of public welfare. Chapter 4
23 addresses the adequacy of the current NOx and SOx secondary NAAQS in addressing the
24 impacts on public welfare from ecological effects. Chapter 5 develops the conceptual design for
25 ecologically relevant multi-pollutant standards and presents options for developing critical
26 components of a secondary NAAQS necessary to implement the conceptual design. Chapter 6
27 describes how secondary NAAQS designed to protect a specific ecological endpoint may also
28 provide protection for other ecological endpoints. Chapter 7 provides an assessment of critical
2 The "indicator" of a standard defines the chemical species or mixture that is to be measured in determining
whether an area attains the standard.
3 The "form" of a standard defines the air quality statistic that is to be compared to the level of the standard in
determining whether an area attains the standard.
September 2010 1-2 Draft -Do Not Quote or Cite
-------
1 uncertainties and sensitivitie considered in developing the options for the components of the
2 standard. Chapter 8 discusses issues related to monitoring of NOx and SOx. Chapter 9 provides
3 initial staff conclusions regarding options for pollutant indicators, averaging times, forms, and
4 ranges of levels for the secondary NOx and SOx NAAQS.
5 In this document we consider how the available scientific evidence and quantitative risk-
6 based analyses, together with related limitations and uncertainties, inform the review of each
7 element of the NAAQS: indicator, averaging times, forms, and levels. These components must be
8 considered collectively in evaluating the welfare protection afforded by the secondary NAAQS
9 standards. This draft document does not contain final staff conclusions as to all the necessary
10 components of an alternative secondary standard for NOx and/or SOx but rather describes the
11 current state of thinking with regard to potential policy options and provides an appropriate
12 context of information for the Administrator to consider in making decisions regarding the
13 standards.
14 While this policy assessment should be of use to all parties interested in the secondary
15 NOx and SOx NAAQS review, it is written with an expectation that the reader has some
16 familiarity with the technical discussions contained in the ISA and REA.
17 EPA's final Policy Assessment will address additional CAS AC comments on this second
18 draft, and will include sufficient information to inform the Administrator on critical components
19 of the standards, and staff conclusions regarding alternative levels of the standards.
20
21 1.1 DEFINITIONS, PARAMETERS, UNITS, AND CONVENTIONS USED FOR THIS
22 ASSESSMENT
23 Throughout this document numerous terms are used that address a variety of atmospheric
24 and ecosystem processes and variables. Some of the more common terms used in the technical
25 community are not always synonymous with definitions imbedded in the CAA. Because of this
26 diversity of terms spanning atmospheric and ecosystem processes along with adherence to
27 scientific and legal conventions, this section provides the terminology as a reference source for
28 the entire report.
29 As discussed in detail in the REA (REA 1.3.1), in the atmospheric science community
30 NOx is typically referred to as the sum of nitrogen dioxide (NO2), and nitric oxide (NO). As
31 defined by the Clean Air Act, the family of NOx includes any gaseous combination of nitrogen
September 2010 1-3 Draft -Do Not Quote or Cite
-------
1 and oxygen (e.g., NO2, NO, nitrous oxide pSPzO], dinitrogen trioxide [N^Os], dinitrogen tetroxide
2 [N2O4], and dinitrogen pentoxide pSkOs]). The term used by the scientific community to
3 represent the complete set of reactive oxidized nitrogen compounds, including those listed in
4 CAA Section 108(c) with the exception of N2O, is total oxidized nitrogen (NOy), commonly
5 defined as NO, NO2 and the all of the oxidation products of NO and NO2 Reactive oxidized
6 nitrogen is defined as NOy = NO2 + NO + HNO3 + PAN +2N2O5 + HONO+ NO3 + organic
7 nitrates + particulate NO3 (Finlayson-Pitts and Pitts, 2000). In this document, unless otherwise
8 indicated, we use the term NOy as the atmospheric indicators associated with the NOx
9 component of the proposed NOx/SOx standard .
10 For this assessment, SOx is defined to include all oxides of sulfur, including multiple
11 gaseous substances (e.g., SO2, sulfur monoxide [SO], sulfur trioxide [SO3], thiosulfate [S2O3],
12 and heptoxide [82©?], as well as particulate species, such as ammonium sulfate [(NH/t^SO/t]).
13 Throughout this text we refer to sulfate as SO4 and nitrate as NO3, recognizing that they have
14 charges of -2 for sulfate and -1 for nitrate. The sum of sulfur dioxide gas (802) and particulate
15 sulfate (804), referred herein as (SO2 + 804) is used throughout this document as the
16 atmospheric indictor for the SOx component of the proposed NOx/SOx standard. From a
17 measurement and modeling perspective we only consider the sum of SO2 and particulate SO4 as
18 the indicator for sulfur. The sum of SO2 and SO4 constitute virtually all of the ambient air sulfur
19 budget and are measured routinely in monitoring networks.
20 Table 1-1 provides further explanation of these indicators, some of which is repeated in
21 Chapters 4, 5 and 8. Table 1 also provides details on the units used throughout the equations
22 and examples in the PAD. Again, because of difference in unit conventions between
23 atmospheric and ecosystem sciences, there are detailed explanations of units as well as
24 procedures for translating between different unit conventions. To facilitate the linkage between
25 atmospheric and ecosystem processes, only the mass (or equivalent charge) associated with
26 sulfur or nitrogen is considered in mass, mixing ratio, and deposition unit conventions.
27
28
29
30
September 2010 1-4 Draft -Do Not Quote or Cite
-------
1 Table 1-1. Description of parameters, units and conventions.
Parameter
Units
Conversions to other
unit conventions used in
figures and calculations
(multiply value in Units
column by:
Explanation
Atmospheric species
CMAQ defined NOy species: NO (nitrogen oxide), NO2 (nitrogen dioxide), HNO3 (nitric acid), p-NO3 (paniculate
bound nitrate), NO3 (sum of HNO3 and p-NO3), PAN (peroxy acetyl nitrate), N2O5 (dinitrogen pentoxide), PANX
(higher order PANs), NTR (organic nitrates), PNA (HNO4); sulfur dioxide (SO2), paniculate sulfate (SO4);
NHX species: NH3 (ammonia), ammonium ion (NH4)
Lumped Atmospheric Species
NOy
The sum of all reactive oxidized nitrogen compounds derived through
summing all nitrogen contributions (i.e., 1-HNO3 + 2-N205 +...) from the
modeled species (HNO3, p-NO3, NO2, NO, PAN, ...) or through direct
measurement which reduces all oxidized nitrogen species to NO and reports
as ppb NO. All references to the quantity NOy refer to the mass, molar or
equivalent charge contribution of nitrogen only. All mass contributions of
oxygen, hydrogen and carbon are not included.
(SO2 +SO4)
Oxidized forms of sulfur defined as sulfate (SO4 + SO2); mass units
maintained for consistency with deposition calculations
Note that only mass as sulfur is counted in state variables; in practice,
individual SO2 and SO4 are measured/modeled and converted to mass of
sulfur atoms or equivalent charge units. Mass contribution of oxygen is
not included.
Reduced nitrogen calculated as the sum of NH3 and NH4. All references to
the quantity NHX used as state variables refer to the mass, molar or
equivalent charge contribution of nitrogen only. Mass contribution of
oxygen is not included.
Atmospheric State Variables used in equations and derivations
NOy concentration
SOX
NHX
Used in various
conventions of:
[ig/m3 as N or S
ppb = (MA/MO- p^) -fig/m3
where pair is the air density
in units of (kg/m3);
pair=28.97(10)-3-P/(R-T)
R = 8.206(10)'
5m3atm/(mol-K)
P = atm
T = degrees K
MA = molecular weight of
air (28.97)
M; = Atomic weight of
nitrogen (14) or sulfur (32)
meq/m3 = (1/M;) -[Jig/m3
September 2010
1-5
Draft -Do Not Quote or Cite
-------
Parameter
NOy deposition
(SO2 +SO4)
NHX
Used in various
conventions of:
Dep,
^-tmax
^(SO4+SO2)
s~i max
^ NOy
s-i min
^ NOy
Dry
V.
Wet
V.
n Dry
Dept
Wet
Dept
Depjotal
TSO*
T^Oy
Units
meq/m2-yr as TV
orS
[ig/m3
l*g/m3
lig/m3
m/yr
m/yr
meq/m2-yr
meq/m2-yr
meq/m2-yr
m/yr
Conversions to other
unit conventions used in
figures and calculations
(multiply value in Units
column by:
hg/ha-yr =(Mi/q)(10)-2 •
meq/m2-yr
where q = charge (1 for N, 2
forS)
Calculated by dividing total
((SO2+SO4)orNOy
deposition (wet and dry) by
the annual average (SO2
+SO4)orNOy
concentration.
Explanation
is the concentration of
(SO2 +SO4) in the
atmosphere consistent
with DL™
is the concentration of
NOy in the atmosphere
consistent with DL™X
is the concentration of
NOy in the atmosphere
consistent with DL™m
dry deposition velocities
wet deposition velocities
dry deposition fluxes
wet deposition fluxes
total (wet+dry)
deposition
the transfer ratio, which
can be considered an
aggregated, "effective"
deposition velocity that
relates total deposition
of(SO2+SO4)orNOyto
the total ambient
concentration, and
represents an average of
the chemical species
specific v1Tot( = v1Dry +
v,Wet) values
September 2010
1-6
Draft -Do Not Quote or Cite
-------
Parameter
Units
Conversions to other
unit conventions used in
figures and calculations
(multiply value in Units
column by:
Explanation
Ecosystem variables
ANC*
ANQimit
CL/ANClimG)
Q
Qws
NECO
Nfeach
DL%eco(0
pvT-max
LJL'N
fieq/L
fieq/L
meq/m2-yr
m/yr
m/yr
meq/m2-yr
meq/m2-yr
meq/m -yr
meq/m -yr
calculated value of ANC
a "target" ANC level
(ueq/L)
depositional load that
does not cause the
catchment to exceed a
given ANQimj where /'
indicates the pollutant of
interest
Average surface water
runoff for an acid
sensitive area (this is
typically equal to
precipitation -
evapotranspiration
Catchment level surface
water runoff (m/yr) (this
is typically equal to
precipitation -
evapotranspiration
nitrogen retention and
denitrification by
terrestrial catchment
N leaching
The deposition metric,
defined as the amount of
deposition that protects
a selected percentage of
individual catchments
for a population of
water bodies from
exceeding their
DLANciim(i), where /
indicates the pollutant
of interest
In the tradeoff curve
for DLo/oeco(/), the
maximum of amount
of N deposition when
S deposition equals
September 2010
1-7
Draft -Do Not Quote or Cite
-------
Parameter
DL™n
DL™X
Units
meq/m2-yr
meq/m2-yr
Conversions to other
unit conventions used in
figures and calculations
(multiply value in Units
column by:
Explanation
zero
In the tradeoff curve
for DLo/oeco(/'), the
amount of N
deposition that will be
captured by the
ecosystem before it
leaches
In the tradeoff curve
for DLo/oeco(/'), the
maximum amount of S
sulfur deposition
considering NECO
2 1.2 POLICY OBJECTIVES
3 In conducting this periodic review of the NOx and SOx secondary NAAQS, EPA has
4 decided to jointly assess the scientific information, associated risks, and standards relevant to
5 protecting the public welfare from adverse effects associated with oxides of nitrogen and sulfur.
6 Although EPA has historically adopted separate secondary standards for oxides of nitrogen
7 (NOx) and oxides of sulfur (SOx), EPA is conducting a joint secondary review of these standards
8 because NOx, SOx, and their associated transformation products are linked from an atmospheric
9 chemistry perspective, as well as from an environmental effects perspective. The National
10 Research Council (NRC) has recommended that EPA consider multiple pollutants, as
11 appropriate, in forming the scientific basis for the NAAQS (NRC, 2004). There is a strong basis
12 for considering these pollutants together, building upon EPA's and CASAC's past recognition of
13 the interactions of these pollutants and on the growing body of scientific information that is now
14 available related to these interactions and associated ecological effects.
15 EPA sets secondary standards for two criteria pollutants related to NOx and SOx: ozone
16 and particulate matter (PM). NOx is a precursor to the formation of ozone in the atmosphere,
17 and under certain conditions, can combine with atmospheric ammonia to form ammonium
18 nitrate, a component of fine PM. SOx is a precursor to the formation of parti culate sulfate,
19 which is a significant component of fine PM in many parts of the U.S. While there are a number
September 2010
Draft -Do Not Quote or Cite
-------
1 of welfare effects associated with ozone and fine PM, including ozone damage to vegetation, and
2 visibility degradation related to PM, protection against those effects is provided by the ozone and
3 fine PM standards. This review focuses on evaluation of the protection provided by NOx and
4 SOx secondary standards for effects associated with direct atmospheric concentrations of NOx
5 and SOx, and effects associated with deposition of NOx and SOx to sensitive ecosystems,
6 including deposition in the form of particulate nitrate and sulfate in their component forms.
7 The ISA highlights the ecological effects associated with deposition of ambient NOx and
8 SOx to sensitive ecosystems other than commercially managed forests and agricultural lands.
9 This assessment primarily focuses on the effects of ambient NOx and SOx via deposition on
10 multiple ecological receptors, but also evaluates information on gas-phase effects of NOx and
11 SOx via stomatal exposure on vegetation, which are the effects that the current secondary
12 standards protect against. The ISA highlighted effects including those associated with
13 acidification and nitrogen nutrient enrichment. Based on these highlighted effects, EPA's policy
14 objective is to develop a framework for NOx and SOx standards that incorporate factors that will
15 lead to standards that are ecologically relevant, and that recognizes the interactions between the
16 two pollutants as they deposit to sensitive ecosystems, with an ultimate goal of setting standards
17 that, based on the ecological criteria described in the ISA, and consistent with the requirements
18 of the Clean Air Act, "are requisite to protect the public welfare from any known or anticipated
19 adverse effects associated with the presence of such air pollutant in the ambient air."
20 In presenting policy options for the Administrator's consideration, we note that the final
21 decision on retaining or revising the current secondary standards for NOx and SOx is largely a
22 public welfare policy judgment based on the Administrator's informed assessment of what
23 constitutes requisite protection against adverse effects to public welfare. A final decision should
24 draw upon scientific information and analyses about welfare effects, exposure and risks, as well
25 as judgments about the appropriate response to the range of uncertainties that are inherent in the
26 scientific evidence and analyses. The ultimate determination as to what level of damage to
27 ecosystems and the services provided by those ecosystems is adverse to public welfare is not
28 wholly a scientific question, although it is informed by scientific studies linking ecosystem
29 damage to losses in ecosystem services, and information on the value of those losses in
30 ecosystem services. Our approach to informing these judgments, as discussed below, is
31 consistent with the requirements of the NAAQS provisions of the Clean Air Act and with how
September 2010 1-9 Draft -Do Not Quote or Cite
-------
1 EPA and the courts have historically interpreted the Act. These provisions require the
2 Administrator to establish secondary NAAQS that, in the Administrator's judgment, are
3 requisite to protect public welfare from any known or anticipated adverse effects associated with
4 the presence of NOx and SOx in the ambient air. In so doing, the Administrator seeks to
5 establish standards that are neither more nor less stringent than necessary for this purpose.
6 For this second draft policy assessment, we have chosen to focus much of our discussion
7 on the effects of ambient NOx and SOx on ecological impacts associated with acidifying
8 deposition of nitrogen and sulfur, which is a transformation product of ambient NOx and SOx.
9 We have the greatest confidence in the causal linkages between NOx and SOx and aquatic
10 acidification effects, and we have the most complete information available with which to develop
11 an ecologically meaningful structure for the standards.
12
13 1.3 CRITICAL POLICY ELEMENTS
14 Our policy objective is guided by the information in the ISA and REA, framed within the
15 legislative requirements of the CAA. This framing leads us to focus on critical policy elements
16 (CPE) consistent with elements of Clean Air Act language.
17 Sections 108 and 109 of the CAA govern the establishment and periodic review of the
18 NAAQS and of the air quality criteria upon which the standards are based. The NAAQS are
19 established for pollutants that are listed under section 108, based on three criteria, including
20 whether emissions of the air pollutant cause or contribute to air pollution which may reasonably
21 be anticipated to endanger public health or welfare and whose presence in the ambient air results
22 from numerous or diverse mobile or stationary sources. The NAAQS are based on air quality
23 criteria that reflect the latest scientific knowledge, useful in indicating the types and extent of
24 identifiable effects on public health or welfare that may be expected from the presence of the
25 pollutant in ambient air. The criteria refer to criteria issued pursuant to §108 of the Clean Air
26 Act, which include "(A) those variable factors (including atmospheric conditions) which of
27 themselves or in combination with other factors may alter the effects on public health or welfare
28 of such air pollutant; (B) the types of air pollutants which, when present in the atmosphere, may
29 interact with such pollutant to produce an adverse effect on public health of welfare; and (C) any
30 known or anticipated adverse effects on welfare."
September 2010 1-10 Draft -Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The following critical policy elements for the design of ecologically relevant secondary
standards forNOx and SOx are identified:
(CPE 1) An evaluation of the effects of ambient NOx and SOx on ecosystems, and the
relationship between those effects and the measure of dose in the ecosystem,
indicated by the deposit!onal loadings of N and S.
(CPE 1.1) Evaluation of the relationship between response of ecological
receptors, e.g. changes in diversity offish species, and the response related to
public welfare, e.g. loss in recreational fishing services.
(CPE 1.2) Evaluation of the extent to which identified effects are occurring
under recent conditions, and the extent to which meeting the current standards
would provide protection against these effects.
(CPE 2) An assessment of how best to characterize, in defining the standards, the
variable ecosystem factors that affect the relationship between ecological
effects and deposit! onal loadings of N and S.
(CPE 2.1) Specification of potential indicators of ecological effects, e.g. acid
neutralizing capacity (ANC) that incorporates variability in ecosystem factors.
(CPE 3) Characterization of the complex relationships between ambient concentrations
of NOx and SOx and deposition of N and S in the specification of a standard.
(CPE 4) Specification of the form for the standard(s), including ambient atmospheric
indicators for NOx and SOx, with consideration of averaging times, and
options for levels of the standard(s).
The development of the conceptual framework for the NOx and SOx standards described
in Section 1.4 will be motivated by these critical policy elements. However, in order to provide a
historical context for this new framework, the next section provides a brief history of previous
reviews of the NOx and SOx secondary NAAQS, as well as other relevant historical reviews of
welfare effects associated with these pollutants.
September 2010
1-11
Draft -Do Not Quote or Cite
-------
1
2 1.4 HISTORICAL CONTEXT
3 1.4.1 History of NOx and SOx NAAQS Review
4 1.4.1.1 NOx NAAQS
5 EPA began the most recent previous review of the NOx secondary standards in 1987 and in
6 November 1991, EPA released an updated draft AQCD for CAS AC and public review and
7 comment (56 FR 59285). This draft document provided a comprehensive assessment of the
8 available scientific and technical information on health and welfare effects associated with NO2
9 and other NOx. CAS AC reviewed the draft document at a meeting held on July 1, 1993, and
10 concluded in a closure letter to the Administrator that the document "provides a scientifically
11 balanced and defensible summary of current knowledge of the effects of this pollutant and
12 provides an adequate basis for EPA to make a decision as to the appropriate NAAQS for NOf
13 (Wolff, 1993). The AQCD Air Quality Criteria for Oxides of Nitrogen was then finalized (U.S.
14 EPA, 1993). EPA also prepared a Staff Paper that summarized and integrated the key studies
15 and scientific evidence contained in the revised NOx AQCD and identified the critical elements
16 to be considered in the review of the NO2 NAAQS. CASAC reviewed two drafts of the Staff
17 Paper and concluded in a closure letter to the Administrator that the document provided a
18 "scientifically adequate basis for regulatory decisions on nitrogen dioxide" (Wolff, 1995). In
19 October 1995, the Administrator announced her proposed decision not to revise either the
20 primary or secondary NAAQS for NO2 (60 FR 52874; October 11, 1995). A year later, the
21 Administrator made a final determination not to revise the NAAQS for NO2 after careful
22 evaluation of the comments received on the proposal (61 FR 52852; October 8, 1996). The level
23 for both the existing primary and secondary NAAQS for NO2 is 0.053 ppm (100 micrograms per
24 cubic meter [|Jg/m3] of air), annual arithmetic average, calculated as the arithmetic mean of the
25 1 -hour NO2 concentrations.
26
27 1.4.1.2 SOx NAAQS
28 Based on the 1970 SOX criteria document (DHEW, 1970), EPA promulgated primary and
29 secondary NAAQS for SO2 on April 30, 1971 (36 FR 8186). The secondary standards included
30 a standard at 0.02 ppm in an annual arithmetic mean and a 3-hour average of 0.5 ppm, not to be
31 exceeded more than once per year. These secondary standards were established solely on the
32 basis of evidence of adverse effects on vegetation. In 1973, revisions made to Chapter 5
September 2010 1-12 Draft -Do Not Quote or Cite
-------
1 ("Effects of Sulfur Oxide in the Atmosphere on Vegetation") of Air Quality Criteria for Sulfur
2 Oxides (U.S. EPA, 1973) indicated that it could not properly be concluded that the vegetation
3 injury reported resulted from the average SO2 exposure over the growing season, rather than
4 from short-term peak concentrations. Therefore, EPA proposed (38 FR 11355) and then finalized
5 (38 FR 25678) a revocation of the annual mean secondary standard. At that time, EPA was aware
6 that SOX air concentrations have other public welfare effects, including effects on materials,
7 visibility, soils, and water. However, the available data were considered insufficient to establish
8 a quantitative relationship between specific ambient SOX concentrations and effects (38 FR
9 25679).
10 In 1979, EPA announced that it was revising the Air Quality Criteria Document (AQCD)
11 for sulfur oxides concurrently with that for particulate matter and would produce a combined
12 particulate matter and sulfur oxides criteria document. Following its review of a draft revised
13 criteria document in August 1980, CAS AC concluded that acid deposition was a topic of
14 extreme scientific complexity because of the difficulty in establishing firm quantitative
15 relationships among (1) emissions of relevant pollutants (e.g., SC>2 and oxides of nitrogen), (2)
16 formation of acidic wet and dry deposition products, and (3) effects on terrestrial and aquatic
17 ecosystems. CAS AC also noted that acid deposition involves, at a minimum, several different
18 criteria pollutants: oxides of sulfur, oxides of nitrogen, and the fine particulate fraction of
19 suspended particles. CAS AC felt that any document on this subject should address both wet and
20 dry deposition, since dry deposition was believed to account for a substantial portion of the total
21 acid deposition problem.
22 For these reasons, CASAC recommended that a separate, comprehensive document on
23 acid deposition be prepared prior to any consideration of using the NAAQS as a regulatory
24 mechanism for the control of acid deposition. CASAC also suggested that a discussion of acid
25 deposition be included in the AQCDs for nitrogen oxides and PM and SOX. Following CASAC
26 closure on the AQCD for SC>2 in December 1981, EPA's Office of Air Quality Planning and
27 Standards published a Staff Paper in November 1982, but the paper did not directly assess the
28 issue of acid deposition. Instead, EPA subsequently prepared the following documents: The
29 Acidic Deposition Phenomenon and Its Effects: Critical Assessment Review Papers, Volumes I
30 and II (U.S. EPA, 1984a, b), and The Acidic Deposition Phenomenon and Its Effects: Critical
31 Assessment Document (U.S. EPA, 1985) (53 FR 14935 -14936). These documents, though they
September 2010 1-13 Draft -Do Not Quote or Cite
-------
1 were not considered criteria documents and did not undergo CASAC review, represented the
2 most comprehensive summary of relevant scientific information completed by EPA at that point.
3 On April 26, 1988 (53 FR 14926), EPA proposed not to revise the existing primary and
4 secondary standards for SC>2 . This proposal regarding the secondary SC>2 NAAQS was due to the
5 Administrator's conclusions that (1) based upon the then-current scientific understanding of the
6 acid deposition problem, it would be premature and unwise to prescribe any regulatory control
7 program at that time, and (2) when the fundamental scientific uncertainties had been decreased
8 through ongoing research efforts, EPA would draft and support an appropriate set of control
9 measures.
10
11 1.4.2 History of Related Assessments and Agency Actions
12 In 1980, the Congress created the National Acid Precipitation Assessment Program
13 (NAPAP) in response to growing concern about acidic deposition. The NAPAP was given a
14 broad 10-year mandate to examine the causes and effects of acidic deposition and to explore
15 alternative control options to alleviate acidic deposition and its effects. During the course of the
16 program, the NAPAP issued a series of publicly available interim reports prior to the completion
17 of a final report in 1990 (NAPAP, 1990).
18 In spite of the complexities and significant remaining uncertainties associated with the
19 acid deposition problem, it soon became clear that a program to address acid deposition was
20 needed. The Clean Air Act Amendments of 1990 included numerous separate provisions related
21 to the acid deposition problem. The primary and most important of the provisions, the
22 amendments to Title IV of the Act, established the Acid Rain Program to reduce emissions of
23 SC>2 by 10 million tons and NOX emissions by 2 million tons from 1980 emission levels in order
24 to achieve reductions over broad geographic regions. In this provision, Congress included a
25 statement of findings that led them to take action, concluding that (1) the presence of acid
26 compounds and their precursors in the atmosphere and in deposition from the atmosphere
27 represents a threat to natural resources, ecosystems, materials, visibility, and public health; (2)
28 the problem of acid deposition is of national and international significance; and (3) current and
29 future generations of Americans will be adversely affected by delaying measures to remedy the
30 problem.
September 2010 1-14 Draft -Do Not Quote or Cite
-------
1 Second, Congress authorized the continuation of the NAPAP in order to assure that the
2 research and monitoring efforts already undertaken would continue to be coordinated and would
3 provide the basis for an impartial assessment of the effectiveness of the Title IV program.
4 Third, Congress considered that further action might be necessary in the long term to
5 address any problems remaining after implementation of the Title IV program and, reserving
6 judgment on the form that action could take, included Section 404 of the 1990 Amendments
7 (Clean Air Act Amendments of 1990, Pub. L. 101-549, § 404) requiring EPA to conduct a study
8 on the feasibility and effectiveness of an acid deposition standard or standards to protect
9 "sensitive and critically sensitive aquatic and terrestrial resources." At the conclusion of the
10 study, EPA was to submit a report to Congress. Five years later, EPA submitted its report,
11 entitled Acid Deposition Standard Feasibility Study: Report to Congress (U.S. EPA, 1995) in
12 fulfillment of this requirement. The Report concluded that establishing acid deposition standards
13 for sulfur and nitrogen deposition may at some point in the future be technically feasible,
14 although appropriate deposition loads for these acidifying chemicals could not be defined with
15 reasonable certainty at that time.
16 Fourth, the 1990 Amendments also added new language to sections of the CAA
17 pertaining to the scope and application of the secondary NAAQS designed to protect the public
18 welfare. Specifically, the definition of "effects on welfare" in Section 302(h) was expanded to
19 state that the welfare effects include effects ".. .whether caused by transformation, conversion, or
20 combination with other air pollutants."
21 In 1999, seven Northeastern states cited this amended language in Section 302(h) in a
22 petition asking EPA to use its authority under the NAAQS program to promulgate secondary
23 NAAQS for the criteria pollutants associated with the formation of acid rain. The petition stated
24 that this language "clearly references the transformation of pollutants resulting in the inevitable
25 formation of sulfate and nitrate aerosols and/or their ultimate environmental impacts as wet and
26 dry deposition, clearly signaling Congressional intent that the welfare damage occasioned by
27 sulfur and nitrogen oxides be addressed through the secondary standard provisions of Section
28 109 of the Act." The petition further stated that "recent federal studies, including the NAPAP
29 Biennial Report to Congress: An Integrated Assessment, document the continued-and increasing-
30 damage being inflicted by acid deposition to the lakes and forests of New York, New England
31 and other parts of our nation, demonstrating that the Title IV program had proven insufficient."
September 2010 1-15 Draft -Do Not Quote or Cite
-------
1 The petition also listed other adverse welfare effects associated with the transformation of these
2 criteria pollutants, including impaired visibility, eutrophication of coastal estuaries, global
3 warming, and tropospheric ozone and stratospheric ozone depletion.
4 In a related matter, the Office of the Secretary of the U.S. Department of Interior
5 requested in 2000 that EPA initiate a rulemaking proceeding to enhance the air quality in
6 national parks and wilderness areas in order to protect resources and values that are being
7 adversely affected by air pollution. Included among the effects of concern identified in the
8 request were the acidification of streams, surface waters, and/or soils; eutrophication of coastal
9 waters; visibility impairment; and foliar injury from ozone.
10 In a Federal Register notice in 2001, EPA announced receipt of these requests and asked
11 for comment on the issues raised in them. EPA stated that it would consider any relevant
12 comments and information submitted, along with the information provided by the petitioners and
13 DOI, before making any decision concerning a response to these requests for rulemaking (65 FR
14 48699).
15 The most recent 2005 NAPAP report states that"... scientific studies indicate that the
16 emission reductions achieved by Title IV are not sufficient to allow recovery of acid-sensitive
17 ecosystems. Estimates from the literature of the scope of additional emission reductions that are
18 necessary in order to protect acid-sensitive ecosystems range from approximately 40-80%
19 beyond full implementation of Title IV.... The results of the modeling presented in this Report to
20 Congress indicate that broader recovery is not predicted without additional emission reductions"
21 (NAPAP, 2005).4
22 Given the state of the science as described in the ISA and in other recent reports, such as
23 the NAPAP's above, EPA believes it is appropriate, in the context of evaluating the adequacy of
24 the current NC>2 and SC>2 secondary standards in this review, to revisit the question of the
25 appropriateness and the feasibility of setting a secondary NAAQS to address remaining known
26 or anticipated adverse public welfare effects resulting from the acidic and nutrient deposition of
27 these criteria pollutants.
28
4 Note that a new NAPAP report is expected to be released later in 2010. The findings of that report will be
considered in the final policy assessment.
September 2010 1-16 Draft -Do Not Quote or Cite
-------
1 1.5 PROPOSED CONCEPTUAL FRAMEWORK FOR COMBINED NOX SOX
2 STANDARDS
3 There is a strong basis for considering NOx and SOx together at this time, building upon
4 EPA's and CAS AC's recognition of the interactions of these pollutants and on the growing body
5 of scientific information that is now available related to these interactions and associated
6 ecological effects. The REA introduced a conceptual framework for ecologically meaningful
7 secondary standards that recognized the complex processes by which ecosystems are exposed to
8 ambient NOx and SOx. That framework provided a flow from ambient concentrations exposures
9 via deposition to ecological indicators and effects (see Figure ES-2 in the REA Executive
10 Summary). This sequence represents the process by which we can determine the risks associated
11 with ambient concentrations of NOx and SOx. However, for the purposes of discussing a
12 conceptual framework for design of standards to protect against those risks, a modified version
13 of the risk framework is needed.
14 Figure 1-1 depicts the framework by which we are considering the structure of an
15 ecologically meaningful secondary standard. It is a conceptual diagram that illustrates how a
16 level of protection related to an indicator of ecological effect(s) equates to atmospheric
17 concentrations of NOx and SOx indicators. This conceptual diagram illustrates the linkages
18 between ambient air concentrations and resulting deposition metrics, and between the deposition
19 metric and the ecological indicator of concern. The Atmospheric Deposition Transformation
20 Function translates ambient atmospheric concentrations of NOx and SOx to nitrogen and sulfur
21 deposition metrics, while the Ecological Effect Function transforms the deposition metric into
22 the ecological indicator.
23 Development of a form for the standard that reflects this structure is a critical step in the
24 overall standard setting process. The atmospheric levels of NOx and SOx that satisfy a particular
25 level of ecosystem protection are those levels that result in an amount of deposition that is less
26 than the amount of deposition that a given ecosystem can accept without defined levels of
27 degradation of the ecological indicator for a targeted effect.
28 The details of this conceptual framework are discussed in Chapter 5, including
29 discussions of modifying factors that alter the relationship between ambient atmospheric
30 concentrations of NOx and SOx and deposit!onal loads of nitrogen and sulfur, and those that
31 modify the relationship between deposition loads and the ecological indicator.
September 2010 1-17 Draft -Do Not Quote or Cite
-------
1 In setting NAAQS to protect public health and welfare, EPA has historically established
2 standards which require the comparison of monitored concentrations of an air pollutant against a
3 numerical metric of atmospheric concentration that does not vary geographically. This approach
4 has appropriately protected public health as at-risk populations are widely distributed throughout
5 the nation. As more is learned about the effects of pollutants such as NOx and SOx and the
6 environment, however, such an approach may not be appropriate to provide the requisite level of
7 protection to public welfare from effects on sensitive ecosystems. EPA is proposing in this
8 review of the secondary standard for NOx and SOx a standard that takes into account variable
9 factors, such as atmospheric variables and location-specific characteristics of ecosystems, as the
10 appropriate approach to protect the public welfare from the effects associated with the presence
11 of these pollutants in the ambient air.
12 EPA must undertake a thorough review of the air quality criteria for the pollutant at issue
13 in reviewing a secondary NAAQS, and determine whether a current standard is requisite to
14 protect the public welfare. Under section 108 of the CAA, air quality criteria are to "reflect the
15 latest scientific knowledge useful in indicating the kind and extent of all identifiable effects"
16 associated with the presence of the pollutant in the ambient air. It is clear from the language of
17 the CAA that where the state of the science provides a basis for considering such effects, the
18 review of the air quality criteria should encompass a broad analysis of "any" known or
19 anticipated adverse effects, as well as the ways in which variable conditions such as atmospheric
20 conditions may impact the effect of a pollutant and the ways in which other air pollutants may
21 interact with the criteria pollutant to produce adverse effects. Specifically, section 108(a)(2) of
22 the CAA provides that:
23 Air quality criteria for an air pollutant shall accurately reflect the latest scientific
24 knowledge useful in indicating the kind and extent of all identifiable effects on public health or
25 welfare which may be expected from the presence of such pollutant in the ambient air, in varying
26 quantities. The criteria for an air pollutant to the extent practicable, shall include information on:
27 (A) Those variable factors (including atmospheric conditions) which of themselves or
28 in combination with other factors may alter the effects on public health or welfare
29 of such air pollutants;
30 (B) The types of air pollutants which, when present in the atmosphere, may interact
31 with such pollutants to produce an adverse effect on public health or welfare; and
September 2010 1-18 Draft -Do Not Quote or Cite
-------
1 (C) Any known or anticipated adverse effects on welfare.
2
3 Based on this extensive review of the air quality criteria for an air pollutant, the
4 Administrator is required to review and to revise, as appropriate, the secondary standard to
5 ensure that the standard "is requisite to protect public welfare from any known or anticipated
6 adverse effects associated with the presence of such air pollutant in the ambient air." CAA §
7 109(b) & (d). "Effects on welfare," in turn, is defined to include a broad array of effects,
8 including effects on soil, water, crops, vegetation, and manmade materials, "whether caused by
9 transformation, conversion, or combination with other air pollutants." CAA § 302(h). Thus, as
10 with the sections of the CAA describing the issuance of air quality criteria, the CAA uses
11 expansive language in describing the scope of EPA's responsibility and the range of effects that
12 EPA should take into account in setting a standard that is requisite to protect public welfare. The
13 term "requisite," however, indicates that section 109 is not open-ended. In considering the
14 meaning of the term "requisite" in the context of the primary standards, the Supreme Court has
15 agreed with EPA that such a standard is one that is "sufficient, but not more than necessary" to
16 protect public health. Whitman v. American Trucking, 531 U.S. 457, 473 (2001).
17 While EPA has most often considered the results of direct exposure to an air pollutant in
18 the ambient air in assessing effects on public health and welfare, such as the health effects on
19 humans when breathing in an air pollutant or the effects on vegetation through the uptake of air
20 pollutants from the ambient air through leaves, EPA has also considered, where appropriate, the
21 effects of exposure to air pollutants through more indirect mechanisms. For example, both in
22 1978 and in 2008, EPA established a NAAQS for lead that addressed the health effects of
23 ambient lead whether the lead particles were inhaled or were ingested after deposition on the
24 ground or other surfaces. 73 FR 66964 (November 12, 2008), Lead Industries v. EPA, 647 F.2d
25 1130 (DC Cir. 1980) (1978 NAAQS). The deposition of ambient NOx and SOx to terrestrial and
26 aquatic environments can impact ecosystems through both direct and indirect mechanisms, as
27 discussed in the REA and this document. Given Congress' instruction to set a standard that "is
28 requisite to protect the public welfare from "any known or anticipated adverse effects associated
29 with the presence of such air pollutant in the ambient air," 42 U.S.C. § 109 (b)(2), this review
30 appropriately attempts to take into consideration widely acknowledged effects, such as
September 2010 1-19 Draft -Do Not Quote or Cite
-------
1 acidification and nutrient enrichment, which are associated with the presence of ambient SOx
2 and NOx.
3 In this review, EPA is also attempting to develop a standard that takes into account the
4 variability in effects from ambient levels of SOx and NOx. The CAA requires EPA to establish
5 "national" standards, based on the air quality criteria that provide the requisite degree of
6 protection, but does not clearly address how to do so under the circumstances present here. One
7 approach is to develop a secondary standard such as the one discussed in this Policy Assessment
8 Document. Such a standard is designed to provide a generally uniform degree of protection
9 throughout the country by allowing for varying concentrations of allowable ambient NOx and
10 SOx, depending on atmospheric conditions and other variabilities, to achieve that degree of
11 protection5. Such a standard protects sensitive ecosystems wherever such ecosystems are found.
12 This approach recognizes that setting a standard that is sufficient to protect the public welfare but
13 not more than is necessary calls for consideration of a standard such as the one discussed in this
14 document.
5 In concept, this approach to setting a national standard using a consistent form that may result in differing levels of
atmospheric concentrations fo NOx and SOx is similar to the current proposal to use PM10 as the national indicator
for protection against the health effects of PM10-2.5, as discussed in the 2010 2nd draft Policy Assessment for the
PM NAAQS primary standards. In that case, EPA is proposing to recognize that the same size fraction of particles
may have different toxicity depending on location, and as a result, it is more appropriate to use an indicator that
reflects that varying toxicity, rather than setting one absolute level of PM10-2.5 which may not be equally protective
in all locations. By proposing this form, EPA is recognizing that in attaining a PM10 standard, the resulting balance
of PM10-2.5 and PM2.5 particles will differ across areas of the U.S. Likewise, setting a joint NOx/SOx standard
that results in differing allowable concentrations of NOx and SOx across the U.S. based on the differing potential for
NOx and SOx to result in ecological damages is appropriate in providing a requisite level of welfare protection.
September 2010 1 -20 Draft -Do Not Quote or Cite
-------
Structure of an Ecologically-based Standard
i
2
3
4
5
6
7
8
9
10
11
12
13
Variable/Fixed
Factors:
Atmospheric
Landscape
Atmospheric
Deposition
Transformation
Function
Deposition
Metric
T
Form of the Standard
Level of the Standard
Figure 1-1. Framework of an alternative secondary standard.
1.6 POLICY RELEVANT QUESTIONS
In this policy assessment, a series of general questions frames our approach to identifying
a range of policy options for consideration by the Administrator regarding secondary NAAQS
for NOx and SOx. These questions are drawn from our Integrated Review Plan with
modifications based on further consideration by staff and comments from CAS AC and the
public. Our policy assessment begins by characterizing "known or anticipated adverse effects"
on public welfare within our conceptual model (CPE 1). As noted earlier, this review is focusing
on effects in sensitive unmanaged ecosystems (not commercial forests or agricultural lands6)
resulting from ambient concentrations of NOx and SOx through deposition of N and S.
The decision to focus on unmanaged ecosystems is based on the weight of evidence of effects in those ecosystems.
The majority of the scientific evidence regarding acidification and nutrient enrichment is based on studies in
unmanaged ecosystems. Non-managed terrestrial ecosystems tend to have a higher fraction of N deposition
resulting from atmospheric N (ISA 3.3.2.5). In addition ,the ISA notes that agricultural and commercial forest lands
are routinely fertilized with amounts of N (100 to 300 kg N/ha) that exceed air pollutant inputs even in the most
polluted areas (ISA 3.3.9). This review recognizes that effect of N deposition in managed areas may be viewed
differently than effects of N deposition in unmanaged ecosystems, largely due to the more homogeneous, controlled
nature of species composition and development in managed ecosystems and the potential for benefits of increased
productivity in those ecosystems.
September 2010
1-21
Draft -Do Not Quote or Cite
-------
1 In Chapter 2, we draw from the information and conclusions presented in the ISA and
2 REA to address the following questions:
3 1. What are the nature and magnitude of ecosystem responses to reactive nitrogen and
4 sulfur deposition, acidification, nutrient depletion and the mobilization of toxic metals
5 in sensitive aquatic and terrestrial ecosystems?
6 a. How are these responses affected by landscape factors?
7 b. What types of ecosystems are sensitive to such responses?
8 2. To what extent can ecosystem responses to nitrogen deposition be separated into
9 responses related to oxidized and reduced forms of reactive nitrogen compounds?
10
11 In Chapter 3, we address the following questions related to linking effects to measures of
12 adversity (CPE 1.1):
13 1. How do we characterize adversity to public welfare? What are the sources of
14 potentially relevant characterization for this policy assessment?
15 2. What is the evidence of effects on ecosystem services, and how can those ecosystem
16 services be linked to ecological indicators?
17 3. To what extent are identified ecosystem effects important from a public welfare
18 perspective, and what are the important uncertainties associated with estimating such
19 effects?
20
21 Once we have described ecological effects, we then provide an assessment of the
22 adequacy of the existing NOx and SOx standards (CPE 1.2). We begin this assessment by
23 drawing from the information and conclusions presented in the ISA and REA to address in
24 Chapter 4 the following questions, which allow us to identify whether the structure of the current
25 standards is appropriate relative to the key ecological effects assessed in the ISA and REA,
26 including acidification and excess nutrient enrichment and whether there is adequate information
27 and analyses available at this time to assess the extent to which potentially adverse effects on
28 aquatic and terrestrial ecosystems can be associated with current levels of atmospheric reactive
29 nitrogen, accounting for the contributions of oxidized and reduced forms, and SOx and with
30 levels that are at or below the current secondary standards:
September 2010 1 -22 Draft -Do Not Quote or Cite
-------
1 1. To what extent are effects that could reasonably be judged to be adverse to public
2 welfare occurring under current conditions and would such effects occur if the nation
3 met the current standards? To what extent do the current NOx and SOx secondary
4 standards provide protection from effects associated with deposition of:
5 a. Sulfur and oxidized nitrogen from atmospheric NOx, and SOx which results in
6 acidification in sensitive aquatic and terrestrial ecosystems?
7 b. Oxidized nitrogen from atmospheric NOx, which results in nutrient enrichment
8 effects in sensitive aquatic and terrestrial ecosystems?
9 c. Sulfur and oxidized nitrogen from atmospheric NOx and SOx which results in
10 other ecological effects (e.g. mercury methylation)?
11 2. In what way are the structures of the current NOx and SOx secondary standards
12 inadequate to protect against public welfare effects?
13
14 In Chapter 5, we follow our adequacy assessment by developing in greater detail the
15 conceptual framework for the design of ecologically relevant multi-pollutant standards
16 introduced in Section 1.4 above. To the extent that the available information calls into question
17 the adequacy of protection afforded by the current standards and/or the appropriateness of the
18 structure of the standards, we explore the extent to which available information supports
19 consideration of alternative standards, in terms of atmospheric and ecological indicators and
20 related averaging times, forms, and levels. This conceptual framework is designed to focus on
21 resolving the following questions:
22 1. (CPE 2.1) Does the available information provide support for the use of ecological
23 indicators to characterize the responses of aquatic and terrestrial ecosystems to
24 oxidized nitrogen and sulfur deposition?
25 2. (CPE 1) Does the available information provide support for the development of
26 appropriate ecological response to deposition relationship(s) that meaningfully relates
27 oxidized nitrogen and sulfur deposition to relevant ecological indicators? Does a
28 quantified relationship exist between the level of a relevant ecological indicator and
29 an amount of nitrogen and sulfur deposition?
September 2010 1 -23 Draft -Do Not Quote or Cite
-------
1 3. (CPE 2) What are the important variables in the ecological response to deposition
2 relationship(s)? Are these relationships applicable nationally? What are the
3 appropriate temporal scales for these relationships?
4 a. How does ecological response to deposition relationship(s) depend upon
5 spatially heterogeneous geologic factors (e.g. bedrock type, weathering rates)
6 that govern sensitivity?
7 b. How do we consider areas with high natural background acidification or
8 nutrient loadings?
9 4. (CPE 3) Does the available information provide support for the development of
10 appropriate functions that characterize the relationships between atmospheric NOx
11 and SOx and the wet and dry deposition of total reactive nitrogen and sulfur?
12 a. What deposition function is appropriate to use for the purpose of relating an
13 amount of nitrogen and/or sulfur deposition in sensitive ecosystems to
14 ambient concentrations of atmospheric reactive nitrogen, including oxides and
15 reduced forms, and/or sulfur? What are the important variables in such a
16 function? What are appropriate spatial and temporal scales to use in
17 specifying such variables?
18
19 Based on the conceptual framework for the structure of the ecologically relevant multi-
20 pollutant standards, we then address in Chapter 6 the components of the standard needed to
21 develop options for consideration by the Administrator. Development of these options will focus
22 on addressing the following questions:
23 1. (CPE 2.1) What ecological indicators are appropriate to use for the purpose of
24 developing an alternative standard for the various ecological effects assessed in this
25 review?
26 2. (CPE 5) What indicators of oxides of nitrogen and sulfur are appropriate to use for
27 the purpose of determining whether the resultant deposition is within the target values
28 needed to achieve the desired degree of protection? What averaging times and forms
29 are appropriate to consider?
September 2010 1 -24 Draft -Do Not Quote or Cite
-------
1 3. (CPE 2) What approaches are available to specify non-atmospheric components of
2 the standard, e.g. weathering rates? Are there approaches that can simplify the
3 structure of the standard?
4 4. What are the available approaches for accounting for reduced N in the structure of the
5 standard?
6 5. What is the most appropriate form for the standards to reflect the relationships
7 between ambient NOx and SOx, acidifying deposition, and the ecological indicator
8 for acidification?
9 Several follow-up questions derive from our assessment of options for specifying the
10 components of a multipollutant standard. In Chapter 6, we address the questions:
11 1. To what extent would a standard specifically defined to protect against one ecological
12 effect (i.e., aquatic acidification) likely provide protection from other relevant
13 ecological effects?
14 2. What are the available approaches for combining multiple indicators into a single
15 standard, e.g. using nitrogen effects to bound the tradeoff curve for NOx/SOx for
16 aquatic acidification effects
17 3. What are the available approaches to integrate potential standards for aquatic and
18 terrestrial acidification and/or aquatic and terrestrial N enrichment?
19
20 In Chapter 7, we provide a range of explorations of uncertainties in the evidence and
21 models as they pertain to the selection of options for components of the standard. In addition,
22 we provide results of sensitivity analyses for components of the proposed AAPI form, as well as
23 characterizing information on variability in those components. The chapter focuses on the
24 following questions:
25 1. What are critical uncertainties in the characterization of pre-industrial levels of
26 ANC?
27 2. What uncertainties are introduced through the use of steady state critical load
28 models relative to dynamic critical loads models?
29 3. What are the critical uncertanties in the modeled relationship between
30 concentrations of NOx and SOx and deposition of N and S?
31 4. What are the critical uncertainties in the modeled values of NHx deposition?
September 2010 1-25 Draft -Do Not Quote or Cite
-------
1 5. How sensitive is the form of the NOx/SOx standard to the primary components?
2 6. How sensitive are the atmospheric transformation ratios calculated from CMAQ
3 to the emissions scenarios, chemical mechanisms, and meteorology in the model?
4 7. How well does CMAQ perform in simulating nitric oxide, nitrogen dioxide, sulfur
5 dioxide, nitrate, ammonium and aerosol nitrate, ammonium, and sulfate relative to
6 observations from different networks for which the data are routinely available?
7 8. How well does CMAQ perform in simulating wet deposition of sulfate, nitrate,
8 and ammonium relative to observations from the National Atmospheric
9 Deposition Program (NADP) network?
10
11 Chapter 8 provides a discussion of several important aspects of monitoring of NOx and
12 SOx, including methods, network design, and frequency.
13 We conclude in Chapter 9 with a discussion of options to consider in selecting pollutant
14 indicators, averaging times, forms, and ranges of levels for the secondary NOx and SOx
15 standards. This discussion is informed by a consideration of the role of ecosystem services in
16 helping to characterize what adversity to public welfare, focused on the following questions:
17 1. (CPE 5) What are the risks of ecosystem service impairment under alternative levels
18 of potential standards for NOx and SOx?
19 2. (CPE 5) To what extent can information about ecosystem services be used to help
20 characterize the extent to which differing levels of relevant ecological indicators
21 reflect impacts that can reasonably be judged to be adverse from a public welfare
22 perspective?
23 3. (CPE 5) Are there relevant benchmarks for adversity to public welfare that can be
24 derived from other sources?
25 4. (CPE 5) Taking into consideration information about ecosystem services and other
26 factors related to characterizing adversity to public welfare for the ecological effects
27 being assessed in this review, what is an appropriate range of levels of protection to
28 be achieved by alternative standards for the Agency to consider?
29
September 2010 1 -26 Draft -Do Not Quote or Cite
-------
1 2 KNOWN OR ANTICIPATED ECOLOGICAL EFFECTS
2
3 In this chapter we address Critical Policy Element 1, evaluation of the effects of ambient
4 NOx and SOx on ecosystems, and the relationship between those effects and the measure of dose
5 in the ecosystem, indicated by the deposit!onal loadings of N and S. In section 302(h) of the
6 Clean Air Act, welfare effects addressed by a secondary NAAQS include, but are not limited to,
7 "effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather,
8 visibility and climate, damage to and deterioration of property, and hazards to transportation, as
9 well as effects on economic values and on personal comfort and well-being". Of these welfare
10 effects categories, the effects of NOx and SOx on aquatic and terrestrial ecosystems, which
11 encompass soils, water, vegetation, wildlife, and contribute to economic value and well-being,
12 are of most concern at concentrations typically occurring in the U.S. Direct effects of NOx and
13 SOx on vegetation are also discussed in this chapter, and have been the focus of previous
14 reviews. However, for this review, the focus of this chapter is on the known and anticipated
15 effects to ecosystems caused by exposure to NOx and SOx through deposition.
16 The information presented here is a concise summary of conclusions from the ISA and
17 the REA. This chapter focuses on effects on specific ecosystems with a brief discussion on
18 critical uncertainties associated with acidification and nutrient enrichment. Those effects are then
19 evaluated in Chapter 3 within the context of alternative definitions of, including assessments of
20 potential impacts on ecosystem services. Effects are broadly categorized into acidification and
21 nutrient-enrichment in the proceeding sections. This is background information intended to
22 support new approaches for the design of ecologically relevant secondary NOx and SOx
23 standards which are protective of U.S. ecosystems. More detailed information on the conceptual
24 design and specific options for the proposed standards are presented in Chapters 5 and 9of this
25 policy assessment document. While we provide a summary of effects for all four of the primary
26 effects categories, we reiterate that the focus of this second draft policy assessment is on effects
27 related to aquatic acidification, without downplaying the potential importance of effects in other
28 categories.
September, 2010 2-1 Do Not Quote or Cite
-------
1 2.1 ACIDIFICATION: EVIDENCE OF EFFECTS ON STRUCTURE AND
2 FUNCTION OF TERRESTRIAL AND FRESHWATER ECOSYSTEMS
3 Sulfur oxides (SOx) and nitrogen oxides (NOx) in the atmosphere undergo a complex mix of
4 reactions in gaseous, liquid, and solid phases to form various acidic compounds. These acidic
5 compounds are removed from the atmosphere through deposition: either wet (e.g., rain, snow),
6 fog or cloud, or dry (e.g., gases, particles). Deposition of these acidic compounds to ecosystems
7 can lead to effects on ecosystem structure and function. Following deposition, these compounds
8 can, in some instances unless buffered by high base soils, leach out of the soils in the form of
9 sulfate (SO42) and nitrate (NO3-), leading to the acidification of surface waters. The effects on
10 ecosystems depend on the magnitude and rate of deposition, as well as a host of biogeochemical
11 processes occurring in the soils and waterbodies (REA 2.1). The chemical forms of nitrogen that
12 may contribute to acidifying deposition include both oxidized and reduced chemical species.
13 When sulfur or nitrogen (NOx, NHx and Nr) leaches from soils to surface waters in the
14 form of SO42 or NO3-, an equivalent amount of positive cations, or countercharge, is also
15 transported. This maintains electroneutrality. If the countercharge is provided by base cations,
16 such as calcium (Ca2+), magnesium (Mg2+), sodium (Na+), or potassium (K+), rather than hydrogen
17 (H+) and dissolved inorganic aluminum, the acidity of the soil water is neutralized, but the base
18 saturation of the soil decreases. Continued SO42 or NO3- leaching can deplete the available base
19 cation pool in soil. As the base cations are removed, continued deposition and leaching of SO42
20 and/or NO3- (with H+and A13+) leads to acidification of soil water, and by connection, surface
21 water. The ability of a watershed to neutralize acidic deposition is determined by a variety of
22 biogeophysical factors including weathering rates, bedrock composition, vegetation and
23 microbial processes, physical and chemical characteristics of soils and hydrologic flowpaths.
24 (REA 2.1) Some of these factors such as vegetation and soil depth are highly variable over
25 small spatial scales such as meters, but can be aggregated to evaluate patterns over larger spatial
26 scales. For the purpose of a national secondary standard, the most relevant characteristics are
27 those that are less variable over small scales.
28 Acidifying deposition of NOx and SOx and the chemical and biological responses
29 associated with these inputs vary temporally. Chronic or long-term deposition processes in the
30 time scale of years to decades result in increases in inputs of N and S to ecosystems and the
31 associated ecological effects. Episodic or short term (i.e., hours or days) deposition refers to
September, 2010 2-2 Do Not Quote or Cite
-------
1 events in which the level of the acid neutralizing capacity (ANC) of a lake or stream is
2 temporarily lowered. In aquatic ecosystems, short-term (i.e., hours or days) episodic changes in
3 water chemistry can have significant biological effects. Episodic acidification refers to
4 conditions during precipitation or snowmelt events when proportionately more drainage water is
5 routed through upper soil horizons that tend to provide less acid neutralizing than was passing
6 through deeper soil horizons (REA 4.2). Some streams and lakes may have chronic or base flow
7 chemistry that is suitable for aquatic biota, but may be subject to occasional acidic episodes with
8 deleterious consequences to sensitive biota.
9 The following summary is a concise overview of the known or anticipated effects caused
10 by acidification to ecosystems within the United States. Acidification affects both terrestrial and
11 freshwater aquatic ecosystems. Terrestrial and aquatic processes are often linked; therefore
12 responses to the following questions address both types of ecosystems unless otherwise noted.
13
14 2.1.1 What is the nature of acidification related ecosystem responses to reactive
15 nitrogen and sulfur deposition?
16 The ISA concluded that deposition of SOX, NOX, and NHX leads to the varying degrees of
17 acidification of ecosystems (EPA 2008). In the process of acidification, biogeochemical
18 components of terrestrial and freshwater aquatic ecosystems are altered in a way that leads to
19 effects on biological organisms. Deposition to terrestrial ecosystems often moves through the
20 soil and eventually leaches into adjacent water bodies.
21 The scientific evidence is sufficient to infer a causal relationship between acidifying
22 deposition and effects on biogeochemistry and biota in aquatic ecosystems (ISA 4.2.2). The
23 strongest evidence comes from studies of surface water chemistry in which acidic deposition is
24 observed to alter sulfate and nitrate concentrations in surface waters, the sum of base cations,
25 ANC, dissolved inorganic aluminum and pH. (ISA 3.2.3.2). Consistent and coherent
26 documentation from multiple studies on various species from all major trophic levels of aquatic
27 systems shows that geochemical alteration caused by acidification can result in the loss of acid-
28 sensitive biological species (ISA 3.2.3.3). For example, in the Adirondacks, of the 53 fish
29 species recorded in Adirondack lakes about half (26 species) were absent from lakes with pH
30 below 6.0 (Baker et al., 1990b). Biological effects are linked to changes in water chemistry
31 including decreases in ANC and pH and increases in inorganic Al concentration. The direct
September, 2010 2-3 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
biological effects are caused by lowered pH and increased inorganic Al concentrations. While
ANC level does not cause direct biological harm it is a good overall indicator of the risk of
acidification (Fig 2-1, See further discussion in Section 2.1.3).
t NCrandSO4
2-
Stream water chemistry
I
^Ecological effects which may include:
Lower biodiversity
Altered species composition
Localized extinction of sensitive species
Individual mortality of sensitive species
Sub-lethal effects to sensitive species and
ecological integrity
Aquatic biota
Figure 2-1. Conceptual model of direct and indirect acidification effects on aquatic biota.
Acidic pollutants (NO3- and SO4-2) lower ANC, resulting in lower pH with
direct toxic effects on fish. The lower pH mobilizes A13+ from soils often
resulting in higher concentration in stream water causing direct toxicity to
fish.
These changes in stream water chemistry contribute to declines in taxonomic richness of
zooplankton, macroinvertebrates, and fish, which are often sources of food for birds and other
animal species in various ecosystems. These fish may also serve as a source of food and
recreation for humans (see Chapter 3). Acidification of ecosystems has been shown to disrupt
food web dynamics causing alteration to the diet, breeding distribution and reproduction of
certain species of birds (ISA Section 4.2.2.2. and Table 3-9). For example, breeding
distribution of the common goldeneye (Bucephala clangula) an insectivorous duck, may be
September, 2010
2-4
Do Not Quote or Cite
-------
1 affected by changes in acidifying deposition (Longcore and Gill, 1993). Similarly, decreases in
2 prey diversity and quantity have been observed to create feeding problems for nesting pairs of
3 loons on low-pH lakes in the Adirondacks (Parker 1988).
4 In terrestrial ecosystems, the evidence is sufficient to infer a causal relationship between
5 acidifying deposition and changes in biogeochemistry (ISA 4.2.1.1). The strongest evidence
6 comes from studies of forested ecosystems, with supportive information on other plant taxa,
7 including shrubs and lichens (ISA 3.2.2.1.). Three useful indicators of chemical changes and
8 acidification effects on terrestrial ecosystems, showing consistency and coherence among
9 multiple studies are: soil base saturation, Al concentrations in soil water and soil C:N ratio (ISA
10 3.2.2.2).
11 In soils with base saturation less than about 15 to 20%, exchange chemistry is dominated
12 by Al (Reuss, 1983). Under these conditions, responses to inputs of sulfuric acid and nitric acid
13 largely involve the release and mobilization of dissolved inorganic Al. The effect can be
14 neutralized by weathering from geologic parent material or base cation exchange. The Ca2+ and
15 Al concentrations in soil water are strongly influenced by soil acidification and both have been
16 shown to have quantitative links to tree health, including Al interference with Ca2+uptake and Al
17 toxicity to roots (Parker et al., 1989; U.S. EPA, 2009). Effects of nitrification and associated
18 acidification and cation leaching have been consistently shown to occur only in soils with a C:N
19 ratio below about 20 to 25 (Aber et al., 2003; Ross et al., 2004).
20 Soil acidification caused by acidic deposition has been shown to cause decreased growth
21 and increased susceptibility to disease and injury in sensitive tree species. Red spruce (Picea
22 rubens) dieback or decline has been observed across high elevation areas in the Adirondack,
23 Green and White mountains (DeHayes et al., 1999). The frequency of freezing injury to red
24 spruce needles has increased over the past 40 years, a period that coincided with increased
25 emissions of S and N oxides and increased acidifying deposition (DeHayes et al., 1999).
26 Acidifying deposition can contribute to dieback in sugar maple {Acer saccharum) through
27 depletion of cations from soil with low levels of available Ca (Horsley et al., 2000; Bailey et al.,
28 2004). Grasslands are likely less sensitive to acidification than forests due to grassland soils
29 being generally rich in base cations (Fenn et al., 2003; Blake et al., 1999).
30
September, 2010 2-5 Do Not Quote or Cite
-------
1 2.1.2 What types of ecosystems are sensitive to such effects? In which ways are these
2 responses affected by atmospheric, ecological, and landscape factors?
3 The intersection between current deposition loading, historic loading, and sensitivity
4 defines the ecological vulnerability to the effects of acidification. Freshwater aquatic and some
5 terrestrial ecosystems, notably forests, are the ecosystem types which are most sensitive to
6 acidification. The ISA reports that the principal factor governing the sensitivity of terrestrial and
7 aquatic ecosystems to acidification from sulfur and nitrogen deposition is geology (particularly
8 surficial geology). Geologic formations having low base cation supply generally underlie the
9 watersheds of acid-sensitive lakes and streams. Other factors that contribute to the sensitivity of
10 soils and surface waters to acidifying deposition include topography, soil chemistry, land use,
11 and hydrologic flowpaths. Episodic and chronic acidification tends to occur in areas that have
12 base-poor bedrock, high relief, and shallow soils (ISA 3.2.4.1).
13
14 2.1.3 What is the magnitude of ecosystem responses to acidifying deposition?
15 Terrestrial and aquatic ecosystems differ in their response to acidifying deposition.
16 Therefore the magnitude of ecosystem response is described separately for aquatic and terrestrial
17 ecosystems in the following sections. The magnitude of response refers to both the severity of
18 effects and the spatial extent of the U.S. which is affected.
19 2.1.3.1 Aquatic
20 Freshwater ecosystem surveys and monitoring in the eastern United States have been
21 conducted by many programs since the mid-1980s, including EPA's Environmental Monitoring
22 and Assessment Program (EMAP), National Surface Water Survey (NSWS), Temporally
23 Integrated Monitoring of Ecosystems (TIME) (Stoddard, 1990), and Long-term Monitoring
24 (LTM) (Ford et al., 1993; Stoddard et al., 1996) programs. Based on analyses of surface water
25 data from these programs, New England, the Adirondack Mountains, the Appalachian Mountains
26 (northern Appalachian Plateau and Ridge/Blue Ridge region), and the Upper Midwest contain
27 the most sensitive lakes and streams (i.e., ANC less than about 50 ueq/L). Portions of northern
28 Florida also contain many acidic and low-ANC lakes and streams, although the role of acidifying
29 deposition in this region is less clear. The western U.S. contains many of the surface waters most
30 sensitive to potential acidification effects, but with the exception of the Los Angeles Basin and
31 surrounding areas, the levels of acidifying deposition are low in most areas. Therefore,
September, 2010 2-6 Do Not Quote or Cite
-------
1 acidification of surface waters by acidic deposition is uncommon in the western U.S., and the
2 extent of chronic surface water acidification that has occurred in that region to date has likely
3 been very limited (ISA 3.2.4.2 and REA 4.2.2).
4 There are a number of species including fish, aquatic insects, other invertebrates and
5 algae that are sensitive to acidification and cannot survive, compete, or reproduce in acidic
6 waters (ISA 3.2.3.3). Decreases in ANC and pH have been shown to contribute to declines in
7 species richness and declines in abundance of zooplankton, macroinvertebrates, and fish (Keller
8 and Gunn 1995; Schindler et al., 1985). Reduced growth rates have been attributed to acid stress
9 in a number offish species including Atlantic salmon (Salmo salar\ Chinook salmon
10 (Oncorhynchus tshawytscha), lake trout (Salvelinus namaycush), rainbow trout (Oncorhynchis
11 mykiss\ brook trout (Salvelinus Fontinalis), and brown trout (Salmo trutta) (Baker et al., 1990).
12 In response to small to moderate changes in acidity, acid-sensitive species are often replaced by
13 other more acid-tolerant species, resulting in changes in community composition and richness.
14 The effects of acidification are continuous, with more species being affected at higher degrees of
15 acidification. At a point, typically a pH <4.5 and an ANC <0 ueq/L, complete to near-complete
16 loss of many taxa of organisms occur, including fish and aquatic insect populations, whereas
17 other taxa are reduced to only acidophilic species. These changes in taxa composition are
18 associated with the high energy cost in maintaining physiological homeostasis, growth, and
19 reproduction at low ANC levels (Schreck, 1981, 1982; Wedemeger et al., 1990; REA appendix
20 2.3). Decreases in species richness related to acidification have been observed in the Adirondack
21 Mountains and Catskill Mountains of New York (Baker et al., 1996), New England and
22 Pennsylvania (Haines and Baker, 1986), and Virginia (Bulger et al., 2000).
23 From the sensitive areas identified by the ISA, further "case study" analyses on aquatic
24 ecosystems in the Adirondack Mountains and Shenandoah National Park were conducted to
25 better characterize ecological risk associated with acidification (REA Chapter 4).
26 ANC is the most widely used indicator of acid sensitivity and has been found in various
27 studies to be the best single indicator of the biological response and health of aquatic
28 communities in acid-sensitive systems (Lien et al., 1992; Sullivan et al., 2006; ISA). In the
29 REA, surface water trends in SC>42" and N(V concentrations and ANC levels were analyzed to
30 affirm the understanding that reductions in deposition could influence the risk of acidification.
31 ANC values were categorized according to their effects on biota, as shown in Figure 2-2. [Need
September, 2010 2-7 Do Not Quote or Cite
-------
1 1
2
3
4
to indicate the source for the categorization] Monitoring data from TIME/LTM and EMAP
programs were assessed for the years 1990 to 2006, and past, present, and future water quality
evels were estimated by both steady-state and dynamic biogeochemical models.
Category Label ANC Levels and Expected Ecological Effects
Acute
Concern
Severe
Concern
Elevated
Concern
Moderate
Concern
Low
Concern
<0 ueq/L
0-20 ueq/L
20-50 ueq/L
50-100
ueq/L
>100 ueq/L
Complete loss offish populations is expected. Planktonic communities
have extremely low diversity and are dominated by acidophilic taxa.
The numbers of individuals in plankton species that are present are
greatly reduced.
Highly sensitive to episodic acidification. During episodes of high
acidifying deposition, brook trout populations may experience lethal
effects. The diversity and distribution of zooplankton communities
decline sharply.
Fish species richness is greatly reduced (i.e., more than half of expected
species can be missing). On average, brook trout populations
experience sublethal effects, including loss of health, ability to
reproduce, and fitness. Diversity and distribution of zooplankton
communities decline.
Fish species richness begins to decline (i.e., sensitive species are lost
from lakes). Brook trout populations are sensitive and variable, with
possible sublethal effects. Diversity and distribution of zooplankton
communities also begin to decline as species that are sensitive to
acidifying deposition are affected.
Fish species richness may be unaffected. Reproducing brook trout
populations are expected where habitat is suitable. Zooplankton
communities are unaffected and exhibit expected diversity and
distribution.
6 Table 2-1. Ecological effects associated with alternative levels of acid
7 neutralizing capacity (ANC)
8
9 The most commonly used models of acidification are presented in Table 2-2 These models are
10 designed to be applied at the spatial scale of the watershed, with the exception of the SMART
11 model. Steady-sate mass balance models, including the steady-state water chemistry (SSWC)
12 model are the most commonly used method for analysis of critical loads of acid deposition. The
13 steady-state models assume steady state conditions. The dynamic models consider how the
14 ecosystem may change through time. These models tend to require more data than the steady-
15 state models.
September, 2010
2-8
Do Not Quote or Cite
-------
Table 2-2 Summary several
A for a more comprehensive
commonly used acidification models (See ISA Annex
list and discussion of acidification models)
Model name
Dynamic
or steady
state
Model description
Steady-sate mass
balance models
/Steady-state
water Chemistry
(SSWC)/
Steady-
state
The basic principle is based on identifying the long-term
average sources of acidity and alkalinity in order to
determine the maximum acid input that will balance the
system at a biogeochemical safe-limit. Several
assumptions have been made in the steady state
calculations. First, it is assumed that ion exchange is at
steady state and there is no net change in base saturation or
no net transfer of ANC from soil solution to the ion
exchange matrix. It is assumed that for N there is no net
denitrification, adsorption or desorption and the N cycle is
at steady state. Sulfate is also assumed to be at steady
state: no sulfide oxidation, sulfate uptake, sulfate
permanent fixation or sulfate reduction are significant.
Simple hydrology is assumed where there is straight
infiltration through the soil profile.
Model of
Acidification of
Groundwater in
Catchment
(MAGIC)
Dynamic
MAGIC is a lumped-parameter model of intermediate
complexity, developed to predict the longterm effects of
acidic deposition on surface water chemistry. The model
simulates soil solution chemistry and surface water
chemistry to predict the monthly and annual average
concentrations of the major ions in these waters. MAGIC
consists of: a section in which the concentrations of major
ions
are assumed to be governed by simultaneous reactions
involving SO4 2- adsorption, cation exchange,
dissolution-precipitation- speciation of aluminum, and
dissolution-speciation of inorganic carbon; and a
mass balance section in which the flux of major ions to
and from the soil is assumed to be controlled by
atmospheric inputs, chemical weathering, net uptake and
loss in biomass and losses to runoff.
PnET-BGC
Dynamic
PnET/BGC simulates major biogeochemical processes,
such as forest canopy element transformations, hydrology,
soil organic matter dynamics, N cycling, geochemical
weathering, and chemical equilibrium reactions in solid
and solution phases, and allows for simulations of land
disturbance. The model uses mass transfer relationships to
describe weathering, canopy interactions and surface
water processes. Chemical equilibrium relationships
describe anion adsorption, cation exchange and soil
solution and surface water speciation. The model operates
September, 2010
2-9
Do Not Quote or Cite
-------
Table 2-2 Summary several
A for a more comprehensive
commonly used acidification models (See ISA Annex
list and discussion of acidification models)
on a monthly time step and is applied at the stand to
small-watershed scale.
DayCent-Chem
Dynamic
DayCent-Chem links two widely accepted and tested
models, one of daily biogeochemistry for forest, grassland,
cropland, and savanna systems, DayCent (Parton et al.,
1998), and the other of soil and water geochemical
equilibrium, PHREEQC (Parkhurst and Appelo, 1999).
The linked DayCent/PHREEQC model was created to
capture the biogeochemical responses to atmospheric
deposition and to explicitly consider those biogeochemical
influences on soil and surface water chemistry. The linked
model expands on DayCent's ability to simulate N, P, S,
and C ecosystem dynamics byincorporating the reactions
of many other chemical species in surface water.
Very Simple
Dynamic (VSD)
soil acidification
model
Dynamic
This model is frequently used in Europe to simulate
acidification effects in soils when observed data are
sparse. The VSD model consists of a set of mass balance
equations, describing the soil input-output relationships,
and a set of equations describing the rate-limited and
equilibrium soil processes.lt only includes weathering,
cation exchange, N immobilization processes, and a mass
balance for cations, sulfur and N. In the VSD model, the
various ecosystem processes have been limited to a few
key processes. Processes that are not taken into account
include canopy interactions; nutrient cycling processes; N
fixation and NH4 adsorption; SO42- transformations
(adsorption, uptake, immobilization, and reduction);
formation and protonation of organic anions; and
complexation of Al.
Simulation
Model for
Acidification's
Regional Trends
(SMART)
Dynamic
The the SMART model consists of a set of mass balance
equations, describing soil input/output relationships, and a
set of equations describing the rate-limited and
equilibrium soil processes. It
includes most of the assumptions and simplifications
given for the VSD model. SMART models the exchange
of Al, H, and divalent base cations using Gaines Thomas
equations. Additionally, SO4
2- adsorption is modeled using a Langmuir equation (as
in MAGIC) and organic acids can be described as mono-,
di-, or tri-protic. The SMART model has been developed
with regional applications in mind, and an early example
of an application to Europe can be found in De Vries et al.
(1994).
September, 2010
2-10
Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
The analyses of the Adirondack Case Study Area indicated that although wet deposition
rates for 862 and NOx have been reduced since the mid-1990s, current concentrations are still
well above simulated pre-acidification (1860) conditions. Modeling predicts NO3" and SO42" are
17- and 5-fold higher, respectively, in 2006 than under simulated pre-acidification conditions.
Based on the 2006 Model of Acidification of Groundwater in Catchment (MAGIC) simulations,
the estimated average ANC across the 44 lakes in the Adirondack Case Study Area is 62.1 ueq/L
(± 15.7 ueq/L); 78 % of all monitored lakes in the Adirondack Case Study Area have a current
risk of Elevated, Severe, or Acute. Of the 78%, 18% are chronically acidic (REA 4.2.4.2).
Based on a N and S deposition scenario that maintains current emission levels to 2020
and 2050, the simulation forecast indicates there would be no improvement in water quality in
the Adirondack Case Study Area. The percentage of lakes within the Elevated to Acute Concern
classes remains the same from 2020 to 2050.
o
'«
o
o
140
120
80
60
40
20
0
1850
1900
1950
2000
2050
Figure 2-2. Average NO3 concentrations (orange), SO42 concentrations
(red), and ANC (blue) across the 44 lakes in the Adirondack
Case Study Area modeled using MAGIC for the period 1850
to 2050.
September, 2010
2-11
Do Not Quote or Cite
-------
ANC Preacidification (1860) and Current Condition (2006)
Preacidification (1860)
Source: EPA 2009
ANC (ueq/L)
• <0
0-20
20-50
• 50-100
• >100
Current (2006)
2 Figure 2-3. ANC concentrations of preacidification (1860) and 2006 conditions
3 based on hindcasts of 44 lakes in the Adirondack Case Study Area
4 modeled using MAGIC
September, 2010
2-12
Do Not Quote or Cite
-------
Current Condition of Acidity
and Sensitivity
Criticial Load
meq/m2/yr
• Highly Sensitive: < 50
Moderately Sensitive: 51 -100
Low Sensitivity: 101 -200
• Not Sensitive: > 201
| | Adirondack Boundary
Source: EPA 2009
2 Figure 2-4. Critical loads of acidifying deposition that each surface water location
3 can receive in the Adirondack Case Study Area while maintaining or
4 exceeding an ANC concentration of 50 ueq/L based on 2002 data.
5 Watersheds with critical load values <100 meq/m2/yr (red and orange
6 circles) are most sensitive to surface water acidification, whereas
7 watersheds with values >100 meq/m2/yr (yellow and green circles) are
8 less sensitive sites.
9
10 Note that studies on fish species richness in the Adirondacks Case Study Area
11 demonstrated the effect of acidification. Of the 53 fish species recorded in Adirondack Case
12 Study Area lakes, only 27 species were found in lakes with a pH <6.0. The 26 species missing
13 from lakes with a pH <6.0 include important recreational species, such as Atlantic salmon, tiger
14 trout (Salmo trutta X Salvelinusfontinalis), redbreast sunfish (Lepomis auritus), bluegill
15 (Lepomis macrochims), tiger musky (Esox masquinongy X Indus), walleye (Sander vitreus),
16 alewife (Alosapseudoharengus), and kokanee (Oncorhynchus nerkd) (Kretser et al., 1989), as
17 well as ecologically important minnows that are commonly consumed by sport fish. A survey of
18 1,469 lakes in the late 1980s found 346 lakes to be devoid offish. Among lakes with fish, there
September, 2010
2-13
Do Not Quote or Cite
-------
1 was a relationship between the number offish species and lake pH, ranging from about one
2 species per lake for lakes having a pH <4.5 to about six species per lake for lakes having a pH
3 >6.5 (Driscoll et al., 2001; Kretser et al., 1989). In the Adirondacks, a positive relationship
4 exists between the pH and ANC in lakes and the number offish species present in those lakes
5 (ISA 3.2.3.4).
6 Since the mid-1990s, streams in the Shenandoah Case Study Area have shown slight
7 declines in NOs- and SC>4 2" concentrations in surface waters. The 2006 concentrations are still
8 above pre-acidification (1860) conditions. MAGIC modeling predicts surface water
9 concentrations of NOs- and SC>42" arelO- and 32-fold higher, respectively, in 2006 than in 1860.
10 The estimated average ANC across 60 streams in the Shenandoah Case Study Area is 57.9 ueq/L
11 (± 4.5 ueq/L). 55% of all monitored streams in the Shenandoah Case Study Area have a current
12 risk of Elevated, Severe, or Acute. Of the 55%, 18% are chronically acidic today (REA 4.2.4.3)
13 Based on a deposition scenario for this study area that maintains current emission levels
14 from 2020 to 2050, the simulation forecast indicates that a large number of streams still have
15 Elevated to Acute problems with acidity. In fact, from 2006 to 2050, the percentage of streams
16 with Acute Concern are predicted to increase by 5%, while the percentage of streams in
17 Moderate Concern decreases by 5%.
18 Biological effects of increased acidification documented in the Shenandoah Case Study
19 Area include a decrease in the condition factor in blacknose dace (Dennis and Bulgar 1995,
20 Bulgar et al., 1999) and a decrease in fish biodiversity associated with decreasing stream ANC
21 (Bulger et al.,1995; Dennis and Bulger, 1995; Dennis et al., 1995; MacAvoy and Bulger, 1995,
22 Bulgar et al., 1999). On average, the fish species richness is lower by one fish species for every
23 21 ueq/L decrease in ANC in Shenandoah National Park streams (ISA 3.2.3.4).
24
25
26
September, 2010 2-14 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2050
Figure 2-5.
2-
Average NOs" concentrations orange), SC>4 "concentrations (red), and
ANC (blue) levels for the 60 streams in the Shenandoah Case Study
Area modeled using MAGIC for the period 1850 to 2050.
ANC Preacidification (1860) and Current Condition (2006)
Pre-acidification (1860) Current (2006)
Source: EPA 2009
ANC (Meq/L)
• <0
0-20
20-50
• 50 -100
• >100
Figure 2-6. ANC levels of 1860 (preacidification) and 2006 (current) conditions
based on hindcasts of 60 streams in the Shenandoah Case Study Area
modeled using MAGIC
September, 2010
2-15
Do Not Quote or Cite
-------
Current Condition of Acidity
and Sensitivity
Criticial Load
meq/m2/yr
• Highly Sensitive: < 50
Moderately Sensitive: 51-100
Low Sensitivity: 101 -200
• Not Sensitive: > 201
Source: EPA 2009
3 Figure 2-7. Critical loads of surface water acidity for an ANC of 50 ueq/L for
4 Shenandoah Case Study Area streams. Each dot represents an
5 estimated amount of acidifying deposition (i.e., critical load) that each
6 stream's watershed can receive and still maintain a surface water
7 ANC >50 ueq/L. Watersheds with critical load values <100 meq/m2/yr
8 (red and orange circles) are most sensitive to surface water
9 acidification, whereas watersheds with values >100 meq/m2/yr (yellow
10 and green circles) are less sensitive sites.
11
12 2.1.3.2 Terrestrial Acidification
13 The ISA identified a variety of indicators that can be used to measure the effects of
14 acidification in soils. Most effects of terrestrial acidification are observed in sensitive forest
15 ecosystem in the U.S. Tree health has been linked to the availability of base cations (Be) in soil
16 (such as Ca2+, Mg2+ and potassium), as well as soil Al content. Tree species show a range of
17 sensitivities to Ca/Al and Bc/Al soil molar ratios, therefore these are good chemical indicators
18 because they directly relate to the biological effects. Critical Bc/Al molar ratios for a large
September, 2010
2-16
Do Not Quote or Cite
-------
1 variety of tree species ranged from 0.2 to 0.8 (Sverdrup and Warfvinge, 1993, a meta-data
2 analysis of laboratory and field studies). This range is similar to critical ratios of Ca/Al. Plant
3 toxicity or nutrient antagonism was reported to occur at Ca/Al molar ratios ranging from 0.2 to
4 2.5 (Cronan and Grigal, 1995; meta-data assessment) (REA pg 4-54, REA Appendix 5).
5 There has been no systematic national survey of terrestrial ecosystems to determine the
6 extent and distribution of terrestrial ecosystem sensitivity to the effects of acidifying deposition.
7 However, one preliminary national evaluation estimated that -15% of forest ecosystems in the
8 U.S. exceed the estimated critical load based on soil ANC leaching for S and N deposition by
9 >250 eq ha"1 yr"1 (McNulty et al., 2007). Forests of the Adirondack Mountains of New York,
10 Green Mountains of Vermont, White Mountains of New Hampshire, the Allegheny Plateau of
11 Pennsylvania, and high-elevation forest ecosystems in the southern Appalachians are the regions
12 most sensitive to terrestrial acidification effects from acidifying deposition (ISA 3.2.4.2). While
13 studies show some recovery of surface waters, there are widespread measurements of ongoing
14 depletion of exchangeable base cations in forest soils in the northeastern U.S. despite recent
15 decreases in acidifying deposition, indicating a slow recovery time.
16 In the REA, a critical load analysis was performed for sugar maple and red spruce forests
17 in the eastern United States by using Bc/Al ratio in acidified forest soils as an indicator to assess
18 the impact of nitrogen and sulfur deposition on tree health. These are the two most commonly
19 studied tree species in North America for effects of acidification. At a Bc/Al ratio of 1.2, red
20 spruce growth can be decreased by 20%. Sugar maple growth can be decreased by 20% at a
21 Bc/Al ratio of 0.6 (REA 4.4). The REA analysis determined the health of at least a portion of the
22 sugar maple and red spruce growing in the United States may have been compromised with
23 acidifying total nitrogen and sulfur deposition. Specifically, total nitrogen and sulfur deposition
24 levels exceeded three selected critical loads for tree growth in 3% to 75% of all sugar maple
25 plots across 24 states. For red spruce, total nitrogen and sulfur deposition levels exceeded three
26 selected critical loads in 3% to 36% of all red spruce plots across eight states (REA 4.4).
27
28 2.1.4 What are the key uncertainties associated with acidification?
29 There are different levels of uncertainty associated with relationships between deposition,
30 ecological effects and ecological indicators. In Chapter 7 of the REA, the case study analyses
31 associated with each targeted effect area were synthesized by identifying the strengths,
September, 2010 2-17 Do Not Quote or Cite
-------
1 limitations, and uncertainties associated with the available data, modeling approach, and
2 relationship between the selected ecological indicator and atmospheric deposition as described
3 by the ecological effect function (Figure 1-1). The key uncertainties were characterized as
4 follows to evaluate the strength of the scientific basis for setting a national standard to protect
5 against a given effect (REA 7.0):
6 • Data Availability: high, medium or low quality. This criterion is based on the
7 availability and robustness of data sets, monitoring networks, availability of data that
8 allows for extrapolation to larger assessment areas, and input parameters for modeling
9 and developing the ecological effect function. The scientific basis for the ecological
10 indicator selected is also incorporated into this criterion.
11 • Modeling Approach: high, fairly high, intermediate, or low confidence. This value is
12 based on the strengths and limitations of the models used in the analysis and how
13 accepted they are by the scientific community for their application in this analysis.
14 • Ecological Effect Function: high, fairly high, intermediate, or low confidence. This
15 ranking is based on how well the ecological effect function describes the relationship
16 between atmospheric deposition and the ecological indicator of an effect.
17
18 2.1.4.1 Aquatic Acidification
19 The REA concludes that the available data are robust and considered high quality. There
20 is high confidence about the use of these data and their value for extrapolating to a larger
21 regional population of lakes. The EPA TEVIE/LTM network represents a source of long-term,
22 representative sampling. Data on sulfate concentrations, nitrate concentrations and ANC from
23 1990 to 2006 used for this analysis as well as EPA EMAP and REMAP surveys, provide
24 considerable data on surface water trends.
25 There is fairly high confidence associated with modeling and input parameters.
26 Uncertainties in water quality estimates (.i.e., ANC) from MAGIC was derived from multiple
27 site calibrations. The 95% confidence interval for pre-acidification of lakes was an average of 15
28 |j,eq/L difference in ANC concentrations or 10% and 8 |j,eq/L or 5% for streams (REA 7.1.2).
29 The use of the critical load model used to estimate aquatic critical loads is limited by the
30 uncertainties associated with runoff and surface water measurements and in estimating the
31 catchment supply of base cations from the weathering of bedrock and soils (McNulty et al.,
September, 2010 2-18 Do Not Quote or Cite
-------
1 2007). To propagate uncertainty in the model parameters, Monte Carlo methods were employed
2 to develop an inverse function of exceedences. There is high confidence associated with the
3 ecological effect function developed for aquatic acidification. In calculating the ANC function,
4 the deposit!onal load for N or S is fixed by the deposition of the other, so deposition for either
5 will never be zero (Figure 7.1-6 REA).
6
7 2.1.4.2 Terrestrial Acidification
8 The available data used to quantify the targeted effect of terrestrial acidification are
9 robust and considered high quality. The USFS-Kane Experimental Forest and significant
10 amounts of research work in the Allegheny Plateau have produced extensive, peer-reviewed data
11 sets. A meta-analysis of laboratory studies showed that tree growth was decreased by 20%
12 relative to controls for BC/A1 ratios (ISA 7.2.1 and Figure 7.2-1). Sugar maple and red spruce
13 were the focus of the REA since they are demonstrated to be negatively affected by soil available
14 Ca2+ depletion and high concentrations of available Al, and occur in areas that receive high
15 acidifying deposition, There is high confidence about the use of the REA terrestrial acidification
16 data and their value for extrapolating to a larger regional population of forests.
17 There is high confidence associated with the models, input parameters, and assessment of
18 uncertainty used in the case study for terrestrial acidification. The Simple Mass Balance (8MB)
19 model, a commonly used and widely applied approach for estimating critical loads, was used in
20 the REA analysis (ISA 7.2.2). There is fairly high confidence associated with the ecological
21 effect function developed for terrestrial acidification (REA 7.2.3).
22
23 2.2 NITROGEN ENRICHMENT: EVIDENCE OF EFFECTS ON STRUCTURE AND
24 FUNCTION OF TERRESTRIAL AND FRESHWATER ECOSYSTEMS
25 The following summary is a concise overview of the known or anticipated effects caused
26 by nitrogen nutrient enrichment to ecosystems within the United States. Nutrient-enrichment
27 affects terrestrial, freshwater and estuarine ecosystems. Nitrogen deposition is a major source of
28 anthropogenic nitrogen. For many terrestrial and freshwater ecosystems other sources of
29 nitrogen including fertilizer and waste treatment are greater than deposition. Nitrogen deposition
30 often contributes to nitrogen-enrichment effects in estuaries, but does not drive the effects since
31 other sources of N greatly exceed N deposition. Both oxides of nitrogen and reduced forms of
September, 2010 2-19 Do Not Quote or Cite
-------
1 nitrogen, (e.g. NHx) contribute to nitrogen deposition. For the most part, nitrogen effects on
2 ecosystems do not depend on whether the nitrogen is in oxidized or reduced form. Thus, this
3 summary focuses on the effects of nitrogen deposition in total. We address the issue of
4 incorporating the relative contributions of oxidized and reduced nitrogen into the standards in
5 Chapters 5 and 8.
6
7 2.2.1 What is the nature of terrestrial and freshwater ecosystem responses to reactive
8 nitrogen and/ sulfur deposition?
9 The ISA found that deposition of nitrogen, including NOX and NHX, leads to the nitrogen
10 enrichment of ecosystems (EPA 2008). In the process of nitrogen enrichment, biogeochemical
11 components of terrestrial and freshwater aquatic ecosystems are altered in a way that leads to
12 effects on biological organisms.
13 The evidence is sufficient to infer a causal relationship between N deposition and the
14 alteration of biogeochemical cycling in terrestrial ecosystems (ISA 4.3.1.1 and 3.3.2.1). This is
15 supported by numerous observational, deposition gradient and field addition experiments in
16 sensitive ecosystems. Stoddard (1994) identified the leaching of NCV in soil drainage waters and
17 the export of NCV in steam water as two of the primary indictors of N enrichment. Several N-
18 addition studies indicate that NCV leaching is induced by chronic additions of N (Edwards et al.,
19 2002b; Kahl et al., 1999; Peterjohn et al., 1996; Norton et al., 1999). Aber et al. (2003) found
20 that surface water NCV concentrations exceeded 1 |j,eq/L in watersheds receiving about 9 to 13
21 kg N/ha/yr of atmospheric N deposition. N deposition disrupts the nutrient balance of
22 ecosystems with numerous biogeochemical effects. The chemical indicators that are typically
23 measured include NCV leaching, soil C:N ratio, rates of N mineralization, nitrification,
24 denitrification, foliar N concentration, and soil water NOs - and NH4+ concentrations. Note that
25 N saturation (N leaching from ecosystems) does not need to occur to cause effects. Substantial
26 leaching of NOs- from forest soils to stream water can acidify downstream waters, leading to
27 effects described in the previous section on aquatic acidification. Due to the complexity of
28 interactions between the N and C cycling, the effects of N on C budgets (quantified input and
29 output of C to the ecosystem) are variable. Regional trends in net ecosystem productivity (NEP)
30 of forests (not managed for silviculture) have been estimated through models based on gradient
31 studies and meta-analysis. Atmospheric N deposition has been shown to cause increased litter
September, 2010 2-20 Do Not Quote or Cite
-------
1 accumulation and carbon storage in above-ground woody biomass. In the West, this has lead to
2 increased susceptibility to more severe fires. Less is known regarding the effects of N deposition
3 on C budgets of non-forest ecosystems.
4 The evidence is sufficient to infer a causal relationship between N deposition on the
5 alteration of species richness, species composition and biodiversity in terrestrial ecosystems
6 (ISA 4.3.1.2). Some organisms and ecosystems are more sensitive to N deposition and effects of
7 N deposition are not observed in all habitats. The most sensitive terrestrial taxa to N deposition
8 are lichens. Empirical evidence indicates that lichens in the U.S. are affected by deposition levels
9 as low as 3 kg N/ha/yr. Alpine ecosystems are also sensitive to N deposition, changes in an
10 individual species (Carex rupestris) were estimated to occur at deposition levels near 4 kg N
11 /ha/yr and modeling indicates that deposition levels near 10 kg N/ha/yr alter plant community
12 assemblages. In several grassland ecosystems, reduced species diversity and an increase in non-
13 native, invasive species are associated with N deposition (Clark and Tillman, 2008; Schwinning
14 et al., 2005).
15 In freshwater ecosystems, the evidence is sufficient to infer a causal relationship between
16 N deposition and the alteration of biogeochemical cycling in freshwater aquatic ecosystems (ISA
17 3.3.2.3). N deposition is the main source of N enrichment to headwater streams, lower order
18 streams and high elevation lakes. The most common chemical indicators that were studied
19 included NOs and dissolved inorganic nitrogen (DIN) concentration in surface waters as well as
20 Chi a:total P ratio. Elevated surface water NO3 concentrations occur in both the eastern and
21 western U.S. Bergstrom and Jansson (2006) report a significant correlation between N deposition
22 and lake biogeochemistry by identifying a correlation between wet deposition and [DIN] and Chi
23 a: Total P. Recent evidence provides examples of lakes and streams that are limited by N and
24 show signs of eutrophication in response to N addition.
25 The evidence is sufficient to infer a causal relationship between N deposition and the
26 alteration of species richness, species composition and biodiversity in freshwater aquatic
27 ecosystems (ISA 3.3.5.3). Increased N deposition can cause a shift in community composition
28 and reduce algal biodiversity, especially in sensitive oligotrophic lakes.
29
30
31
September, 2010 2-21 Do Not Quote or Cite
-------
1 2.2.2 What types of ecosystems are sensitive to such effects? How are these responses
2 affected by atmospheric, ecological, and landscape factors?
3 The numerous ecosystem types that occur across the U.S. have a broad range of
4 sensitivity to N deposition (Clark and Tilman 2008; Aber et al., 2003; Fenn et al., 2003; Rueth et
5 al., 2003; Egerton-Warburton and Allen 2000; Williams et al., 1996; and additional studies
6 summarized in Table 4-4 ISA). Increased deposition to N-limited ecosystems can lead to
7 production increases that may be either beneficial or adverse depending on the system and
8 management goals.
9 Organisms in their natural environment are commonly adapted to a specific regime of
10 nutrient availability. Change in the availability of one important nutrient, such as N, may result
11 in imbalance in ecological stoichiometry, with effects on ecosystem processes, structure and
12 function (Sterner and Elser, 2002). In general, N deposition to terrestrial ecosystems causes
13 accelerated growth rates in some species deemed desirable in commercial forests but may lead to
14 altered competitive interactions among species and nutrient imbalances, ultimately affecting
15 biodiversity. The onset of these effects occurs with N deposition levels as low as 3 kg N/ha/yr in
16 sensitive terrestrial ecosystems to N deposition. In aquatic ecosystems, N that is both leached
17 from the soil and directly deposited to the water surface can pollute the surface water. This
18 causes alteration of the diatom community at levels as low as 1.5 kg N/ha/yr in sensitive
19 freshwater ecosystems.
20 The degree of ecosystem effects lies at the intersection of N loading and N-sensitivity.
21 N-sensitivity is predominately driven by the degree to which growth is limited by nitrogen
22 availability. Grasslands in the western United States are typically N-limited ecosystems
23 dominated by a diverse mix of perennial forbs and grass species (Clark and Tilman, 2008;
24 Suding et al., 2005). A meta-analysis by Lebauer and Treseder (2008) indicated that N
25 fertilization increased aboveground growth in all non-forest ecosystems except for deserts. In
26 other words, almost all terrestrial ecosystems are N-limited and will be altered by the addition of
27 anthropogenic nitrogen (LeBauer and Treseder, 2008). Likewise, a freshwater lake or stream
28 must be N-limited to be sensitive to N-mediated eutrophication. There are many examples of
29 fresh waters that are N-limited or N and phosphorous (P) co-limited (ISA 3.3.3.2). In a meta-
30 analysis that included 653 datasets, Elser et al. (2007) found that N-limitation occurred as
31 frequently as P-limitation in freshwater ecosystems. Additional factors that govern the sensitivity
September, 2010 2-22 Do Not Quote or Cite
-------
1 of ecosystems to nutrient enrichment from N deposition include rates and form of N deposition,
2 elevation, climate, species composition, plant growth rate, length of growing season, and soil N
3 retention capacity. (ISA 4.3). Less is known about the extent and distribution of the terrestrial
4 ecosystems in the U.S. that are most sensitive to the effects of nutrient enrichment from
5 atmospheric N deposition compared to acidification.
6
7 2.2.3 What is the magnitude of ecosystem responses to nitrogen deposition?
8 2.2.3.1 Terrestrial
9 Little is known about the full extent and distribution of the terrestrial ecosystems in the
10 U.S. that are most sensitive to impacts caused by nutrient enrichment from atmospheric N
11 deposition. As previously stated, most terrestrial ecosystems are N-limited, therefore they are
12 sensitive to perturbation caused by N additions (LeBauer and Treseder, 2008). Effects are most
13 likely to occur where areas of relatively high atmospheric N deposition intersect with N-limited
14 plant communities. The alpine ecosystems of the Colorado Front Range, chaparral watersheds of
15 the Sierra Nevada, lichen and vascular plant communities in the San Bernardino Mountains and
16 the Pacific Northwest, and the southern California coastal sage scrub (CSS) community are
17 among the most sensitive terrestrial ecosystems. There is growing evidence that existing
18 grassland ecosystems in the western United States are being altered by elevated levels of N
19 inputs, including inputs from atmospheric deposition (Clark and Tilman, 2008; Suding et al.,
20 2005).
21 In the eastern U.S., the degree of N saturation of the terrestrial ecosystem is often
22 assessed in terms of the degree of NOs leaching from watershed soils into ground water or
23 surface water. Stoddard (1994) estimated the number of surface waters at different stages of
24 saturation across several regions in the eastern U.S. Of the 85 northeastern watersheds examined
25 60% were in Stage 1 or Stage 2 of N saturation on a scale of 0 (background or pretreatment) to 3
26 (visible decline). Of the northeastern sites for which adequate data were available for assessment,
27 those in Stage 1 or 2 were most prevalent in the Adirondack and Catskill Mountains. Effects on
28 individual plant species have not been well studied in the U.S. More is known about the
29 sensitivity of particular plant communities. Based largely on results obtained in more extensive
30 studies conducted in Europe, it is expected that the more sensitive terrestrial ecosystems include
September, 2010 2-23 Do Not Quote or Cite
-------
1 hardwood forests, alpine meadows, arid and semi-arid lands, and grassland ecosystems (ISA
2 3.8.2).
3 The REA used published research results (REA 5.3.1 and ISA Table 4.4) to identify
4 meaningful ecological benchmarks associated with different levels of atmospheric nitrogen
5 deposition. These are given by Figure 2-8. The sensitive areas and ecological indicators
6 identified by the ISA were analyzed further in the REA to create a national map that illustrates
7 effects observed from ambient and experimental atmospheric nitrogen deposition loads in
8 relation to Community Multi-scale Air Quality (CMAQ) 2002 modeling results and NADP
9 monitoring data. This map, reproduced in Figure 2-9, depicts the sites where empirical effects of
10 terrestrial nutrient enrichment have been observed and site proximity to elevated atmospheric N
11 deposition.
September, 2010 2-24 Do Not Quote or Cite
-------
1
2
Rocky Mountain alpine lakes: shift in diatom community dominance (Baron, 2006)
• Southern California: N growth requirement threshold (Wood et al., 2006)
• San Bernardino Mountains and Sierra Nevada Mountains: acidophytic lichen
decline in MCF (Fenn et al., 2008)
• Eastern Rocky Mountain Slope: low carbon:nitrogen; low lignin:nitrogen (Baron et
al.,2000)
• Eastern Rocky Mountain Slope: increased foliar nitrogen; increased mineralization
(Baronet al.,2000)
• San Bernardino Mountains and Sierra Nevada Mountains: shift from acidophytic
to neutral or nitrogen-tolerant lichen in MCF (Fenn et al., 2008)
• Minnesota grasslands: decreased plant species (Clark and Tilman, 2008)
• Northeast U.S.: NO3 leaching (Aber et al., 2003)
• Bay Area, CA: Increased cover of nonnative grasses; decreased native
grasses (Weiss, 1999)
San Bernardino Mountains and Sierra Nevada Mountains: loss of acidophytic
lichen in MCF (Fenn et al., 2008)
Southern California: shift in mycorrhizal species in CSS (Egerton-Warburton
and Allen, 2000)
Southern California: shift from native species to invasive grasses in CSS (Allen,
2008)*
• San Bernardino Mountains: high dissolved organic nitrogen (Meixnerand
Fenn, 2004)
• San Bernardino Mountains: nitrogen saturation (Meixner and Fenn, 2004)
Increased nitrogen in lichen (Fenn et al., 2007)
17
• MCF: N03 leaching (Fenn et al., 2008)
> MCF: 25% decrease in fine-root biomass (Fenn et al., 2008)
• Southern California: NO3 leaching (Fenn et al., 2003)
• Southern California: high foliar nitrogen (Bytnerawicz and
Fenn, 1996)
• Los Angeles Basin, California: High NO emissions
(Bytnerowicz and Fenn, 1996)
Fraser Experimental Forest, CO:
increased foliar nitrogen; increased
mineralization (Rueth et al., 2003)**
35
40
45
Nitrogen Deposition, kg/ha/yr
* Personal communication, 2008. Also referenced in Bcbbink et. al ,2010, Ecological Applications,^ 1):3Q-59 and USDS FS, 2010,
httpi/i-WAw. rt rs.fs. fed. ua^clean_ali_wal6r^ciean_waEer/critlcaljQadsflocal-resourc6s/dc3cs^Empirlcal_CLS_af_N_1DD414.pdf
"Nitrogen deposition levels include ambient and experimental additions.
Figure 2-8. Benchmarks of atmospheric nitrogen deposition for several ecosystem
indicators (REA 5.3.1.2) MCF-Mixed Conifer Forest, CSS-Coastal Sage
Srnih
September, 2010
2-25
Do Not Quote or Cite
-------
Total N Deposition p> Effgcts ^ Observatlorvaj Research
(kgfha/yr)
•• High 66507
Low: 0.761
/ \ Eff&clsfrorn Experimental Research
^H National Parks
j^B National Forests
1. Impacts on lichen communilies
(CA rmxed conifer forests. 3_1 kg/ha/yr.
Columbia R. Gorge. GR/VWA: 11.5-29.4)
Z Evidence of threatened and endangered
species impacted (San Francisco Bay. CA
10-15 kg/ha/yr)
3 N saturation, high nitrate in streanwater. soil
leaves; htgh NO emissions (LA CA Air Basm
25-45 kg/ha/yr. northeast US: 7-10 kg/ha/yr)
4. N saturation, hign nitrate in streamwaEer
(San Bernardino Wins _ CA 18 8 kt^ha/yr)
5 N saturation, high Dissolved Inorganic N
(San Bernardino Mlns . CA. 11-40 kg/ha/yr)
t (Ambient N Deposition onrv>
6 Increased tree mortality and beetle activity
(San Bernardino Mtns., CA
7 N enrichment or eulrophication of lakes
(Loch Vale. CO. 0.5-1.5 kg/ha/yr.
Niwot Ridge. CO 4.71-16. 5 kg/ha/yr)
8. Increased soil and foliar N concentration
(Eastern slope of Rocky Mtns
3 5-3 6 kgtis/yr)
9. Decrease in prtcher plant (Hawiey Bog. MA
and Malty Bog. VA: 10-14 kg/ha/yr)
10. Nitrate leaching (Adirondack lakes.
New England. 8-10 kg/ha/yr)
Effects from Experimental Research (Ambient H deposition * Experimental N Additional
11. Decreased diversity of mycorrhtzal communities
(southern California: -10 kg/ha/yr. Northern
Michigan-35-38 kg/heJyr)
12 Decreased desert creosote bush,
increased non native grasses
(Mojave Desert. CA 46-62 fcgrtia/yr)
13 Community structure changes
(Chihuahuan Desert. CA 21 7 kg/ha/yr and up)
14. Alpine meadows elevated nitrate levels
in runoff (Colo. Front Range: 20 kgftia/yr)
15. Alpine meadows' shift toward hairgrass
r u^ing inodelt-
iiowLvur BIU observed ftfleets w«re produced as a
re&ult of ehron.it additions greater ihan 10
2 Figure 2-9 Observed effects from ambient and experimental atmospheric nitrogen deposition loads in relation to
3 using CMAQ 2002 modeling results and NADP monitoring data. Citations for effect results are from the ISA, Table
4 4.4 (U.S. EPA, 2008) 1= Fenn et. al. (2008), 2=Weiss (1999), 3=Bytnerowicz and Fenn (1996), 4=Fenn et al. (2000), 5= Meixner and
5 Fenn (2004), 6=Jones et al. 2004, 7=Baron (2006), 8=Baron et al. (2000), 9=Gotelli and Ellison (2002), 10=Stoddard et al. (1994),
6 1 l=Egerton Warburton and Allen (2000), 12=Brooks (2003), 13=Baez et al. (2007), 14=Bowman et al. (2006), 15=Bowman et al.
7 (1995), 16=Rueth et al. (2003), 17=DeWalle et al. (2006), 18=Clark and Tillman (2008), 19=Rueth et al. 2003
September, 2010
2-26
Do Not Quote or Cite
-------
1 Based on information in the ISA and initial analysis in the REA, further case
2 study analyses on terrestrial nutrient enrichment of ecosystems were developed for the
3 CCS community and Mixed Conifer Forest (MCF) (EPA 2009). Geographic information
4 systems (GIS) analysis supported a qualitative review of past field research to identify
5 ecological benchmarks associated with CSS and mycorrhizal communities, as well as
6 MCF's nutrient-sensitive acidophyte lichen communities, fine-root biomass in Ponderosa
7 pine, and leached nitrate in receiving waters.
8 The ecological benchmarks that were identified for the CSS and the MCF are
9 included in the suite of benchmarks identified in the ISA (ISA 3.3). There are sufficient
10 data to confidently relate the ecological effect to a loading of atmospheric nitrogen. For
11 the CSS community, the following ecological benchmarks were identified:
12 • 3.3 kg N/ha/yr - the amount of nitrogen uptake by a vigorous stand of CSS; above
13 this level, nitrogen may no longer be limiting
14 • 10 kg N/ha/yr - mycorrhizal community changes
15 For the MCF community, the following ecological benchmarks were identified:
16 • 3.1 kg N/ha/yr - shift from sensitive to tolerant lichen species
17 • 5.2 kg N/ha/yr-dominance of the tolerant lichen species
18 • 10.2 kg N/ha/yr-loss of sensitive lichen species
19 • 17 kg N/ha/yr - leaching of nitrate into streams.
20
21 These benchmarks, ranging from 3.1 to 17 kg N/ha/yr, were compared to 2002
22 CMAQ/NADP data to discern any associations between atmospheric deposition and
23 changing communities. Evidence supports the finding that nitrogen alters CSS and MCF.
24 Key findings include the following: 2002 CMAQ/NADP nitrogen deposition data show
25 that the 3.3 kg N/ha/yr benchmark has been exceeded in more than 93% of CSS areas
26 (654,048 ha). These deposition levels are a driving force in the degradation of CSS
27 communities. Although CSS decline has been observed in the absence of fire, the
28 contributions of deposition and fire to the CSS decline require further research. CSS is
29 fragmented into many small parcels, and the 2002 CMAQ/NADP 12-km grid data are not
30 fine enough to fully validate the relationship between CSS distribution, nitrogen
31 deposition, and fire. 2002 CMAQ/NADP nitrogen deposition data exceeds the 3.1 kg
September, 2010 2-27 Do Not Quote or Cite
-------
1 N/ha/yr benchmark in more than 38% (1,099,133 ha) of MCF areas, and nitrate leaching
2 has been observed in surface waters. Ozone effects confound nitrogen effects on MCF
3 acidophyte lichen, and the interrelationship between fire and nitrogen cycling requires
4 additional research.
5
6 2.2.3.2 Freshwater
7 The magnitude of ecosystem response may be thought of on two time scales,
8 current conditions and how ecosystems have been altered since the onset of
9 anthropogenic N deposition. As noted previously, Elser et al. (2008) found that N-
10 limitation occurs as frequently as P-limitation in freshwater ecosystems (ISA 3.3.3.2).
11 Recently, a comprehensive study of available data from the northern hemisphere surveys
12 of lakes along gradients of N deposition show increased inorganic N concentration and
13 productivity to be correlated with atmospheric N deposition (Bergstrom and Jansson
14 2006). The results are unequivocal evidence of N limitation in lakes with low ambient
15 inputs of N, and increased N concentrations in lakes receiving N solely from atmospheric
16 N deposition (Bergstrom and Jansson, 2006). These authors suggested that most lakes in
17 the northern hemisphere may have originally been N-limited, and that atmospheric N
18 deposition has changed the balance of N and P in lakes.
19 Available data suggest that the increases in total N deposition do not have to be
20 large to elicit an ecological effect. For example, a hindcasting exercise determined that
21 the change in Rocky Mountain National Park lake algae that occurred between 1850 and
22 1964 was associated with an increase in wet N deposition that was only about 1.5 kg
23 N/ha (Baron, 2006). Similar changes inferred from lake sediment cores of the Beartooth
24 Mountains of Wyoming also occurred at about 1.5 kg N/ha deposition (Saros et al.,
25 2003). Pre-industrial inorganic N deposition is estimated to have been only 0.1 to 0.7 kg
26 N/ha based on measurements from remote parts of the world (Galloway et al., 1995;
27 Holland et al., 1999). In the western U.S., pre-industrial, or background, inorganic N
28 deposition was estimated by (Holland et al., 1999) to range from 0.4 to 0.7 kg N/ha/yr.
29 Eutrophication effects from N deposition are most likely to be manifested in
30 undisturbed, low nutrient surface waters such as those found in the higher elevation areas
31 of the western U.S. The most severe eutrophi cation from N deposition effects is expected
September, 2010 2-28 Do Not Quote or Cite
-------
1 downwind of major urban and agricultural centers. High concentrations of lake or
2 streamwater NOs , indicative of ecosystem saturation, have been found at a variety of
3 locations throughout the U.S., including the San Bernardino and San Gabriel Mountains
4 within the Los Angeles Air Basin (Fenn et al., 1996), the Front Range of Colorado
5 (Baron et al., 1994; Williams et al., 1996), the Allegheny mountains of West Virginia
6 (Gilliam et al., 1996), the Catskill Mountains of New York (Murdoch and Stoddard,
7 1992; Stoddard, 1994), the Adirondack Mountains of New York (Wigington et al., 1996),
8 and the Great Smoky Mountains in Tennessee (Cook et al., 1994) (ISA 3.3.8).
9
10 2.2.3.3 Estuaries
11 In contrast to terrestrial and freshwater systems, atmospheric N load to estuaries
12 contributes to the total load but does not necessarily drive the effects since other sources
13 of N greatly exceed N deposition. In estuaries, N-loading from multiple anthropogenic
14 and non-anthropogenic pathways leads to water quality deterioration, resulting in
15 numerous effects including hypoxic zones, species mortality, changes in community
16 composition and harmful algal blooms that are indicative of eutrophication. The
17 following summary is a concise overview of the known or anticipated effects of nitrogen
18 enrichment on estuaries within the United States.
19
20 What is the nature of estuary responses to reactive nitrogen deposition?
21 In the ISA, the evidence is sufficient to infer a causal relationship between Nr
22 deposition and the biogeochemical cycling of N and carbon (C) in estuaries (ISA 4.3.4.1
23 and 3.3.2.3). In general, estuaries tend to be nitrogen-limited, and many currently receive
24 high levels of nitrogen input from human activities (REA 5.1.1). It is unknown if
25 atmospheric deposition alone is sufficient to cause eutrophication; however, the
26 contribution of atmospheric nitrogen deposition to total nitrogen load is calculated for
27 some estuaries and can be >40% (REA 5.1.1).
28 The evidence is sufficient to infer a causal relationship between N deposition and
29 the alteration of species richness, species composition and biodiversity in estuarine
30 ecosystems (ISA 4.3.4.2 and 3.3.5.4). Atmospheric and non-atmospheric sources of N
31 contribute to increased phytoplankton and algal productivity, leading to eutrophication.
September, 2010 2-29 Do Not Quote or Cite
-------
1 Shifts in community composition, reduced hypolimnetic DO, decreases in biodiversity,
2 and mortality of submerged aquatic vegetation are associated with increased N deposition
3 in estuarine systems.
4
5 What types of ecosystems are sensitive to such effects? How are these
6 responses affected by atmospheric, ecological, and landscape factors?
7 Because the productivity of estuarine and near shore marine ecosystems is
8 generally limited by the availability of N, they are susceptible to the eutrophication effect
9 of N deposition (ISA 4.3.4.1). A recent national assessment of eutrophic conditions in
10 estuaries found the most eutrophic estuaries were generally those that had large
11 watershed-to-estuarine surface area, high human population density, high rainfall and
12 runoff, low dilution, and low flushing rates (Bricker et al., 2007). In the REA, the
13 National Oceanic and Atmospheric Administration's (NOAA) National Estuarine
14 Eutrophication Assessment (NEEA) assessment tool, Assessment of Estuarine Tropic
15 Status (ASSETS) categorical Eutrophication Index (El) (Bricker et al., 2007) was used to
16 evaluate eutrophication due to atmospheric loading of nitrogen. ASSETS El is an
17 estimation of the likelihood that an estuary is experiencing eutrophication or will
18 experience eutrophication based on five ecological indicators: chlorophyll a, macroalgae,
19 dissolved oxygen, nuisance/toxic algal blooms and submerged aquatic vegetation (SAV)
20 (Bricker et al., 2007).
21 In the REA, two regions were selected for case study analysis using ASSETS El,
22 the Chesapeake Bay and Pamlico Sound. Both regions received an ASSETS El rating of
23 Bad indicating that the estuary had moderate to high pressure due to overall human
24 influence and a moderate high to high eutrophic condition (REA 5.2.4.1 and 5.2.4.2).
25 These results were then considered with SPAtially Referenced Regression (SPARROW)
26 modeling to develop a response curve to examine the role of atmospheric nitrogen
27 deposition in achieving a desired decrease in load. To change the Neuse River Estuary's
28 El score from Bad to Poor not only must 100% of the total atmospheric nitrogen
29 deposition be eliminated, but considerably more nitrogen from other sources as well must
30 be controlled (REA section 5.2.7.2). In the Potomac River estuary, a 78% decrease of
31 total nitrogen could move the El score from Bad to Poor (REA 5.2.7.1). The results of
32 this analysis indicated decreases in atmospheric deposition alone could not eliminate
33 coastal eutrophication problems due to multiple non-atmospheric nitrogen inputs (REA
September, 2010 2-30 Do Not Quote or Cite
-------
1 7.3.3). However, by decreasing atmospheric contributions, it may help avoid the need for
2 more costly controls on nitrogen from other sources. In addition, the somewhat arbitrary
3 discreteness of the El scale can mask the benefits of decreases in nitrogen between
4 categories.
5 In general, estuaries tend to be N-limited (Elser et al., 2008), and many currently
6 receive high levels of N input from human activities to cause eutrophication (Howarth et
7 al., 1996; Vitousek and Howarth, 1991). Atmospheric N loads to estuaries in the U.S. are
8 estimated to range from 2-8% for Guadalupe Bay, TX on the lowest end to as high as
9 72% for St Catherines-Sapelo estuary, GA (Castro et al., 2003). The Chesapeake Bay is
10 an example of a large, well-studied and severely eutrophic estuary that is calculated to
11 receive as much as 30% of its total N load from the atmosphere.
12
13 What is the magnitude of ecosystem responses to eutrophication?
14 There is a scientific consensus that nitrogen-driven eutrophication in shallow
15 estuaries has increased over the past several decades and that the environmental
16 degradation of coastal ecosystems due to nitrogen, phosphorus, and other inputs is now a
17 widespread occurrence (Paerl et al., 2001). For example, the frequency of phytoplankton
18 blooms and the extent and severity of hypoxia have increased in the Chesapeake Bay
19 (Officer et al., 1984) and Pamlico estuaries in North Carolina (Paerl et al., 1998) and
20 along the continental shelf adjacent to the Mississippi and Atchafalaya rivers' discharges
21 to the Gulf of Mexico (Eadie et al., 1994).
22 A recent national assessment of eutrophic conditions in estuaries found that 65%
23 of the assessed systems had moderate to high overall eutrophic conditions and generally
24 received the greatest N loads from all sources, including atmospheric and land-based
25 sources (Bricker et al., 2007). Most eutrophic estuaries occurred in the mid-Atlantic
26 region and the estuaries with the lowest degree of eutrophication were in the North
27 Atlantic (Bricker et al., 2007). Other regions had mixtures of low, moderate, and high
28 degrees of eutrophication (ISA 4.3.4.3).
29 The mid-Atlantic region is the most heavily impacted area in terms of moderate or
30 high loss of submerged aquatic vegetation due to eutrophication (ISA 4.3.4.2).
31 Submerged aquatic vegetation is important to the quality of estuarine ecosystem habitats
32 because it provides habitat for a variety of aquatic organisms, absorbs excess nutrients,
September, 2010 2-31 Do Not Quote or Cite
-------
1 and traps sediments (ISA 4.3.4.2). It is partly because many estuaries and near-coastal
2 marine waters are degraded by nutrient enrichment that they are highly sensitive to
3 potential negative impacts from nitrogen addition from atmospheric deposition.
4
5 2.2.4 What are the key uncertainties associated with nutrient enrichment?
6 There are different levels of uncertainty associated with relationships between
7 deposition, ecological effects and ecological indicators. The criteria used in the REA to
8 evaluate the degree of confidence in the data, modeling and ecological effect function are
9 detailed in Chapter 7 of the REA and summarized in section 2.1.4 of this chapter (REA
10 7.0).
11
12 2.2.4.1 Aquatic
13 The approach for assessing atmospheric contributions to total nitrogen loading in
14 the REA, was to consider the main-stem river to an estuary (including the estuary) rather
15 than an entire estuary system or bay. The biological indicators used in the NOAA
16 ASSETS El required the evaluation of many national databases including the US
17 Geological Survey National Water Quality Assessment (NAWQA) files, EPA's
18 STORage and RETrival (STORET) database, NOAA's Estuarine Drainage Areas data,
19 and EPA's water quality standards nutrient criteria for rivers and lakes (REA Appendix
20 6, Table 1.2.-1). Both the SPARROW modeling for nitrogen loads and assessment of
21 estuary conditions under NOAA ASSETS El, have been applied on a national scale. The
22 REA concludes that the available data are medium quality with intermediate confidence
23 about the use of these data and their values for extrapolating to a larger regional area
24 (REA 7.3.1). Intermediate confidence is associated with the modeling approach using
25 ASSETS El and SPARROW. The REA states there is low confidence with the
26 ecological effect function due to the results of the analysis which indicated that
27 reductions in atmospheric deposition alone could not solve coastal eutrophication
28 problems due to multiple non-atmospheric nitrogen inputs (REA 7.3.3).
29
30
31
September, 2010 2-32 Do Not Quote or Cite
-------
1 2.2.4.2 Terrestrial
2 Ecological thresholds are identified for CSS and MCF and these data are
3 considered to be of high quality, however, the ability to extrapolate these data to larger
4 regional areas is limited (REA 7.4.1). No quantitative modeling was conducted or
5 ecological effect function developed for terrestrial nutrient enrichment reflecting the
6 uncertainties associated with these depositional effects.
7
8 2.3 WHAT ECOLOGICAL EFFECTS ARE ASSOCIATED WITH GAS-
9 PHASE NOX AND SOX?
10 Acidifying deposition and nitrogen enrichment are the main focus of this policy
11 assessment; however, there are other known ecological effects are attributed to gas-phase
12 NOx and SOx. Acute and chronic exposures to gaseous pollutants such as sulfur dioxide
13 (802), nitrogen dioxide (NO2), nitric oxide (NO), nitric acid (HNOs) and peroxyacetyl
14 nitrite (PAN) are associated with negative impacts to vegetation. The current secondary
15 NAAQS were set to protect against direct damage to vegetation by exposure to gas-phase
16 NOx or SOx, such as foliar injury, decreased photosynthesis, and decreased growth. The
17 following summary is a concise overview of the known or anticipated effects to
18 vegetation caused by gas phase N and S. Most phototoxic effects associated with gas
19 phase NOx and SOx occur at levels well above ambient concentrations observed in the
20 U.S. (ISA 3.4.2.4).
21
22 2.3.1 What is the nature of ecosystem responses to gas-phase nitrogen and sulfur?
23 The 2008 ISA found that gas phase N and S are associated with direct phytotoxic
24 effects (ISA 4.4). The evidence is sufficient to infer a causal relationship between
25 exposure to SO2 and injury to vegetation (ISA 4.4.1 and 3.4.2.1). Acute foliar injury to
26 vegetation from SO2 may occur at levels above the current secondary standard (3-h
27 average of 0.50 ppm). Effects on growth, reduced photosynthesis and decreased yield of
28 vegetation are also associated with increased SO2 exposure concentration and time of
29 exposure.
30 The evidence is sufficient to infer a causal relationship between exposure to NO,
31 NO 2 and PAN and injury to vegetation (ISA 4.4.2 and 3.4.2.2). At sufficient
September, 2010 2-33 Do Not Quote or Cite
-------
1 concentrations, NO, NC>2 and PAN can decrease photosynthesis and induce visible foliar
2 injury to plants. Evidence is also sufficient to infer a causal relationship between
3 exposure toHNOs and changes to vegetation (ISA 4.4.3 and 3.4.2.3). Phytotoxic effects
4 of this pollutant include damage to the leaf cuticle in vascular plants and disappearance of
5 some sensitive lichen species.
6 2.3.2 What types of ecosystems are sensitive to such effects? How are these
7 responses affected by atmospheric, ecological, and landscape factors?
8 Vegetation in ecosystems near sources of gaseous NOx and SOx or where
9 ambient concentrations of SC>2, NO, NO2, PAN and HNOs are higher are more likely to
10 be impacted by these pollutants. Uptake of these pollutants in a plant canopy is a complex
11 process involving adsorption to surfaces (leaves, stems and soil) and absorption into
12 leaves (ISA 3.4.2). The functional relationship between ambient concentrations of gas
13 phase NOx and SOx and specific plant response are impacted by internal factors such as
14 rate of stomatal conductance and plant detoxtification mechanisms, and external factors
15 including plant water status, light, temperature, humidity, and pollutant exposure regime
16 (ISA 3.4.2).
17 Entry of gases into a leaf is dependent upon physical and chemical processes of
18 gas phase as well as to stomatal aperature. The aperature of the stomata is controlled
19 largely by the prevailing environmental conditions, such as water availability, humidity,
20 temperature, and light intensity. When the stomata are closed, resistance to gas uptake is
21 high and the plant has a very low degree of susceptibility to injury. Mosses and lichens
22 do not have a protective cuticle barrier to gaseous pollutants or stomata and are generally
23 more sensitive to gaseous sulfur and nitrogen than vascular plants (ISA 3.4.2).
24 The appearance of foliar injury can vary significantly across species and growth
25 conditions affecting stomatal conductance in vascular plants (REA 6.4.1). For example,
26 damage to lichens from SO2 exposure include decreases in photosynthesis and
27 respiration, damage to the algal component of the lichen, leakage of electrolytes,
28 inhibition of nitrogen fixation, decreased K+ absorption, and structural changes (Belnap
29 et al., 1993; Farmer et al., 1992, Hutchinson et al., 1996).
30
31
September, 2010 2-34 Do Not Quote or Cite
-------
1
2 2.3.3 What is the magnitude of ecosystem responses to gas phase effects of NOx
3 and SOx?
4 The phytotoxic effects of gas phase NOx and SOx are dependent on the exposure
5 concentration and duration and species sensitivity to these pollutants. Effects to
6 vegetation associated with NOx and SOx, are therefore, variable across the U.S. and tend
7 to be higher near sources of photochemical smog. For example, SO2 is considered to be
8 the primary factor contributing to the death of lichens in many urban and industrial areas,
9 with fruticose lichens being more susceptible to SO2 than many foliose and crustose
10 species (Hutchinson et al., 1996).
11 The ISA states there is very limited new research on phytotoxic effects of NO,
12 NO2, PAN and HNOs at concentrations currently observed in the United States with the
13 exception of some lichen species (ISA 4.4). Past and current HNOs concentrations may
14 be contributing to the decline in lichen species in the Los Angeles basin (Boonpragob and
15 Nash 1991; Nash and Sigal, 1999; Riddell et al., 2008). PAN is a very small component
16 of nitrogen deposition in most areas of the United States (REA 6.4.2). Current
17 deposition of HNOs is contributing to N saturation of some ecosystems close to sources
18 of photochemical smog (Fenn et al., 1998) such as the MCF's of the Los Angeles basin
19 mountain (Bytnerowicz et al., 1999). Most phototoxic effects associated with gas phase
20 NOx and SOx occur at levels well above ambient concentrations observed in the U.S.
21 (ISA 3.4.2.4).
22
23 2.4 SUMMARY
24 In summary, NOx and SOx in the atmosphere contribute to effects on individual
25 species and ecosystems through direct contact with vegetation, and more significantly
26 through deposition to sensitive ecosystems. The ISA concludes that the evidence is
27 sufficient to conclude causal relationships between acidifying deposition of N and S and
28 effects on freshwater aquatic ecosystems and terrestrial ecosystems, and between
29 nitrogen nutrient enrichment and effects on sensitive terrestrial and freshwater aquatic
30 ecosystems. The ISA also concludes that a causal relationship is supported between
31 nitrogen nutrient enrichment and effects on estuarine ecosystems; however, the
September, 2010 2-35 Do Not Quote or Cite
-------
1 contribution of atmospheric oxidized nitrogen relative to reduced nitrogen and non-
2 atmospheric nitrogen is more difficult to determine.
3 The REA provides additional support that under recent conditions, deposition
4 levels have exceeded benchmarks for ecological indicators of acidification and nutrient
5 enrichment that indicate that effects are likely to be widespread in lakes and streams
6 within sensitive ecosystems.
7
September, 2010 2-36 Do Not Quote or Cite
-------
1 2.5 REFERENCES
2
3 Aber JD; Goodale CL; Ollinger SV; Smith ML; Magill AH; Martin ME; Hallett RA;
4 Stoddard JL. (2003). Is nitrogen deposition altering the nitrogen status of
5 northeastern forests? Bioscience, 53, 375-389.
6 Allen, E.B. (2008). Personal communication from E.B. Allen, U.S. Department of
7 Agriculture Forest Service.
8 Bailey SW; Horsley SB; Long RP; Hallett RA. (2004). Influence of edaphic factors on
9 sugar maple nutrition and health on the Allegheny Plateau. Soil Sci Soc Am J, 68,
10 243-252.
11 Baker JP; Bernard DP; Christensen SW; Sale MJ. (1990). Biological effects of changes in
12 surface water acid-base chemistry. (State of science / technology report
13 #13).Washington DC: National Acid Precipitation Assessment Program
14 (NAPAP).
15 Baker JP; Van Sickle J; Gagen CJ; DeWalle DR; Sharpe WE; Carline RF; Baldigo BP;
16 Murdoch PS; Bath DW; Kretser WA; Simonin HA; Wigington PJ Jr. (1996).
17 Episodic acidification of small streams in the northeastern United States: Effects
18 on fish populations. Ecol Appl, 6, 423-437.
19 Baron JS; Ojima DS; Holland EA; Parton WJ. (1994). Analysis of nitrogen saturation
20 potential in Rocky Mountain tundra and forest: Implications for aquatic systems.
21 Biogeochemistry, 27, 61-82.
22 Baron, J.S., H.M. Rueth, A.M. Wolfe, K.R. Nydick, E.J. Allstott, J.T. Minear, and B.
23 Moraska.(2000). Ecosystem responses to nitrogen deposition in the Colorado
24 Front Range. Ecosystems 3:352-368.
25 Baron JS. (2006). Hindcasting nitrogen deposition to determine ecological critical load.
26 Ecol Appl, 16, 433-439.
27 Belnap J; Sigal L; Moir W; Eversman S. (1993). Identification of sensitive species, in
28 lichens as bioindicators of air quality. In: Huckaby LS (Ed.), Lichens as
29 bioindicators of air quality (pp. 67-88). Fort Collins, Colorado: Rocky Mountain
30 Forest and Range Experimental Station, U.S. Forest Service, U. S. Department of
31 Agriculture.
32 Bergstrom A; Jansson M. (2006). Atmospheric nitrogen deposition has caused nitrogen
33 enrichment and eutrophication of lakes in the northern hemisphere. Glob Chang
34 Biol, 12, 635-643.
35 Blake L; Goulding KWT; Mott CJB; Johnston AE. (1999). Changes in soil chemistry
36 accompanying acidification over more than 100 years under woodland and grass
37 at Rothamsted Experimental Station, UK. Eur J Soil Sci, 50, 401-412.
38 Boonpragob K; Nash THI. (1991). Physiological responses of the lichen Ramalina
39 menziesii Tayl. to the Los Angeles urban environment. Environ Exp Bot, 31, 229-
40 238.
41 Bricker S; Longstaff B; Dennison W; Jones A; Boicourt K; Wicks C; Woerner J. (2007).
42 Effects of nutrient enrichment in the nation's estuaries: A decade of change.
43 http://ccmaserver.nos.noaa.gov/publications/eutroupdate/. (NOAA Coastal Ocean
44 Program Decision Analysis Series No. 26). Silver Spring, MD: National Centers
45 for Coastal Ocean Science, National Oceanic and Atmospheric
46 Administration(NOAA).
September, 2010 2-37 Do Not Quote or Cite
-------
1 Bulger AJ; Dolloff CA; Cosby BJ; Eshleman KN; Webb JR; Galloway JN. (1995). The
2 "Shenandoah National Park: fish in sensitive habitats" (SNP: FISH) Project. An
3 integrated assessment offish community responses to stream acidification. Water
4 Air Soil Pollut, 85, 309-314.
5
6 Bulger AJ; Cosby, BJ; Dolloff, CA; Eshleman, KN; Webb, JR; Galloway, JN. (1999).
7 SNP:FISH, Shenandoah National Park: Fish in sensitive habitats. Project final
8 report -Volume I: Project description and summary of results; Volume II: Stream
9 water chemistry and discharge, and synoptic water quality surveys. Volume III:
10 Basin-wide habitat and population inventories, and behavioral responses to acid in
11 a laboratory stream. Volume IV: Stream bioassays, aluminum toxicity, species
12 richness and stream chemistry, and models of susceptibility to acidification.
13 (Project Completion Report to the National Park Service). Charlottesville, VA,
14 Department of Environmental Sciences; University of Virginia.
15 Bulger AJ; Cosby BJ; Webb JR. (2000). Current, reconstructed past, and projected future
16 status of brook trout (Salvelinus fontinalis) streams in Virginia. Can J Fish Aquat
17 Sci, 57, 1515-1523.
18 Bytnerowicz, A., and M.E. Fenn. (1996). Nitrogen deposition in California forests: a
19 review Environmental Pollution 92(2): 127-146.
20 Bytnerowicz A; Padgett P; Percy K; Krywult M; Riechers G; Horn J. (1999). Direct
21 effects of nitric acid on forest vegetation. In: Miller PR, McBride J (Eds.),
22 Oxidant air pollution impacts in the Montane forests of Southern California: The
23 San Bernardino Mountains case study, Ecological Series 134. (pp. 270-287). New
24 York: Springer.
25 Castro MS; Driscoll CT; Jordan TE; Reay WG; Boynton WR. (2003). Sources of
26 nitrogen to estuaries in the United States. Estuaries, 26, 803-814.
27 Clark CM; Tilman D. (2008). Loss of plant species after chronic low-level nitrogen
28 deposition to prairie grasslands. Nature, 451, 712-715.
29 Cook RB; Elwood JW; Turner RR; Bogle MA; Mulholland PJ; Palumbo AV. (1994).
30 Acid-base chemistry of high-elevation streams in the Great Smoky Mountains.
31 Water Air Soil Pollut, 72, 331-356.
32 Cronan CS; Grigal DF. (1995). Use of calcium/aluminum ratios as indicators of stress in
33 forest ecosystems. J Environ Qual, 24, 209-226.
34 DeHayes DH; Schaberg PG; Hawley GJ; Strimbeck GR. (1999). Acid rain impacts on
35 calcium nutrition and forest health. Bioscience, 49, 789-800.
36 Dennis TD; MacAvoy SE; Steg MB; Bulger AJ. (1995). The association of water
37 chemistry variables and fish condition in streams of Shenandoah National Park
38 (USA). Water Air Soil Pollut, 85, 365-370.
39 Dennis TE; Bulger AJ. (1999). The susceptibility of blacknose dace (Rhinichthys
40 atratulus).to acidification in Shenandoah National Park. In: Bulger AJ; Cosby BJ;
41 Dolloff CA; Eshleman KN; Galloway JN; Webb. JR (Eds.), Shenandoah National
42 Park: Fish in sensitive habitats. Project final report, Volume IV: Stream bioassays,
43 aluminum toxicity, species richness and stream chemistry, and models of
44 susceptibility to acidification. Chapter 6B. Project completion report to the
45 National Park Service. Charlottesville, VA; Department of Environmental
46 Sciences, University of Virginia.
September, 2010 2-38 Do Not Quote or Cite
-------
1 Driscoll CT; Lawrence GB; Bulger AJ; Butler TJ; Cronan CS; Eagar C; Lambert KF;
2 Likens GE; Stoddard JL; Weather KC. (200Ib). Acidic deposition in the
3 northeastern United States: Sources and inputs, ecosystem effects, and
4 management strategies. Bioscience, 51, 180-198.
5
6 Eadie BJ; McKee BA; Lansing MB; Robbins JA; Metz S; Trefry JH. (1994). Records of
7 nutrient-enhanced coastal productivity in sediments from the Louisiana
8 continental shelf. Estuaries, 17, 754-765.
9 Edwards PJ; Kochenderfer JN; Coble DW; Adams MB. (2002). Soil leachate responses
10 during 10 years of induced whole-watershed acidification. Water Air Soil Pollut,
11 140,99-118.
12 Egerton-Warburton, L.M., and E.B. Allen. (2000). Shifts in arbuscular mycorrhizal
13 communities along an anthropogenic nitrogen deposition gradient. Ecological
14 Applications 70(2):484-496.
15 Elser JJ; Bracken MES; Cleland EE; Gruner DS; Harpole WS; Hillebrand IIH; Ngai JT;
16 Seabloom EW; Shurin JB; Smith JE. (2007). Global analysis of nitrogen and
17 phosphorus limitation of primary producers in freshwater, marine, and terrestrial
18 ecosystems. Ecol Lett, 10, 1135-1142
19 Farmer AM; Bates JW; Bell JNB. (1992). Ecophysiological effects of acid rain on
20 bryophytes and lichens. In: Bates JW; Farmer AM (Eds.), Bryophytes and lichens
21 in a changing environment. Oxford, UK: Claredon Press.
22 Fenn ME; Poth MA; Johnson DW. (1996). Evidence for nitrogen saturation in the San
23 Bernardino Mountains in southern California. For Ecol Manage, 82, 211-230.
24 Fenn, M.E., M.A. Poth, S.L. Schilling, and D.B. Grainger. (2000). Throughfall and fog
25 deposition of nitrogen and sulfur at an N-limited and N-saturated site in the San
26 Bernardino Mountains, southern California. Canadian Journal of Forest Research
27 30:1476-1488.
28 Fenn, M.E., J.W. Baron, E.B. Allen, H.M. Rueth, K.R. Nydick, L. Geiser, W.D. Bowen,
29 J.O. Sickman, T. Meixner, D.W. Johnson, and P. Neitlich. (2003). Ecological
30 effects of nitrogen deposition in the western United States. Bioscience 53(4):404-
31 420.
32 Fenn, M.E., L. Geiser, R. Bachman, T.J. Blubaugh, and A. Bytnerowicz. (2007).
33 Atmospheric deposition inputs and effects on lichen chemistry and indicator
34 species in the Columbia River Gorge, USA. Environmental Pollution 146:11-91.
35 Fenn, M.E., S. Jovan, F. Yuan, L. Geiser, T. Meixner, and B.S. Gimeno. (2008).
36 Empirical and simulated critical loads for nitrogen deposition in California mixed
37 conifer forests. Environmental Pollution 755(3):492-511.
38 Ford J; Stoddard JL; Powers CF. (1993). Perspectives in environmental monitoring: an
39 introduction to the U.S. EPA long-term monitoring (LTM) project. Water Air Soil
40 Pollut, 67, 247-255.
41 Galloway JN; Schlesinger WH; Levy H; Michaels AF; Schnoor JL. (1995). Nitrogen-
42 fixation: anthropogenic enhancement, environmental response. Glob Biogeochem
43 Cycles, 9, 235-252.
44 Gilliam FS; Adams MB; Yurish BM. (1996). Ecosystem nutrient responses to chronic
45 nutrient inputs at Fernow Experimental Forest, West Virginia. Can J For Res, 26,
46 196-205.
September, 2010 2-39 Do Not Quote or Cite
-------
1 Haines TA; Baker JP. (1986). Evidence offish population responses to acidification in
2 the eastern United States. Water Air Soil Pollut, 31, 605-629.
3 Holland EA; Dentener FJ; Braswell BH; Sulzman JM. (1999). Contemporary and pre-
4 industrial global reactive nitrogen budgets. Biogeochemistry, 46, 7-43.
5 Horsley SB; Long RP; Bailey SW; Hallett RA; Hall TJ. (2000). Factors associated with
6 the decline disease of sugar maple on the Allegheny Plateau. Can J For Res, 30,
7 1365-1378.
8 Howarth RW; Billen G; Swaney D; Townsend A; Jaworski N; Lajtha K; Downing JA;
9 Elmgren R; Caraco N; Jordan T; Berendse F; Freney J; Kudeyarov V; Murdoch
10 PS; Zhao-Liang Z. (1996). Regional nitrogen budgets and riverine N & P fluxes
11 for the drainages to the North Atlantic Ocean: natural and human influences.
12 Biogeochemistry, 35, 75-139.
13 Hutchinson J; Maynard D; Geiser L. (1996). Air quality and lichens - a literature review
14 emphasizing the Pacific Northwest, USA. Pacific Northwest Region Air Resource
15 Management Program; U.S. Forest Service; U.S. Department of Agriculture
16 (USD A).
17 Kahl J; Norton S; Fernandez I; Rustad L; Handley M. (1999). Nitrogen and sulfur input-
18 output budgets in the experimental and reference watersheds, Bear Brook
19 Watershed in Maine (BBWM). Environ Monit Assess, 55,113-131
20 Keller W; Gunn JM. (1995). Lake water quality improvements and recovering aquatic
21 communities. In: Gunn JM (Ed.), Restoration and recovery of an industrial region:
22 progress in restoring the smelter-damaged landscape near Sudbury, Canada (pp.
23 67-80). New York, NY: Springer-Verlag.
24 Kretser W; Gallagher J; Nicolette J. (1989). Adirondack Lakes Study, 1984-1987: An
25 Evaluation of Fish Communities and Water Chemistry: Ray Brook, NY; prepared
26 for: Adirondack Lakes Survey (ALS) Corporation.
27 LeBauer DS; Treseder KK. (2008). Nitrogen limitation of net primary productivity in
28 terrestrial ecosystems is globally distributed. Ecology, 89,371-379.
29 Lien L; Raddum GG; Fjellheim A. (1992). Critical loads for surface waters: invertebrates
30 and fish. (Acid rain research report no 21). Oslo, Norway: Norwegian Institute for
31 Water Research.
32 Longcore JR; Gill JD (Eds.), (1993). Acidic depositions: effects on wildlife and habitats.
33 (Wildlife Society technical review no 93-1). Bethesda, MD: The Wildlife Society.
34 MacAvoy SW; Bulger AJ. (1995). Survival of brook trout (Salvelinus fontinalis)
35 embryos and fry in streams of different acid sensitivity in Shenandoah National
36 Park, USA. Water Air Soil Pollut, 85, 445-450.
37 McNulty SG; Cohen EC; Myers JAM; Sullivan TJ; Li H. (2007). Estimates of critical
38 acid loads and exceedances for forest soils across the conterminous United States.
39 Environ Pollut, 149, 281-292.
40 Meixner, T., and M. Fenn. (2004). Biogeochemical budgets in a Mediterranean
41 catchment with high rates of atmospheric N deposition - Importance of scale and
42 temporal asynchrony. Biogeochemistry 70:3 31 -3 5 6.
43 Murdoch PS; Stoddard JL. (1992). The role of nitrate in the acidification of streams in the
44 Catskill Mountains of New York. Water Resour Res, 28, 2707-2720
45 Nash TH; Sigal LL. (1999). Epiphytic lichens in the San Bernardino mountains in
46 relation to oxidant gradients. In: Miller PR, McBride JR (Eds.), Oxidant air
September, 2010 2-40 Do Not Quote or Cite
-------
1 pollution impacts on the montane forests of southern California: A case study of
2 the San Bernardino mountains. Ecological Studies, 134, (pp. 223-234). New
3 York, NY: Springer-Verlag.
4 Norton S; Kahl J; Fernandez I. (1999). Altered soil-soil water interactions inferred from
5 stream water chemistry at an artificially acidified watershed at Bear Brook
6 Watershed, Maine USA. Environ Monit Assess, 55, 97-111.
7 Officer CB; Biggs RB; Taft JL; Cronin LE; Tyler MA; Boynton WR. (1984). Chesapeake
8 Bay anoxa. Origin, development and significance. Science, 223, 22-27.
9 Paerl HW; Boynton WR; Dennis RL; Driscoll CT; Greening HS; Kremer JN; Rabalais
10 NN; Seitzinger SP. (2001). Atmospheric deposition of nitrogen in coastal waters:
11 Biogeochemical and ecological implications. In: Valigura RW; Alexander RB;
12 Castro MS; Meyers TP; Paerl HW; Stacey PE; Turner RE (Eds.), Nitrogen
13 loading in coastal water bodies: an atmospheric perspective. (Coastal and
14 estuarine series; Volume 57, pp 11-52). Washington, DC: American Geophysical
15 Union.
16 Paerl H; Pinckney J; Fear J; Peierls B. (1998). Ecosystem responses to internal and
17 watershed organic matter loading: Consequences for hypoxia in the eutrophying
18 Neusse River Estuary, NC, USA. Mar Ecol Prog Ser, 166,17-25.
19 Parker KE. (1988). Common loon reproduction and chick feeding on acidified lakes in
20 the Adirondack Park, New York. Canadian Journal of Zoology, 66, 804-810.
21 Parker DR; Zelazny LW; Kinraide TB. (1989). Chemical speciation and plant toxicity of
22 aqueous aluminum. In: Lewis TE (Ed.), Environmental chemistry and toxicology
23 of aluminum (pp. 117-145). American Chemical Society.
24 Peterjohn WT; Adams MB; Gilliam FS. (1996). Symptoms of nitrogen saturation in two
25 central Appalachian hardwood forest ecosystems.Biogeochemistry, 35, 507-522.
26 Reuss JO. (1983). Implications of the calcium-aluminum exchange system for the effect
27 of acid precipitation on soils. J Environ Qual, 12, 591-595.
28 Riddell J; Nash lii TH; Padgett P. (2008). The effect of HNO3 gas on the lichen
29 Ramalina menziesii. Flora - Morphology, Distribution, Functional Ecology of
30 Plants, 203, 47-54.
31 Ross DS; Lawrence GB; Fredriksen G. (2004). Mineralization and nitrification patterns at
32 eight northeastern USA forested research sites. For Ecol Manage, 188, 317-335.
33 Rueth, H.M., J.S. Baron, and E.J. Allstott. 2003. Responses of Engelmann spruce forests
34 to nitrogen fertilization in the Colorado Rocky Mountains. Ecological
35 Applications 13:664-673.
36 Saros JE; Interlandi SJ; Wolfe AP; Engstrom DR. (2003). Recent changes in the diatom
37 community structure of lakes in the Beartooth Mountain Range, USA. Arct
38 Antarct Alp Res, 35,18-23.
39 Schindler DW; Mills KH; Malley DF; Findlay MS; Schearer JA; Davies IJ; Turner MA;
40 Lindsey GA; Cruikshank DR. (1985). Long-term ecosystem stress: Effects of
41 years of experimental acidification. Science, 228, 1395-1401.
42 Schreck CB. (1981).Stress and rearing of salmonids. Aquaculture, 28, 241-249.
43 Schreck CB. (1982). Stress and compensation in teleostean fishes: response to social and
44 physical factors. In: Pickering AD (Ed.), Stress and fish (pp. 295-321). London:
45 Academic Press.
September, 2010 2-41 Do Not Quote or Cite
-------
1 Schwinning S; Starr BI; Wojcik NJ; Miller ME; Ehleringer JE; Sanford RL Jr. (2005).
2 Effects of nitrogen deposition on an arid grassland in the Colorado Plateau cold
3 desert. Journal of Rangeland Ecology and Management, 58, 565-574.
4 Sterner RW; Elser JJ. (2002). Ecological stoichiometry: the biology of elements from
5 molecules to the biosphere. Princeton, NJ: Princeton University Press.
6 Stoddard JL. (1990). Plan for converting the NAPAP aquatic effects long-term
7 monitoring (LTM) project to the temporally integrated monitoring of ecosystems
8 (TIME) project. (International Report). Corvallis, OR; U.S. Environmental
9 Protection Agency.
10 Stoddard JL. (1994). Long-term changes in watershed retention of nitrogen: its causes
11 and aquatic consequences. In Baker LA (Ed.), Environmental chemistry of lakes
12 and reservoirs, (pp. 223-284). Washington, D.C.: American Chemical Society.
13 Stoddard JL; Urquhart NS; Newell AD; Kugler D. (1996). The Temporally Interated
14 Monitoring of Ecosystems (TIME) project design 2. Detection of regional
15 acidification trends. Water Resour Res, 32, 2529-2538.
16 Suding KN; Collins SL; Gough L; Clark C; Cleland EE; Gross KL; Milchunas DG;
17 Pennings S. (2005). Functional- and abundance-based mechanisms explain
18 diversity loss due to N fertilization. Proc Natl Acad Sci USA, 102, 4387-4392.
19 Sullivan TJ; Driscoll CT; Cosby BJ; Fernandez IJ; Herlihy AT; Zhai J; Stemberger R;
20 Snyder KU; Sutherland JW; Nierzwicki-Bauer SA; Boylen CW; McDonnell TC;
21 Nowicki NA. (2006). Assessment of the extent to which intensively studied lakes
22 are representative of the Adirondack Mountain region. (Final Report no 06-
23 17).Corvallis, OR; prepared by Environmental Chemistry, Inc. for: Albany, NY;
24 Environmental Monitoring Evaluation and Protection Program of the New York
25 State Energy Research and Development Authority (NYSERDA).
26 Sverdrup H; Warfvinge P. (1993). The effect of soil acidification on the growth of trees,
27 grass and herbs as expressed by the (Ca+ Mg+ K)/A1 ratio. Rep in Ecol & Eng, 2,
28 1993.
29 US EPA (2008) U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen
30 and Sulfur Ecological Criteria (Final Report). U.S. Environmental Protection
31 Agency, Washington, D.C., EPA/600/R-08/082F, 2008.
32 US EPA (2009) Risk and Exposure Assessment for Review of the Secondary National
33 Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur-Main
34 Content - Final Report. U.S. Environmental Protection Agency, Washington,
35 D.C., EPA-452/R-09-008a
36 Vitousek PM; Howarth RW. (1991). Nitrogen limitation on land and in the sea: how can
37 it occur? Biogeochemistry, 13, 87-115.
38 Weiss, S.B. 1999. Cars, cows, and checkerspot butterflies: Nitrogen deposition and
39 management of nutrient-poor grasslands for a threatened species. Conservation
40 5/o/ogy 73:1476-1486.
41 Williams MW; Baron JS; Caine N; Sommerfeld R; Sanford JR. (1996). Nitrogen
42 saturation in the Rocky Mountains. Environ Sci Technol, 30, 640-646.
43 Wigington PJ Jr; DeWalle DR; Murdoch PS; Kretser WA; Simonin HA; Van Sickle J;
44 Baker JP. (1996b). Episodic acidification of small streams in the northeastern
45 United States: Ionic controls of episodes. Ecol Appl, 6, 389-407.
September, 2010 2-42 Do Not Quote or Cite
-------
1 Wedemeyer GA; Barton BA; MeLeay DJ. (1990).Stress and acclimation. In: Schreck CB,
2 Moyle PB (Eds.), Methods for fish biology (pp. 178-196). Bethesda, MD:
3 American Fisheries Society.
4 Wood, Y., T. Meixner, PJ. Shouse, and E.B. Allen. (2006). Altered Ecohydrologic
5 response drives native shrub loss under conditions of elevated N-deposition.
6 Journal of Environmental Quality 35:76-92.
September, 2010 2-43 Do Not Quote or Cite
-------
1 3 CONSIDERATIONS OF ADVERSITY TO PUBLIC WELFARE
2
3 3.1 How do we characterize adversity to public welfare? What are the relevant factors
4 and how are they addressed in this document?
5
6 Characterizing a known or anticipated adverse effect to public welfare is an important
7 component of developing any secondary NAAQS. According to the Clean Air Act, welfare
8 effects include:
9 effects on soils, water, crops, vegetation, manmade materials, animals, wildlife,
10 weather, visibility, and climate, damage to and deterioration of property, and
11 hazards to transportation, as well as effect on economic values and on personal
12 comfort and well-being, whether caused by transformation, conversion, or
13 combination with other air pollutants (CAA, Section 302(h)).
14
15 While the text above lists a number of welfare effects, these effects do not define public
16 welfare in and of themselves.
17 Although there is no specific definition of adversity to public welfare, the paradigm of
18 adversity to public welfare as deriving from disruptions in ecosystem structure and function has
19 been used broadly by EPA to categorize effects of pollutants from the cellular to the ecosystem
20 level. An evaluation of adversity to public welfare might consider the likelihood, type,
21 magnitude, and spatial scale of the effect as well as the potential for recovery and any
22 uncertainties relating to these considerations.
23 Similar concepts were used in past reviews of secondary NAAQS for ozone, PM (relating
24 to visibility), as well as in initial reviews of effects from lead deposition. Because NOy and SOx
25 are deposited from ambient sources into ecosystems where they affect changes to organisms,
26 populations and ecosystems, the concept of adversity to public welfare as related to impacts on
27 the public from alterations in structure and function of ecosystems is appropriate for this review.
28 Other information that may be helpful to consider includes the role of critical loads and
29 ecosystem service impacts as benchmarks or measures of impacts on ecosystems that may affect
30 public welfare. Ecosystem services can be related directly to concepts of public welfare to
31 inform discussions of societal adverse impacts. In subsequent sections we will discuss each of
32 these concepts as they relate to adversity.
33
34
September, 2010 3-1 Do Not Quote or Cite
-------
1 3.1.1 What are the benchmarks for adversity from other sources?
2 3.1.1.1 Ozone NAAQS Review
3 The evaluation of adversity from a public welfare perspective in the context of ozone and
4 particulate matter (PM) are relevant to this current review. Both ozone and PM have
5 documented effects on ecological receptors. These criteria pollutants are being reviewed on a
6 schedule as part of the NAAQS process. The ozone secondary standard is currently under
7 reconsideration from the 2008 ruling with a proposal was published January 6, 2010. The final
8 Policy Assessment for PM is being developed and is expected to be finalized in the fall of 2010.
9
10 For the purposes of the reconsideration of the secondary standard for ozone welfare
11 effects of ozone are primarily limited to vegetation. These effects begin at the level of the
12 individual cell and accumulate up to the level of whole leaves and plants. If effects occur on
13 enough individual plants within the population, communities and ecosystems may be impacted.
14 Prior to the 2008 ozone review, Ozone vegetation effects were classified as either "injury" or
15 "damage" (FR 72 37889). "Injury" was defined as; encompassing all plant reactions, including
16 reversible changes or changes in plant metabolism, quality or reduced growth that does not
17 impair the intended use of the plant while "damage" includes those injury effects that reach
18 sufficient magnitude as to decrease or impair the intended use of the plant (FR 72 37890). The
19 "intended use" of the plant was imbedded with the concept of adversity to public welfare.
20 Ozone-associated "damage" was considered adverse if the intended use of the plant was
21 compromised (i.e. crops, ornamentals, plants located in Class I areas). Effects of ozone on
22 single plants or species grown in monocultures such as agricultural crops and managed forests
23 were evaluated without consideration of potential effects on natural forests or entire ecosystems.
24 In the 2008 rulemaking, EPA expanded the characterization of adversity beyond the
25 individual plant level and this language is continued in the 2010 ozone reconsideration. The
26 2008 final rule and 2010 proposal conclude that a determination of what constitutes an "adverse"
27 welfare effect in the context of secondary NAAQS review can appropriately occur by
28 considering effects at higher ecological levels (populations, communities, ecosystems) as
29 supported by recent literature. The ozone review uses the example of the construct presented in
30 Hogsett et al. (1997) as a model for assessing risks to forests. This study suggests that adverse
31 effects could be classified into one or more of the following categories: (1) economic production,
September, 2010 3-2 Do Not Quote or Cite
-------
1 (2) ecological structure, (3) genetic resources, and (4) cultural values". Another recent
2 publication, "A Framework for Assessing and Reporting on Ecological Condition: an SAB
3 report" (Young and Sanzone, 2002) provides additional support for expanding the consideration
4 of adversity beyond the species level and at higher levels by making explicit the linkages
5 between stress-related effects at the species level and at higher levels within an ecosystem
6 hierarchy (See Figure 3-1).
9
10
11
12
13
14
Figure 3-1.
Hydrologic alteration
Habitat conversion
Habitat fragmentation
Climate change
Invasive non-native species
Turbidity/sedimentation
Pesticides
Disease/pest outbreaks
Nutrient pukes
Modi
Dissolved oxygen depletion
Ozone (tropoipheric)
Hfdrologic alteration
Hah At t conversion
Habita t fragmentation
Climate change
Over-harvesting of vegetation
Large-scale invasive
specifs introductions
Large-scale disease/pest outbreaks
Landscape
Condition
Biotlc
Condition
itural
Hydrologic alteration
Habitat conversion
Climate change
Over-harvesting of vegetation
Disease/pest outbreaks
Altered fire regime
Altered flood regime
Hydrologic alteration
Habitat conversion
Climate change
Turbidity/sedimentation
Pesticides
Nutrient pulses
Metali
Dissohjed oxygen depletion
Ozone (tropoipheric)
Nitrogen oxides
Natural
Disturbance
Hydrology/
GeoniDrphology
Hydrologic alteration
Habitat conversion
Habitat fragmentation
C&rtiate change
Turbidity/sedimentation
Hydrologic alteration
Halfita t conversion
Climate change
Pesticides
Disease/pest outbreaks
Nutrient pulses
Dissolve/a oxygen depletion
Nitrogen oxides
Common anthropogenic stressors and the essential ecological attributes they
affect. Modified from Young and Sanzone (2002)
In the 2008 ozone NAAQS review and current ozone NAAQS proposal, the
interpretation of what constitutes an adverse effect on public welfare can vary depending on the
location and intended use of the plant. The degree to which Os-related effects are considered
adverse to public welfare depends on the intended use of the vegetation and its significance to
September, 2010
$-3
Do Not Quote or Cite
-------
1 public welfare (73 FR 16496). Therefore, effects on vegetation (e.g., biomass loss, foliar injury,
2 impairment of intended use) may be judged to have a different degree of impact on public
3 welfare depending, for example, on whether that effect occurs in a Class I area, a city park,
4 commercial cropland or private land.
5 In the proposed ozone reconsideration in 2010 the Administrator has found that the types
6 of information most useful in informing the selection of an appropriate range of protective levels
7 is appropriately focused on information regarding exposures and responses of sensitive trees and
8 other native species that occur in protected areas such as Class I areas or on lands set aside by
9 States, Tribes and public interest groups to provide similar benefits to the public welfare. She
10 further notes that while direct links between O3 induced visible foliar injury symptoms and other
11 adverse effects (e.g., biomass loss) are not always found, visible foliar injury in itself is
12 considered by the National Park Service (NFS) to affect adversely air quality related values
13 (AQRV) in Class I areas, while the Administrator recognizes that uncertainty remains as to what
14 level of annual tree seedling biomass loss when compounded over multiple years should be
15 judged adverse to the public welfare, she believes that the potential for such anticipated effects
16 should be considered in judging to what degree a standard should be precautionary (73 FR
17 16496). The range of proposed levels from 7-15 ppb includes at the maximum level of 15 ppb
18 protection of approximately 75% of seedlings from more than 10% biomass loss.
19
20 3.1.2 Other EPA Programs and Federal Agencies
21 Various federal laws and policies exist to protect ecosystem health. How other federal
22 agencies and EPA offices consider ecosystem effects in carrying out their programs can help
23 inform the Administrator when she evaluates the adversity of ecosystem impacts on public
24 welfare. From the 1996 National Acid Precipitation Assessment Program Report to Congress: "
25 The 1990 Clean Air Act Amendments require that the National Acid Precipitation Assessment
26 Program (NAPAP) prepare biennial reports to Congress, and that "every four years ... the report
27 ... shall include the reduction in deposition rates that must be achieved in order to prevent
28 adverse ecological effects" (Public Law 101-549, Title IX, Section 903 (j)(3)(F)(i), codified as
29 amended at 42 USC §7403(j)(3)(F)(I)). Although the term adverse ecological effects is not
30 specifically defined in the Clean Air Act Amendments, a working definition can be derived from
31 relevant statements at various locations in the statute. Congress expresses its concern with
September, 2010 3-4 Do Not Quote or Cite
-------
1 ecological components (the scope is broad and inclusive, since ecology encompasses the
2 interrelationships of organisms and their environment) in the preceding subsection (E) of the
3 statute. That subsection requires reporting on "the status of ecosystems (including forest and
4 surface waters) ... affected by acid deposition ... including changes in surface water quality and
5 forest and soil conditions ... [and] high elevation watersheds" (42 USC §7403(j)(3)(E)(iii-v)).
6 The adverse effects of concern to Congress, as evidenced in its findings and declaration of
7 purpose, are the "dangers to the public health and welfare ... including injury ... damage ... and
8 ... deterioration" (42 USC §7401 (a)). Based on the intent of Congress, as expressed above and
9 elsewhere in the Clean Air Act, and shaped by indications of intent expressed in other relevant
10 environmental statutes and regulations, the following working definition of adverse ecological
11 effects has been derived and is used in the preparation of the NAPAP report:
12 any injury (i.e., loss of chemical or physical quality or viability) to any
13 ecological or ecosystem component, up to and including at the regional
14 level, over both long and short terms. Similarly, adverse effects for other
15 areas of concern addressed in this report—i.e., visibility, materials, and
16 human health—consist of loss of quality up to and including at the
17 regional level, over both long and short terms."
18
19 As another example, an effect may be considered adverse to public welfare if it
20 contributes to the inability of areas to meet water quality objectives as defined by the Clean
21 Water Act. The following federal statutes and policies may prove helpful to consider.
22
23 3.1.21. Prevention of Significant Deterioration Program
24 The Clean Air Act's Prevention of Significant Deterioration (PSD) program (42 U.S.C.
25 7470) purposes include to "preserve, protect and enhance the air quality in national parks,
26 wilderness areas and other areas of natural, recreational, scenic or historic value . . . ." Also, the
27 PSD program charges the Federal Land Managers, including the NPS, with ". . . an affirmative
28 responsibility to protect the air quality related values . . . "within federal Class I lands. (42 U.S.C.
29 7475(d)(2)(B)).
30
31
September, 2010 3-5 Do Not Quote or Cite
-------
1
2 3.1.2.2 EPA Office of Water
3 Section 101 of the Clean Water Act (CWA) (Declaration of Goals and Policy) states that
4 the objective of the CWA is to restore and maintain the chemical, physical, and biological
5 integrity of the Nation's waters and to attain, where possible, water quality that protects fish,
6 shellfish, wildlife and provides for water-based recreation.
7 The CWA also authorizes EPA to develop water quality criteria as a guide for the states
8 to set water quality standards to protect aquatic life. In consideration of acidification effects,
9 EPA's Redbook, Quality Criteria for Water., published originally in 1976, recommends that
10 alkalinity be 20 mg/l\L or more as CaCOs for freshwater aquatic life except where natural
11 concentrations are less. Alkalinity is the sum total of components in the water that tend to
12 el evate the pH of the water ab ove a value of ab out 4.5.
13 As mentioned in the Redbook, alkalinity is expressed as CaCOs in mg/L. Alkalinity
14 differs slightly from ANC in that ANC includes other buffering compounds (Na, Mg, and K) as
15 well and includes buffering capacity of particulates in the water sample. Since alkalinity is
16 expressed as mg/L and ANC is expressed as ueq/L, alkalinity must be multiplied by 20 to be
17 converted to ueq/L. Thus a recommended criterion of 20 mg/L alkalinity is equivalent to an
18 ANC of 400 ueq/L.
19
20 3.1.2.3 National Park Service
21 The National Park Service (NFS) is responsible for the protection of all resources within
22 the national park system. These resources include those that are related to and/or dependent
23 upon good air quality, such as whole ecosystems and ecosystem components. The NFS, in its
24 Organic Act (16 U.S.C. 1), is directed to conserve the scenery, natural and historic objects and
25 wildlife and to provide for the enjoyment of these resources unimpaired for current and future
26 generations.
27 The Wilderness Act of 1964 asserts wilderness areas will be administered in such a
28 manner as to leave them unimpaired and preserve them for the enjoyment of future generations.
29 NFS Management Policies (2006) guide all NFS actions including natural resources
30 management. In general, the NFS Management Policies reiterate the NFS Organic Act's
31 mandate to manage the resources "unimpaired."
September, 2010 3-6 Do Not Quote or Cite
-------
1 3.1.2.3 U.S. Fish and Wildlife Service
2 On endangered species, Title 16 USC Chapter 35 Section 1531 states "The Congress
3 finds and declares that— these species offish, wildlife , and plants are of esthetic , ecological,
4 educational, historical, recreational, and scientific value to the Nation and its people and that all
5 Federal departments and agencies will use their authorities to conserve threatened and
6 endangered species.
7 The United States Fish and Wildlife Service (FWS) manages the National Wildlife
8 Refuge System lands to "...ensure that the biological integrity, diversity, and environmental
9 health of the Systems are maintained for the benefit of present and future generations of
10 Americans." 16 U.S.C. Section 668dd(a)(4)(B)(1997).
11
12 3.1.2.4 U.S. Forest Service
13 The National Forest units are managed consistent with Land and Resource Management
14 Plans (LRMPs) under the provisions of the National Forest Management Act (NFMA). 16
15 §U.S.C. 1604 (1997). LRMPs are, in part, specifically based on recognition that the National
16 Forests are ecosystems and their management for goods and services requires an awareness and
17 consideration of the interrelationships among plants, animals, soil, water, air, and other
18 environmental factors within such ecosystems. 36 C.F.R. §219.1(b)(3)
19 Any measures addressing Air Quality Related Values (AQRV) on National Forest
20 System lands will be implemented through, and be consistent with, the provisions of an
21 applicable LRMP or its revision (16 U.S.C. §1604(i)). Additionally, the Secretary of Agriculture
22 must prepare a Renewable Resource Program that recognizes the need to protect and, if
23 necessary, improve the quality of air resources. 16 U.S.C. §1602(5)(C).
24 AQRVs in Wilderness areas may receive further protection by the previously mentioned
25 1964 Wilderness Act. For Wilderness Areas in the National Forest System, the Act's
26 implementing regulations are found at 36 C.F.R. §293 requiring these Wilderness Areas be
27 administered to preserve and protect [their] wilderness character.
28
29 3.1.25Chesapeake Bay Total Maximum Daily Loads
30 Under section 303(d) of the Clean Water Act, states, territories, and authorized tribes are
31 required to develop lists of impaired waters. These are waters that are too polluted or otherwise
September, 2010 3-7 Do Not Quote or Cite
-------
1 degraded to meet the water quality standards set by states, territories, or authorized tribes. The
2 law requires that these jurisdictions establish priority rankings for waters on the lists and develop
3 TMDLs for these waters. A Total Maximum Daily Load, or TMDL, is a calculation of the
4 maximum amount of a pollutant that a waterbody can receive and still safely meet water quality
5 standards. EPA is developing a TMDL for the Chesapeake Bay and its tributaries. The
6 Chesapeake Bay Program has modeled the level of nitrogen that can reach the Bay and still meet
7 the Bay's water quality standards. The TMDL, with full public participation, will set waste load
8 allocations for point source discharges and load allocations for nonpoint sources of nitrogen. Air
9 deposition to the Bay and its watershed, as a source category, will have a specific allocation.
10 The allocation can be used to calculate the level of ambient air concentrations of reactive
11 nitrogen that are likely to meet the deposition allocation. To find the NOy portion of the
12 allocation one would subtract the reduced forms from the total allocation. If the total load to the
13 Bay of nitrogen from all the allocated source categories remains below the allocations, then the
14 Bay is expected to meet the water quality standards, which are set to protect the designated uses
15 of the Bay. Since the designated uses are set by the states with public input, not meeting the
16 designated uses can be seen as having an adverse effect to public welfare.
17
18 3.1.2.6 Critical Loads
19 The term critical load is used to describe the threshold of air pollution deposition that
20 causes a specified level of harm to sensitive resources in an ecosystem. A critical load is
21 technically defined as "the quantitative estimate of an exposure to one or more pollutants below
22 which significant harmful effects on specified sensitive elements of the environment are not
23 expected to occur according to present knowledge" (Nilsson and Grennfelt, 1988). The
24 determination of when a harmful effect becomes "significant" may be in the view of a researcher
25 or through a policy development process. Researchers often use the term "critical loads" to
26 describe when particular detrimental effects are realized, as is the case in Figure 2-1.
27 Harmful effects due to acidification have been defined here as those that occur below a
28 given ANC for aquatic systems and below a given Be: Al ratio for terrestrial systems. However,
29 the level at which an effect becomes adverse to public welfare is determined by the
30 Administrator, informed by available scientific information.
31
September, 2010 3-8 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
3.1.2.7 United Nations Economic Commission for Europe (UNECE)
In many European countries a critical loads framework is used to determine a level of
damages to ecosystem services from pollution that is legally allowed. These critical loads are
determined through a policy process. Indeed critical loads have been modeled by individual
countries and submitted to the UNECE (in cases where countries have not submitted their own
critical loads those loads have been calculated for them) and are being used to support
international emissions reduction agreements including the 1999 Gothenburg protocol and the
National Emission Ceiling Directive of the European Commission. Figure 3-2 shows critical
loads for eutrophication and acidification that protect 95% of forests, seminatural vegetation or
surface waters in Europe. For comparison to the U.S maps presented in this document the units
of deposition convert to a range of 3.2 kg ha'V1 to > 24 kg ha'V1'
CLnut(N) (5th percentile }
AII ecosystem s CLmax(S) (5th percentile )
eq ha"'a"'
• <200
H 200 - 400
D400-700
D 700 -1000
• 1000-1500
D>1500
All ecosystem
eq ha"'a"'
m<2QQ
D 200 - 400
D400-700
D700-1000
n1000-isoo
n> 1500
Figure 3-2 European maps of eutrophication (left) and acidification (right) which
protect 95% of natural areas in 50x50 km2 European Monitoring and
Evaluation Programme grid. Red shaded areas illustrate grid cells where
deposition needs to be lower than 200 eq ha"1 a"1 to reach this protection target.
Source: Critical Load, Dynamic Modelling and Impact Assessment in Europe CCE Status Report 2008 available at
http://www.pbl.nl/en/publications/2009/Critical-load-dynamic-modelling-and-impact-assessment-in-
Europe-CCE-Status-Report-2008.html
The Coordination Center for Effects, a working center for the Working Group on Effects
of the Convention on Long Range Transboundary Pollution, in the 2008 status report shows
calculated critical loads based on an ANC target of 20 jieq/L for the protection of brown trout.
September, 2010
$-9
Do Not Quote or Cite
-------
1 Individual countries have also set ANC targets for other species for example Norway uses a
2 critical limit of 30|ieq/L for Atlantic salmon (Jenkins et al, 2003).
3
4 3.2 What are ecosystem services and how does this concept relate to public welfare?
5 An additional concept that may be useful in considering the issue of adversity to public
6 welfare is ecosystem services. In the next section the concept of ecosystem services, its
7 relationship to adversity and public welfare within the context of this review are explained.
8 Ecosystem services can be generally defined as the benefits individuals and organizations
9 obtain from ecosystems. Ecosystem services can be classified as provisioning (food and water),
10 regulating (control of climate and disease), cultural (recreational), and supporting (nutrient
11 cycling) (MEA 2005). Conceptually, changes in ecosystem services may be used to aid in
12 characterizing a known or anticipated adverse effect to public welfare. In the context of this
13 review, ecosystem services may also aid in assessing the magnitude and significance to the
14 public of a resource and in assessing how NOy and SOx concentrations and deposition may
15 impact that resource. The relationship between ecosystem services and public welfare effects is
16 illustrated in Table 3-1.
September, 2010 3-10 Do Not Quote or Cite
-------
1 Table 3-1. Crosswalk between Ecosystem Services and Public Welfare Effects
Public Welfare Effect Ecosystem Service Service Category
Soils
Water
Crops
Vegetation
Wildlife
Climate
*Personal Comfort and
Wellbeing
Nutrient Cycling
Drinking water, Recreation,
Aesthetic
Food, Fuel Production, Forest
Products
Food, Recreation, Aesthetic,
Nonuse
Recreation, Food, Nonuse
Climate Control
Supporting
Provisioning, Cultural
Provisioning
Provisioning, Cultural
Cultural, Provisioning
Regulating
2
O
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
*A11 ecosystem services contribute to personal comfort and wellbeing.
EPA has defined ecological goods and services for the purposes of a Regulatory Impact
Analysis as the "outputs of ecological functions or processes that directly or indirectly contribute
to social welfare or have the potential to do so in the future. Some outputs may be bought and
sold, but most are not marketed" (US EPA 2006). Additionally Executive Order 12866 requires
a regulatory Impact Analysis for any rule considered "economically significant" and defines
significant as a rule having $100 million or more in impacts. Though this is not a definition
specifically for use in the NAAQS process it may be a useful one in considering the scope of
ecosystem services and the effects of air pollutants upon those services. Especially important is
the acknowledgement that it is difficult to measure and/or monetize the goods and services
supplied by ecosystems. Valuing ecological benefits, or the contributions to social welfare
derived from ecosystems, can be challenging as noted in EPA's Ecological Benefits Assessment
Strategic Plan (US EPA 2006) and the Science Advisory Board report "Valuing the Protection of
Ecological Systems and Services" (US EPA, 2009). It can be informative in characterizing
adversity to public welfare to attempt to place an economic valuation on the set of goods and
September, 2010
3-11
Do Not Quote or Cite
-------
1
2
3
4
5
services that have been identified with respect to a change in policy however it must be noted
that this valuation will be
incomplete and illustrative only. The stepwise concept leading to the
valuation of ecosystem services is graphically depicted in Figure 3-3.
^m
^m
^
EPA action
6
7
8
9
*
^H
^^^^^^^
Ecosyst
ems
—
;
^^M
Ecological goods
and services
affected by the policy
Planning and problem formulation
in
1 \J
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
|
^
^N
Goods and services
identified
V
V
Ecological analysis
Goods and servi:
quantified
;es
^V
^•»
L
^-~™
•-— ^
Goods and
^_ J> sen/ices not
identified
. Identified
goods and
services not
quantified
Economic analysis
Goods and
services
monetized
\ ^
\^
ti
V.
jf
X^
^
_Ts
~~l/
Quantified
.
fc. goods and
services not
monetized
Figure 3-3. Representation of the benefits assessment process indicating where some
ecological
benefits may remain
unrecognized, unquantified, or unmonetized.
(Source: EBASP USEPA 2006).
A conceptual model integrating the role of ecosystem services in characterizing known or
anticipated adverse effects to public welfare
CAA, the secondary standard is to specify a
welfare. For this review, the relevant air
is shown in Figure 3-3. Under Section 109 of the
level of air quality that is requisite to protect public
quality
indicator is interpreted as ambient NOy and SO
X
concentrations that can be linked to levels of deposition for which there are ecological effects
that are adverse to public
welfare. The case study analyses (described in Chapters 4 and 5 of the
REA and summarized in Chapter 2 of this document) link deposition in sensitive ecosystems
(e.g., the exposure pathway) to changes in a given ecological indicator (e.g., for aquatic
acidification, changes in acid neutralizing capacity [ANC]) and then to changes in ecosystems
September, 2010
3-12
Do Not Quote or Cite
-------
1 and the services they provide (e.g., fish species richness and its influence on recreational
2 fishing). To the extent possible for each targeted effect area, ambient concentrations of nitrogen
3 and sulfur (i.e., ambient air quality indicators) were linked to deposition in sensitive ecosystems
4 (i.e., exposure pathways), and then deposition was linked to system response as measured by a
5 given ecological indicator (e.g., lake and stream acidification as measured by ANC). The
6 ecological effect (e.g., changes in fish species richness) was then, where possible, associated
7 with changes in ecosystem services and their public welfare effects (e.g., recreational fishing).
8 We recognize that there is a certain amount of natural change in ecosystems over time that can
9 affect the level of acidity and the response of the ecosystem to additional acid and nutrient
10 inputs. However, this review is focused on the impact of anthropogenic nitrogen and sulfur
11 deposition given the existing state of non-anthropogenically determined ecosystem
12 characteristics, and as such we essentially hold these other factors as "fixed" for the purposes of
13 the review.
14 Knowledge about the relationships linking ambient concentrations and ecosystem
15 services can be used to inform a policy judgment on a known or anticipated adverse public
16 welfare effect. The conceptual model outlined for aquatic acidification in Figure 3-3 can be
17 modified for any targeted effect area where sufficient data and models are available. For
18 example, a change in an ecosystem structure and process, such as foliar injury, would be
19 classified as an ecological effect, with the associated changes in ecosystem services, such as
20 primary productivity, food availability, forest products, and aesthetics (e.g., scenic viewing),
21 classified as public welfare effects. Additionally, changes in biodiversity would be classified as
22 an ecological effect, and the associated changes in ecosystem services—productivity,
23 recreational viewing and aesthetics—would be classified as public welfare effects. This
24 information can then be used by the Administrator to determine whether or not the changes
25 described are adverse to public welfare. In subsequent sections these concepts are applied to
26 characterize the ecosystem services potentially affected by nitrogen and/or sulfur for each of the
27 effect areas assessed in the REA.
28
September, 2010 3-13 Do Not Quote or Cite
-------
Ambient Air Quality
Indicator
Exposure Pathway
Affected Ecosystem
Ecological Response
(ecological indicator )
NOX/SOX
Concentrations
Atmospheric
Deposition
N&S
an
Aquatic
I
Acidification
(lake/stream ANC )
1
2
3
4
5
6
7
9
10
11
12
13
Ecological Effect
Ecological Benefit/
Welfare Effect
Change in Ecosystem
Structure & Processes
(fish species richness )
Change in
Ecosystem Services
(recreational fishing
;es
ng )
Figure 3-4. Conceptual model showing the relationships among ambient air quality
indicators and exposure pathways and the resulting impacts on ecosystems,
ecological responses, effects and benefits to characterize known or
anticipated adverse effects to public welfare
These concepts can also be applied to the programs described in section 3.1. National
parks represent areas of nationally recognized ecological and public welfare significance, which
are afforded a higher level of protection. Therefore, staff has also focused on air quality and
deposition in the subset of national park sites and important natural areas. The spatial
relationships between sensitive regions, Class 1 areas and nitrogen deposition levels are
illustrated in Figures 3-5 and 3-6. Please note that the scale of deposition levels is different for
the two maps to allow greater differentiation of the deposition in the western U.S.
September, 2010
3-14
Do Not Quote or Cite
-------
Weight in grams of elemental nitrogen and sulfur were
extracted from the various species reported by both
CMAQ (dry) and NADP (wet) by dividing by the species'
molecular weight. These values were then converted
to equivalents of either sulfur or nitrogen and combined.
The source of the federal and state public recreation
areas is the Conservation Biology Institute. The lands
were filtered to remove non-recreation areas as well
as those less than 64,000 acres.
Combined N and S
eq/ha/yr
H 1,219-2,652
| 1,054-1,218
| | 926- 1,053
| | 835 - 925
^] 753 - 834
607 - 752
^B 322 - 606
250
500
1
2
3
Figure 3-5. Locations of Eastern U.S. Public Lands relative to deposition of nitrogen and
sulfur in sensitive aquatic areas. Source 2005 CMAQ and NADP.
September, 2010
3-15
Do Not Quote or Cite
-------
1
2
3
4
5
Combined N and S
eq/ha/yr
^B 300-1,337
250 - 299
220 - 249
195-219
175- 194
140 - 174
69- 139
Weight in grams of elemental nitrogen and sulfur were
extracted from the various species reported by both
CMAQ (dry) and NADP (wet) by dividing by the species
molecular weight. These values were then converted
to equivalents of either sulfur or nitrogen and combined
The source of the federal and state public recreation
areas is the Conservation Biology Institute. The lands
were filtered to remove non-recreation areas as well
as those less than 64.000 acres.
Figure 3-6. Location of Western U.S. Public Lands relative to deposition of nitrogen and
sulfur. Source 2005 CMAQ and NADP.
September, 2010
3-16
Do Not Quote or Cite
-------
1 3.3 Applying Economic Valuation to Ecosystem Services
2 As discussed earlier in this document, a secondary NAAQS is required to be set at the
3 "level(s) of air quality necessary to protect the public welfare from any known or anticipated
4 adverse effects". As part of the effort to determine the standard, EPA linked the changes in the
5 ambient air concentrations of NOy and SOx to the changes in ecosystem services and ultimately
6 to changes in public welfare (U.S. EPA, 2009). The difficulty in the monetization for ecosystem
7 services has been previously emphasized. This difficulty necessitates focusing on a subset of
8 services for economic valuation. And although economics on its own cannot determine what
9 level of impact on public welfare is "adverse," economics can be helpful in the context of a
10 secondary NAAQS for determining the degree to which improvements are beneficial to public
11 welfare and illustrating and aggregating those impacts.l
12
13 3.3.1 Ecosystem Services and Links to Public Welfare
14 An ecosystem service framework provides a structure to measure changes in public
15 welfare from changes in ecosystem functions affected by air pollution. EPA's Risk Assessment
16 for this rulemaking defines ecosystem services as "the ecological processes or functions having
17 monetary or nonmonetary value to individuals or society at large" (EPA 2009) The discipline of
18 economics provides a useful approach for summarizing how the public values changes in the
19 services provided by the environment. An ecosystem services framework (with or without
20 valuation) can provide measures of changes in public welfare.
21
22 3.3.2 Economics as a Framework to Illustrate Changes in Public Welfare
23 Economics can provide a framework to illustrate how public welfare changes in response
24 to changes in environmental quality by quantitatively linking changes in ecosystem services to
25 preferences. Economics assumes that the choices that individuals make reflect their preferences
26 over certain outcomes and that, generally speaking, they will make choices that, in expectation,
27 will make them as well off as possible given their resources. In economics revealed and stated
1 Section 109 of the Clean Air Act forbids consideration of the compliance costs of reducing pollution when setting
a NAAQS. However, there is no prohibition regarding the consideration of the monetized impacts of welfare effects
occurring due to levels of pollution above alternative standards in evaluating the adversity of the impacts to public
welfare. Ecosystem services can be characterized as a method of monetizing the impacts of the air pollution.
Although a separate regulatory document quantifying the costs and benefits of attaining a NAAQS is prepared
simultaneously, this document is not considered when selecting a standard.
September, 2010 3-17 Do Not Quote or Cite
-------
1 preference methods are used to observe the choices individuals make to understand the outcomes
2 individuals prefer. What individuals are willing to give up for an outcome is their willingness-to-
3 pay (WTP) for that outcome. An example of an outcome is an improvement in an ecosystem
4 service. Often, to provide comparability to other goods and services, in economics these
5 tradeoffs are framed relative to dollars for convenience.
6 Economics could inform the Administrator by valuing and characterizing the changes in
7 public welfare from changes in the quantity and quality of ecosystem services. Overall, this
8 assessment intends to characterize changes in ecosystem services from a scientific perspective
9 using effects on ecosystem structures and functions or ecosystem integrity. Economics then
10 estimates the effect on public welfare of these changes in the quantity and quality of ecosystem
11 services using willingness to pay as a measure of this effect. For example, a decrease in a
12 particular bird species can be characterized by its effect on the ecosystem's structure and
13 function, while from an economic perspective, the effects would be based on the impact on
14 public welfare or the value the public places on that species. A simple example is a comparison
15 between a decrease in a bird species that is relatively unknown compared to a decrease in a very
16 prominent species (e.g. bald eagle). The public is likely to have a higher WTP to avoid the latter,
17 and thus the decrease would affect the public welfare more, even if the changes in the two bird
18 species generally have the same impact on an ecosystem's structure or function.
19 There are important complications with using preferences to understand the effect of
20 pollution on public welfare. For example, while the field of economics generally assumes that
21 public preferences are the paramount consideration; care must be taken that these preferences
22 may change when the public receives new information. Evaluation of public preferences should
23 take place under conditions of full information. If individuals do not understand how pollution
24 will affect ecosystem services, or even how those ecosystem services affect their quality of life,
25 then they will have a difficult time valuing changes in those services. Similarly, it may be very
26 difficult and time-consuming for individuals to learn and understand how changes in particular
27 ecosystem services may affect them, in part because typically there are significant
28 interdependences within an ecosystem. Because of this complexity, individuals may implicitly
29 value a species, or habitat, or ecosystem function because it supports an ecosystem service that
30 they do clearly value. Furthermore, the public also has limited understanding regarding
31 irreversibilities, tipping points, and other more complex aspects of ecosystems, which limits the
September, 2010 3-18 Do Not Quote or Cite
-------
1 ability to adequately value these ecosystems.2 In addition, where and when a change in an
2 ecosystem takes places is crucial for characterizing the associated change in an ecosystem
3 service, and will also affect the value the public places on that change.
4
5 3.3.3 The Role of Economics in Defining "Adversity"
6 If economic valuation can establish a significant loss to public welfare, then this can
7 provide strong support for a determination of an "adverse " effect. However, there is neither an
8 economic definition of how much loss in public welfare is adverse nor an economic definition of
9 adversity. While an economist might consider a particular scenario adverse because it might
10 imply some harm or potential for improvement, there is no specific threshold level when a loss in
1 1 welfare (e.g. loss in dollars) becomes adverse. An individual might be willing to give up some
12 of their resources to avoid a threat or negative outcome (i.e., willing to pay to avoid a particular
13 outcome). According to economic theory, if an individual is willing to give up something to
14 avoid the outcome, then imposing the outcome on the individual must make them worse off, at
15 which point an economist might describe the outcome as adverse. However, the amount an
16 individual is willing to pay to avoid the outcome may or may not rise to the level of harm that the
17 Administrator interprets as "adverse" to public welfare. At the same time, an economic
1 8 valuation that shows that there are substantial damages from current levels of acidification or
19 nutrient enrichment would provide strong evidence for finding that current impacts are adverse
20 to public welfare. In summary, economic analysis (particularly valuation) can provide useful
21 information for the Administrator as to the interpretation of the word "adverse" in the context of
22 public welfare, but it does not provide a complete set of information needed to make that
23 determination.
24
25 3.3.44 Collective Action as an Indicator of Public Preferences
26 Typically economics uses information on willingness to pay for improved environmental
27 quality that is gathered from observing individuals' market behavior (revealed preference) or that
28 they provide through surveys (i.e., stated preference methods). The analyses in the following
2 While the public may not fully appreciate the interdependencies within ecosystems, they can learn them, but again
it may be costly to do so. It is possible for individuals to value outcomes that are irreversible or result in discrete
changes (i.e., tipping points) in the quality and quantity of ecosystem services. Avoiding irreversible outcomes
should be and are more valued by individuals than outcomes that are not irreversible (Arrow and Fischer, 1974).
September, 2010 3-19 Do Not Quote or Cite
-------
1 sections use revealed and stated-preference information to quantify a portion of the social costs
2 of current levels of acidification and nutrient enrichment. However, the studies supporting these
3 analyses evoke specific contexts and thus the findings may not be generalizable across all of
4 those affected by acidification or nutrient enrichment.3 An alternative source of revealed
5 information on the damages caused by acidification can be found in the behavior of groups.
6 Often groups collectively make choices to engage in activities that improve the collective
7 welfare of the group. For example, a community around an acidified lake might undertake
8 activities designed to improve the quality of that lake, including purchasing lime, to use as a tool
9 to reduce the acidity of the lake. These collective decisions can be used to gain insights into how
10 people value improvements to ecosystem services. However, there are many obstacles to
11 collective actions, including problems of organization, free ridership and others (Olson 1965)
12 that make it difficult to use the actions of organizations to interpret individuals' preferences.
13 In addition to communities, states may also take actions to increase the quality of their
14 impaired lakes. Non-Governmental Organizations (NGOs) or advocacy groups, as well may
15 organize support for, and/or directly undertake, activities to improve lake and stream quality on
16 behalf of its members/donors. How individual's preferences are expressed through these
17 collective actions is discussed below. For brevity, this discussion will focus on collective efforts
18 to reduce acidity of lakes and streams by Communities, Nongovernmental Organizations and
19 States.
20
21 3.3.4.1 Communities
22 In cases where property rights to a resource are well defined, collective action is more
23 likely to take place, as individuals have greater ownership and control over the affected resource.
24 Rights to use a lake, as well as, mandatory membership in a lake association is often written right
25 into the deed of properties which abut or surround a lake, giving these property owners more
26 control over the resource. This mechanism of granting rights and responsibilities to the lake
27 encourages better management of the lake resource by remedying unrestricted access and free
28 rider problems. This coupling of the costs of resource improvements with their benefits
29 encourages individuals to maintain the quality of the resource.
3 Even in the case where the existing studies provide a reliable characterization of the effects of acidification or
nutrient enrichment on a limited number of individuals, it is advisable to make use of corroborating data and studies
when such information is available.
September, 2010 3-20 Do Not Quote or Cite
-------
1 There have been several documented instances where communities (particularly
2 Homeowners Associations) have spent time and money to improve the quality of a lake. These
3 include actions to combat acidity, eutrophication, invasive species (e.g. Zebra Mussels) and other
4 problems. The Lake Wononscopomuc Association in Salisbury, Connecticut is a typical
5 example (Mayland 2009.) They spend their own funds to hire scientific consultants to survey
6 and test the lake water (for e.g. pH., dissolved oxygen, visibility, and many other factors related
7 to the lake's condition) and recommend management strategies to improve the quality of the
8 lake. Likewise, in Georgia, the Berkeley Lake Homeowners Association (BLHA) is a non-profit
9 homeowner association dedicated to protecting Berkeley Lake. BLHA is typical of many other
10 home owners associations with access to a lake, in that they are also concerned with and
11 managing acidity, eutrophication, invasive species and a whole host of more mundane upkeep
12 and maintenance issues (Hunkapiller 2006.) BLHA recognizes the relationship between lake
13 acidity and resident's enjoyment of the lake's fishing swimming and aesthetics.
14
15 3.3.4.2 Nongovernmental Organizations
16 Nongovernmental Organizations (NGOs) or Advocacy Groups organize individuals and
17 smaller groups thereby reducing the transaction costs associated with individual's desires to
18 advance a specific goal. For example, Living Lakes, Inc. (LLI) is a not-for profit organization
19 which sponsors an applied aquatic resources restoration demonstration program for acidified
20 waters. In the late 1980's LLI began evaluation of seven different liming technologies on 22
21 lakes and 10 streams in 6 states. Lakes and streams have been treated in the states of
22 Massachusetts, Rhode Island, New York, Pennsylvania, Maryland and West Virginia (Brocksen
23 and Emler 1988.) Likewise, sportsman groups such as Trout Unlimited, as well as, smaller local
24 groups, have an interest in improving or maintaining the quality of lakes and streams. Trout
25 Unlimited is well known for these activities and is discussed further later on. However, several
26 smaller, localized groups also work to decrease aquatic acidification. One of these is the
27 Mosquito Creek Sportsman's Association in Pennsylvania. Mosquito Creek and its main
28 tributary Gifford Run were once famous for naturally reproducing wild brook and brown trout.
29 However, since the early 1960's, the pH of the stream steadily declined due to acid rain. As a
30 result, wild brook trout, as well as, wild brown trout have substantially declined in the watershed
31 (Hoover and Rightnour 2002.) Aerial liming was undertaken as part of an overall watershed
32 restoration program that included constructed wetlands, forest liming, and in-stream liming to
September, 2010 3-21 Do Not Quote or Cite
-------
1 improve this fishery and provide increased opportunities for public recreation in the region.
2 Fifty tons of lime were applied in the headwaters of Mosquito Creek Watershed. This liming
3 project was part of the Mosquito Creek Sportsman's Association's efforts to improve the water
4 quality of the 90 square mile watershed located in Clearfield and Elk counties. However, while
5 the project first phase and the other ongoing phases of the overall restoration project have been
6 initiated by the Mosquito Creek Sportsman Association, they received technical support from
7 multiple public, private and other non-profit groups.4 "A benefit/cost analysis was prepared on
8 the four implementation phases of this project. Costs were based on alkaline deficiencies and the
9 additional costs determined for the technologies. Benefits were estimated as returns on direct
10 recreational use losses and community willingness-to-pay. Full restoration of the watershed is
11 estimated to cost approximately $3.4 million over 15 years, for an annualized cost of $229,000,
12 or $5,400 per mile per year for 42 miles of potential improvements. Expected returns range from
13 $1.2 million per year for recreational use to $6.1 million per year for total community
14 willingness-to-pay. It was concluded that restoration is technically feasible and economically
15 beneficial for the Mosquito Creek watershed, and it is recommended that planned projects and
16 the remainder of the progressive restoration plan be implemented." (Hoover and Rightnour
17 2002.)
18
19 3.3.4.3 States
20 Several states including: Vermont, New Hampshire, New York and Tennessee have
21 developed Total Maximum Daily Loads (TMDLs) for lakes impaired for acidification in their
22 jurisdictions. As mentioned in the previous section regarding the TMDL for the Chesapeake Bay
23 the applicable water quality standards and designated uses are set by the states with public
24 participation. Although most states set their standard either by legislation or regulations in at
25 least one case, specifically New York, the designated uses and water quality standard are part of
26 the state constitution. The Adirondack Forest Preserve is required to be "forever kept as wild
27 forest lands". New York has interpreted this to mean that the waters included in the preserve are
These included: the Pennsylvania State University Environmental Resource Research Institute, Pennsylvania
Game Commission, Pennsylvania DCNR Bureau of Forestry, Pennsylvania Fish and Boat Commission,
Pennsylvania Department of Environmental Protection, Pennsylvania Department of Corrections Quehanna Boot
Camp, Wood Duck Chapter Trout Unlimited, Canaan Valley Institute, Clearfield County Conservation District and
Pennsylvania USDA Natural Resource Conservation Service.
September, 2010 3-22 Do Not Quote or Cite
-------
1 required to be kept in natural conditions. To this end New York has chosen to set a tiered TMDL
2 that allows interim water quality targets in order to address the reality that some lakes in the
3 Adirondacks will naturally have a pH that does not meet the state's water quality standards. For
4 lakes that can meet the standards the state has chosen to set the water quality standard for pH is
5 above 6.5. New Hampshire has chosen to set their water quality target at an ANC of 60|ieq/L
6 that, according to the TMDL document, corresponds to a pH of 6.5. Vermont, in a similar
7 process has chosen a target ANC of 50|ieq/L. In Tennessee the state faces a similar problem as
8 New York in trying to set levels to protect streams Great Smoky Mountains National Park which
9 include some naturally acidic streams. Accordingly they have set site specific ANC targets
10 where data is available to do so and chosen to target an ANC of 50|ieq/L as a default value
11 where data is not available. The Tennessee TMDL is a partnership between the state and the
12 National Park Service which is sharing the data collection and modeling activities with academic
13 institutions.
14 In each case the implementation sections of these TMDLs cites the fact that the sources
15 of pollution responsible for the degradation of water quality in the named lakes and streams are
16 not located within the jurisdiction of the state. Each state has called on EPA to require regional
17 or national decreases in acidifying deposition. Vermont goes so far as to say "In short,
18 implementation of this TMDL is primarily the responsibility of EPA.... This TMDL sets out
19 clear endpoints to guide EPA actions. However, in the absence of vigorous efforts by EPA to
20 bring about reductions in acid emissions from out-of-state sources this TMDL will merely have
21 been a paper exercise."
22
23 3.3.4.4 Public-Private Collaborations
24 In some cases, industry, government and private efforts partner to reduce the acidity of a
25 lake or stream. In one such instance in 2005, The U.S. Forest Service used helicopters to apply
26 200 tons of limestone sand into the St. Marys River and its tributaries to lower acid levels in one
27 of Virginia's prime trout fisheries to mitigate the impacts of acidification until a long-term
28 solution to acidification is found. The NGO, Trout Unlimited was one of the partners in the
29 liming project, while Dominion Virginia Power provided $10,000 for the liming project
30 (Associated Press 2005.) In another partnership, Living Lakes participated in a project in the
31 Woods Lake Watershed in the Adirondack region of the state of New York that was co-
September, 2010 3-23 Do Not Quote or Cite
-------
1 sponsored by the Electric Power Research Institute (EPRI.) There are many such examples,
2 where all three of these types of groups partner on the same project.
3
4 3.3.4.5 Using Information on Collective Actions to Estimate Welfare Impacts
5 The fact that collective action activities are being undertaken by communities, NGO's
6 and States underscores the fact that there is a societal demand for further improvement to the
7 quality of many water bodies which have been impaired by acidic deposition.5 However, they
8 provide insufficient quantitative evidence as to what society willingness to pay to reduce lake
9 and stream acidity because it is difficult to separately identify individual preferences from the
10 actions of the group. For example, groups suffer from the problem of free-ridership. When there
11 are free-riders we know that the activity of the group understates the preferences of the
12 community for an improvement in environmental quality, however it is difficult to estimate the
13 magnitude of its impact on the total activity of the group.
14
15 3.4 Effects of Acidification and Nutrient Enrichment on Ecosystem Services
16 The process used to link ecological indicators to ecosystem services is discussed
17 extensively in Appendix 8 of the REA. In brief, for each case study area assessed the ecological
18 indicators were linked to an ecological response that was subsequently linked, to the extent
19 possible, to associated services. For example in the case study for aquatic acidification the
20 chosen ecological indicator is ANC which can be linked to the ecosystem service of recreational
21 fishing as illustrated in the conceptual model shown in Figure 3-7. Although recreational fishing
22 losses are the only service effects that can be independently quantified or monetized at this time,
23 there are numerous other ecosystem services that may be related to the ecological effects of
24 acidification.
However, one must recognize that often times reducing acidity is often part of a larger effort to generally improve
the quality of a water body. Therefore, separating out the portion of people's desire to just to reduce acidity from
the more general improvements is difficult.
September, 2010 3-24 Do Not Quote or Cite
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Acidifying Inputs
to Surface Water
N+S Deposition
Ecological
Symptoms
impacts on ecosystem tndooints
Af'ectec En- systeri Services
Acidific;
Leat
Surta
hate to
ce Water
Surf ace Water
Acidification:
Low oH and
AiVC
i
Mobilization of
Alum num:
Elevated Af
Declines in
Aquatic Biota
Fitness:
Reduced
growth,
development,
and
reproduction
— »
Decl
Aquat
Wee
S/x
Atxm
Divers
Rid
nes in
c Biota:
uced
jcies
dance,
ity, and
wess
Decl nes in
Terrestrial
Nearshore
Biota
^^^^H
*
Provisioning Services
•production for commercial
and subsistence fishing
Cultural Services
•recreational fishing
•waterfowl hunting
•jest"s:i: en c yne-'t
•no nuse services
•biological control
Figure 3-7. Conceptual model linking ecological indicator (ANC) to affected ecosystem
services The red arrows highlight the path to monetization of recreational
fishing effects.
While aquatic acidification is the focus of this policy assessment, the other effect areas
analyzed in the REA still merit some discussion in view of the fact that these ecosystems are
being harmed by nitrogen and sulfur deposition and will obtain some measure of protection with
any decrease in that deposition regardless of the reason for the decrease. In next four sections we
summarize the current levels of specific ecosystem services for aquatic and terrestrial
acidification, and aquatic and terrestrial nutrient enrichment. We also present results of analyses
that have attempted to quantify and monetize the harms to public welfare, as represented by
ecosystem services, due to nitrogen and sulfur deposition.
For the purposes of the following sections nutrient enrichment refers only to that due to
NOy deposition. Additionally these sections focus on the detrimental effects of that deposition.
Staff acknowledges that a certain amount of NOy deposition in managed terrestrial ecosystems
may have a beneficial effect, specifically increased growth (a fertilization effect). However no
attempt has been made to quantify those beneficial effects since this document and preceding
analyses are focused on unmanaged sensitive ecosystems.
September, 2010
3-25
Do Not Quote or Cite
-------
1 3.4.1 Evidence for Adversity Related to Aquatic Acidification
2 Acidification primarily affects the ecosystem services that are derived from the fish and
3 other aquatic life found in these surface waters (REA, Section 5.2.1.3). In the northeastern
4 United States, the surface waters affected by acidification are not a major source of
5 commercially raised or caught fish; however, they are a source of food for some recreational and
6 subsistence fishers and for other consumers. Although data and models are available for
7 examining the effects on recreational fishing, relatively little data are available for measuring the
8 effects on subsistence and other consumers. For example, although there is evidence that certain
9 population subgroups in the northeastern United States, such as the Hmong and Chippewa ethnic
10 groups, have particularly high rates of self-caught fish consumption (Hutchison and Kraft, 1994;
11 Peterson et al., 1994), it is not known if and how their consumption patterns are affected by the
12 reductions in available fish populations caused by surface water acidification.
13 Inland surface waters support several cultural services, such as recreational fishing,
14 aesthetic and educational services; however, Banzhaf et al (2006) has shown that non-use
15 services are arguably a significant source of benefits from reduced acidification. The areas of the
16 country containing the most sensitive lakes and streams are New England, the Adirondack
17 Mountains, the Appalachian Mountains (northern Appalachian Plateau and Ridge/Blue Ridge
18 region), northern Florida, and the Upper Midwest. Within the Adirondack Mountains
19 approximately 8% of the lakes were considered acidic and in the northern Appalachian Plateau
20 and Ridge/Blue Ridge 6 - 8% of the streams (ISA 3.2.4.2 and REA 4.2.2). Recreational fishing
21 in lakes and streams is among the most popular outdoor recreational activities in the northeastern
22 United States. Data from the 2006 National Survey of Fishing, Hunting, and Wildlife Associated
23 Recreation (FHWAR) indicate that more than 9% of adults in this part of the country participate
24 annually in freshwater fishing with 140 million freshwater fishing days. Based on studies
25 conducted in the northeastern United States, Kaval and Loomis (2003) estimated average
26 consumer surplus values per day of $35 for recreational fishing (in 2007 dollars). Therefore, the
27 implied total annual value of freshwater fishing in the northeastern United States was $5 billion
28 in 2006. We recognize that embedded in these numbers is a degree of harm to recreational
29 fishing services due to acidification that has occurred over time. These harms have not been
30 quantified on a regional scale. However given the magnitude of the resource, the length of time
31 nitrogen and sulfur have been affecting freshwaters in the northeast and the level of monetary
September, 2010 3-26 Do Not Quote or Cite
-------
1 damages calculated for the case study in the Adirondacks described in the next section we would
2 expect these damages to be significant.
3 In general, inland surface waters such as lakes, rivers, and streams provide a number of
4 regulating services, playing a role in hydrological regimes and climate regulation. There is little
5 evidence that acidification of freshwaters in the northeastern United States has significantly
6 degraded these specific services; however, freshwater ecosystems also provide biological control
7 services by providing environments that sustain delicate aquatic food chains.
8 The toxic effects of acidification on fish and other aquatic life impair these services by
9 disrupting the trophic structure of surface waters (Driscoll et al., 2001). Although it is difficult
10 to quantify these services and how they are affected by acidification, it is worth noting that some
11 of these services may be captured through measures of provisioning and cultural services. For
12 example, these biological control services may serve as "intermediate" inputs that support the
13 production of "final" recreational fishing and other cultural services.
14
15 3.4.1.1 What is the value of the impaired recreational fishing and other cultural services?
16 In the previous section we described the ecosystem services that are most likely to be
17 affected by N and S deposition and summarized evidence regarding the current magnitude and
18 values of recreational fishing services, the degree to which these services are impaired by
19 existing NOy/SOx levels has not been quantified. To address this limitation, the REA
20 (Appendix 8) provides insights into the magnitude of ecosystem service impairments. The
21 REA provides quantitative estimates of selected ecosystem services impairments or
22 enhancements for three main categories of ecosystem effects - aquatic acidification, terrestrial
23 acidification, and aquatic nutrient enrichment6. Within these three categories, the selection of
24 specific ecosystem services for more in-depth analysis depended primarily on the expected
25 magnitude of impairments and on the availability of appropriate data and modeling tools.
26 The analysis of ecosystem service impairments due to aquatic acidification builds on the
27 case study analysis of lakes in the New York Adirondacks. In this study estimates of changes in
28 recreational fishing services are determined, as well as changes more broadly in "cultural"
29 ecosystem services (including recreational, aesthetic, and nonuse services). First, the MAGIC
6 Estimates for terrestrial nutrient enrichments were not generated due to the limited availability of necessary data
and models for this effect category.
September, 2010 3-27 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
model (REA, Appendix 8, Sec 2.2) was applied to 44 lakes to predict what ANC levels would
be under both "business as usual" conditions (i.e., allowing for some decline in deposition due to
existing regulations) and pre-emission (i.e., background) conditions. These model runs assumed
a 2010 "zero-out" emissions scenario (where all N and S deposition is eliminated) with a
projected lag time between the elimination of emissions to observed improvement in ANC of 10
years thus benefits results were calculated for the year 2020. These predictions were then
extrapolated to the full universe of Adirondack lakes. Table 3-2 reports the number of
"impacted" lakes in each year, where impact means that the lake is predicted to be below the
ANC threshold under business-as-usual and above the threshold under pristine conditions.
Table 3-2. Count of Impacted Lakes
ANC Threshold
(in jieq/L)
20
20
20
20
50
50
50
50
100
100
100
100
Year
2005
2020
2050
2100
2005
2020
2050
2100
2005
2020
2050
2100
Lake Count
0
107
95
74
0
244
222
200
0
430
404
354
Note: There are 1,076 lakes in the "Adirondack Region".
Second, to estimate the recreational fishing impacts of aquatic acidification in these lakes,
an existing model of recreational fishing demand and site choice was applied. This model
predicts how recreational fishing patterns in the Adirondacks would differ and how much higher
the average annual value of recreational fishing services would be for New York residents if lake
ANC levels corresponded to background (rather than business as usual) conditions. Table 3-3
September, 2010
3-28
Do Not Quote or Cite
-------
1 summarizes the results and the present value of benefits and annualized benefits at 3 and 7%
2 discount rates.
Table 3-3. Present Value and Annualized Benefits, Adirondack Region
ANC
Threshold
(in (leq/L)
20
50
100
Present Value Benefits"
(in million of 2007 dollars)
3% Discount
Rate
$142.59
$285.15
$298.67
7% Discount
Rate
$60.05
$114.18
$120.61
Annualized Benefitsb
(in million of 2007 dollars)
3% Discount
Rate
$4.46
$8.91
$9.33
7% Discount
Rate
$3.94
$7.49
$7.91
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
a Annual benefits for 2010 to 2100 discounted to 2010.
b Present value benefits annualized over 2009-2100.
Current annual impairments are most likely of a similar magnitude because, although
current NOy/SOx levels are somewhat higher than those expected in 2020 (under business as
usual - given expected emissions controls associated with Title IV regulations but no additional
nitrogen or sulfur controls), and the affected NY population is also somewhat smaller (based on
U.S. Census Bureau projections).
To estimate impacts on a broader category of cultural (and some provisioning)
ecosystem services, results from the Banzhaf et al (2006) valuation survey of New York
residents were adapted and applied to this context. The survey used a contingent valuation
approach to estimate the average annual household WTP for future reductions (20% and 45%) in
the percent of Adirondack lakes impaired by acidification. The focus of the survey was on
impacts on aquatic resources. Pretesting of the survey indicated that respondents nonetheless
tended to assume that benefits would occur in the condition of birds and forests as well as in
recreational fishing. The survey that measured the benefits of 20% of the lakes improving
indicated that terrestrial benefits were minor and econometric controls were used to adjust the
willingness to pay estimate for those that suspected that terrestrial improvements were greater
than described in the survey. The survey that measured the benefits of improving 45% of the
total number of lakes also indicated that the benefits to forests and birds were significant. .
The WTP estimates from the two versions of the survey were then (1) scaled to reflect
predicted changes between business-as-usual and background conditions in 2020 (MAGIC lake
September, 2010
3-29
Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
modeling results indicate that impaired lakes would decrease from 22 to 31% using background
conditions with ANC increasing from 20 to 50|ieq/L), and (2) aggregated across New York
households. The scaling entails converting the average household willingness-to-pay for the
improvements described in the Adirondacks surveys to an average household willingness-to-pay
per percentage point of the total population of lakes improved.7 The results are summarized in
Table 3-4. The range of average household willingness to pay reflects the range in willingness to
pay per percentage point of lakes improved described in the two versions of the survey.
Estimates are provided at ANC 20, 50, and 100 to reflect the range of ANC discussed
throughout the REA and this document and for consistency with the Random Utility Model
analysis.
Table 3-4. Aggregate Annual Benefit Estimates for the Zero-Out Scenario
ANC
Threshold
20 |ieq/L
50 |ieq/L
100 |ieq/L
Reduction in
Percentage of
Unhealthy
Lakes
22%
31%
26%
Range of Average
Household WTP
per Percentage
Reduction
$2.63
$1.32
$2.63
$5.87
$3.76
$5.87
Number of NY
Households
(in millions)
7.162
7.162
7.162
Range of Annual
Benefits
(in millions of 2007$)
$410.6
$291.2
$491.6
$916.4
$829.4
$1,097.2
11
12
13
14
15
16
17
18
19
20
21
22
These results suggest that the value of avoiding current impairments to ecosystem
services from Adirondack lakes are even higher than the estimate, because the estimates assume
a lag of 10 years in which no benefits accrue and because the percent of impaired lakes is slightly
higher today than expected in 2020 under business-as-usual. These results imply significant
value to the public in addition to those derived from recreational fishing services. Note that the
results are only applicable to improvements in the Adirondacks valued by residents of New
York. If similar benefits exist in other acid-impacted areas, benefits for the nation as a whole
could be substantial. The analysis provides results on only a subset of the impacts of acidification
on ecosystem services and suggests that the overall impact on these services is likely to be
substantial.
7 Scaling is required because neither of the surveys administered by Banzhaf et al. (2006) describe improvements
that correspond exactly to the improvement scenario modeled here.
September, 2010
3-30
Do Not Quote or Cite
-------
1 3.4.2 Evidence for Adversity Related to Terrestrial Acidification
2 In the previous chapter of this document we discussed the effects of acidifying deposition
3 on terrestrial ecosystems, especially forests. These include the observed decline and/or dieback
4 in red spruce and sugar maple. These species are particularly sensitive to acidifying deposition
5 and have ranges that overlap the areas of the U.S. where some of the highest levels of acidifying
6 deposition occur. Additionally these species are present in the case study areas examined in the
7 REA. As a result we chose to focus on red spruce and sugar maple as the species of interest for
8 the analysis of ecosystem services presented in this section.
9 A similar model to Figure 3-6 can be drawn for terrestrial acidification that links Be: Al
10 molar ratio to reduced tree growth to decreases in timber harvest, although we have less
11 confidence in the significance of this linkage than we do for aquatic acidification. There are
12 numerous services expected to be affected but the data and methods to adequately describe those
13 losses does not as yet exist. These services include effects to forest health, water quality, and
14 habitat, including decline in habitat for threatened and endangered species, decline in forest
15 aesthetics, decline in forest productivity, increases in forest soil erosion and decreases in water
16 retention (ISA, 2009; REA, 2009; Krieger, 2001). Forests in the northeastern United States
17 provide several important and valuable provisioning services, which are reflected in the
18 production and sales of tree products.
19 Sugar maples are a particularly important commercial hardwood tree species in the United
20 States, producing timber and maple syrup that provide hundreds of millions of dollars in
21 economic value annually (NASS, 2008). Red spruce is also used in a variety of wood products
22 and provides up to $100 million in economic value annually (USFS, 2006).
23 Forests in the northeastern United States are also an important source of cultural
24 ecosystem services, including nonuse (existence value for threatened and endangered species),
25 recreational, and aesthetic services (ISA, 2009; REA, 2009). Red spruce forests are home to two
26 federally listed species, the spruce-fir moss spider and the rock gnome lichen. The value of these
27 two endangered species has not been estimated.
28 Although we do not have the data to link acidification damages directly to economic
29 values of lost recreational services in forests, these resources are valuable to the public. For
30 example, most recent data from the National Survey on Recreation and the Environment (NSRE)
31 indicate that, from 2004 to 2007, 31% of the U.S. adult (16 and older) population visited a
September, 2010 3-31 Do Not Quote or Cite
-------
1 wilderness or primitive area during the previous year, and 32% engaged in day hiking (Cordell et
2 al., n.d.). A recent study suggests that the total annual value of off-road driving recreation was
3 more than $9 billion, total and value of hunting and wildlife viewing was more than $4 billion
4 each in the Northeastern United States in 2006 (Kaval and Loomis, 2003). Table 3-5
5 summarizes data from the NSRE and the Fishing, Hunting, and Wildlife Related Activity Survey
6 (U.S. DOT, 2007) along with average WTP estimates from Kaval and Loomis to estimate the
7 total value of these services in the northeast.
9
10
Table 3-5. Annual participation and value of outdoor (forest related)
activity in the northeast.
Recreational
Activity
Off Road Driving
Hunting
Wildlife Viewing
Participation
Rate
16
5.5
10
Activity
Days
(in
Thousands)
366,336
83,821
122,200
Avg.
WTP Per
Activity
Day
($2007)
$25.25
$52.36
$34.46
Total
Value
(in millions)
$9,250
$4,380
$4,210
11
12
13
14
15
16
17
18
19
20
In addition, fall color viewing is a recreational activity that is directly dependent on
forest conditions. Sugar maple trees, in particular, are known for their bright colors and are,
therefore, an essential aesthetic component of most fall color landscapes. Statistics on fall color
viewing are much less available than for the other recreational and tourism activities; however, a
few studies have documented the extent and significance of this activity. For example, Spencer
and Holecek (2007) found that approximately 30% of residents in the Great Lakes area reported
at least one trip in the previous year involving fall color viewing. In a separate study conducted
in Vermont, Brown (2002) reported that more than 22% of households visiting Vermont in 2001
made the trip primarily for the purpose of viewing fall colors.
September, 2010
3-32
Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Two studies have estimated values for protecting high-elevation spruce forests in the
Southern Appalachians. Kramer et al. (2003) conducted a contingent valuation study estimating
households' WTP for programs to protect remaining high-elevation spruce forests from damages
associated with air pollution and insect infestation (Haefele et al., 1991; Holmes and Kramer,
1995). The survey presented respondents with a sheet of color photographs representing three
stages of forest decline and explained that, without forest protection programs, high-elevation
spruce forests would all decline to worst conditions (with severe tree mortality) and two potential
forest protection programs. Median household WTP was estimated to be roughly $29 (in 2007
dollars) for the minimal program and $44 for the more extensive program. Another study by
Jenkins, Sullivan, and Amacher (2002) estimated an aggregate annual value of $3.4 billion for
avoiding a significant decline in the health of high-elevation spruce forests in the Southern
Appalachian region.
Table 3-6. Summary of Studies of Select Terrestrial Ecosystem Services
Fall Color Viewing
Protection of spruce
30%
22%
$29
$44
$3.4 b
Great Lakes area residents
Vermont visitors
Southern Appalachians
WTP per household for minimal program
WTP per household for extensive program
Aggregate annual value
Spencer (2007)
Brown (2002)
Kramer et al. (2003)
Jenkins (2002)
15
16
17
18
19
20
21
Forests in the northeastern United States also support and provide a wide variety of
valuable regulating services, including soil stabilization and erosion control, water regulation,
and climate regulation (Krieger, 2001). Forest vegetation plays an important role in maintaining
soils in order to reduce erosion, runoff, and sedimentation that can adversely impact surface
waters. In addition to protecting the quality of water in this way, forests also help store and help
regulate the quantities and temporal discharge patterns of water in watersheds. Forests also play
September, 2010
3-33
Do Not Quote or Cite
-------
1 an important role in carbon sequestration at both regional and global scales. The total value of
2 these ecosystem services is very difficult to quantify.
3 3.4.2.1 What is the value of current ecosystem service impairments? The REA
4 Appendix 8 describes an analysis of ecosystem service impairments associated with the impacts
5 of terrestrial acidification on the forest product provisioning services from two commercially
6 important tree species on unmanaged forests - sugar maple and red spruce - that are particularly
7 sensitive to the effects of acidification. Evidence of effects due to terrestrial acidification is
8 particularly strong for these two species whose range includes the northeastern U.S. where levels
9 of nitrogen and sulfur deposition have historically been relatively high, however more
10 widespread impacts that include other tree species are also possible. We acknowledge that there
11 may be some beneficial fertilization effects of nitrogen deposition however given the complexity
12 of the nitrogen cycle it is not possible to quantify all those effects here. There is a detailed
13 discussion of nitrogen fertilization effects in Chapter 4.
14 In an exploratory study that is still under development we used data from the USFS
15 Forest Inventory and Analysis (FIA) database, to estimate an exposure-response relationship for
16 each species to measure the average negative effect of critical load exceedances (CLEs) of
17 nitrogen and sulfur deposition on annual tree growth. These estimated relationships were then
18 applied to sugar maple and red spruce stocks in the Northeast and North central regions to
19 estimate the average percent increase in annual tree growth that would occur if all CLEs were
20 eliminated. To estimate the aggregate-level forest market impacts of eliminating CLEs starting
21 in the year 2000, the tree-level growth adjustments were applied using the Forest and
22 Agricultural Sector Optimization Model (FASOM), which is a dynamic optimization model of
23 the U.S. forest and agricultural sectors. The model results are reported as the present discounted
24 values of future welfare changes in the forestry sector (in 5-year increments from 2000 to 2080)
25 due to increased tree growth. Summing over this 80-year period, the total present value of these
26 welfare gains is $40.705 million (in 2006 dollars, using a 4% discount rate). On an annualized
27 basis (at 4%), this is equivalent to $1.64 million per year. These estimates can also be
28 interpreted as the current value of impairments to forest provisioning services provided by red
29 spruce and sugar maple due to acidification effects from nitrogen and sulfur. These results
30 should be considered very uncertain due to the pending revision of the exposure - response curve
31 and release of an updated version of the FASOM model.
September, 2010 3-34 Do Not Quote or Cite
-------
1 3.4.3 Evidence for Adversity Related to Aquatic Nutrient Enrichment
2 Estuaries in the eastern United States are important for fish and shellfish production. The
3 estuaries are capable of supporting large stocks of resident commercial species, and they serve as
4 the breeding grounds and interim habitat for several migratory species (U.S. EPA, 2009). To
5 provide an indication of the magnitude of provisioning services associated with coastal fisheries,
6 from 2005 to 2007, the average value of total catch was $1.5 billion per year in 15 East Coast
7 states. It is not known, however, what percentage of this value is directly attributable to or
8 dependent upon the estuaries in these states. Based on commercial landings in Maryland and
9 Virginia, the values for three key species—blue crab, striped bass, and menhaden- totaled nearly
10 $69 million in 2007 in the Chesapeake Bay alone.
11 Assessing how eutrophication in estuaries affects fishery resources requires bioeconomic
12 models (i.e., models that combine biological models offish population dynamics with economic
13 models describing fish harvesting and consumption decisions), but relatively few exist (Knowler,
14 2002). Kahn and Kemp (1985) estimated that a 50% decline in submerged aquatic vegetation
15 (SAV) from levels existing in the late 1970s (similar to current levels [Chesapeake Bay Program,
16 2008]) would decrease the net social benefits from striped bass by $16 million (in 2007 dollars).
17 In a separate analysis, Anderson (1989) modeled blue crab harvests under baseline conditions
18 and under conditions with "full restoration" of SAV. In equilibrium, the increase in annual
19 producer surplus and consumer surplus with full restoration of SAV was estimated to be $7.9
20 million (in 2007 dollars) or an 11% increase from current service provision from blue crab alone.
21 Mistiaen et al. (2003) found that reductions in dissolved oxygen (DO) cause a statistically
22 significant reduction in commercial harvest and revenues crab harvests. For the Patuxent River
23 alone, a simulated reduction of DO from 5.6 to 4.0 mg/L was estimated to reduce crab harvests
24 by 49% and reduce total annual earnings in the fishery by $275,000 (in 2007 dollars). While
25 these values do not quantify the increase in terms of atmospheric loadings alone, the estimated
26 20% loading to the Potomac River watershed (REA 5.2.4) from atmospheric deposition indicates
27 that the benefits apportioned to deposition are significant.
28 In addition, eutrophi cation in estuaries may also affect the demand for seafood. For
29 example, a well-publicized toxic pfiesteria bloom in the Maryland Eastern Shore in 1997 led to
30 an estimated $56 million (in 2007 dollars) in lost seafood sales for 360 seafood firms in
September, 2010 3-35 Do Not Quote or Cite
-------
1 Maryland in the months following the outbreak (Lipton, 1999). Surveys by Whitehead, Haab,
2 and Parsons (2003) and Parsons et al. (2006) indicated a reduction in consumer surplus due to
3 eutrophication-related fish kills ranging from $2 to $5 per seafood meal.8 As a result, they
4 estimated aggregate consumer surplus losses of $43 million to $84 million (in 2007 dollars) in
5 the month after a fish kill.
6 As mentioned in the REA (5.2.1.3), estuaries in the eastern United States also provide an
7 important and substantial variety of cultural ecosystem services, including water-based
8 recreational and aesthetic services. For example, FHWAR data indicate that 4.8% of the
9 population in coastal states from North Carolina to Massachusetts participated in saltwater
10 fishing, with a total of 26 million saltwater fishing days in 2006 (U.S. DOT, 2007). Based on
11 estimates in Section 5.2.1.3 of the REA, total recreational consumer surplus value from these
12 saltwater fishing days was approximately $1.3 billion (in 2007 dollars). Recreational
13 participation estimates for several other coastal recreational activities are also available for
14 1999-2000 from the NSRE. Almost 6 million individuals participated in motorboating in coastal
15 states from North Carolina to Massachusetts. Again, based on analysis in the REA, the aggregate
16 value of these coastal motorboating outings was $2 billion per year. Almost 7 million people
17 participated in birdwatching, for a total of almost 175 million days per year, and more than 3
18 million participated in visits to nonbeach coastal waterside areas, for a total of more than 35
19 million days per year.
20 Estuaries and marshes have the potential to support a wide range of regulating services,
21 including climate, biological, and water regulation; pollution detoxification; erosion prevention;
22 and protection against natural hazards (MEA, 2005c). The relative lack of empirical models and
23 valuation studies imposes obstacles to the estimation of ecosystem services affected by nitrogen
24 deposition. While atmospheric deposition contributes to eutrophication there is uncertainty in
25 separating the effects of atmospheric nitrogen from nitrogen reaching the estuaries from many
26 other sources.
8 Surprisingly, these estimates were not sensitive to whether the fish kill was described as major or minor or to the
different types of information included in the survey.
September, 2010 3-36 Do Not Quote or Cite
-------
1
2
Table 3-7. Summary of Values for Current Levels of Services and Changes in
Service Levels in $2007
Ecosystem Service
Total Catch -
Commercial Fishing
Change in Ecosystem
Service
50% decline in SAV
Full restoration SAV
0.4% mg/L decrease DO
HAB
Ecosystem Service
Saltwater fishing
Motorboating
Bird watching
Non-beach coastal visits
Area or Population
Affected
14 east coast states
MD/VA
Chesapeake Bay
Chesapeake Bay
Patuxent River
1 997 MD eastern
shore
4.6% pop. MA-NC
6 million
7 million
3 million
Value
($2007)
$1.5b/yr
$69 m/yr
Value of
Change
($2007)
1 $16 m/yr
T $ 8 m/yr
| $275 th/yr
| $56 m
| $43-84 m
Participation
Days
26m days
175 m days
35m days
Species
blue crab, striped bass,
menhaden
striped bass
blue crab
| 49% blue crab harvest
loss to seafood industry
sustained loss over months
Value ($2007)
$1.3b/yr
2b/yr
4
5
6
7
September, 2010
3-37
Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
3.4.3.1 Value of aquatic ecosystem service impairments from current levels nutrient
enrichment
The aquatic nutrient enrichment case study relied on the NOAA Eutrophication Index as
the indicator, which includes dissolved oxygen, HABs, loss of SAV and loss of water clarity.
There are methods available to link some of the components to ecosystem services, most notably
loss of SAV and reductions in DO. The REA analysis estimates the change in several ecosystem
services including recreational fishing, boating, beach use, aesthetic services and nonuse
services. The REA focuses on two major East Coast estuaries - the Chesapeake Bay and the
Neuse River. Both estuaries receive between 20%-30% percent of their annual nitrogen loadings
through atmospheric deposition and both are showing symptoms of eutrophication. The analysis
uses and adapts results from several existing studies to approximate effects on several ecosystem
services, including commercial fishing, recreation, aesthetic enjoyment, and nonuse values. For
example, it is estimated that atmospheric nitrogen decreases the annual benefits of recreational
fishing, boating, and beach use in the Chesapeake Bay by $43-$217 million, $3-8 million, and
$124 million respectively, and reduces annual aesthetic benefits to nearshore residents by $39-
102 million. In the Neuse River, the value of annual commercial crab fishing services would be
between $0.1-1 million higher without the contribution of atmospheric nitrogen, and recreation
fishing services in the larger Albermarle Pamlico Sound estuary system (which includes the
Neuse) would be $1-8 million greater per year.
Table 3-8. Summary of Annual Damages to Services due to Atmospheric Loading
Ecosystem Service
Recreational Saltwater Fishing
Beach Use
Boating
Commercial Crab Fishing
Annual Value ($2007)
$43-217 b
$1-8 m
$39-102 m
$3-8 m
$0.1-1 m
Waterbody Affected
Chesapeake Bay
Albemarle Pamlico Sound
Chesapeake Bay
Chesapeake Bay
Neuse River
22
23
24
September, 2010
3-38
Do Not Quote or Cite
-------
1 3.4.4 Evidence for Adversity Related to Terrestrial Nutrient Enrichment
2 The ecosystem service impacts of terrestrial nutrient enrichment in unmanaged ecosystems
3 include primarily cultural and regulating services. In CSS areas, concerns focus on a decline in
4 CSS and an increase in nonnative grasses and other species, impacts on the viability of threatened
5 and endangered species associated with CSS, and an increase in fire frequency. Changes in MCF
6 include changes in habitat suitability and increased tree mortality, increased fire intensity, and a
7 change in the forest's nutrient cycling that may affect surface water quality through nitrate
8 leaching (EPA, 2008).
9 The terrestrial nutrient enrichment case study relies on benchmark deposition levels for
10 various species and ecosystems as indicators of ecosystem response. While it would be expected
11 that deposition above those levels would have deleterious effects on the provision of ecosystem
12 services in those areas, at this time it is possible only to describe the magnitude of the some of
13 the services currently being provided. Methods are not yet available to allow estimation of
14 changes in services due to nitrogen deposition.
15 The value that California residents and the U.S. population as a whole place on CSS and
16 MCF habitats is reflected in the various federal, state, and local government measures that have
17 been put in place to protect these habitats. Threatened and endangered species are protected by
18 the Endangered Species Act. The State of California passed the Natural Communities
19 Conservation Planning Program (NCCP) in 1991, and CSS was the first habitat identified for
20 protection under the program (seewww.dfg.ca.gov/habcon/nccp). It is estimated that only 10 -
21 15% of the original extent of CSS habitat remains (NPS.gov/cabr/naturescience/coastal-sage-
22 scrub-and-southern-chaparrel-communities.htm). Private organizations such as The Nature
23 Conservancy, the Audubon Society, and local land trusts also protect and restore CSS and MCF
24 habitat.
25 CSS and MCF are found in numerous recreation areas in California. Three national parks
26 and monuments in California contain CSS, including Cabrillo National Monument, Channel
27 Islands National Park, and Santa Monica National Recreation Area. All three parks showcase
28 CSS habitat with educational programs and information provided to visitors, guided hikes, and
29 research projects focused on understanding and preserving CSS. Over a million visitors traveled
30 through these three parks in 2008. MCF is highlighted in Sequoia and Kings Canyon National
September, 2010 3-39 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Park, Yosemite National Park, and Lassen Volcanic National Park, where more than 5 million
people visited in 2008.
The 2006 FHWAR for California (DOT, 2007) reports on the number of individuals
involved in fishing, hunting, and wildlife viewing in California. Millions of people are involved
in just these three activities each year. The quality of these trips depends in part on the health of
the ecosystems and their ability to support the diversity of plants and animals found in important
habitats found in CSS or MCF ecosystems and the parks associated with those ecosystems.
Based on analyses in Section 5.3.1.3 of the REA (U.S.EPA, 2009), average values of the total
benefits in 2006 from fishing, hunting, and wildlife viewing away from home in California were
approximately $947 million, $169 million, and $3.59 billion, respectively. In addition, data from
California State Parks (2003) indicate that in 2002, 68.7% of adult residents participated in trail
hiking for an average of 24.1 days per year. The analyses in the REA (U.S.EPA, 2009) indicate
that the aggregate annual benefit for California residents from trail hiking in 2007 was $11.59
billion. It is not currently possible to quantify the loss in value of services due to nitrogen
deposition as those losses are already reflected in the estimates of the contemporaneous total vale
of these recreational activities. Restoration of services through decreases in nitrogen deposition
would likely increase the total value of recreational services.
Table 3-9. Summary of Current Levels of Ecosystem Services
20
21
22
Activity
Trail Hiking
Fishing
Wildlife
Viewing
Hunting
Participation
68.7% of CA
population
1.7m
6.2m
0.28m
#of
days/yr
453m
19m
45m
3.3m
Average
WTP
$25.59
$48.86
$79.81
$50.10
Annual Aggregate
Value ($2007 in
millions)
115,900
947
3,600
169
Sources: 2006 FHWAR for California (DOI, 2007), California State Parks (2003),
Kaval and Loomis(2003)
September, 2010
3-40
Do Not Quote or Cite
-------
1 CSS and MCF are home to a number of important and rare species and habitat types. CSS
2 displays richness in biodiversity with more than 550 herbaceous annual and perennial species. Of
3 these herbs, nearly half are endangered, sensitive, or of special status (Burger et al., 2003).
4 Additionally, avian, arthropod, herpetofauna, and mammalian species live in CSS habitat or use
5 the habitat for breeding or foraging. Communities of CSS are home to three important federally
6 endangered species: the Quino checkerspot butterfly, the kangaroo rat and the California
7 gnatcatcher. MCF is home to one federally endangered species (mountain yellow-legged frog)
8 and a number of state-level sensitive species. The Audubon Society lists 28 important bird areas
9 in CSS habitat and at least 5 in MCF in California (http://ca.audubon.org/iba/index.shtml).9
10 The terrestrial enrichment case study in Section 5.3.1.3 of the REA and Section 3.3.5 of
11 the ISA identified fire regulation as a service that could be affected by nutrient enrichment of the
12 CSS and MCF ecosystems by encouraging growth of more flammable grasses, increasing fuel
13 loads, and altering the fire cycle. Over the 5-year period from 2004 to 2008, Southern California
14 experienced, on average, over 4,000 fires per year burning, on average, over 400,000 acres per
15 year (National Association of State Foresters [NASF], 2009). It is not possible at this time to
16 quantify the contribution of nitrogen deposition, among many other factors, to increased fire risk.
17 The CSS and MCF were selected as case studies for terrestrial enrichment because of the
18 potential that these areas could be adversely affected by excessive N deposition. To date, the
19 detailed studies needed to identify the magnitude of the adverse impacts due to N deposition
20 have not been completed. Based on available data, this report provides a qualitative discussion of
21 the services offered by CSS and MCF and a sense of the scale of benefits associated with these
22 services. California is famous for its recreational opportunities and beautiful landscapes. CSS
23 and MCF are an integral part of the California landscape, and together the ranges of these
24 habitats include the densely populated and valuable coastline and the mountain areas. Through
25 recreation and scenic value, these habitats affect the lives of millions of California residents and
26 tourists. Numerous threatened and endangered species at both the state and federal levels reside
27 in CSS and MCF. Both habitats may play an important role in wildfire frequency and intensity,
28 an extremely important problem for California. The potentially high value of the ecosystem
29 services provided by CSS and MCF justify careful attention to the long-term viability of these
30 habitats.
9 Important Bird Areas are sites that provide essential habitat for one or more species of bird.
September, 2010 3-41 Do Not Quote or Cite
-------
1 3.5 REFERENCES
2 Adams, D., R. Alig, B.A. McCarl, and B.C. Murray. February 2005. FASOMGHG
3 Conceptual Structure, and Specification: Documentation. Available at
4 http://agecon2.tamu.edu/people/faculty/mccarl-bruce/FASOM.html. Accessed on
5 October 22, 2008.
6 Anderson, E. 1989. "Economic Benefits of Habitat Restoration: Seagrass and the Virginia Hard-
7 Shell Blue Crab Fishery." North American Journal of Fisheries Management 9:140-149.
8 Acharya, G., and L.L. Bennett. 2001. "Valuing Open Space and Land-Use Patterns in Urban
9 Watersheds." Journal of Real Estate Finance and Economics 22(2/3):221-237. Arrow K.
10 and A.C. Fisher. 1974. "Environmental Preservation, Uncertainty, and Irreversibility."
11 The Quarterly Journal of Economics, 88(2):312-319.
12 Arrow K. and A.C. Fisher. 1974. "Environmental Preservation, Uncertainty, and Irreversibility."
13 The Quarterly Journal of Economics, 88(2):312-319.
14 Associated Press, Prime Trout Fishery gets Limestone Treatments to Blunt Acid Rain.
15 November 4, 2005
16 Banzhaf, S., D. Burtraw, D. Evans, and A. Krupnick. 2006. "Valuation of Natural Resource
17 Improvements in the Adirondacks." Land Economics 82:445-464.
18 Bartik, TJ. 1988. "Measuring the Benefits of Amenity Improvements in Hedonic Price
19 Models "LandEconomics 64(2): 172-183.
20 Bockstael, N.E., K.E. McConnell, and I.E. Strand. 1988. "Benefits from Improvements in
21 Chesapeake Bay Water Quality, Volume III." Environmental Protection Agency
22 Cooperative Agreement CR-811043-01-0.
23 Bockstael, N.E., K.E. McConnell, and I.E. Strand. 1989. "Measuring the Benefits of
24 Improvements in Water Quality: The Chesapeake Bay." Marine Resource Economics
25 6:1-18.
26 Boyd, J., and S. Banzhaf. 2007. "What are Ecosystem Services? The Need for Standardized
27 Environmental Accounting Units." Ecological Economics 63:616-626.
28 Bricker, S., D. Lipton, A. Mason, M. Dionne, D. Keeley, C. Krahforst, J. Latimer, and
29 J.Pennock. 2006. Improving Methods and Indicators for Evaluating Coastal Water
30 Eutrophication: A Pilot Study in the Gulf of Maine. NOAA Technical Memorandum
31 NOS NCCOS 20. Silver Spring, MD: National Ocean Service, National Centers for
32 Coastal Ocean Science, Center for Coastal Monitoring and Assessment.
33 Bricker, S., Longstaff, B., Dennison, W., Jones, A., Boicourt, K., Wicks, C., and Woerner, J.
34 2007. Effects of Nutrient Enrichment in the Nation's Estuaries: A Decade of Change.
35 NOAA Coastal Ocean Program Decision Analysis Series No. 26. National Centers for
36 Coastal Ocean Science, Silver Spring, MD.
37 Bricker, S.B., J.G. Ferreira, and T. Simas. 2003. "An Integrated Methodology for Assessment of
38 Estuarine Trophic Status." Ecological Modelling 169:39-60.
39 Brocksen, Robert W. and Emler , Paul W. Jr. "Living Lakes: an Aquatic Liming and Fish
40 Restoration Demonstration Program." Water, Air, & Soil Pollution. Springer
41 Netherlands Volume 41, Numbers 1-4 / pp. 85-93. March, 1988.
42 Brown, L.H. 2002. Profile of the Annual Fall Foliage Tourist in Vermont: Travel Year
43 2007.Report prepared for the Vermont Department of Tourism and Marketing.
44 Burlington, VT: Vermont Tourism Data Center, University of Vermont.
45 Burger, J.C., R.A. Redak, E.B. Allen, J.T. Rotenberry, and M.F. Allen. 2003. "Restoring
46 Arthropod Communities in Coastal Sage Scrub." Conservation Biology 17(2):460-467.
September, 2010 3-42 Do Not Quote or Cite
-------
1 California Department of Fish and Game. n.d.
2 http://www.dfg.ca.gov/habcon/nccp/images/region.gif
3 Chen, L., and C.T. Driscoll. 2004. "Modeling the Response of Soil and Surface Waters in the
4 Adirondack and Catskill Regions of New York to Changes in Atmospheric Deposition
5 and Historical Land Disturbance." Atmospheric Environment 38:4099-4109.
6 Chesapeake Bay Program, n.d. "Chesapeake Bay Program: A Watershed Partnership."
7 Available at http://www.chesapeakebay.net/.
8 Chesapeake Bay Program. "Bay Grass Abundance (Baywide)." Available at
9 http://www.chesapeakebay.net/ status_baygrasses.aspx?menuitem=19652. Last updated
10 May 21, 2008.
11 Chesapeake Eco-Check. 2007. "Indicator Details: 2007. Water Quality Index." Available at
12 http://www.eco-check.org/reportcard/chesapeake/2007/indicators/water_quality_index/.
13 Environmental Systems Research Institute, Inc. (ESRI). 2002. "Census Block Groups: DC,
14 DE,MD, and VA." [CD-ROM]. ESRI Data & Maps 2002.
15 Cordell, H.K., CJ. Betz, G. Green, and M. Owens. 2005. Off-Highway Vehicle Recreation in the
16 United States, Regions and States: A National Report from the National Survey on
17 Recreation and the Environment (NSRE). Report prepared for the Forest Service's
18 National OHV Policy & Implementation Teams. USDA Forest Service.
19 Cordell, K., B. Leeworthy, G.T. Green, C. Betz, and B. Stephens, n.d. The National Survey
20 onRecreation & the Environment. Research Work Unit 4953. Athens, GA: Pioneering
21 Research on Changing Forest Values in the South and Nation USDA Forest Service
22 Southern Research Station. Available at www.srs.fs.fed.us/trends.
23 Daily, G.C., S. Alexander, P.R. Ehrlich, L. Goulder, J. Lubchenco, P.A. Matson, H.A.
24 Mooney,S. Postel, S.H. Schneider, D. Tilman, and G.M. Woodwell. 1997. "Ecosystem
25 Services:Benefits Supplied to Human Societies by Natural Ecosystems." Issues in
26 Ecology 2:1-16
27 DeHayes, D.H., P.G. Schaberg, GJ. Hawley, and G.R. Strimbeck, 1999. "Acidic Rain Impacts
28 on Calcium Nutrition and Forest Health." BioScience 49:789-800.
29 Driscoll, C.T. et al. 2003. "Effects of Acidic Deposition on Forest and Aquatic Ecosystems in
30 New York State." Environmental Pollution 123:327-336.
31 Driscoll, C.T., G.B. Lawrence,AJ. Bulger, TJ. Butler, C.S. Cronan, C. Eagar, K.F. Lamber,
32 G.E. Likens, J.L. Stoddard,and K.C. Weathers. 2001. Acidic Deposition in the
33 Northeastern United States: Sources and Inputs, Ecosystem Effects and Management
34 Strategies. BioScience 51:180-198.
35 Federal Register (2007) National Ambient Air Quality Standards for Ozone; Proposed Rule. 40
36 CFR 50; Federal Register 72: 37818
37 Federal Register (2008) National Ambient Air Quality Standards for Ozone; Final Rule. 40 CFR
38 parts 50 and 58; Federal Register 73:16436
39 Fenn, M.E., T.G. Huntington, S.B. McLaughlin, C. Eager, A. Gomez, and R.B. Cook.
40 2006."Status of Soil Acidification in North America." Journal of Forest Science 52
41 (special i ssue): 3 -13.
42 Greene, W.H. 2003. Econometric Analysis, 5th Ed. New Jersey: Prentice Hall.
43 Haefele, M., R.A. Kramer, and T.P. Holmes. 1991. Estimating the Total Value of a Forest
44 Quality in High-Elevation Spruce-Fir Forests. The Economic Value of Wilderness:
45 Proceedings of the Conference. Gen. Tech. Rep. SE-78 (pp. 91-96). Southeastern For
46 Exper. Station. Asheville, NC: USDA Forest Service.
September, 2010 3-43 Do Not Quote or Cite
-------
1 Hogsett, W. E., Weber, J. E., Tingey, D., Herstrom, A., Lee, E. H., Laurence, J. A. (1997)
2 "Environmental auditing:an approach for characterizing tropospheric ozone risk to
3 forests." Environ. Manage. 21: 105-120
4 Holmes, T.P. 1992. Economic Welfare Impacts of Air Pollution Damage to Forests in the
5 Southern United States. Asheville, NC: U.S. Dept. of Agriculture, Forest Service,
6 Southeastern Forest Experiment Station, pp. 19-26.
7 Holmes, T., and R. Kramer. 1995. "An Independent Sample Test of Yea-Saying and Starting
8 Point Bias in Dichotomous-Choice Contingent Valuation. " Journal of Environmental
9 Economics and Management 28:121-132.Final
10 Hoover, Kevin L. and Terry A. Rightnour. Atmospheric Acidification Abatement Demonstration
11 Projects -Mosquito Creek Phase 1.. Mosquito Creek Sportsmen's Association. Clearfield
12 County, PA. Pennsylvania Growing Greener Final Report. May 2002.
13 Hunkapiller, Bill. State of the Lake. 2010 Berkeley Lake Homeowners Association. March,
14 2006.
15 Hutchison, R., and C.E. Kraft. 1994. "Hmong Fishing Activity and Fish Consumption." Journal
16 of Great Lakes Research 20(2):471-487.
17 Industrial Economics, Inc. June 2008. "The Economic Impact of the Clean Air Interstate Rule on
18 Recreational Fishing in the Adirondack Region of New York State." Prepared for the
19 Clean Air Markets Division, Office of Air and Radiation, U.S. EPA.
20 Irwin, E.G. 2002. "The Effects of Open Space on Residential Property Values." Land Economics
21 78(4):465-480.
22 J. Nilsson and Grennfelt, P. (Eds) (1988) Critical loads for sulphur and nitrogen. UNECE/Nordic
23 Council workshop report, Skokloster, Sweden. March 1988. Nordic Council of Ministers:
24 Copenhagen.
25 Jenkins, D.H., J. Sullivan, and G.S. Amacher. 2002. "Valuing High Altitude Spruce-Fir
26 Forestlmprovements: Importance of Forest Condition and Recreation Activity." Journal
27 of Forest Economics 8:77-99.
28 Johnston, R.J., E.Y. Besedin, R. lovanna, C.J. Miller, R.F. Wardwell, and M.H. Ranson.
29 2005."Systematic Variation in Willingness to Pay for Aquatic Resource Improvements
30 and Implications for Benefit Transfer: A Meta-Analysis." Canadian Journal of
31 Agricultural Economics 53:221-248.
32 Johnston, R.J., M.H. Ranson, E.Y. Besedin, and E.C. Helm. 2006. "What Determines
33 Willingness to Pay per Fish? A Meta-Analysis of Recreational Fishing Values." Marine
3 4 Resource Economics 21:1-32.
35 Kahn, J.R., and W.M. Kemp. 1985. "Economic Losses Associated with the Degradation of an
36 Ecosystem: The Case of Submerged Aquatic Vegetation in Chesapeake Bay." Journal of
3 7 Environmental Economics and Management 12:246-263.
38 Kaoru, Y. 1995. "Measuring Marine Recreation Benefits of Water Quality Improvements by the
3 9 Nested Random Utility Model." Resource and Energy Economics 17(2):119-136.
40 Kaoru, Y., V. Smith, and J.L. Liu. 1995. "Using Random Utility Models to Estimate the
41 Recreational Value of Estuarine Resources." American Journal of Agricultural
42 Economics 77:141-151.
43 Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
44 National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
45 1200-99-009, Project number EVIDE-02-0070.
September, 2010 3-44 Do Not Quote or Cite
-------
1 Kennedy, P. 2003. A Guide to Econometrics., pp. 312-313. Cambridge, MA: MIT Press.
2 Millennium Ecosystem Assessment (MEA). 2005. Ecosystems and Human Well-being:
3 Wetlands and Water. Synthesis. A Report of the Millennium Ecosystem Assessment.
4 Washington, DC: World Resources Institute.
5 Knowler, D. 2002. "A Review of Selected Bioeconomic Models: With Environmental Influences
6 in Fisheries." Journal of Bioeconomics 4:163-181.
7 Kramer, A., T. Holmes, and M. Haefele. 2003. "Contingent Valuation of Forest Ecosystem
8 Protection." In Forests in a Market Economy, E.O. Sills and K.L. Abt, eds., pp. 303-320.
9 Dordrecht, The Netherlands: Kluwer Academic Publishers.
10 Krieger, DJ. 2001. Economic Value of Forest Ecosystem Services: A Review. Washington,
11 DC:The Wilderness Society.
12 Krupnick, A. 1988. "Reducing Bay Nutrients: An Economic Perspective." Mary land Law Review
13 47(2):453-480.
14 Land Trust Alliance. 2006. The 2005 National Land Trust Census Report. Washington,
15 D.C.:Land Trust Alliance, November 30, 2006.
16 Leeworthy, V.R., and P.C. Wiley. 2001. Current Participation Patterns inMarine
17 Recreation.Silver Spring, MD: U.S. Department of Commerce, National Oceanic and
18 Atmospheric Administration, National Ocean Service, Special Projects.
19 Lipton, D. W. 1999. "Pfiesteria's Economic Impact on Seafood Industry Sales and Recreational
20 Fishing." In Proceedings of the Conference, Economics of Policy Options for Nutrient
21 Management andPfiesteria. B. L. Gardner and L. Koch, eds., pp. 35-38. College Park,
22 MD: Center for Agricultural and Natural Resource Policy, University of Maryland.
23 Lipton, D. 2004. "The Value of Improved Water Quality to Chesapeake Bay Boaters." Marine
24 Resource Economics 19:265-270.
25 Lipton, D. 2006. "Economic Impact of Maryland Boating in 2005." University of Maryland Sea
26 Grant Extension Program. Available at
27 ftp://ftp.mdsg.umd.edu/Public/MDSG/rec_boat05.pdf.
28 Lipton, D. 2008. "Economic Impact of Maryland Boating in 2007." University of Maryland Sea
29 Grant Extension Program. Available at
30 ftp://ftp.mdsg.umd.edu/Public/MDSG/rec_boat07.pdf.
31 Lipton, D.W., and R.W. Hicks. 1999. "Linking Water Quality Improvements to Recreational
32 Fishing Values: The Case of Chesapeake Bay Striped Bass." In Evaluating the Benefits of
33 Recreational Fisheries. Fisheries Centre Research Reports 7(2), TJ. Pitcher, ed., pp.105-
34 110. Vancouver: University of British Columbia.
35 Lipton, D.W., and R.W. Hicks. 2003. "The Cost of Stress: Low Dissolved Oxygen and the
36 Economic Benefits of Recreational Striped Bass Fishing in the Patuxent River." Estuaries
37 26(2A):310-315.
38 Loomis, J., D. Griffin, E. Wu, and A. Gonzalez-Caban. 2002. "Estimating the Economic Value
39 of Big Game Habitat Production from Prescribed Fire Using a Time Series Approach.
40 Journal of Forest Economics 2:119-29.
41 Loomis, J.B., and D.S. White. 1996. "Economic Benefits of Rare and Endangered Species:
42 Summary and Meta-Analysis." Ecological Economics 18(3): 197-206.
43 Luzadis, V. A. and E.R. Gossett. 1996. "Sugar Maple." In Forest Trees of the Northeast,
44 J.P.Lassoie, V.A. Luzadis, and D.W. Grover, eds., pp. 157-166. Cooperative Extension
45 Bulletin 235. Cornell Media Services. Available at
46 http ://maple. dnr. Cornell. edu/pub s/trees .htm.
September, 2010 3-45 Do Not Quote or Cite
-------
1 Mansfield, C.A., S.K. Pattanayak, W. McDow, R. MacDonald, and P. Halpin. 2005. "Shades of
2 Green: Measuring the Value of Urban Forests in the Housing Market." Journal of Forest
3 Economics ll(3):m-199.
4 Massey, M., S. Newbold, and B. Gentner. 2006. "Valuing Water Quality Changes Using a
5 Bioeconomic Model of a Coastal Recreational Fishery." Journal of
6 EnvironmentalEconomics and Management 52:482-500.
7 Mayland, Don. 2009 State of the Lake Report. Lake Wononscopomuc Association, Marine Study
8 Program, Inc. 2009.
9 Millennium Ecosystem Assessment (MEA). 2005a. Ecosystems and Human Well-being: Current
10 State and Trends, Volume 1. R. Hassan, R. Scholes, and N. Ash, eds. Washington,
11 DCisland Press. Available at
12 http://www.millenniumassessment.org/documents/document.766.aspx.pdf.
13 Millennium Ecosystem Assessment (MEA). 2005b. Ecosystems and Human Well-
14 being:Synthesis. Washington, DC: World Resources Institute.
15 Mistiaen, J.A., I.E. Strand, and D. Lipton. 2003. "Effects of Environmental Stress on Blue Crab
16 (Callinectes sapidus) Harvests in Chesapeake Bay Tributaries." Estuaries 26(2a): 316-
17 322.
18 Montgomery, M. and M. Needelman. 1997. "The Welfare Effects of Toxic Contamination in
19 Freshwater Fish." Land Economics 73(2):211-223.
20 Morgan, C., and N. Owens. 2001. "Benefits of Water Quality Policies: The Chesapeake Bay."
21 Ecological Economics 39:271-284.
22 Murray, T., and J. Lucy. 1981. "Recreational Boating in Virginia: A Preliminary Analysis."
23 Special Report in Applied Marine Science and Ocean Engineering No. 251. Available at
24 http://www.vims.edu/GreyLit/VIMS/sramsoe251 .pdf.
25 National Agricultural Statistics Service (NASS). June 12, 2008. "Maple Syrup Production Up 30
26 Percent Nationwide." New England Agricultural Statistics, NASS, USDA. Spencer, D.M.,
27 and D.F. Holecek. 2007. "Basic Characteristics of the Fall Tourism Market." Tourism
28 Management 28:491-504
29 National Association of State Foresters (NASF). 2009. Quadrennial Fire Review
30 2009.Washington, DC: NASF. Quadrennial Fire and Fuel Review Final Report 2009.
31 National Marine Manufacturers Association (NMMA). 2008. "2007 Recreational Boating
32 Statistical Abstract." Available at
33 http://www.nmma.org/facts/boatingstats/2007/files/Abstract.pdf.
34 National Oceanic and Atmospheric Administration (NOAA). (2007, August). "Annual
35 Commercial Landing Statistics." Available at http://www.st.nmfs.noaa.gov/
36 stl/commercial/landings/annual_landings.html.
37 National Oceanic and Atmospheric Administration (NOAA). "Recreational Fisheries."
38 Available at www.st.nmfs.noaa.gov/stl/recreational/overview/overview.html. Last
39 updated February 12, 2009.
40 National Research Council. 2005. Valuing Ecosystem Services: Toward Better Environmental
41 Decision-Making. Washington, DC: National Academies Press.
42 New Hampshire Department of Environmental Services, Prepared by Robert H. Estabrook.
43 Total Maximum Daily Load (TMDL) for 65 Acid Impaired New Hampshire Ponds,
44 FINAL, September, 2004.
45 New York State, Department of Environmental Conservation Division of Water, Bureau of
46 Water Assessment and Management. Impaired Waters Restoration Plan For Acid Rain
September, 2010 3-46 Do Not Quote or Cite
-------
1 Lakes (NYS Forest Preserve) Adirondack Region, New York and Proposed Total
2 Maximum Daily Load (TMDL) for pH/AcidRain Impacts, September 2006.
3 Olson, Mancur. The Logic of Collective Action : Public Goods and the Theory of Groups
4 (Revisededition ed.). Harvard University Press. ISBN 0-674-53751-3. 1971. [1965.]
5 Parsons, G., A.O. Morgan, J.C. Whitehead, and T.C. Haab. 2006. "The Welfare Effects of
6 Pfiesteria-Related Fish Kills: A Contingent Behavior Analysis of Seafood Consumers."
7 Agricultural and Resource Economics Review 3 5 (2): 1 -9.
8 Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Diedrich, H.A. Anderson, P.L.
9 Remington, and T.B. Sheffy. 1994. "Fish Consumption Patterns and Blood Mercury
10 Levels in Wisconsin Chippewa Indians." Archives of Environmental Health 49(1):53 58.
11 Poor, P.I, K.L. Pessagno, and R.W. Paul. 2007. "Exploring the Hedonic Value of Ambient
12 Water Quality: A Local Watershed-Based Study." Ecological Economics 60:797-806.
13 Smith, M.D. 2007. "Generating Value in Habitat-Dependent Fisheries: The Importance of
14 Fishery Management Institutions." Land Economics 83(l):59-73.
15 Sweeney, J. 2007. "Impacts of CAMD 2020 CAIR on Nitrogen Loads to the Chesapeake Bay."
16 University of Maryland, Chesapeake Bay Program Office.
17 Tennessee Department of Environment and Conservation Division of Water Pollution Control.
18 Proposed Total Maximum Daily Load (TMDL) For Low pH In The Great Smoky
19 Mountains National Park Located In The Pigeon River Watershed (HUC 06010106)
20 Lower French Broad River Watershed (HUC 06010107) Ft. Loudoun Lake Watershed
21 (HUC 06010201) Cocke andSevier County, Tennessee DRAFT. April 12, 2010
22 U.S. Department of Agriculture, Forest Service. Forest Inventory and Analysis National
23 Program, RPA Assessment Tables. 2002. Available at
24 http://Ncrs2.Fs.Fed.Us/4801/Fiadb/Rpa_Tabler/Draft_RPA_2002_Forest_Resource_Tabl
25 es.pdf.
26 U.S. Census Bureau. 2008a. "Annual Social and Economic (ASEC) Supplement." Available at
27 http://pubdb3.census.gov/macro/032008/hhinc/new02_001.htm.
28 U.S. Census Bureau. 2008b. "Housing Unit Estimates for Counties of MD and VA: April 1/2000
29 to July 1/2007." Available at http://www.census.gov/popest/housing/HU-EST2007-
30 CO.html.
31 U.S. Department of the Interior, Fish and Wildlife Service, and U.S. Department of Commerce,
32 U.S. Census Bureau. 2007. 2006 National Survey of Fishing, Hunting, and Wildlife-
33 Associated Recreation.
34 U.S. EPA. 1998.. EPA/630/R-95/002F. Risk Assessment Forum, Washington, DC. Available
35 from: http://www.epa.gov/ncea/ecorsk.htm.
36 U.S. Environmental Protection Agency (EPA), Office of Air and Radiation. November 1999.The
37 Benefits and Costs of the Clean Air Act 1990 to 2010: EPA Report to Congress.EPA-
38 410-R-99-001. Washington, DC: U.S. Environmental Protection Agency.
39 U.S. Environmental Protection Agency (EPA). 2002. December. Environmental and Economic
40 Benefit Analysis of Final Revisions to the National Pollutant Discharge Elimination
41 System Regulation and the Effluent Guidelines for Concentrated Animal Feeding
42 Operations (EPA-821-R-03-003). Washington, DC: U.S. Environmental Protection
43 Agency, Office of Water, Office of Science and Technology.
44 U. S. EPA (Environmental Protection Agency). 2006. Ecological Benefits Assessment Strategic
45 Plan. EPA-240-R-06-001. Office of the Administrator. Washington, DC. Available at
46 http://www.epa.gov/economics.
September, 2010 3-47 Do Not Quote or Cite
-------
1 U.S. Environmental Protection Agency (EPA). 2008. Integrated Science Assessment for Oxides
2 of Nitrogen and Sulfur-Environmental Criteria.EPA/600/R-08/082. U.S. Environmental
3 Protection Agency, Office of Research and Development, National Center for
4 Environmental Assessment - RTF Division, Research Triangle Park, NC
5 US Environmental Protection Agency (EPA). 2009. Risk and Exposure Assessment for the
6 Review of the Secondary national Ambient Air Quality Standards for Oxides of Nitrogen
7 and Oxides of Sulfur. EPA/452/R/09/008a. U.S environmental Protection Agency,
8 Office of Air Quality planning and Standards, Health and Environmental impacts
9 Division, Research Triangle Park, NC.
10 U.S. EPA (Environmental Protection Agency). 2009. Valuing the Protection of Ecological
11 Systems and Services: A report of the EPA Science Advisory Board. EPA-SAB-09-012.
12 Office of the Administrator. Washington, DC. Available at
13 http://yosemite.epa.gov/sab/sabproduct.nsf/WebBOARD/ValProtEcolSys&Serv?
14 U.S. EPA. 1996. National Acid Precipitation Assessment Program Report to Congress.
15 Available at:
16 http://dwb4.unl.edu/chem/chem869v/chem869vlinks/www.nnic.noaa.gov/CENR/NAPAP
17 /NAP AP_96.htm
18 Valigura, R.A., R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacy, and
19 R.E.Turner. 2001. Nitrogen Loading in Coastal Water Bodies: An Atmospheric
20 Perspective.Washington, DC: American Geophysical Union.
21 Van Houtven, G. and A. Sommer. December 2002. Recreational Fishing Benefits: A Case Study
22 of Reductions in Nutrient Loads to the Albemarle-Pamlico Sounds Estuary. Final Report.
23 Prepared for the U.S. Environmental Protection Agency. Research Triangle Park,
24 NC:RTI International.
25 Van Houtven, G.L., J. Powers, and S.K. Pattanayak. 2007. "Valuing Water Quality
26 Improvements Using Meta-Analysis: Is the Glass Half-Full or Half-Empty for National
27 Policy Analysis?" Resource and Energy Economics 29:206-228.
28 Vaughan, WJ. 1986. "The Water Quality Ladder." Included as Appendix B in R.C. Mitchell,
29 and R.T. Carson, eds. The Use of Contingent Valuation Data for Benefit/Cost Analysis in
30 Water Pollution Control. CR-810224-02. Prepared for the U.S. Environmental Protection
31 Agency, Office of Policy, Planning, and Evaluation.
32 Vermont Department of Environmental Conservation Water Quality Division, Total Maximum
33 Daily Loads, (Approved by EPA Region 1 September 30, 2003)
34 Virginia Department of Conservation and Recreation. 2007. "2006 Virginia Outdoors
35 Survey."Available at
36 http://www.dcr. virginia.gov/recreational_planning/documents/vopsurvey06.pdf
37 Wallace, KJ. 2007. "Classification of Ecosystem Services: Problems and Solutions." Ecological
38 Conservation 139:235-246.
39 Whitehead, J.C., T.C. Haab, and G.R. Parsons. 2003. "Economic Effects of Pfiesteria." Ocean &
40 Coastal Management 46(9-10):845-858.
41 Young, T. F.; Sanzone, S., eds. (2002) A framework for assessing and reporting on ecological
42 condition: an SABreport. Washington, DC: U.S. Environmental Protection Agency,
43 Science Advisory Board; report no. EPASAB-EPEC-02-009. Available:
44 http://www.epa.gov/sab/pdf/epec02009.pdf [9 December, 2003].
September, 2010 3-48 Do Not Quote or Cite
-------
2 4 ADDRESSING THE ADEQUACY OF THE CURRENT STANDARDS
o
4 Based on the information in Chapters 2 and 3, we conclude that there is support in the
5 available effects-based evidence for consideration of secondary standards for NOx and SOx that
6 are protective against adverse ecological effects associated with deposition of NOx and SOx to
7 sensitive ecosystems. Having reached this general conclusion, we then to the extent possible
8 evaluate the adequacy of the current NOx and SOx secondary standards by considering to what
9 degree risks to sensitive ecosystems would be expected to occur in areas that meet the current
10 standards. Staff conclusions regarding the adequacy of the current standards are based on the
11 available ecological effects, exposure and risk-based evidence. In evaluating the strength of this
12 information, staff have taken into account the uncertainties and limitations in the scientific
13 evidence. This chapter addresses key policy relevant questions that inform our determination
14 regarding the adequacy of the structure and levels of the current secondary standards. The
15 chapter begins with a discussion of the structure of the current standards, followed by a
16 presentation of information on recent air quality relative to the existing standards, recent NOx
17 and SOx deposition levels, evaluation of recent deposition levels relative to levels where adverse
18 ecological effects have been observed, and a set of conclusions regarding the adequacy of the
19 current structure and levels of the standards. Acidification occurs over extended periods and
20 the ability of both terrestrial and aquatic systems to recover is dependent upon not only the
21 decrease in acidic deposition, but the ability of these ecosystems to generate cations needed for
22 nutrients and base cation supply. As a result, given the same decrease in deposition, ecosystems
23 with high levels of base cation replacement will recover faster than those with low levels.
24
25 4.1 Are The Structures Of The Current NOX And SOX Secondary Standards Based On
26 Relevant Ecological Indicators Such That They Are Adequate To Determine And
27 Protect Public Welfare Against Adverse Effects On Ecosystems?
28
29 The current secondary NOx and SOx standards are intended to protect against adverse
30 effects to public welfare. For NOX, the current secondary standard was set identical to the
September, 2010 4-1 Do Not Quote or Cite
-------
1 primary standard1, e.g. an annual standard set for NC>2 to protect against adverse effects on
2 vegetation from direct exposure to ambient NOx. For SOx, the current secondary standard is a
3 3-hour standard intended to provide protection for plants from the direct foliar damage
4 associated with atmospheric concentrations of SC>2. It is appropriate in this review to consider
5 whether the current standards are adequate to protect against the direct effects on vegetation
6 resulting from ambient NC>2 and SC>2 which were the basis for the current secondary standards.
7 The ISA concluded that there was sufficient evidence to infer a causal relationship between
8 exposure to 862, NO, NC>2 and PAN and injury to vegetation. Additional research on acute
9 foliar injury has been limited and there is no evidence to suggest foliar injury below the levels of
10 the current secondary standards for SOx and NOx. There is sufficient evidence to suggest that
11 the levels of the current standards are likely adequate to protect against direct phytotoxic effects.
12 The ISA however, has established that the major effects of concern for this review of the
13 NOx and SOx standards are associated with deposition of N and S caused by atmospheric
14 concentrations of NOx and SOx (see Chapter 2). As discussed in the following sections, the
15 current standards are not directed toward deposit onal effects, and none of the elements of the
16 current NAAQS - indicator, form, averaging time, and level - are suited for addressing the
17 effects of N and S deposition. Thus, by using atmospheric NO2 and SO2 concentrations as
18 indicators, the current standards address only a fraction of total atmospheric NOx and SOx, and
19 do not take into account the effects from deposition of total atmospheric NOx and SOx. By
20 addressing short-term concentrations, the current SO2 standard, while protective against direct
21 foliar effects from gaseous SOx, does not take into account the findings of effects in the ISA,
22 which notes the relationship between annual deposition of S and acidification effects which are
23 likely to be more severe and widespread than phytotoxic effects under current ambient
24 conditions, and include effects from long term deposition as well as short term.. Acidification is
25 a process which occurs over time, as the ability of an aquatic system to counteract acidic inputs
26 is reduced as natural buffers are used more rapidly than they can be replaced through geologic
27 weathering. The relevant period of exposure for ecosystems is therefore not the exposures
28 captured in the short averaging time of the current SO2 standard.
1 The current primary NO2 standard has recently been changed to the 3 year average of the 98thpercentile of the
annual distribution of the 1 hour daily maximum of the concentration of NO2. The current secondary standard
remains as it was set in 1971.
September, 2010 4-2 Do Not Quote or Cite
-------
1 The levels of the current standards also are not well suited to dealing with deposit!on-
2 based effects of NOx and SOx. Current standards are specified as allowable single atmospheric
3 concentration levels for NO2 or SO2. This type of structure does not take into account variability
4 in the atmospheric and ecological factors that may alter the effects of NOx and SOx on public
5 welfare. Consistent with section 108 of the CAA, the ISA includes in the air quality criteria
6 consideration of how these variable factors impact the effects of ambient NOx and SOx on public
7 welfare. See CAA section 108 (a)(2)(A) requiring air quality criteria to include information on
8 "those variable factors (including atmospheric conditions) which of themselves or in
9 combination with other factors may alter the effects on ... welfare of such air pollutant".
10 Secondary standards are intended to address a wide variety of effects occurring in different types
11 of environments and ecosystems. Ecosystems are not uniformly distributed either spatially or
12 temporally in their sensitivity to air pollution. Therefore, failure to account for the major
13 determinants of variability, including geological and soil characteristics related to the sensitivity
14 to acidification as well as atmospheric and landscape characteristics that govern rates of
15 deposition, may lead to standards that do not provide requisite levels of protection across
16 ecosystems. Finally, given the mismatch of all of the other elements of the current secondary
17 NAAQS with deposition-based effects, the form of those standards will also be mismatched.
18 Because most areas of the U.S. are in attainment with the current NO2 and SOx standards,
19 it is possible to evaluate current conditions, and evaluate the impact on public welfare from the
20 current effects on ecosystems from NOX and SOX deposition in areas that attain the current
21 standards that use NO2 and SO2 as indicators. In addition, this chapter qualitatively addresses
22 the adequacy of the structures of the existing standards relative to ecologically relevant standards
23 for NOx and SOx, and sets up arguments for developing an ecologically relevant structure for the
24 standards as described in Chapter 5.
25
26 4.2 To What Extent Are The Structures Of The Current NOX AND SOX Secondary
27 Standards Meaningfully Related To Relevant Ecological Indicators Of Public
28 Welfare Effects?
29
30 The current secondary standard for NOx, set in 1971, using NO2 as the atmospheric
31 indicator, is 0.053 parts per million (ppm) (100 micrograms per cubic meter of air [|jg/m3]),
September, 2010 4-3 Do Not Quote or Cite
-------
1 annual arithmetic average, calculated as the arithmetic mean of the 1-hour NO2 concentrations.
2 This standard was selected to provide protection to the public welfare against acute injury to
3 vegetation from direct exposure and resulting phytoxicity. During the last review of the NOX
4 standards, impacts associated with chronic acidification and eutrophication from NOx deposition
5 were acknowledged, but the relationships between atmospheric concentrations of NOx and levels
6 of acidification and eutrophication and associated welfare impacts were determined to be too
7 uncertain to be useful as a basis for setting a national secondary standard (USEPA 1995).
8 The current secondary standard for SOx, set in 1971, uses 862 as the atmospheric
9 indicator, is a 3-hour average of 0.5 ppm, not to be exceeded more than once per year. This
10 standard was selected to provide protection to the public welfare against acute injury to
11 vegetation. In the last review of the SOX secondary standard, impacts associated with chronic
12 acidification were acknowledged, but the relationships between atmospheric concentrations of
13 SOx and levels of acidification, along with the complex interactions between SOx and NOx in
14 acidification processes, were cited as critical uncertainties which made the setting of secondary
15 NAAQS to protect against acidification inappropriate at that time (USEPA 1982).
16 In the previous separate reviews of the NOx and SOx secondary standards, EPA
17 acknowledged in each review the additional impacts of NOx and SOx on public welfare through
18 the longer term impact of the pollutants once deposited to ecosystems. However, the previous
19 reviews cited numerous uncertainties as the basis for not directly addressing those impacts in the
20 setting of secondary standards. In addition, these previous reviews did not consider the common
21 pathways of impact for both nitrogen and sulfur acting on the same ecosystem endpoints.
22 Three issues arise that call into question the ecological relevance of the current structure
23 of the secondary standards for NOx and SOx. One issue is the exposure period that is relevant
24 for ecosystem impacts. The majority of deposition related impacts are associated with
25 depositional loads that occur over periods of months to years. This differs significantly from
26 exposures associated with hourly concentrations of NO2 and SO2 as measured by the current
27 standards. Even though the NO2 standard uses an annual average of NO2, it is focused on the
28 annual average of 1-hour NO2 concentrations, rather than on a cumulative metric or an averaging
29 metric based on daily or monthly averages. A second issue is the choice of atmospheric
30 indicators. NO2 and SO2 are used as the component of oxides of nitrogen and sulfur that are
31 measured, but they do not provide a complete link to the direct effects on ecosystems from
September, 2010 4-4 Do Not Quote or Cite
-------
1 deposition of NOx and SOx as they do not capture all relevant chemical species of oxidized
2 nitrogen and oxidized sulfur that contribute to deposition. The ISA provides evidence that
3 deposition related effects are linked with total nitrogen and total sulfur deposition, and thus all
4 forms of oxidized nitrogen and oxidized sulfur that are deposited will contribute to effects on
5 ecosystems. This suggests that more comprehensive atmospheric indicators should be
6 considered in designing ecologically relevant standards. Further discussions of the need for
7 more ecologically relevant atmospheric indicators as well as the relative contributions to
8 deposition from various species of NOx and SOx can be in found in Chapters 5 and 6 below.
9 The third issue is that the current standards reflect separate assessments of the two individual
10 pollutants, NO2 and SO2, rather than assessing the joint impacts of deposition of NOx and SOx to
11 ecosystems, recognizing the role that each pollutant plays in jointly affecting ecosystem
12 indicators, functions, and services. The clearest example of this interaction is in assessment of
13 the impacts of acidifying deposition on aquatic ecosystems.
14 Acidification in an aquatic ecosystem depends on the total acidifying potential of the
15 deposition of both N and S from both atmospheric deposition of NOx and SOx as well as the
16 inputs from other sources of N and S such as reduced nitrogen and non-atmospheric sources. It is
17 the joint impact of the two pollutants that determines the ultimate effect on organisms within the
18 ecosystem, and critical ecosystem functions such as habitat provision and biodiversity.
19 Standards that are set independently are less able to account for the contribution of the other
20 pollutant. This suggests that interactions between NOX and SOX should be a critical element of
21 the conceptual framework for ecologically relevant standards. There are also important
22 interactions between NOx and SOx and reduced forms of nitrogen, which also contribute to
23 acidification and nutrient enrichment. Although the standards do not directly address reduced
24 forms of nitrogen in the atmosphere, e.g. they do not require specific levels of reduced nitrogen,
25 it is important that the structure of the standards address the role of reduced nitrogen in
26 determining the ecological effects resulting from deposition of atmospheric NOx and SOx.
27 Consideration will also have to be given to account for loadings coming from non-atmospheric
28 sources as ecosystems will respond to these sources as well.
29 In addition to the fundamental issues discussed above, the current structures of the
30 standards do not address the complexities in the responses of ecosystems to deposition of NOx
31 and SOx. Ecosystems contain complex groupings of organisms that respond in various ways to
September, 2010 4-5 Do Not Quote or Cite
-------
1 the alterations of soil and water that result from deposition of nitrogen and sulfur compounds.
2 Different ecosystems therefore respond in different ways depending on a multitude of factors
3 that control how deposition is integrated into the system. For example, the same levels of
4 deposition falling on limestone dominated soils have a very different effect than those falling on
5 shallow glaciated soils underlain with granite. One system may over time display no obvious
6 detriment while the other may experience a catastrophic loss in fish communities. This degree
7 of sensitivity is a function of many atmospheric factors which control rates of deposition as well
8 as ecological factors which control how an ecosystem responds to that deposition. The current
9 standards do not take into account spatial and seasonal variations not only in depositional
10 loadings but also in sensitivity of ecosystems exposed to those loadings.
11
12 4.3 To What Extent Do Current Monitoring Networks Provide A Sufficient Basis For
13 Determining The Adequacy of Current Secondary NOx and SOx Standards?
14
15 Staff have closely evaluated whether levels of N and S allowed by current standards are
16 requisite to protect public welfare. Doing so requires relating atmospheric concentrations with
17 deposition, exposure pathways, and measured effects on ecologic receptors. A combination of
18 monitoring and air quality model applications is useful in linking atmospheric concentrations to
19 ecological effects. There are over 1000 ground level monitoring platforms (Figures 4-1 and 4-2
20 and Table 4-1) that provide measurements of some form of atmospheric nitrogen or sulfur. The
21 key pollutants for this assessment are total oxidized nitrogen (NOy), total reduced nitrogen
22 (NHX), and total oxidized sulfur which is referenced herein as (SOx) and defined as the sum of
23 SO2 (gas) and particulate sulfate.. Total reactive oxidized atmospheric nitrogen, NOy, is defined
24 as NOX (NO and NO2) and all oxidized NOX products: NOy = NO2 + NO + HNO3 + PAN
25 +2N2O5 + HONO+ NO3 + organic nitrates + particulate NO3 (Finlayson-Pitts and Pitts, 2000).
26 This definition of NOY reflects the operational principles of standard measurement techniques in
27 which all oxidized nitrogen species are converted to nitrogen oxide (NO) through catalytic
28 reduction and the resulting NO is detected through luminescence. Thus, NOy is truly defined as
29 total oxidized nitrogen as converted to NO, essentially representing all oxidized nitrogen atoms.
30 NOy is not a strict representation of the all moles of oxidized nitrogen as the diatomic
31
September, 2010 4-6 Do Not Quote or Cite
-------
Table 4-1. Summary of Monitoring Networks.
Network
Number
of Sites
Species Measured
Sampling
Frequenc
y
Comments
All Sulfur Sites
NCore
SEARCH
S02
PM Speciation
IMPROVE
CASTNET
82
8
751
242
215
88
S02
S02
S02
Sulfates
Sulfates
Sulfates
Hourly
Hourly
Hourly
24-hour
24-hour
Weekly
Ave.
Includes 20 rural sites
Includes 3 rural sites
NAMS/SLAMS/PAM
S for 2008
Measurements of
Sulfates (88403)
identified in AQS for
Trends and
Supplemental
Speciation monitoring
type for 2008
IMPROVE
Monitoring Sites with
Measurements of
Sulfates (88403)
identified in AQS
EPA & NFS
September, 2010
4-7
Do Not Quote or Cite
-------
1
2
Table 4-1 Summary of Monitoring Networks (continued)
Network
Number of Species Sampling Comments
Sites Measured Frequency
All Nitrogen Sites
NCore
SEARCH
PAMS
SLAMS
NOY
IMPROVE
CASTNET
82
8
119
643
59
214
88
NO/NOV
NO/NO2/NOy/HNO
3
NC-2/NOx
NO/NO2/NOx/NOy
NOy
Nitrates
Nitrates
Hourly
Hourly
Hourly
Hourly
Hourly
24-hour
Weekly
Ave.
Includes 20 rural sites
Includes 3 rural sites
Official sites as of
12/09
All SLAMS
Monitoring Sites with
Measurements of NO,
NO2,NOXorNOYin
2009 identified in
AQS
All Monitoring Sites
with Measurements of
NOY in 2009
identified in AQS,
regardless of
Monitoring Type
MPROVE Monitoring
Sites with
Measurements of
Nitrates (88306)
identified in AQS
EPA & NFS
September, 2010
4-8
Do Not Quote or Cite
-------
1
2
3
4
MAPI (AIIN)
NCore, NOY(2009), SEARCH, PAMS/SLAMS, CASTNET, IMPROVE
Rural NCoie
* SEARCH
EH Rural SEARCH
PAMS_NO- HO2- NQX-NQY_20Q!
SLAMS_N 0-NQ2-N OX- HOY
CA3TNET-NPS
CASTNET-EPA
IMPRQVE_Hrtr3tes_20Q6
6 Figure 4-1 Routinely operating surface monitoring stations measuring forms of
7 atmospheric nitrogen.
September, 2010
4-9
Do Not Quote or Cite
-------
1
2
3
4
5
MAP 6-1 a (All S)
NCore, SO2(2008), SEARCH, CASTNET, IMPROVE, and
Trends/Supplemental Speciation Sites (2008)
- SO2(2008) includes NAMS / SLAMS / PAMS --
NCore
Rural NCore
SEARCH
Rural SE ARCH
CASTNET-NPS
CASTNET-EPA
Speciation_Sulfates_2008
IMPROVE SulMes 2006
Figure 4-2 Routinely operating surface monitoring stations measuring forms of
atmospheric sulfur. All site locations measure both 862 and sulfate except for
the green 862 only sites.
September, 2010
4-10
Do Not Quote or Cite
-------
1 nitrogen species such as N2Os yield 2 moles of NO. This definition is consistent with the
2 relationship between atmospheric nitrogen and acidification processes as the reported NOy
3 provides a direct estimate of the potential equivalents available for acidification. Total reduced
4 nitrogen (NHx) includes ammonia, NH?, plus ammonium, NH4 (EPA, 2008). Reduced nitrogen
5 plus oxidized nitrogen is referred to as total reactive nitrogen. Total oxidized sulfur (SOx)
6 includes SO2 gas and particulate sulfate, SO/t. These species are converted to mass of sulfur
7 which is used directly, or converted to charge equivalents, in deposition analyses linking
8 atmospheric deposition and ecosystem models. Ammonium and sulfate are components of
9 atmospheric particulate matter as well as directly measured and modeled in precipitation as
10 direct deposition components. As discussed in this section, there are only very limited routine
11 measurements of total oxidized and reduced nitrogen. In addition, existing monitoring networks
12 do not provide adequate geographic coverage to fully assess concentrations and deposition of
13 reactive nitrogen and sulfur in and near sensitive ecosystems.
14 The principal monitoring networks include the regulatory based State and Local Air
15 Monitoring Stations (SLAMS) providing mostly urban-based SO2, NO and NOX, the PM2 5
16 chemical speciation networks Interagency Monitoring of Protected Visual Environments
17 (IMPROVE) and EPA's Chemical Speciation Network (CSN) providing particle bound sulfate
18 and nitrate, and the Clean Air Status and Trends Network (CASTNET) providing weekly
19 averaged values of SO2, nitric acid, and particle bound sulfate, nitrate and ammonium. The
20 private sector supported South Eastern Aerosol Research and Characterization (SEARCH) Study
21 network of 4-8 sites in the Southeast provides the only routinely operating source of true
22 continuous NO2, ammonia, and nitric acid measurements. SEARCH also provides PM2 5 size
23 fractions of nitrate and sulfate. Collectively, the SLAMS, Photochemical Assessment
24 Measurement Stations (PAMS), SEARCH and NCore networks will provide over 100 sites
25 measuring NOy (Figure 4-3). The NCore network (Scheffe et al., 2009) is a multiple pollutant
26 network with co-located measurements of key trace gases (CO, SO2, Os, NO and NOy), PM2.5
27 and PM(i0-2.5) mass and PM2.5 chemical speciation. Additional air pollutants, particularly volatile
28 organic compounds (VOCs), will be measured at those sites that are part of the existing PAMS
29 and National Air Toxics Trends (NATTS) platforms. The NATTS (EPA, 2008) include 27
30 stations across the U.S. that monitor for a variety of hazardous air pollutants and are intended to
31 remain in place to provide a long-term record. Additional
September, 2010 4-11 Do Not Quote or Cite
-------
MAP 4
Currents Planned Routine NOY Monitoring Sites
NCore, NOY{2009), SEARCH
2 Figure 4-3. Anticipated network of surface based NOY stations based on 2009 network
3 design plans. The NCore stations are scheduled to be operating by January,
4 2011.
5
6 measurements of ammonia and possibly true NC>2 are under consideration. True NC>2 is noted to
7 differentiate from the NC>2 determined through routine regulatory networks that have known
8 variable positive bias for NC>2. The network currently is being deployed and expected to be
9 operational with nearly 75 sites by January, 2011. The sites are intended to serve as central site
10 monitors capturing broadly representative (e.g., not strongly influenced by nearby sources) air
11 quality in a suite of major and mid size cities, and approximately 20 sites are located in rural
12 locations.
13 There are significant measurement gaps for characterizing NOY, NHX and SO2 in the
14 nations ambient air observation networks (EPA, 2008) that lead to greater reliance on air quality
15 modeling simulations to describe current conditions. National design of routinely operating
16 ambient air monitoring networks is driven mostly by data uses associated with implementing
September, 2010
4-12
Do Not Quote or Cite
-------
1 primary NAAQS, with noted exceptions of the CASTNET and IMPROVE networks. In
2 addition to significant spatial gaps in sensitive ecosystem areas that arise from a population
3 oriented network design, the current measurements for primary and secondary nitrogen are
4 markedly different and in some instances of negligible value for secondary NOx and SOx
5 standards. For example, a true NOx (NO plus NO2) measurement typically would capture less
6 than 50% (see discussion below) of the total regional NOy mass in rural locations as the more
7 aged air masses contain significant oxidized nitrogen products in addition to NOx. Note that the
8 NOx monitors used for NAAQS primary compliance purposes capture varying amounts of
9 transformed nitrogen species; however, the method provides biased low estimates with
10 significant airshed induced variability relative to true NOy. With the exception of the SEARCH
11 network in the Southeast, there are virtually no routine networks that measure ammonia,
12 although EPA is considering options for ammonia sampling in CASTNET and NCORE
13 networks. Ammonium is reported in EPA chemical speciation networks, although the values are
14 believed to be biased low due to ammonia volatization.
15 CASTNET provides mostly rural measurements of SO2, total nitrate, and ammonium, and
16 affords an existing infrastructure useful for future monitoring in support of a potential NOx and
17 SOx secondary standard. However, the lack of NOy, SOx and NHX measurements in sensitive
18 ecosystems will require attention in the N/S secondary standard proposal.
19 As a result of the limited monitoring networks for NOy and SOx in sensitive ecosystems,
20 we are unable to use current ambient monitoring data to adequately link measured current
21 atmospheric concentrations to ecological effects transmitted through deposition. At this time for
22 the purpose of illustrating current atmospheric conditions, we supplement the available
23 monitoring data with the use of sophisticated atmospheric modeling conducted using EPA's
24 CMAQ model (as discussed in Chapter 7).
25
26 4.3.1 To what extent does the NADP monitoring network provide an adequate
27 characterization of deposition and what are the major limitations?
28 The National Atmospheric Deposition Program (NADP) includes approximately 250 sites
29 (Figure 4-4) across the U.S. providing annual total wet deposition based on weekly averaged
30 measures of wet deposition of nitrate, ammonium, sulfate and other ions based on the
31 concentrations of these ions in precipitation samples. Meteorological models have difficulty in
September, 2010 4-13 Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
capturing the correct spatial and temporal features of precipitation events, raising the importance
of the NADP as a principal source of precipitation chemistry. The NADP has enabled several
organizations to participate in a measurement program with a centralized laboratory affording
measurement and analysis protocol consistency nationwide. Virtually every CASTNET site is
located at an NADP site and the combined NADP/CASTNET infrastructure is a starting point for
discussions addressing future NOx and SOx monitoring needs. The organic bound nitrogen is
not analyzed routinely in NADP samples. Consideration might be given to adding NADP sites
in locations where ambient air monitoring is conducted to assess compliance with a secondary
NOX/ SOX standard.
Ammonium ion wet deposition, 2005
Ammonium as NH4*
Sites not pictured:
AK03 0.3 kg/ha
VI01 0.4 kg/ha
1
National Atmospheric Deposition Program/National Trends Network
http://nadp.sws.uiuc.edu
Figure 4-4. Location of approximately 250 National Atmospheric Deposition Monitoring
(NADP) National Trends Network (NTN) sites illustrating annual
ammonium deposition for 2005. Weekly values of precipitation based nitrate,
sulfate and ammonium are provided by NADP.
September, 2010
4-14
Do Not Quote or Cite
-------
1 4.3.2 What are the relative strengths and important gaps in existing atmospheric
2 monitoring networks to address a combined NOx/SOx secondary standard?
O
4 Of the currently operating monitoring networks, precipitation based sulfate, ammonium
5 and nitrate measurements provided by the NADP are the most relevant measurements that would
6 support the secondary standard as they provide atmospheric deposition inputs that drive
7 ecosystem models, and NADP site locations generally include acid sensitive areas. However,
8 there are significant gaps in ambient air (aerosols and gases) monitoring networks for the
9 measurement of the likely ambient indicators of NOy, SC>2, and SO/t. CASTNET filter packs
10 provide the most relevant source of ambient sulfate (804) measurements as the open inlet of the
11 filter packs incorporates the full range of particle sizes that contribute to deposition. The 862
12 measurements from CASTNET represent about 10% of all 862 sites nationally, but are
13 especially relevant based on their locations in rural and regional settings, although CASTNET is
14 not as spatially extensive (breadth and resolution) as the NADP network of precipitation sites.
15 Although CASTNET does provide measurements of total ambient nitrate, other oxidized
16 nitrogen species constituting a more complete NOy budget are not captured. In their current
17 configuration, the State and local monitoring networks virtually offer no support for a secondary
18 NOx/ SOx standard due to their urban based site orientation and exclusion of important oxidized
19 nitrogen species (e.g., nitrates and PAN). The chemical speciation networks, including rural
20 based IMPROVE, all provide ambient sulfate measurements based on on a 2.5ji size cut. While
21 the sulfate mass within that size fraction may constitute 80% or greater of the ambient sulfate
22 budget, the missing larger size particles can contribute significantly to sulfate deposition due to
23 their relatively high gravitationally driven deposition velocities. Finally, there are virtually no
24 ambient ammonia measurements routinely collected in acid sensitive areas. CASTNET does
25 provide ammonium measurements, but the routine speciation networks that report ammonium
26 have expected artifacts due to ammonia offgassing from nylon filters.
27
28 Although this summary of existing networks suggests significant challenges in meeting
29 the monitoring needs of a new NOx/ SOx standard, the networks do serve as a useful building
30 block for moving forward. The site locations of NADP and CASTNET offer an infrastructure
31 to accommodate additional instruments. The NCORE network has introduced nearly 75 NOY
September, 2010 4-15 Do Not Quote or Cite
-------
1 trace level 862 monitors that are establishing operational familiarity and a basis for instrument
2 performance characterization. In many cases, acid sensitive areas will be strongly influenced
3 by regional transport of pollutants which typically is associated with relatively homogeneous
4 spatial concentration patterns which allows for a correspondingly greater range of spatial
5 representativeness of monitoring sites. Consequently, the expected burden on monitoring
6 resources may be realistically dampened by the available infrastructure and expected
7 homogeneity of air concentration patterns. A more thorough assessment of the adequacy of
8 existing networks is predicated on identification of the area wide boundaries of the acid sensitive
9 areas of concern which will initially e developed in the second PAD.
10
11 4.3.3 How do we characterize deposition through monitoring and models?
12
13 Routinely available directly measured precipitation to quantify wet deposition of sulfur
14 and nitrogen species is provided through the NADP. Dry deposition is not a directly measured
15 variable in routine monitoring efforts. It is important to pursue the development of direct dry
16 deposition measurements to improve model parameterizations of deposition processes and
17 possibly evolve into routine operations. Estimates of dry deposition based on observations are
18 provided through the CASTNET program. However, dry deposition is a calculated value
19 represented as the product of ambient concentration (either observed or estimated through air
20 quality modeling) and deposition velocity, DeptDry = vf^ • C^mb
21 Deposition velocity is modeled as a mass transfer process through resistance layers
22 associated with the canopy, uptake by vegetation, water and soil which collectively are
23 influenced by micrometeorology, land surface and vegetation types and species specific
24 solubility and reactivity. Dry deposition is calculated through deposition velocity models
25 capturing these features and using species specific ambient air concentrations. This approach
26 conceptually is similar using either observed or modeled air concentrations. Dry deposition
27 estimates from the Community Multi-scale Air Quality (CMAQ) model (EPA, 1999) have been
28 used in this assessment to provide spatially more resolved and extensive estimates of dry
29 deposition for sulfur and all reactive nitrogen (oxidized and reduced) species (CASTNET does
30 not capture important gases such as nitrogen dioxide, ammonia and peroxyacetyl nitrate). All
31 of the relevant meteorological, land use, vegetation and elevation data required to estimate
September, 2010 4-16 Do Not Quote or Cite
-------
1 deposition velocities are generated or accessible in the CMAQ and/or meteorological pre-
2 processors.
3
4 4.3.2.1 Why are we using CMAQ to model deposition? How are we using it? Why is
5 CMAQ the right model to use? What is the spatial and temporal resolution of
6 CMAQ? What are the model years? What are the limitations to CMAQ?
7
8 CMAQ provides a platform that allows for a consistent mass accounting approach across
9 ambient concentrations and dry and wet deposition values. Recognizing the limitations of
10 ambient air networks, CMAQ was used to estimate dry deposition to complement NADP wet
11 deposition for MAGIC modeling and for the first-order acidity balance (FAB) critical load
12 modeling. CMAQ promotes analytical consistency and efficiency across analyses of multiple
13 pollutants. EPA's Office of Research and Development continues to enhance the underlying
14 deposition science in CMAQ. For the purposes of this policy assessment, CMAQ provides a
15 consistent platform incorporating the atmospheric and deposition species of interest over the
16 entire United States. The caveats and limitations of the use of model predictions are largely
17 associated with the general reliance on calculated values, rather than on measurements. Model
18 evaluation addressing the comparison of predictions with observed values is addressed in the
19 REA and summarized in Chapter 7 of this PA. Currently, there are efforts to improve a number
20 of nitrogen related processes in CMAQ, recognizing comparatively less uncertainty with the
21 treatment of sulfur. Active areas of model process improvement are in the treatment of
22 lightning generated NOx and the transference of nitrogen between atmospheric and terrestrial
23 and aquatic media, often referred to as bi-directional flux. Lightning NOx potentially provides
24 a significant contribution to wet deposition as the resulting NOx is rapidly entrained into aqueous
25 cloud processes. Both the thermodynamics of soil processes and mass transfer of nitrogen
26 species across the surface-atmosphere interface is governed by an assortment of temperature,
27 moisture, advection and concentration patterns. These processes and mass transfer relationships
28 are coupled within the emissions, meteorological, and chemical simulation processes and
29 associated surface/vegetation and terrain information incorporated in or accessed by the CMAQ.
30 In addition to research activities to improve the characterization of nitrogen-related processes in
31 CMAQ, efforts are also underway to improve the general characterization of ammonia emissions
September, 2010 4-17 Do Not Quote or Cite
-------
1 which remains as an area of large uncertainty due to limited source data and the ubiquitous
2 nature of these emissions. Another challenge for regional/national air quality modeling is
3 properly representing the effects on pollutant concentrations, precipitation and therefore
4 deposition of variable terrain features, particularly steep mountain-valley gradients and the
5 interfaces to wide open basins encountered in the Western United States.
6 The CMAQ was used in this assessment because it is the state of science model for
7 simulating sources, formation, and fate of nitrogen and sulfur species. In addition to undergoing
8 periodic independent scientific peer review, CMAQ bridges the scientific and regulatory
9 communities as it is used extensively by EPA for regulatory air quality assessments and rules.
10 CMAQ provides hourly estimates of the important precursor, intermediate and secondarily
11 formed species associated with atmospheric chemistry and deposition processes influencing
12 ozone, particulate matter concentrations and sulfur and nitrogen deposition. Simulations based
13 on horizontal spatial scale resolutions of 12 km and 36 km were used in this policy assessment
14 for 2002-2005.
15
16 4.4 What Is Our Best Characterization of Atmospheric Concentrations Of NOy and
17 SOX, and Deposition Of N And S?
18
19 Air quality models and blending of model results and observations are used to
20 characterize current environmental state conditions due to the relative sparseness of monitoring
21 coverage in sensitive ecosystems as well as gaps in coverage for specific atmospheric species of
22 N and S most relevant to deposition, such as NOy, in available monitoring platforms.
23
24 4.4.1 What are the current atmospheric concentrations of reactive nitrogen, NOy,
25 reduced nitrogen, NHX, sulfur dioxide, SOi, and sulfate, SO4?
26
27 To provide information for use in characterizing the adequacy of the current standards, we assess
28 the best available data for estimating the ambient concentrations of the major sources of
29 atmospheric nitrogen and sulfur across the U.S. Acidification and nutrient enrichment processes
30 are largely dependent on the cycling of total nitrogen and sulfur species. From an atmospheric
31 perspective, it is convenient and consistent with current measurement and modeling frameworks
September, 2010 4-18 Do Not Quote or Cite
-------
1 to consider the reduced and oxidized forms of atmospheric nitrogen. Virtually all atmospheric
2 sulfur is considered oxidized sulfur in the forms of particulate bound sulfate and gaseous sulfur
3 dioxide. In order to assess current concentrations of reactive nitrogen and sulfur, we evaluated
4 data available from the existing monitoring networks as well as from the CMAQ model.
5 Regarding the monitoring data, there are a number of important issues in understanding the
6 measurements of NOy provided by different monitoring networks. In principle, measured NOy is
7 based on catalytic conversion of all oxidized species to NO followed by chemiluminescence NO
8 detection. We recognize the caveats associated with instrument conversion efficiency and
9 possible inlet losses. The CMAQ treats the dominant NOy species as explicit species while the
10 minor contributing non-PAN organic nitrogen compounds are aggregated. Atmospheric
11 nitrogen and sulfur are largely viewed as regional air quality issues due to the importance of
12 chemical conversion of primary emissions into secondarily formed species, a combination of
13 ubiquitous sources, particularly mobile source emissions of NOx, and elevated emissions of NOx
14 and SO2 that aid pollutant mass dispersal and broader physical transport over large distances. In
15 effect, the regional nature is due to both transport processes as well as the relatively ubiquitous
16 nature of sources combined with chemical processes that tend to form more stable species with
17 extended atmospheric lifetimes. This regionalized effect, particularly throughout the eastern
18 United States, dominates the overall patterns discussed below of secondarily formed species such
19 as sulfate or NOy, which is an aggregate of species with the more aged air masses consisting
20 largely of chemically processed air dominated by secondarily formed peroxyacetyl nitrate
21 (PAN), parti culate nitrate and nitric acid.
22 Nationwide maps of CMAQ-predicted 2005 annual average NOY, NHX (NH3 and NH4),
23 NHa, NH/t, SOx, SO/t, and SO2 are provided in Figures 4-5 through 4-11 respectively. Given the
24 considerable gaps in air quality observation networks as discussed in the REA and ISA (2008),
25 modeled concentration patterns are used here to illustrate national representations of current air
26 quality conditions for nitrogen and sulfur. The 2005 model year reflects the most recent
27 available simulation for inclusion in this policy assessment. In addition, Figures 4-12 and 4-13
28 provide maps of 2005 annual average SO2 and SO4, respectively based on CASTNET
29 observations. Site specific annual average 2005 NOy measured concentrations at SLAMS
30 (Figure 4-14) are typically are less than 40 ppb. The spatial patterns for the 2005 modeled and
31 observed NOy and SO2 concentrations are similar to the 2002 CMAQ-based maps provided in
September, 2010 4-19 Do Not Quote or Cite
-------
1 the REA, largely capturing the influence of major source regions throughout the nation. A
2 spreading of the oxidized sulfur fields (Figures 4-5 and 4-6), relative to 862, is consistent with
3 sulfate transformation and associated air mass aging and transport. Ammonia and ammonium
4 concentration patterns (Figure 4-4) are influenced strongly by the ammonia emissions
5 distribution, with marginal spreading associated with the formation of NFLi. The NHx fields are
6 more strongly influenced by source location, relative to sulfur, based on the fast removal of
7 atmospheric ammonia through deposition. However, recent incorporation of ammonia bi-
8 directional flux treatment (see Chapter 7) does reduce NFL? spatial gradients. Total deposition
9 for nitrogen and sulfur (Figures 4-15 and 4-16) basically follow the patterns of ambient air
10 concentrations. The contribution of reduced nitrogen to total nitrogen deposition (Figure 4-17)
11 illustrates the strong influence of agricultural based ammonia emissions, particularly in upper
12 midwest and eastern North Carolina.
13
14 The 2005 ambient conditions indicate that the current SC>2 and NC>2 secondary standards
15 are not exceeded (Figures 4-18 and 4-19) in locations where ecological effects have been
16 observed, and where critical loads of nitrogen and sulfur are exceeded. This information is
17 consistent with the fact that NC>2 accounts for only a fraction of NOy, and thus decreases in NO
18 and NC>2 emissions that result in attaining the current secondary NC>2 standard would not be
19 expected to fully address deposition of NOy in acid sensitive areas that generally are not
20 represented by urban oriented NC>2 monitoring locations. The map in Figure 4-20 further
21 illustrates this point by showing that the contribution of NC>2 to NOy is often less than 50% in
22 rural areas. Neither NOy nor NC>2 concentrations correlate well with total oxidized nitrogen
23 deposition (Figure 4-21), based on the annual average values in each 12 km CMAQ grid cell.
24 The lack of correlation between NOy and nitrogen deposition is largely due to the inclusion of
25 NOy species with low deposition velocity, primarily NC>2. While NC>2 does not reflect the
26 majority of oxidized nitrogen in rural environments, the species remains a significant contributor
27 to the total ambient NOy budget. In contrast, species with high deposition velocities such as
28 nitric acid correlate well with oxidized nitrogen deposition. The temporal correlations between
29 NOy and deposition are likely to be similar to the nitric acid to deposition relationship when
30 averaged over larger a spatial area as the influence of species with low deposition velocities is
September, 2010 4-20 Do Not Quote or Cite
-------
1 minimized in rural locations associated with acid sensitive areas. Those relationships will be
2 explored in the final PAD.
3
4
5
6
AMAD 2005af CMAQ — NOy (ppb)
>= 1.0 to < 30
>= 3.0 to < 5.0
>= 5.0 to < 7.0
>= 7,0 to < 10.0
>= 10.0 to < 25 0
>= 25.0
Figure 4-5. 2005 CMAQ modeled annual average NOy (ppb; see Table 1-1 for unit
conversions).
September, 2010
4-21
Do Not Quote or Cite
-------
>=1.0to<3.0
>=3.0to-=S.Q
>= 50 to < 7.0
>= 7.0 to < 10.0
>= 10.0
AMAD 2005af CMAQ — NHx (ug/rn3)
3 Figure 4-6. 2005 CMAQ modeled annual average total reduced nitrogen (NHX) (as ug/m3
4 nitrogen - see Table 1-1 for unit conversions)
September, 2010
4-22
Do Not Quote or Cite
-------
2
3
4
AMAD 2005af CMAQ — NH3 (ug/m3)
Figure 4-7. 2005 CMAQ modeled annual average ammonia, NHs, (as ug/m N; see Table
1-1 for unit conversions )
September, 2010
4-23
Do Not Quote or Cite
-------
2
3
4
>=1.0to<3.0
>=3.0to-=S.Q
>= 50 to < 7.0
>= 7.0 to < 10.0
>= 10.0
AMAD 2005af CMAQ — NH4 (ug/rn3)
Figure 4-8. 2005 CMAQ modeled annual average ammonium, NELt, (as ug/m3 N; see
Table 1-1 for unit conversions)
September, 2010
4-24
Do Not Quote or Cite
-------
2
3
4
5
6
7
226 -
211 -
196 -
181 -
166 -
151 -
136 -
121 -
106 -
91 -
76 -
61 -
46 -
31 -
16 -
1 -
[1]=cctm_N1 a_2005af_Q5b_v3.4beta3.12EUS1 .yearly.avgaconc.sumdepl
u
1 40 79 113 157 196 235
April 5, 3 00:00:00 UTC
Min (3, 100) = 0.188, Max (84, 49) = 55.956
274
12.000
10.524
n
9.047-
7.571 -
6.094-
4.618-
3.141
1.665-
0.188-
Figure 4-9. 2005 CMAQ modeled annual average SOX, (as ug/m S from SO2 and SO4;
see Table 1-1 for unit conversions).
September, 2010
4-25
Do Not Quote or Cite
-------
2
3
4
226
211
196
181
166
151
136
121
106
91
76
61
46
31
16
[1]=cctm_N1a_2005af_05b_v3.4beta3.12EUS1.yearly.avgaconc.sumdep1
a"
12.000
10.504
I
9.003-
7.512
E
di
6.016-
4.520-
3.024-
1.523-
0.032-
Figure 4-10. 2005 CMAQ modeled annual average SOi (as ug/m3 S; see Table 1-1 for unit
conversions )
September, 2010
4-26
Do Not Quote or Cite
-------
[1]=cctm_N1 a_2005af_Q5b_v3.4beta3.1 2EUS1 .yearly.avgaconc.sumdepl
226 -
211 -
196 -
161 -
166 -
151 -
136 -
121 -
106 -
91 -
76 -
61 -
46 -
31 -
16 -
1 -
2.781
2.449-
2.117-
1.785-
.453
1.121
0.789-
0.457-
0.124-
2 1 40 79 113 157 196 235 274
3 Figure 4-11. 2005 CMAQ modeled annual average SC>4 (as ug/m3 S; see Table 1-1 for unit
4 conversions).
September, 2010
4-27
Do Not Quote or Cite
-------
1
2
3
4
5
6
1
Figure 4-12. 2005 annual average sulfur dioxide concentrations based on CASTNET
generated by the Visibility Information Exchange Web Sysytem (VIEWS).
See Table 1-1 for unit conversions
September, 2010
4-28
Do Not Quote or Cite
-------
1
2
3
4
5
6
Figure 4-13 2005 annual average sulfate concentrations based on CASTNET generated
by the Visibility Information Exchange Web Sysytem (VIEWS), [interpolating
relative sparse data can produce unrealistic concentration plumes as demonstrated
in the central U.S.] see Table 1-1 for unit conversions.
September, 2010
4-29
Do Not Quote or Cite
-------
Annual Average NOY Concentrations (2005)
1
2
3
Figure 4-14. Annual average 2005 NOy concentrations from reporting stations in the Air
Quality System (AQS). see Table 1-1 for unit conversions.
September, 2010
4-30
Do Not Quote or Cite
-------
2
O
4
5
<2.0
>= 2.0 to < 3.0
>= 3.0 to < 4.0
>- 4.0 to < 5.0
>= 5.0 to < 7.0
>=7.0to<9.0
>= 9.0 to < 14.0
>= 14.0 to < 20.0
>= 20.0
AMAD 2005af CMAQ —
Oxidized Nitrogen Deposition ( kgN/Ha/Yr)
Figure 4-15. 2005 CMAQ modeled oxidized nitrogen deposition (kgN/ha-yr). see Table 1-
1 for unit conversions
September, 2010
4-31
Do Not Quote or Cite
-------
1
2
3
»=1.0to<2.0
«= 2.0 lo < 3.0
>=3.0(o<6.0
>=6.0IO< 10.0
>= 10.0 to < 16.0
>=1S-Oto<24.0
>= 24 0 10 < 30.0
>= 30.0
AMAD 200Saf CMAQ —
Oxidized Sulfur Deposition { kgS/Ha/Yr ) „
Figure 4-16. 2005 CMAQ modeled oxidized sulfur deposition (kgS/ha-yr). see Table 1-1
for unit conversions
September, 2010
4-32
Do Not Quote or Cite
-------
Fraction of Reduced to Total N Deposition
239 -
225 -
211 -
197 -
183 -
169 -
155 -
141 •
127 •
113 •
99 -
85 •
71 -
57 •
43 -
29 -
15 •
1
0.895
0.793
0.486-
0.334-
0.282-
0.180-
0.078-
1 1 40 79 118 157 196 235 274
2 Figure 4-17 2005 CMAQ derived annual average ratio of reduced to total nitrogen
3 deposition.
September, 2010
4-33
Do Not Quote or Cite
-------
3-hr Max SO2 Concentrations (2005)
2
3 Figure 4-18. Three hour average maximum 2005 SOi concentrations based on the SLAMS
4 reporting to EPA's Air Quality System (AQS) data base The current SO2
5 secondary standard based on the maximum 3 hour average value is 500 ppb, a
6 value not exceeded. While there are obvious spatial gaps, the majority of these
7 stations are located to capture maximum values generally in proximity to major
8 sources and high populations. Lower relative values are expected in more remote
9 acid sensitive areas, see Table 1-1 for unit conversions
September, 2010
4-34
Do Not Quote or Cite
-------
Annual Average NO2 Concentrations (2005)
\f
2 Figure 4-19 Annual average 2005 NOi concentrations based on the SLAMS reporting to
3 EPA's Air Quality System (AQS) data base. The current NC>2 secondary
4 standard is 53 ppb, a value well above those observed. While there are obvious
5 spatial gaps, the stations are located in areas of relatively high concentrations in
6 highly populated areas. Lower relative values are expected in more remote acid
7 sensitive areas, see Table 1-1 for unit conversions
September, 2010
4-35
Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
Layer 1 (NOY[1 J- NO2[2])/ NOY[1]
231
221 -
211
201
191
181
171
161
151
141
131 •
>- 121 •
111
101
91
81
71
Figure 4-20
v
121 141 161 181
X
December 31,0002 00:00:00 UTC
Min (10,34) = 0.175, Max (5,10) = 0.915
0.915
0.822
D.730
0.637
Q. 0.545
Q.
0.453
0.360
D.268
0.175
I
2005 CMAQ derived annual average ratio of (NOy - NOiXNOy. The fraction
of NC>2 contributing to total NOy generally is less than 50% in the Adirondack
and Shenandoah case study areas. The ratio (dimensionless - scale units of ppbv
are an automated output) reflects the relative air mass aging associated with
transformation of oxidized nitrogen beyond NO and NC>2 as one moves from
urban to rural locations.
September, 2010
4-36
Do Not Quote or Cite
-------
Adirondacks
60
30
4 5
Concentrationippb)
3 Figure 4-21 Scatter plots of total oxidized nitrogen deposition with average annual NOi,
4 HNO3 and NOY concentrations on each 12 km2 grid based on 2005 CMAQ
5 results for the Adirondack region.
6
September, 2010 4-37 Do Not Quote or Cite
-------
1
2 4.5 Are Adverse Effects On The Public Welfare Occurring Under Current Air Quality
3 Conditions For NO2 And SO2 And Would They Occur If The Nation Met The
4 Current Secondary Standards?
5
6 In the previous sections we have established that almost all areas of the U.S. were at
7 concentrations of SO2 and NO2 below the levels of the current standards. In many locations, SO2
8 and NO2 concentrations are substantially below the levels of the standards. This pattern suggests
9 that levels of deposition and any effects on ecosystems due to deposition of NOx and SOx under
10 recent conditions are occurring even though areas meet or are below current standards. In this
11 section we focus on summarizing the evidence of effects occurring at deposition levels consistent
12 with recent conditions.
13 The ISA summarizes the available studies of relative nitrogen contribution and finds that
14 in much of the U.S., NOx contributes from 50 to 75 percent of total atmospheric deposition
15 relative to total reactive nitrogen that includes oxidized and reduced nitrogen species. [ISA
16 Section 2.8.4]. Although the proportion of total nitrogen loadings associated with atmospheric
17 deposition of nitrogen varies across locations (N deposition in the eastern U.S. includes locations
18 with greater than 9 kg N/ha-yr, and in the central U.S. high deposition locations with values on
19 the order of 6 to 7 kg N/ha-yr), the ISA indicates that atmospheric N deposition is the main
20 source of new anthropogenic N to most headwater streams, high elevation lakes, and low-order
21 streams. Atmospheric N deposition contributes to the total N load in terrestrial, wetland,
22 freshwater, and estuarine ecosystems that receive N through multiple pathways. In several large
23 estuarine systems, including the Chesapeake Bay, atmospheric deposition accounts for between
24 10 and 40 percent of total nitrogen loadings (U.S. EPA, 2000).
25 Atmospheric concentrations of SOx account for nearly all S deposition in the US. For the
26 period 2004-2006, mean S deposition in the U.S. was greatest east of the Mississippi River with
27 the highest deposition amount, 21.3 kg S/ha-yr, in the Ohio River Valley where most recording
28 stations reported 3 year averages >10 kg S/ha-yr. Numerous other stations in the East reported S
29 deposition >5 kg S/ha-yr. Total S deposition in the U.S. west of the 100th meridian was
30 relatively low, with all recording stations reporting <2 kg S/ha-yr and many reporting <1 kg
September, 2010 4-38 Do Not Quote or Cite
-------
1 S/ha-yr. S was primarily deposited in the form of wet 864 2 followed in decreasing order by a
2 smaller proportion of dry 862 and a much smaller proportion of deposition as dry SC>42 .
3 New scientific evidence exists to address each of the areas of uncertainty raised in the
4 previous reviews (summarized above in section 1.4). Based on the new evidence, the current
5 ISA concludes that:
6 (1) The evidence is sufficient to infer a causal relationship between acidifying
7 deposition (to which both NOx and SOx contribute) and effects on
8 biogeochemistry related to terrestrial and aquatic ecosystems; and biota in
9 terrestrial and aquatic ecosystems.
10 (2) The evidence is sufficient to infer a causal relationship between N deposition,
11 to which NOX and NHX contribute, and the alteration of A) biogeochemical
12 cycling of N and carbon in terrestrial, wetland, freshwater aquatic, and coastal
13 marine ecosystems; B) biogenic flux of methane (CH4), and N2O in terrestrial
14 and wetland ecosystems; and C) species richness, species composition, and
15 biodiversity in terrestrial, wetland, freshwater aquatic and coastal marine
16 ecosystems.
17 (3) The evidence is sufficient to infer a causal relationship between S deposition
18 and increased Hg methylation in wetlands and aquatic environments.
19 Subsequent to the previous review of the NOx secondary standard, a great deal of
20 information on the contribution of atmospheric deposition associated with ambient NOX has
21 become available. In Chapter 3 of the REA a thorough assessment is provided of the
22 contribution of NOx to nitrogen deposition throughout the U.S., and the relative contributions of
23 ambient NOx and reduced forms of nitrogen. Staff concludes that based on that analysis,
24 ambient NOx is a significant component of atmospheric nitrogen deposition, even in areas with
25 relatively high rates of deposition of reduced nitrogen. In addition, staff concludes that
26 atmospheric deposition of oxidized nitrogen contributes significantly to total nitrogen loadings in
27 nitrogen sensitive ecosystems.
28 As discussed throughout the REA document, there are several key areas of risk that are
29 associated with ambient concentrations of NOx and SOx. As noted earlier, in previous reviews
30 of the NOx and SOx secondary standards, the standards were designed to protect against direct
31 exposure of plants to ambient concentrations of the pollutants. A significant shift in
September, 2010 4-39 Do Not Quote or Cite
-------
1 understanding of the effects of NOx and SOx has occurred since the last reviews, reflecting the
2 large amount of research that has been conducted on the effects of deposition of nitrogen and
3 sulfur to ecosystems. The most significant risks of adverse effects to public welfare are those
4 related to deposition of NOx and SOx to both terrestrial and aquatic ecosystems. These risks fall
5 into two categories: acidification and nutrient enrichment. These made up the emphasis of the
6 REA, and are most relevant to evaluating the adequacy of the existing standards in protecting
7 public welfare from adverse ecological effects.
8
9 4.5.1 To what extent do the current NOx and SOx secondary standards provide
10 protection from adverse effects associated with deposition of atmospheric NOx and SOx
11 which results in acidification in sensitive aquatic and terrestrial ecosystems?
12
13 The focus of the REA case studies was on determining whether deposition of sulfur and
14 oxidized nitrogen in locations where ambient NOx and SOx was at or below the current
15 standards was resulting in acidification and related effects. This review has focused on
16 identifying ecological indicators that can link atmospheric deposition to ecological effects
17 associated with acidification. NOx and SOx contribute to acidification in both aquatic and
18 terrestrial ecosystems, although the indicators of effects differ. Although there are some
19 geographic areas with both terrestrial and aquatic ecosystems that are vulnerable to acidification,
20 the case study areas do not fully overlap. The locations of the case studies evaluated in the REA
21 are shown on Figure 4-20.
September, 2010 4-40 Do Not Quote or Cite
-------
750 1,000
1
2 Figure 4-22 National map highlighting the nine case study areas evaluated in the REA.
3 4.5.1.1 Aquatic Acidification
4 Based on the case studies conducted for lakes in the Adirondacks and streams in
5 Shenandoah National Park, staff concludes that there is significant risk to acid sensitive aquatic
6 ecosystems at atmospheric concentrations of NOx and SOx at or below the current standards.
7 This conclusion is based on application of the MAGIC model to estimate the effects of
8 deposition at levels consistent with atmospheric NOx and SOx concentrations that are at or
9 below the current standards. An important ecological indicator for aquatic acidification effects is
10 acid neutralizing capacity (ANC) of a waterbody, and the case study focused on evaluating
11 whether locations were likely to be below critical values of ANC given deposition levels
12 associated with NOx and SOx atmospheric concentrations that meet the current standards. In
13 addition, the case studies assessed the ecological effects and some of the known ecosystem
14 services that are associated with different levels of ANC in order to associate levels of ANC with
September, 2010
4-41
Do Not Quote or Cite
-------
1 measures of public welfare that may be adversely affected by deposition levels consistent with
2 atmospheric concentrations of NOx and SOx that meet the current standards.
3 Staff concludes that the evidence and risk assessment support strongly a relationship
4 between atmospheric deposition of NOx and SOx and loss of ANC in sensitive ecosystems, and
5 that ANC is an excellent indicator of aquatic acidification. Staff also concludes that at levels of
6 deposition associated with NOx and SOx concentrations at or below the current standards, ANC
7 levels are expected to be below benchmark values that are associated with significant losses in
8 fish species richness (REA Section 4).
9 Many locations in sensitive areas of the U.S. have ANC levels below benchmark levels
10 for ANC classified as severe, elevated, or moderate concern (see Figure 2-1). The average
11 current ANC levels across 44 lakes in the Adirondack case study area is 62.1 jieq/L (moderate
12 concern). However, 44 percent of lakes had deposition levels exceeding the critical load for an
13 ANC of 50 |ieq/L, and 28 percent of lakes had deposition levels exceeding the critical load for an
14 ANC of 20 neq/L (REA Section 4.2.4.2). This information indicates that almost half of the 44
15 lakes in the Adirondacks case study area are at an elevated concern levels, and almost a third are
16 at a severe concern level. These levels are associated with greatly diminished fish species
17 diversity, and losses in the health and reproductive capacity of remaining populations. Based on
18 assessments of the relationship between number offish species and ANC level in both the
19 Adirondacks and Shenandoah areas, the number offish species is decreased by over half at an
20 ANC level of 20 |ieq/L relative to an ANC level at 100 |ieq/L (REA Figure 4.2-1). At levels
21 below 20 |ieq/L, populations of sensitive species, such as brook trout, may decline significantly
22 during episodic acidification events. When extrapolated to the full population of lakes in the
23 Adirondacks area using weights based on the EMAP probability survey (REA 4.2.6.1), 36
24 percent of lakes exceeded the critical load for an ANC of 50 jieq/L and 13 percent of lakes
25 exceeded the critical load for an ANC of 20 jieq/L.
26 Many streams in the Shenandoah case study area also have levels of deposition that are
27 associated with ANC levels classified as severe, elevated, or moderate concern. The average
28 ANC under recent conditions for the 60 streams evaluated in the Shenandoah case study area is
29 57.9 |ieq/L, indicating moderate concern. However, 85 percent of streams had recent deposition
30 exceeding the critical load for an ANC of 50 |ieq/L, and 72 percent exceeded the critical load for
31 an ANC of 20 |ieq/L. As with the Adirondacks area, this information suggests that significant
September, 2010 4-42 Do Not Quote or Cite
-------
1 numbers of sensitive streams in the Shenandoah area are at risk of adverse impacts on fish
2 populations under recent conditions. Many other streams in the Shenandoah area are likely to
3 experience conditions of elevated to severe concern based on the prevalence in the area of
4 bedrock geology associated with increased sensitivity to acidification suggesting that effects due
5 to stream acidification could be widespread in the Shenandoah area (REA 4.2.6.2).
6 In the ISA it is noted that significant portions of the U.S. are acid sensitive, and that
7 current deposition levels exceed those that would allow recovery of the most acid sensitive lakes
8 in the Adirondacks (ISA ES). In addition, because of past loadings, areas of the Shenandoah are
9 sensitive to current deposition levels (ISA ES). Parts of the West are naturally less sensitive to
10 acidification and subjected to lower deposition (particularly SOx) levels relative to the eastern
11 United States, and as such, less focus in the ISA is placed on the adequacy of the existing
12 standards in these areas, with the exception of the mountainous areas of the West, which
13 experience episodic acidification due to deposition.
14 While most (99 percent) of stream kilometers in the U.S. are not chronically acidified
15 under current conditions, a recent survey found sensitive streams in many locations in the U.S.,
16 including the Appalachian mountains, the Coastal Plain, and the Mountainous West (ISA
17 Section 4.2.2.3). In these sensitive areas, between 1 and 6 percent of stream kilometers are
18 chronically acidified.
19 The ISA notes that "consideration of episodic acidification greatly increases the extent
20 and degree of estimated effects for acidifying deposition on surface waters." (ISA Section
21 3.2.1.6) Some studies show that the number of lakes that could be classified as acid-impacted
22 based on episodic acidification is 2 to 3 times the number of lakes classified as acid-impacted
23 based on chronic ANC. These episodic acidification events can have long term effects on fish
24 populations (ISA Section 3.2.1.6). Under recent conditions, episodic acidification has been
25 observed in locations in the eastern U.S. and in the mountainous western U.S. (ISA Section
26 3.2.1.6).
27 It can therefore be concluded that recent levels of NOx and SOx are associated with
28 deposition that leads to ANC values below benchmark values known to cause ecological harm in
29 sensitive aquatic systems, including lakes and streams in multiple areas of the U.S. These
30 changes are known to have impacts on ecosystem services including recreational fishing which is
31 discussed along with other services in Chapter 3. While other ecosystem services (e.g. habitat
September, 2010 4-43 Do Not Quote or Cite
-------
1 provisioning, subsistence fishing, and biological control as well as many others) are potentially
2 affected by reductions in ANC, confidence in the specific translation of ANC values to these
3 additional ecosystem services is much lower.
4
5 4.5.1.2 Terrestrial Acidification
6 Based on the case studies on sugar maple and red spruce habitat, staff concludes that
7 there is significant risk to sensitive terrestrial ecosystems from acidification at atmospheric
8 concentrations of NOx and SOx at or below the current standards. This conclusion is based on
9 application of the simple mass balance model to deposition levels associated with NOx and SOx
10 concentrations at or below the current standards. The ecological indicator selected for terrestrial
11 acidification is the base cation to aluminum ratio (BC:A1), which has been linked to tree health
12 and growth. The results of the REA strongly support a relationship between atmospheric
13 deposition of NOx and SOx and BC:A1, and that BC:A1 is a good indicator of terrestrial
14 acidification. At levels of deposition associated with NOx and SOx concentrations at or below
15 the current standards, BC:A1 levels are expected to be below benchmark values that are
16 associated with significant effects on tree health and growth. Such degradation of terrestrial
17 ecosystems could affect ecosystem services such as habitat provisioning, endangered species,
18 goods production (timber, syrup, etc.) and many others.
19 Many locations in sensitive areas of the U.S. have BC:A1 levels below benchmark levels
20 classified as providing low to intermediate levels of protection to tree health. At a BC:A1 ratio of
21 1.2 (intermediate level of protection), red spruce growth can be reduced by 20 percent. At a
22 BC:A1 ratio of 0.6 (low level of protection), sugar maple growth can be decreased by 20 percent.
23 The REA did not evaluate broad sensitive regions. However, in the sugar maple case study area
24 (Kane Experimental Forest), recent deposition levels are associated with a BC: Al ratio below
25 1.2, indicating between intermediate and low level of protection, which would indicate the
26 potential for a greater than 20 percent reduction in growth. In the red spruce case study area
27 (Hubbard Brook Experimental Forest), recent deposition levels are associated with a BC: Al ratio
28 slightly above 1.2, indicating slightly better than an intermediate level of protection (REA
29 Section 4.3.5.1)
30 Over the full range of sugar maple, 12 percent of evaluated forest plots exceeded the
31 critical loads for a BC: Al ratio of 1.2, and 3 percent exceeded the critical load for a BC: Al ratio
September, 2010 4-44 Do Not Quote or Cite
-------
1 of 0.6. However, there was large variability across states. In New Jersey, 67 percent of plots
2 exceeded the critical load for a BC: Al ratio of 1.2, while in several states on the outskirts of the
3 range for sugar maple (e.g. Arkansas, Illinois) no plots exceeded the critical load for a BC:A1
4 ratio of 1.2. For red spruce, overall 5 percent of plots exceeded the critical load for a BC:A1 ratio
5 of 1.2, and 3 percent exceeded the critical load for a BC:Al ratio of 0.6. In the major red spruce
6 producing states (Maine, New Hampshire, and Vermont), critical loads for a BC:A1 ratio of 1.2
7 were exceeded in 0.5, 38, and 6 percent of plots.
8 The ISA reported one study (McNulty, 1997) that estimated 15 percent of U.S. forest
9 ecosystems exceeded the critical loads for acidity for N and S deposition by >250 eq/ha/year
10 under current conditions (ISA Section 4.2.1.3) . Staff concludes that this represents a significant
11 portion of sensitive terrestrial ecosystems.
12 It can therefore be concluded that recent levels of NOx and SOx are associated with
13 deposition that leads to BC:A1 values below benchmark values that cause ecological harm in
14 some sensitive terrestrial ecosystems. While effects are more widespread for sugar maple, there
15 are locations with low to intermediate levels of protection from effects on both sugar maple and
16 red spruce. While there are many other ecosystem services, including timber production, natural
17 habitat provision, and regulation of water, climate, and erosion, potentially affected by
18 reductions in BC:A1, linkages of BC:A1 values to these additional ecosystem services is on the
19 whole not well understood.
20 4.5.2 To what extent does the current NOx secondary standard provide protection from
21 adverse effects associated with deposition of atmospheric NOx, which results in nutrient
22 enrichment effects in sensitive aquatic and terrestrial ecosystems?
23 Nutrient enrichment effects are due to nitrogen loadings from both atmospheric and non-
24 atmospheric sources. Evaluation of nutrient enrichment effects requires an understanding that
25 nutrient inputs are essential to ecosystem health. The specific long term levels of nutrients in a
26 system affect the types of species that occur over long periods of time. Short term additions of
27 nutrients can affect species competition, and even small additions of nitrogen in areas that are
28 traditionally nutrient poor can have significant impacts. In certain limited situations, additions of
29 nitrogen can increase rates of growth, and these increases can have short term benefits in certain
30 managed ecosystems. As noted earlier, this review of the standards is focused on unmanaged
September, 2010 4-45 Do Not Quote or Cite
-------
1 ecosystems. As a result, in assessing adequacy of the current standards, we are focusing on the
2 adverse effects of nutrient enrichment in unmanaged ecosystems. However, the following
3 discussion provides a brief assessment of effects in managed ecosystems.
4 Impacts of nutrient enrichment in managed ecosystems may be positive or negative
5 depending on the levels of nutrients from other sources in those areas. Positive effects can occur
6 when crops or commercial forests are not receiving enough nitrogen nutrients. Nutrients
7 deposited on crops from atmospheric sources are often referred to as passive fertilization.
8 Nitrogen is a fundamental nutrient for primary production in both managed and unmanaged
9 ecosystems. Most productive agricultural systems require external sources of nitrogen in order
10 to satisfy nutrient requirements. Nitrogen uptake by crops varies, but typical requirements for
11 wheat and corn are approximately 150 kg/ha-yr and 300 kg/ha-yr, respectively (NAPAP, 1990).
12 These rates compare to estimated rates of passive nitrogen fertilization in the range of 0 to 5.5
13 kg/ha-yr (NAPAP, 1991).
14 Information on the effects of changes in passive nitrogen deposition on forestlands and
15 other terrestrial ecosystems is very limited. The multiplicity of factors affecting forests, including
16 other potential stressors such as ozone, and limiting factors such as moisture and other nutrients,
17 confound assessments of marginal changes in any one stressor or nutrient in forest ecosystems.
18 The ISA notes that only a fraction of the deposited nitrogen is taken up by the forests, most of
19 the nitrogen is retained in the soils (ISA 3.3.2.1). In addition, the ISA indicates that forest
20 management practices can significantly affect the nitrogen cycling within a forest ecosystem, and
21 as such, the response of managed forests to NOx deposition will be variable depending on the
22 forest management practices employed in a given forest ecosystem (ISA Annex C C.6.3)
23 Increases in the availability of nitrogen in N-limited forests via atmospheric deposition could
24 increase forest production over large non-managed areas, but the evidence is mixed, with some
25 studies showing increased production and other showing little effect on wood production (ISA
26 3.3.9). Because leaching of nitrate can promote cation losses, which in some cases create nutrient
27 imbalances, slower growth and lessened disease and freezing tolerances for forest trees, the net
28 effect of increased N on forests in the U.S. is uncertain (ISA 3.3.9).
29 In managed agricultural ecosystems, nitrogen inputs from atmospheric NOx comprise a
30 small fraction (less than 3 percent) of total nitrogen inputs, which include commercially applied
31 fertilizers as well as applications of composted manure. And because of the temporal and spatial
September, 2010 4-46 Do Not Quote or Cite
-------
1 variability in atmospheric deposition of NOx, it is unlikely that farmers would alter their
2 fertilization decisions based on expected nitrogen inputs from NOx. And, in some locations,
3 farmers need less nitrogen inputs due to production of excess nitrogen through livestock. In
4 some locations, nitrogen production through livestock waste exceeds the absorptive capacity of
5 the surrounding land, and as such, excess nitrogen from deposition of NOx in those locations
6 reduces the capacity of the system to dispose of excess nitrogen, potentially increasing the costs
7 of waste management from livestock operations (Letson and Gollehon, 1996). A USD A
8 Economic Research Service report found that in 1997, 68 counties with high levels of confined
9 livestock production had manure nitrogen levels that exceed the assimilative capacity of the
10 entire county's crop and pasture land (Gollehon et al, 2001). In those locations, additional
11 nitrogen inputs from NOX deposition will result in excess nitrogen, leading to nitrogen leaching
12 and associated effects that adversely effect ecosystems.
13
14 4.5.3 Aquatic Nutrient Enrichment
15 The REA case studies focused on coastal estuaries and revealed that while current
16 ambient loadings of atmospheric NOx are contributing to the overall deposit!onal loading of
17 coastal estuaries, other non-atmospheric sources are contributing in far greater amounts in total,
18 although atmospheric contributions are as large as some other individual source types. The
19 ability of current data and models to characterize the incremental adverse impacts of nitrogen
20 deposition is limited, both by the available ecological indicators, and by the inability to attribute
21 specific effects to atmospheric sources of nitrogen. The REA case studies used as the ecological
22 indicator for aquatic nutrient enrichment an index of eutrophication known as the Assessment of
23 Estuarine Trophic Status Eutrophication Index (ASSETS El). This index is a six level index
24 characterizing overall eutrophi cation risk in a waterbody. This indictor is not sensitive to
25 relatively large changes in nitrogen deposition. In addition, this type of indicator does not reflect
26 the impact of nitrogen deposition in conjunction with other sources of nitrogen.
27 For example, if NOx deposition is contributing nine tenths of the nitrogen loading
28 required to move a waterbody from an ASSETS El category of "moderate" to a category of
29 "poor", zeroing out NOx deposition will have no impact on the ASSETS El value. However, if
30 an area were to decide to put in place decreases in nitrogen loadings to move that waterbody
September, 2010 4-47 Do Not Quote or Cite
-------
1 from "poor" to "moderate," the area would have to reduce the full amount of the loadings
2 through other sources if atmospheric deposition were not considered. Thus, the adverse impact
3 of atmospheric nitrogen is in its contribution to the overall loading, and reductions in NOX will
4 decrease the amount of reductions from other sources of nitrogen loadings that would be required
5 to move from a lower ASSETS El category to a higher category. NOx deposition can also be
6 characterized as reducing the risk of a waterbody moving from a higher ASSETS El category to
7 a lower category, by reducing the vulnerability of that waterbody to increased loadings from
8 non-atmospheric sources.
9 Based on the above considerations, staff preliminarily concludes that the ASSETS El is
10 not an appropriate ecological indicator for estuarine aquatic eutrophication. Staff further
11 concludes that additional analysis is required to develop an appropriate indicator for determining
12 the appropriate levels of protection from N nutrient enrichment effects in estuaries related to
13 deposition of NOx. As a result, staff is unable to make a determination as to the adequacy of the
14 existing secondary NOx standard in protecting public welfare from N nutrient enrichment effects
15 in estuarine aquatic ecosystems.
16 Additionally, nitrogen deposition can alter species composition and cause eutrophi cation
17 in freshwater systems. In the Rocky Mountains, for example, deposition loads of 1.5 to 2 kg/ha-
18 yr which are well within current ambient levels are known to cause changes in species
19 composition in diatom communities indicating impaired water quality (ISA Section 3.3.5.3). It
20 then seems apparent then that the existing secondary standard for NOX does not protect such
21 ecosystems and their resulting services from impairment.
22 4.5.4 Terrestrial Nutrient Enrichment
23 The scientific literature has many examples of the deleterious effects caused by excessive
24 nitrogen loadings to terrestrial systems. Several studies have set benchmark values for levels of
25 N deposition at which scientifically adverse effects are known to occur. These benchmarks are
26 discussed more thoroughly in Chapter 5 of the REA. Large areas of the country appear to be
27 experiencing deposition above these benchmarks for example, Fenn et al. (2008) found that at
28 3.1 kg N/ha-yr, the community of lichens begins to change from acidophytic to tolerant species;
29 at 5.2 kg N/ha-yr, the typical dominance by acidophytic species no longer occurs; and at 10.2 kg
30 N/ha-yr, acidophytic lichens are totally lost from the community. Additional studies in the
September, 2010 4-48 Do Not Quote or Cite
-------
1 Colorado Front Range of the Rocky Mountain National Park support these findings and are
2 summarized in Chapter 6.0 of the Risk and Exposure Assessment. These three values (3.1, 5.2,
3 and 10.2 kg/ha-yr) are one set of ecologically meaningful benchmarks for the mixed conifer
4 forest (MCF) of the pacific coast regions. Nearly all of the known sensitive communities receive
5 total nitrogen deposition levels above the 3.1 N kg/ha-yr ecological benchmark according to
6 the 12 km, 2002 CMAQ/NADP data, with the exception of the easternmost Sierra Nevadas.
7 MCFs in the southern portion of the Sierra Nevada forests and nearly all MCF communities in
8 the San Bernardino forests receive total nitrogen deposition levels above the 5.2 N kg/ha-yr
9 ecological benchmark.
10 Coastal Sage Scrub communities (CSS) are also known to be sensitive to community
11 shifts caused by excess nitrogen loadings. Wood et al. (2006) investigated the amount of
12 nitrogen utilized by healthy and degraded CSS systems. In healthy stands, the authors estimated
13 that 3.3 kg N/ha-yr was used for CSS plant growth (Wood et al., 2006). It is assumed that 3.3 kg
14 N/ha-yr is near the point where nitrogen is no longer limiting in the CSS community. Therefore,
15 this amount can be considered an ecological benchmark for the CSS community. The majority of
16 the known CSS range is currently receiving deposition in excess of this benchmark. Thus, staff
17 concludes that recent conditions where NOx ambient concentrations are at or below the current
18 NOx secondary standards are not adequate to protect against anticipated adverse impacts from N
19 nutrient enrichment in sensitive ecosystems.
20
21 4.6 To What Extent Do The Current NOX And/Or SOX Secondary Standards Provide
22 Protection From Other Ecological Effects (Eg. Mercury Methylation) Associated
23 With The Deposition Of Atmospheric NOX, And/Or SOX?
24
25 It is stated in the ISA (ISA Sections 3.4.1 and 4.5) that mercury is a highly neurotoxic
26 contaminant that enters the food web as a methylated compound, methylmercury. Mercury is
27 principally methylated by sulfur-reducing bacteria and can be taken up by microorganisms,
28 zooplankton and macroinvertebrates. The contaminant is concentrated in higher trophic levels,
29 including fish eaten by humans. Experimental evidence has established that only inconsequential
30 amounts of methylmercury can be produced in the absence of sulfate. Once methylmercury is
31 present, other variables influence how much accumulates in fish, but elevated mercury levels in
September, 2010 4-49 Do Not Quote or Cite
-------
1 fish can only occur where substantial amounts of methylmercury are present. Current evidence
2 indicates that in watersheds where mercury is present, increased SOx deposition very likely
3 results in additional production of methylmercury which leads to greater accumulation of MeHg
4 concentrations in fish (Munthe et al, 2007; Drevnick et al., 2007).
5 The production of meaningful amounts of methylmercury (MeHg) requires the presence
6 of SC>42" and mercury, and where mercury is present, increased availability of SC>42" results in
7 increased production of MeHg. There is increasing evidence on the relationship between sulfur
8 deposition and increased methylation of mercury in aquatic environments; this effect occurs only
9 where other factors are present at levels within a range to allow methylation. The production of
10 methylmercury requires the presence of sulfate and mercury, but the amount of methylmercury
11 produced varies with oxygen content, temperature, pH, and supply of labile organic carbon (ISA
12 Section 3.4). In watersheds where changes in sulfate deposition did not produce an effect, one or
13 several of those interacting factors were not in the range required for meaningful methylation to
14 occur (ISA Section 3.4). Watersheds with conditions known to be conducive to mercury
15 methylation can be found in the northeastern United States and southeastern Canada. The
16 relationship between sulfur and methylmercury production is addressed qualitatively in Chapter
17 6 of the Risk and Exposure Assessment.
18 With respect to sulfur deposition and mercury methylation, the final ISA determined: The
19 evidence is sufficient to infer a causal relationship between sulfur deposition and increased
20 mercury methylation in wetlands and aquatic environments. However, staff did not conduct a
21 quantitative assessment of the risks associated with increased mercury methylation under current
22 conditions. As such, staff are unable to make a determination as to the adequacy of the existing
23 SC>2 standards in protecting against welfare effects associated with increased mercury
24 methylation.
25
September, 2010 4-50 Do Not Quote or Cite
-------
1 4.7 REFERENCES:
2
3 BJ. Finlayson-Pitts and J.N. Pitts, 2000, Chemistry of the Upper and Lower Troposhere,
4 Academic Press, San Diego, CA
5 Drevnick, P.E., D.E. Canfield, P.R. Gorski, A.L.C. Shinneman, D.R. Engstrom, D.C.G. Muir,
6 G.R. Smith, PJ. Garrison, L.B. Cleckner, J.P. Hurley, R.B. Noble, R.R. Otter, and J.T.
7 Oris. 2007. Deposition and cycling of sulfur controls mercury accumulation in Isle
8 Royale fish. Environmental Science and Technology ¥7(21):7266-7272.
9 Fenn, M.E., S. Jovan, F. Yuan, L. Geiser, T. Meixner, and B.S. Gimeno. 2008. Empirical and
10 simulated critical loads for nitrogen deposition in California mixed conifer forests.
11 Environmental Pollution 755(3 ): 492-511.
12 Munthe, 1, R.A. Bodaly, B.A. Branfireun, C.T. Driscoll, C.C. Gilmour, R. Harris, M. Horvat, M.
13 Lucotte, and O. Malm. 2007. Recovery of mercury-contaminated fisheries. Ambio 36:33-
14 44.
15 Scheffe, R.D., P. A. Solomon, R. Husar, T. Hanley, M. Schmidt, M. Koerber , M. Gilroy, J.
16 Hemby, N. Watkins, M. Papp, J. Rice, J. Tikvart andR. Valentinetti, The National
17 Ambient Air Monitoring Strategy: Rethinking the Role of National Networks, JAWMA,
18 ISSN: 1047-3289 J. Air & Waste Manage. Assoc. 2009, 59:579-590 DOT: 10.3155/1047-
19 3289.59.5.579
20 U.S. EPA (Environmental Protection Agency). 1982. Review of the National Ambient Air Quality
21 Standards for Sulfur Oxides: Assessment of Scientific and Technical Information.
22 OAQPS Staff Paper. EPA-450/5-82-007. U.S. Environmental Protection Agency, Office
23 of Air Quality Planning and Standards, Research Triangle Park, NC.
24 U.S. EPA (Environmental Protection Agency). 1995. Review of the National Ambient Air Quality
25 Standards for Nitrogen Dioxide: Assessment of Scientific and Technical Information.
26 OAQPS Staff Paper. EPA-452/R-95-005. U.S. Environmental Protection Agency, Office
27 of Air Quality Planning and Standards, Research Triangle Park, NC. September.
28 U.S. EPA. (Environmental Protection Agency). 2000. Deposition of Air Pollutants to the Great
29 Waters: Third report to Congress. (EPA-453/R-00-005)
30 http://www.epa.gov/air/oaqps/gr8water/3rdrpt/index.html; Research Triangle Park, NC;
31 Office of Air Quality Planning and Standards;
32 U.S. Environmental Protection Agency
33 U.S. EPA (Environmental Protection Agency). 2008. Integrated Science Assessment (ISA) for
34 Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report). EPA/600/R-
35 08/082F. U.S. Environmental Protection Agency, National Center for Environmental
36 Assessment-RTF Division, Office of Research and Development, Research Triangle
37 Park, NC. Available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201485.
38 Wood, Y., T. Meixner, PJ. Shouse, and E.B. Allen. 2006. Altered Ecohydrologic response
39 drives native shrub loss under conditions of elevated N-deposition. Journal of
40 Environmental Quality 35:76-92.
41
September, 2010 4-51 Do Not Quote or Cite
-------
1 5 OPTIONS FOR ELEMENTS OF THE STANDARD
2
3
4 The elements of the standard include the ambient air indicator, the form, the averaging
5 time and the level. The "indicator" of a standard defines the chemical species or mixture of
6 criteria air pollutants that is to be measured in determining whether an area attains the standard.
7 The "form" of a standard defines the air quality statistic that is to be compared to the level of the
8 standard in determining whether an area attains the standard. The "averaging time" defines the
9 period of time over which the air quality indicator is averaged, e.g. annual average. The "level"
10 is the specific quantity to which the air quality statistic will be compared.
11 Historically, EPA has established NAAQS so that the locally-monitored ambient
12 concentration of an air pollutant indicator is compared against a specified numerical level of
13 atmospheric concentration, using a specified statistical form and averaging time. For example,
14 the current secondary standard for oxides of nitrogen uses ambient concentrations of NC>2 as the
15 indicator. Attainment is determined by comparing the annual arithmetic mean of the measured
16 maximum dailyl-hour NOi concentrations, for a calendar year, against the level of 0.053 ppm.
17 As discussed in Chapters 4, a standard using this kind of approach for defining indicator, form,
18 averaging time, and level is not the most appropriate way to protect sensitive ecosystems from
19 effects associated with ambient concentrations of NOx and SOx. This is because the ecological
20 effects of NOx and SOx are a result of deposition of these air pollutions. The inherently
21 complex and variable linkages between ambient concentrations of NOX and SOX, their deposited
22 forms of nitrogen and sulfur, and the ecological responses that are associated with public welfare
23 effects call for consideration of a more complex and ecologically relevant design of the standard
24 that reflects these linkages. In this chapter, we present a set of potential options for defining the
25 elements of the NOx and SOx secondary standard including indicator, form, averaging time, and
26 certain aspects of the level, with additional discussion of the options for specifying a range of
27 levels is discussed in Chapter 9.
28 After review of the ISA and REA, CASAC concluded that aquatic acidification should be
29 the focus for developing a multi-pollutant standard, based on the quantity and quality of
30 available data. CASAC also recommended that, in addition to aquatic acidification, the EPA
September, 2010 5-1 Draft - Do Not Quote or Cite
-------
1 should consider multiple ecological indicators and made the following statement in their letter to
2 the EPA on August 28, 2009:
3
4 ".. .the Panel finds the information in the current REA sufficient to inform setting
5 separate standards for terrestrial acidification, eutrophication of western alpine lakes and
6 terrestrial nutrient enrichment. However, the Panel believes that setting a standard for
7 coastal nutrient enrichment would be difficult because of the substantial inputs of non-
8 atmospheric sources of N to these systems."
9
10 As a result of our assessment of the science, and reflecting the comments of CAS AC, this policy
11 assessment is focused on developing a standard specifically designed to protect against the
12 effects of aquatic acidification in sensitive ecosystems, while recognizing that such a standard
13 may also provide co-protection against effects of terrestrial acidification, eutrophication of high
14 elevation western lakes and terrestrial nutrient enrichment. Co-protection against these effects is
15 discussed in Chapter 6.
16 Our development of options for the standards recognizes the need for nationally
17 applicable standard for protection against adverse effects to public welfare, while recognizing the
18 complex and heterogeneous interactions between atmospheric concentrations of NOy and SOx,
19 deposition from of NOy and SOx, and ecological response. Our approach also recognizes that
20 while a standard is national in scope and coverage, the effects to public welfare from aquatic
21 acidification may not occur to the same extent in all locations in the U.S. - in fact, protection
22 may vary among locations according to sensitivity. As noted in Chapters 2, 3, and 4, many
23 locations in the U.S. are naturally protected against acid deposition due to underlying geological
24 conditions. Likewise, some locations in the U.S., including lands managed for commercial
25 agriculture and forestry, are not likely to be negatively impacted by current levels of nitrogen and
26 sulfur deposition. As a result, our design for the standards is intended to protect sensitive
27 ecosystems and the services provided by those sensitive ecosystems.
28 In this chapter we present our reasoning for suggesting a standard that employs (1) NOy
29 and SOx as the atmospheric indicators, (2) a multi-year averaging time, (3) a form of a secondary
30 standard that takes into account variable factors, such as atmospheric and ecosystem conditions
31 that modify the amounts of deposited NOx and SOx, and the associated effects of deposited N
32 and S on ecosystems, and (4) a target ANC level and target percentages of water bodies to
33 protect to the target ANC level. Our goal in developing the form of the standard is to create an
September, 2010 5-2 Draft - Do Not Quote or Cite
-------
1 index, directly expressed in terms of atmospheric concentrations of NOy and SOx, which can be
2 applied across the nation to convey the allowable levels of ambient NOy and SOx, based on
3 various factors such as the sensitivity of an area and the desired degree of protection from
4 acidification caused by atmospheric deposition. This chapter is structured around questions
5 related to the various elements of a standard. The chapter begins in section 5.1 with a discussion
6 of atmospheric indicators. In Section 5.2 the averaging times for the atmospheric indicators is
7 presented. In Section 5.3 a suggested ecologically relevant form of the standard is presented.
8 In Section 5.4 the issues regarding the spatial area over which a standard might be evaluated are
9 discussed, along with related issues regarding spatial averaging within areas. Section 5.5
1.0 provides a discussion of specific target ANC levels. Section 5.6 discusses the selection of target
11 percentages of water bodies to protect to this ANC level. As noted previously, additional
12 discussion related to options for ranges of the level of the standard are discussed in Chapter 9.
13
14 5.1 What atmospheric indicators of oxidized nitrogen and sulfur are appropriate for
15 use in a secondary NAAQS that provides protection for public welfare from
16 exposure related to deposition of NOx and SOx?
17
18 Staff concludes that indicators other than NO2 and SO2 should be considered as the
19 appropriate pollutant indicators for protection against the acidification effects associated with
20 deposition of NOX and SOX. This conclusion is based on the recognition that all forms of
21 oxidized nitrogen and sulfur in the atmosphere contribute to deposition and resulting
22 acidification, and as such NO2 and SO2 are incomplete indicators. Furthermore, staff
23 concludes that NOy (total oxidized nitrogen) should be considered as an appropriate indicator
24 for oxides of nitrogen. NOy is defined as NOx (NO and NO2) and all oxidized NOx products:
25 including NO, NO2, and all other oxidized N-containing compounds transformed from
26 NO and NO2 (Finlayson-Pitts and Pitts, 2000). As described in Chapter 4, this set of
27 compounds includes NO2 + NO + HNO3 + PAN +2N2O5 + HONO + NO3 + organic nitrates +
28 particulate NO3. SOx includes sulfur monoxide (SO), sulfur dioxide, sulfur trioxide (SO3),
29 and disulfur monoxide (S2O), and particulate-phase S compounds that result from gas-
30 phase sulfur oxides interacting with particles. Staff concludes that SO2 + SO4 should be
31 considered as an appropriate indicator for oxides of sulfur.
September, 2010 5-3 Draft - Do Not Quote or Cite
-------
1 In principle, measured NOy based on catalytic conversion of all oxidized species to NO,
2 followed by chemiluminescence NO detection is consistent with this definition. We recognize
3 the caveats associated with instrument conversion efficiency and possible inlet losses which are
4 discussed in Chapter 8. The development of the function that converts atmospheric
5 concentrations of NOy and SOx to N and S deposition incorporates NOy estimates based on the
6 Community Multi-scale Air Quality (CMAQ) model (EPA, 1999). CMAQ treats the dominant
7 NOy species (NO, NO2, nitric acid, PAN and particulate NOa) as explicit species while the
8 minor contributing non-PAN organic nitrogen compounds are aggregated. From a
9 measurement and modeling perspective we only consider the sum of SO2 and particulate SO4 as
10 the indicator for sulfur. The sum of SO2 and SO4 constitute virtually all of the ambient air
11 sulfur budget and are measured routinely in monitoring networks. In addition to accounting for
12 virtually the entire oxidized sulfur budget, SO2 and particulate SO4 are routinely measured in
13 ambient air monitoring networks, although only CASTNET filter packs capture the entire
14 particle size range. The CMAQ treatment of SOx is the simple addition of both species which
15 are treated explicitly in the model formulation. All particle size fractions are included in the
16 CMAQ SOx estimates. Consistent with units and the charge balance relationships applied in
17 ecosystem acidification models, only mass as sulfur or nitrogen is considered when aggregating
18 the species constituting NOy or SOx.
19
20 5.2 What is the appropriate averaging time for the air quality indicators NOy and SOx
21 to provide protection of public welfare from adverse effects from aquatic
22 acidification?
23
24 Based on the review of the scientific evidence, welfare effects associated with
25 acidification result from event based to annual cumulative deposition of N and S. Annual
26 cumulative deposition of N and S reflects the chronic acid base balance of the surface water as
27 indicated by the ANC level (measured as annual ANC). Also, critical loads for acidity are in
28 terms of annual cumulative deposition of N and S. Aquatic acidification can occur over both
29 long- and short-term timescales. Short-term (i.e., hours or days) episodic changes in water
30 chemistry, often due to changes in the hydro logic flow paths (Chen et al. 1984), can have
31 significant biological effects. Short-term change in chemistry is termed "episodic acidification."
September, 2010 5-4 Draft - Do Not Quote or Cite
-------
1 Some streams may have chronic or base flow chemistry that is suitable for aquatic biota, but may
2 be subject to occasional acidic episodes with lethal consequences.
3 Episodic declines in pH and ANC are nearly ubiquitous in drainage waters throughout the
4 eastern United States. Episodic acidification can result from several mechanisms related to
5 changes in hydrologic flow paths. For example, snow can store N deposited throughout the
6 winter, snowmelt releases this stored N in a pulse that leads to episodic acidification in the
7 absence of increased deposition during the actual episodic acidification event. However, inputs
8 of nitrogen and sulfur from snowpack and atmospheric deposition largely cycle through soil. As
9 a result short-term direct deposition inputs are not important in episodic acidification. As noted
10 in Chapter 3 of the ISA, protection against episodic acidity events can be achieved by
11 establishing a higher chronic ANC level (See 5.3.2.1). Protection against a low chronic ANC
12 level is provided by reducing overall annual average deposition levels for nitrogen and sulfur.
13 This supports the conclusion that long term NOy and SOx concentrations are appropriate to
14 provide protection against low chronic ANC levels, which protects against both long term
15 acidification and acute acidic episodes.
16 Staff suggests a 3 to 5 year averaging time of the cumulative annual average for the
17 air quality indicators NOy and SOx, to account for interannual variations.
18
19 5.3 What form(s) of the standard are most appropriate to provide protection of
20 sensitive ecosystems from the effects of acidifying deposition related to ambient NOx
21 and SOx concentrations?
22
23 Based on the evidence of the aquatic acidification effects caused by NOy and SOx, staff
24 concludes that it is appropriate to consider changes to the form of the existing NOX and SOX
25 secondary standards to provide protection to ecosystems. EPA staff has developed a conceptual
26 design for the form of the standard that includes four main components: atmospheric and
27 ecological indicators, deposition metrics, functions that relate indicators to deposition metrics
28 and factors that modify the functions. These components of the design are illustrated in Fig 5.1.
29 The rectangles represent indicators. Ecological indicators are chemical or biological components
30 of the ecosystem that can be linked to N and S deposition based on scientific evidence. Air
31 quality indicators are the chemical species of the criteria air pollutants that best represent the
32 atmospheric pollutants that cause ecological harm in the criteria pollutant categories of oxides of
September, 2010 5-5 Draft - Do Not Quote or Cite
-------
1 nitrogen and oxides of sulfur (selection of air quality indicators was discussed in Section 5.1).
2 Diamonds indicate functions in which two variables are related. The ecological effect function is
3 the relationship between the ecological indicator and deposition. The atmospheric deposition
4 transference ratio is the relationship between deposition and the atmospheric concentration of an
5 air quality indicator. The ovals represent factors which will modify the functions. Modifying
6 factors can vary across the landscape, such as soil depth, catchment size, etc. The spatial
7 heterogeneity of modifying factors can be challenging to characterize, and therefore in some
8 cases we present multiple options for how to incorporate them into the design.
9
10
11
12
13
14
15
Ecological
Indicator
Location
specific
Modifying
Factors
Location
specific
Modifying
Ecological
Response
to
Deposition
Function
Atmospheric
Deposition
Transference
Ratio
Indicator(s)
Figure 5-1 Conceptual design of the form of NOX and SOx secondary standard
5.3.1 Conceptual Design of the Form: General Overview
16 A summary of the design of the form based on aquatic acidification is given here to help
17 provide context and support for the more detailed discussions that follow. The conceptual design
18 for a form that is specifically based on aquatic acidification effects is illustrated by Fig 5.2.
19 Starting on the left side of the figure, ANC is a chemical indicator of sensitivity to aquatic
20 acidification that is tied to the probability of biological harm that may occur to the system from
21 aquatic acidification. ANC is often considered the best chemical indicator of sensitivity to
22 aquatic acidification (Section 5.3.2.1). Ambient NOy and SOx add to the total deposition of N
23 and S that lead to aquatic acidification. NHx is a component comprising a varying fraction of the
September, 2010
5-6
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
total N deposition. The load of N and S deposition that causes a specific level of ANC will vary
depending on the characteristics of the receiving ecosystem, but can be quantified at the
catchment scale with acid-base balance acidification models. The acidification models that
characterize the relationship between ANC and N and S deposition that are most appropriate to
inform the NAAQS are discussed in Sections 5.3.2.
Modifying
Factors
ANC
level
selected
based on
biological
effects
Ecosyste m\
-------
1 exposure. In addition, there are uncertainties in our estimates of risk and exposure for individual
2 waterbodies. Therefore, we propose to examine various alternative spatial groupings of sensitive
3 populations of waterbodies to inform the development of requisite protection. EPA staff
4 proposes to categorize the landscape nationally, such that within a category there are generally
5 similar acid-sensitivity characteristics. Each national acid-sensitivity category is represented by
6 a population of catchments for which critical loads at a specified ANC limit are calculated.
7 In our proposed approach, for each acid-sensitive category of the landscape the
8 distribution of critical loads from a population of catchments would be evaluated to develop a
9 deposition metric. The deposition metric would be defined as the amount of deposition that
10 reasonably assures a specified percentage of individual catchments in a population would
11 achieve a target ANC level.
12 The level of the ANC target is tied to the described degree of protection for aquatic biota.
13 It will be selected as one element of the set of information that informs the Administrator's
14 selection of the allowable level(s) of ambient NOy and SOx under the secondary standard.
15 Options for target ANC levels are discussed in Section 5.5. Taken together, the secondary
16 standard would be set with a goal of specifying allowable ambient levels of NOy and SOx such
17 that there is a desired degree of confidence that a targeted percentage of aquatic ecosystems
18 would achieve a specified ANC level within each acid-sensitivity category (see Chapter 9).
19 Finally, the deposition of reduced N should be taken into account because reduced forms
20 of N are often quickly converted to NO3~ by the processes of nitrification. For this reason, the
21 amount of reduced N needs to be accounted for in the standard because some fraction of the acid
22 buffering capacity of an ecosystem is used up by reduced nitrogen, leaving less protection
23 against deposition from NOy and SOx. Spatial characterization of reduced N is discussed in
24 Section 5.3.2. The deposition tradeoff curves are then multiplied by the deposition to
25 concentration transference ratio to calculate the atmospheric concentration tradeoff curves.
26 Development of these functions is discussed in Section 5.3.3. Thus far we have summarized our
27 conceptual design by walking through the components of the standard shown in Fig 5.2 from the
28 left to the right, and a more detailed illustration of these steps is given by Fig 5.3.
29 The bidirectional arrows shown in Fig 5.2 emphasize that the order in which one
30 considers the links between ANC and deposition resulting from of NOy and SOx is conceptually
31 important. Moreover, different questions may be answered by working through Fig 5.2 right to
September, 2010 5-8 Draft - Do Not Quote or Cite
-------
1 left versus the left to right. For example, if the amount of N and S deposited to a given
2 catchment is known, the level of ANC that would result may be calculated. The calculated ANC
3 could then be compared to a benchmark value of ANC. Working through Fig 5.2 from the right
4 to the left is the basis for the Atmospheric Acidification Protection Index (AAPI, Fig 5.3 and
5 Section 5.3.4). The AAPI is essentially a function that determines the allowable levels of
6 ambient NOy and SOx based on the target ANC limit, given uncertainties in the parameters used
7 to calculate an ANC equivalent at the national scale, and weighing other factors such as time to
8 recovery for ecosystems, based on populations of catchments that represent acid-sensitive areas
9 in the U.S. The AAPI is designed to be a more ecologically relevant form of the standard
10 relative to the current form . The intent of the AAPI is to weight atmospheric concentrations of
11 NOx and SOx by their propensity to contribute to acidification through deposition, given the
12 fundamental acidifying potential of each pollutant, and the ecological factors that govern acid
13 sensitivity in different ecosystems, as well as the contribution of reduced nitrogen. Thus the
14 AAPI is more relevant to protecting ecosystems from acidifying deposition compared to simple
15 ambient concentration forms which do not reflect factors that affect acidifying potential.
September, 2010 5-9 Draft - Do Not Quote or Cite
-------
CO
§,<
i
to
c '•"
u
•=
1"
CO .
c •
o .,
*\
CU
0
3-j
L
VI —
tO a .42
« •§ 1 ^
t E a 1 1
i i * « "5
= 2 i 1 -B
5 W | | *
5 g *S X e -n X
Hi °P!
_ -rf -S + r5 "« 0
3 '5 J» « £5
- £ -£ ^ ¥ •
i£ U 3 fc -f^
1 - eo 4
i
I -g *
s >
O 'S
•si
OfB
W ._
s *? S
o 15 4=
*S A O)
.. &
|I|
5 te
*= CL
Q. u.
£2
3 /" \
\
lili 1^
pfl f«l|
r 1 1|.»
« "3 J* S S ^ J
1 tg •§ < 1 '6
— Q "* To
c S E £
0 (J "
o
v\
•^ n^ f^
f^| /^|J|\ ijl
1 — g \ UJ — ^ ^/
H \ — •
CO
%S
l\
O j.
*
"2 !
E E
_^ f£
en c
y .v
11
S •»
•p
"*n *
o t
A fl-
at •'
*c •-
II
& ;"
j. ^T
T 'E
Sg
to ^w
O S
c-*i
II «
W '* c/l
O '•" "~>
"^ i£
X ^
:jii
123
I
*> ^
O
1
1
^
S
5.S
*l
c
1 ^
(C
5 »
P .>* i-
^P -Q O
in -n t3
•S 5 fe
|1|
.1 Si
.« <
0 E
Z. g 0)
< a, = -g
~° £ n ">
i * -i i g
p (fl to js d>
1 1 1 8 I
J^ = T3 -0 «
_ ^ • O> f= CO
g 1 i £ S ?•
^T ^ J3 iC1 O C
< -f S.IS S 'i.
S c o o" «
•g « « z §
i^li
i
i
T
1 'C 3
j s^
1 ?l
1 = 1
o
0
1
;
\ 1
rft: S."'H S Tt °S ^ £^ ^
fill N^iili
Ifll-X If II
l°! fg<^
\. «v °
I
1
t/i
J
•^*
o
g
«2
i
IM d
O O
.11
09 ^
"3 '§
-S «
O H|
H
il
«3 M
O A
v'j
.i
O
1
O
(N
I
f
00
•—i (N CO
-------
1 Staff notes two important concepts illustrated by the 63 and PMio NAAQS that
2 lend support to using this index as the form of a NAAQS. First, in recent reviews
3 of the secondary ozone standards, EPA has considered use of a form of the
4 standard that reflects ecologically relevant exposures, by using a cumulative index
5 which weights exposures at higher concentrations greater than those at lower
6 concentrations based on scientific literature demonstrating the cumulative nature
7 of Oa-induced plant effects and the need to give greater weight to higher
8 concentrations (EPA, 2007; See 75 FR 2938, 2999 January 19, 2010). Staff also
9 notes that PMio is the indicator for the coarse PM NAAQS standard (PMio =
10 PM2.5 +PM 10-2.5). Although the standard has a single level (150 /xg/m3), the actual
11 amount of coarse PM that is allowed varies depending on how much fine PM
12 (PM2.5) is present. By its nature, the PMioNAAQS provides the appropriate level
13 of protection from exposure to coarse PM across locations using a related index of
14 PM (PMio) as the indicator, while allowing the, level of coarse PM to vary across
15 locations. The proposed form for the NOx and SOx standard builds on the
16 concept of using a related index (the AAPI) to provide a homogeneous level of
17 protection across the nation, while allowing ambient air concentrations of NOx
18 and SOx to vary based on ecosystem sensitivity and other relevant factors.
19
20 5.3.2 Conceptual design of the form: Linking the ecosystem indicator to deposition
21 The relationship between the ecosystem indicator, acidification models, and
22 deposition metric components of the form are discussed in this section. The relationship
23 between deposition metric and atmospheric concentrations is discussed in Section 5.3.3.
24 and the derivation of the AAPI equation is presented in Section 5.3.4.
25
26 5.3.2.1 Ecological Indicator: Does the available information provide support to use
27 ANC as the ecological indicator in the conceptual design of the NOy and SOx standard
28 based on aquatic acidification?
29
30 Ecological indicators of acidification in aquatic ecosystems can be chemical or
31 biological components of the ecosystem that are altered by the acidifying effects of N and
32 S deposition. A desirable ecological indicator for aquatic acidification is one that is
September, 2010 5-11 Draft - Do Not Quote or Cite
-------
1 measurable or estimable, linked causally to deposition of N and S, and linked causally,
2 either directly or indirectly to ecological effects known or anticipated to adversely affect
3 public welfare.
4 As summarized in Chapter 2, atmospheric deposition of NOy and SOx causes
5 aquatic acidification through the input of strong acid anions (e.g. NCV and SC>4 ") or
6 chemical forms that are transformed by the ecosystem into strong acid ions. The anions
7 are deposited either directly to the aquatic ecosystem, or indirectly via drainage from
8 terrestrial ecosystems. In other words, when these anions are mobilized in the terrestrial
9 soil, they can leach into adjacent waterbodies. Aquatic acidification is indicated by
10 changes in the surface water chemistry of ecosystems. In turn, the alteration of surface
11 water chemistry has been linked to negative effects on the biotic integrity of freshwater
12 ecosystems. There are a suite of chemical indicators that could be used to assess the
13 effects of acidifying deposition on lake or stream acid-base chemistry. These indicators
r\
14 include acid neutralizing capacity (ANC), surface water pH and concentrations of SC>4 ",
15 NCV, Al, and Ca2+; the sum of base cations; and the base cation surplus. ANC is the
16 most widely used chemical indicator of acid sensitivity and has been found in various
17 studies to be the best single indicator of the likelihood of biological response and health
18 of aquatic communities in acid-sensitive systems (Lien et al, 1992; Sullivan et al., 2006).
19 The utility of the ANC criterion lies in the association between ANC and the surface
20 water constituents that directly contribute to or ameliorate acidity-related stress, in
21 particular pH, Ca2+, and Al. For example, there is an extensive literature documenting
22 threshold pH levels for various aquatic species mortality and health. ANC and pH are
23 directly related under equilibrium conditions (Fig 5.4A). In the field there is a
24 relationship between decreasing ANC and increasing risk to exposure of a threshold pH
25 level (Fig 5.4B). ANC is also used because it integrates overall acid status (ISA 3.2.3
26 and REA 5.2.1) and the acid-related stress for biota that occupies the water that can be
27 directly related to biological impairment, including the number offish species (ISA
28 3.2.3).
29
September, 2010 5-12 Draft - Do Not Quote or Cite
-------
B
•o
o>
5
2
Q.
309 -i
100-
-JW) -
-300
3
1,0-1
08-
06-
0.4-
0.2-
pH
0.0-
avemjp
year
100 200 300
ANC Omol/L)
1
2
3
4
5
6
7
8
9
10
11
12
ANC
Figure 5-4. A. The relationship between pH and ANC under equilibrium conditions
with mineral phase gibbsite. Triangles indicate calculated values while
circles indicate measurements (Bi and Liu 2001). B. The relationship
between precipitation (wet and average year), ANC and the risk to
exposure of a pH below 5.5 (Gerritsen et al. 1996).
EPA staff thus concludes that the available information provides support for the
use of ecological indicators to characterize the responses of aquatic ecosystems to
acidification effects associated with N and S deposition, and that ANC is the most
supportable indicator.
September, 2010
5-13
Draft - Do Not Quote or Cite
-------
1 5.3.2.2 Relating the ecological indicator to atmospheric deposition: Does a
2 quantified relationship exist between the ecological indicator and atmospheric
3 deposition of nitrogen and sulfur?
4
5 There is evidence to support a quantified relationship between deposition of N
6 and S, and ANC. This relationship was analyzed in the REA to determine current risk for
7 two case study areas, the Adirondack and Shenandoah Mountains, with three techniques
8 (1) a time series analysis to evaluate long-term trends using MAGIC modeling to
9 reconstruct past trends, (2) a time series analysis using monitoring data to evaluate recent
10 observed trends and (3) a critical load approach using water chemistry data from the
11 Temporally Integrated Monitoring of Ecosystems (TIME) program and Long-term
12 Monitoring (LTM) to calculate critical loads with the SSWC and FAB aquatic
13 acidification models. A technical summary of the methods used in the REA are provided
14 in Appendix A.
15 Modeled lone -term trends over time
16 In the REA analysis, long-term trends in surface water nitrate, sulfate and ANC
17 were modeled using Model of Acidification of Groundwater in Catchment (MAGIC) for
18 the two case study areas. These data were used to compare recent surface water
19 conditions (2006) with preindustrial conditions (i.e. preacidification 1860). The results
20 showed a marked increase in the number of acid impacted lakes, characterized as a
21 decrease in ANC levels, since the onset of anthropogenic N and S deposition (REA
22 appendix 4 section 5)
23 Observed recent trends over time
24 In the REA, more recent trends in ANC, over the period from 1990 to 2006, were
25 assessed using monitoring data collected at the two case study areas. In both case study
26 areas, nitrate and sulfate deposition decreased over this time period. In the Adirondack
27 Mountains, this corresponded to a decreased concentration of nitrate and sulfate in the
28 surface waters and an increase in ANC (REA 4.2.4.2). In the Shenandoah Mountains,
29 there was a slight decrease in nitrate and sulfate concentration in surface waters
30 corresponding to modest increase in ANC from 50 ueq/L in 1990 to 67 ueq/L in 2006
31 (REA 4.2.4.3 and REA appendix 4 section 3.4).
32
September, 2010 5-14 Draft - Do Not Quote or Cite
-------
1 Critical loads
2 In the REA, the quantified relationship between deposition and ANC was
3 investigated using ecosystem acidification models, also referred to as acid balance
4 models or critical loads models (REA Chapter 4 and REA Appendix 4). These
5 models quantify the relationship between deposition of N and S and the ability of an
6 ecosystem to counterbalance or buffer the deposition. The ecosystem acidification
7 models simulate a variety of water and soil acidification responses at the laboratory, plot,
8 hillslope, and catchment scales. For example, the level of deposition that causes a
9 specified level of an ecosystem endpoint could be calculated (e.g. a critical load for
10 ANC=50/Jeq/L). A summary of acidification models is given in ISA appendix A and
11 further discussed in section 5.3.2.3. In the REA analysis, critical loads and their
12 exceedances were calculated for four values of ANC (i.e., ANC of 0, 20, 50, and 100
13 /^eq/L) for 169 lakes in the Adirondack Mountains and 60 streams in the Shenandoah
14 Mountains.
15 In summary, EPA staff concludes from the REA analysis, in combination with
16 information presented in the ISA, that a quantitative relationship exists between the level
17 of surface water ANC and an amount of nitrogen and sulfur deposition. These
18 relationships are shown by long-term trends going back to preindustrial conditions in the
19 1860s, recent trends since the 1990s and critical loads modeling.
20
21 5.3.2.3 Relating the ecological indicator to deposition: How do steady state models
22 compare to dynamic models?
23
24 Models are important tools to evaluate how multiple environmental factors alter
25 the relationship between ANC atmospheric deposition. There are two general types of
26 acidification models: steady-state and dynamic. These models make different
27 assumptions, indicate different time horizons for their critical loads and they have
28 different data requirements.
29
30 Basic approach of steady-state vs. dynamic acidification models
31 The basic principle of the steady-state approach of aquatic acidification models is
32 to determine the maximum acid input that will result in adequate biogeochemical
September, 2010 5-15 Draft - Do Not Quote or Cite
-------
1 conditions to sustain ecosystem health. Adequate biogeochemical conditions is a subjective
2 term that relates to a particular benchmark (e.g. ANC = 20, 50, 100), representing
3 different degrees of protection of aquatic ecosystems against acidic deposition. The steady-
4 state models relate an aquatic ecosystem's critical load to the weathering rate of its
5 drainage basin expressed in terms of the base cation flux. The weathering of soil minerals
6 is often a major source of base cation supply to an ecosystem. It is considered one of the
7 governing factors of ecosystem critical loads.
8 A dynamic model includes mathematical descriptions of processes that are
9 important in controlling the chemical response of a catchment. One of the most well-
10 known dynamic models of aquatic acidification is MAGIC (Cosby et al., 1985a; 1985b;
11 1985c). It is a lumped-parameter model of soil and surface water acidification in
12 response to atmospheric deposition based on process-level information about
13 acidification. "Lumped-parameter" refers to the extent that spatially distributed physical
14 and chemical processes in the catchment are averaged or lumped together without
15 affecting the model's reproduction of catchment response. Process-level information
16 refers to how the model characterizes acidification into (1) a section in which the
17 concentrations of major ions are assumed to be governed by simultaneous reactions
f\
18 involving 864 " adsorption, cation exchange, dissolution-precipitation- speciation of
19 aluminum, and dissolution-speciation of inorganic carbon; and (2) a mass balance section
20 in which the flux of major ions to and from the soil is assumed to be controlled by
21 atmospheric inputs, chemical weathering, net uptake and loss in bio mass and losses to
22 runoff. At the heart of MAGIC is the size of the pool of exchangeable base cations in the
23 soil. As the fluxes to and from this pool change over time owing to changes in
24 atmospheric deposition, the chemical equilibria between soil and soil solution shift to
25 give changes in surface water chemistry. The degree and rate of change of surface water
26 acidity thus depend both on flux factors and the inherent characteristics of the affected
27 soils.
28
29 Trajectory of recovery for ecosystems from CL calculated by steady-state vs. dynamic
30 acidification models
September, 2010 5-16 Draft - Do Not Quote or Cite
-------
1 Steady-state models assume that the ecosystem is in equilibrium with the critical
2 load of deposition; therefore the long-term sustainable deposition is indicated. This is the
3 relevant information needed to provide protection from deposition in perpetuity as the
4 system comes into equilibrium with the pollutant critical load (ISA Appendix D). In the
5 U.S., few (if any) ecosystems qualify as steady-state systems. Therefore the assumption
6 of equilibrium in the steady-state model is often false. This has implications for the
7 temporal aspects of ecosystem recovery. The steady-state models give no information
8 concerning the time to achieve the equilibrium or what may happen to the receptor along
9 the path to equilibrium. The recovery of an ecosystem based on a critical load from a
10 steady state model may take several hundred years. In other words the assumption that
11 attainment of a deposition values below the steady-state critical load will result in
12 biological recovery within a specified time period may not be valid.
13 Dynamic models calculate time-dependent critical loads and therefore do not
14 assume an ecosystem is in equilibrium. This is the relevant information needed to provide
15 protection from damage by the pollutant within a specified time frame. As a general rule,
16 the shorter the time frame selected, the lower the critical load.
17 The most comprehensive study done in the United States is Holdren et al. 1992
18 that compared critical loads calculated by the dynamic MAGIC model versus SSWC
19 steady-state approach. A 50-yr simulation critical load was obtained from the MAGIC
20 model. Holdren et al. 1992 found that both models yielded the same general trends. The
21 critical load estimates projected using the dynamic versus steady-state models are
22 consistently higher. Both model produced critical load values approximately equal for
23 systems with critical loads of about zero. However, at higher critical load values the two
24 model outputs diverge rapidly, implying that watersheds with larger inherent buffering
25 capacities respond more slowly to a given level of acidic deposition. The apparent
26 reason for this is that the watersheds represented by the dynamic model retain a larger
27 fraction of their buffering capacity in the base cation exchange pool for the 50-yer time
28 scale of the simulation. In the Steady-state models, the cation exchange pool is assumed
29 to be in equilibrium and does not provide additional buffering be on what results during
30 equilibrium conditions.
31
September, 2010 5-17 Draft - Do Not Quote or Cite
-------
1 Data Requirements of steady-state vs. dynamic acidification models
2 There are various factors that modify the ANC to deposition relationship, which
3 are described by models that parameterize ecosystems to simulate the process of
4 acidification. The steady-state models used for critical loads analysis in the REA
5 required input data for between 17 and 20 variables, including water chemistry data from
6 the TIME and LTM programs, which are part of the Environmental Monitoring and
7 Assessment Program (EMAP). A summary of the variables for steady state models (and
8 data sources for the calculations made in the REA) is given in Appendix A.
9 The data requirements required to run dynamic models, such as MAGIC, are
10 greater. The equations that characterize the chemical composition of soil water in
11 MAGIC contain 33 variables and 21 parameters (Cosby et al. 1985). Data required to
12 conduct dynamic modeling are not available for as many places as the data required to
13 conduct steady-state modeling.
14
15 Comparison of two steady-state models: FAB andSSWC
16 The steady state models used in the REA were the Steady State Water Chemistry
17 model (SSWC), and the First-order Acid Balance model (FAB). The SSWC and FAB
18 models were used to calculate critical loads for specified ANC levels in the case study
19 areas.
20 The SSWC and FAB make different assumptions of ecosystem function. Most
21 notably, biogeochemical pathways of N deposition are considered differently in the two
22 models. In the SSWC model, sulfate is assumed to be a mobile anion (i.e. S leaching = S
23 deposition), while nitrogen is retained in the catchment by various processes. The
24 assumption that all N is retained by the ecosystem and does not contribute to acidification
25 is incorrect because in many ecosystems nitrate leaching is observed. If nitrate is leaching
26 out of an ecosystem, it cannot also be true that it has all been retained. Nitrate leaching is
27 determined from the sum of the measured concentrations of nitrate in the runoff. The
28 critical load for sulfur that is calculated by SSWC can be corrected for the amount of
29 nitrogen that contributes to acidification. When an exceedence value for the critical load
30 is calculated, the critical load is subtracted from S deposition plus the amount of nitrate
September, 2010 5-18 Draft - Do Not Quote or Cite
-------
1 leaching, as it represents the difference between N deposition and N retention by the
2 ecosystem. N leaching data used in this calculation are considered robust.
3 In contrast to the SSWC approach, the FAB model includes more explicit
4 modeling of N processes including soil immobilization, denitrification, and wood
5 removal, in-lake retention of N and S, as well as lake size. Although N cycling is more
6 detailed in the FAB model, there is greater uncertainty in the input data needed to
7 characterize the components of the N cycle. The FAB model yields a deposition load
8 function for a specified level of an endpoint. This function is characterized by three
9 nodes that are illustrated on Fig 5.5: 1. the maximum of amount of N deposition
10 when S deposition equals zero (DLmax (N)); 2. the amount of N deposition that
11 will be captured by the ecosystem before it leaches (DLmin(N)); and 3. the
12 maximum amount of S deposition considering the N captured by the ecosystem
13 (DLmax (S)). The function represents many unique pairs of N and S deposition that
14 will equal the critical load for acidifying deposition. The slope portion of the
15 function will vary according to attributes of the water body that is modeled,
16 including lake size and in-lake retention.
17
18
19
20
21
22
23
24
H Dspoeltion
Figurer 5-5 Illustration of a generalized N + S deposition tradeoff curve that is
calculated by using the FAB approach
The EPA staff concludes that the available information supports using the steady-
state acidification models to characterize the relationship between the ANC ecological
indicator and total nitrogen and sulfur deposition. The steady state models take a simple
September, 2010
5-19
Draft - Do Not Quote or Cite
-------
1 mass balance approach to characterizing ecosystem acidification and data is available for
2 over 8,000 locations to conduct steady-state modeling.
3
4 5.3.2.4 Relating the ecological indicator to deposition: What is the appropriate
5 ecosystem acidification model to represent the relationship between nitrogen and sulfur
6 deposition, and the ecological indicator?
7
8 A combination of the SSWC and FAB models is proposed by EPA staff for
9 catchment scale modeling for use in developing critical loads to specify the form of the
1 0 standards. A modified SSWC model is used because there is high confidence in the
1 1 availability and quality of the input data that is required by this model. The SSWC model
1 2 for aquatic acidification is expressed as equation 1 .
13
14 CLN + SBC-ANCQ (1)
15
16 Where,
1 7 CLANciim(N+S) = depositional load of S and N that does not cause the ecosystems to
/\ i
1 8 exceed a given ANQim (meq/m yr)
19 [BC]o* = the preindustrial concentration of base cations (/Jeq/L)
20 ANCiimit = a ANC limit (peq/L)
21 Q= surface water runoff (m/yr) (this is typically equal to precipitation -
22 evapotranspiration
23
24 This model is further constrained by a quantity of N which would be taken up,
25 immobilized or denitrified by ecosystems and used to adjust the quantity of deposition
26 required to meet a specified critical load. This term is represented as CLmin(N) in the
27 FAB model and illustrated in Fig. 5.5. For application in the form of the standard and in
28 the following discussion, the parameter is designated with the abbreviation Neco. The
29 acid-base model constrained by Neco is expressed by equation 2.
30
3 1 CLANC,m (N + S) = ([BC]0 - [ANCUm Jg + Neco (2)
September, 201 0 5-20 Draft - Do Not Quote or Cite
-------
1
2 Where,
3 Neco= nitrogen retention and denitrification by terrestrial catchment and nitrogen retention
4 in the lake
5 Two options to derive the term Neco were considered, each yielding different
6 results. The first approach is to calculate the long-term amount of N an ecosystem can
7 immobilize and denitrify before leaching (i.e. N saturation) that is derived from the FAB
8 model [denoted as DLmin(N) in the FAB model]. This approach requires the input of
9 multiple ecosystem parameters and is expressed by eq 3.
10
11 Neco = JNupt + Nret + (1 - r\Nimm + Nden) (3)
12
13 Where,
14 Nupt= nitrogen uptake by the catchment (meq/m2/yr)
15 Njmm= nitrogen immobilization by the catchment (meq/m2/yr)
16 Nden=denitrification of nitrogen in the catchment (meq/m2/yr)
17 Nret= in-lake retention of nitrogen (meq/m2/yr)
18 f =forest cover in the catchment (dimensionless parameter)
19 r = fraction lake/catchment ratio (dimensionless parameter)
20
21 The second approach for estimating Neco is to take the difference between N
22 deposition and measured N leaching in a catchment as expressed by eq 4.
23
24 Nem=DepNTotal-Nleach (4)
25
26 CAS AC has advised EPA to use the second approach to derive the Neco term:,
27 "In principal, Equation 3, captures many of the major landscape and
28 ecological factors that influence the processing of atmospherically deposited
29 N within an ecosystem, including biological and abiotic retention of N
30 (immobilization, uptake and sedimentation), and gaseous loss (denitrification)
31 after N has been deposited and/or transported. This approach opens the "black
32 box" and attempts to estimate some of the component parts of N processing
33 and loss. It is best modeled as a dynamic process because such factors as age
September, 2010 5-21 Draft - Do Not Quote or Cite
-------
1 and stage of vegetation, soil moisture, moisture regime, and nutrient demand
2 will affect various components of N cycling. For many ecosystems, complete
3 supporting data may not be available.
4 Equation 4 is basically a mass balance approach that keeps track of
5 inputs and outputs. It does not include the detailed biogeochemical processes
6 identified in Equation 3. Data for this latter approach may be much more
7 readily available than the approach using Equation 3. Watershed N retention
8 can be estimated using this approach (e.g., Lovett et al. 2000, and many
9 others), but the processes involved in the retention are not detailed. This
10 approach is most effectively used when hydrologic boundaries (e.g.,
11 watershed) are defined. Hence, the Panel recommends that the mass-balanced
12 (i.e., Equation 4) approach should be used." (CASAC, April 29, 2010)
13
14 (A comparison of calculating Neco using both methods is provided in Appendix A)
15
16 Two methods to calculatepre-industrial base cation weathering: F factor and MAGIC
17 The preindustrial concentration of base cations ([BC]o*) is calculated to represent
18 conditions prior to industrialization (-about 1860). It incorporates the main source of
19 base cations to an ecosystem including preindustrial weathering from soil and pre-
20 industrial base cation deposition. It is therefore considered one of the governing factors
21 of critical loads. [BC]o is commonly calculated using one of two approaches: dynamic
22 modeling (i.e. MAGIC) and calculation by the F-factor approach (Henriksen and Posch
23 2001). Staff propose that the base cation weathering rates may be calculated by either
24 method.
25 In this section, critical load estimates obtained from two steady-state approaches
26 were compared. The exercise is not intended to provide an assessment of the accuracy of
27 the two models, but rather to provide a means for evaluating the relative performance of
28 the two different models. The EPA conducted analysis compared steady-state CL values
29 based on Henriksen and Posch (2001) F-factor approach and output from the MAGIC
30 model. The primary purpose of the F-factor is to obtain estimates of preindustrial surface
31 water base cation concentrations ([BC]o ) for equation 1. The MAGIC (Cosby et al.,
32 1985) model can also be used to derive preindustrial surface water base cation
33 concentrations. MAGIC is a process-based model designed to mimic the geochemical
34 effects of mineral weathering, soil cation exchange, and other watershed processes. Once
35 the model has been calibrated for a watershed, it can be run to simulate how surface
September, 2010 5-22 Draft - Do Not Quote or Cite
-------
1 water chemistry changes with time and to predict preindustrial cation concentrations
2 ([BC]o* ) to be used in the steady-state SSWC CL model in equation 1.
3 The comparison of CLs between the F-factor and MAGIC approached was done
4 for two regions, streams in Southern Appalachia and lakes in the Adirondack Mountains.
5 For 67 streams and 99 lakes, [BC]o* were determined for both approaches and CLs were
6 calculated using the same value of Q. The results are show in Fig 5.6. For this analysis,
7 the steady-state MAGIC model yielded critical load values that show the same trend for
8 both regions, but were on average 16 meq S /m2-yr for the Adirondack Lakes and 5 meq
9 S/m2"yr for Southern Appalachia streams higher than those from the SSWC F-factor
10 approach. The two models converge at low critical, but diverge as the buffering potential
11 for watersheds increase. This is particularly the case for CL values above 80 meq/m2-
12 yr for lakes in the Adirondack Mountains. These results are consistent with similar
13 comparison of critical load done by Holdren et al. 1992. Holdren et al 1992
14 found on average that the MAGIC model yielded CL values that were on average
15 29 meq S /m2-yr higher than those from the SSWC model. Holdren et al 1992 also
16 found that as the buffering potential for watersheds increased, as indicated by increasing
17 CLs, the results from the two models gradually diverge.
September, 2010 5-23 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
y = 0.9976X - 4.591
R2 = 0.8453
-100
-100
100
200
300
400
SSWC CL
MAGIC - BCo
meq/m2/yr
B
o
_l o
o
-------
1 applied at the spatial scale of the catchment. Response to N and S deposition will vary
2 among catchments. However, modeling every catchment in a region (i.e. a spatial area
3 that includes a large population of individual catchments) is implausible at this time due
4 to the data limitations. A method to extrapolate watershed-scale analysis to a region was
5 developed in the REA. In that method, the critical loads (combined N+S load for a
6 selected ANC level) developed for the case study sites were applied over a region using
7 water chemistry data. Critical load exceedance (i.e., the amount of actual deposition
8 above the critical load, if any) was calculated for each waterbody in the region to quantify
9 the number of lakes or streams that receive levels of deposition exceeding the CL. Lakes
10 and streams with positive exceedance values, where actual deposition was above its
11 critical load, were not protected at that critical limit (e.g. ANC= 20, 50, 100 /ieq/L; REA
12 Appendix 4).
13 The relationship between ANC and N and S deposition is also applicable
14 nationally. Areas that have similar geologic underpinnings, mineral weathering rates, and
15 hydrology should show similar sensitivity to NOy and SOx deposition. Weathering rate
16 of geologic parent material is the main source of base cations to an ecosystem, and it is
17 therefore considered one of the governing factors of ecosystem critical loads. Landscape
18 features that are linked to ecosystem acid-sensitivity include lithology, elevation, percent
19 forested watershed, and watershed area (Sullivan et al. 2007).
20 How do we use catchment scale acidification modeling to inform the NAAQS? The
21 goal is to have a national applicable standard, but still take into account catchment base
22 processes. As previously noted, acidification models are parameterized for catchments
23 and their critical loads vary at the spatial scale of the catchment. However, aggregating
24 critical loads from multiple catchments will allow for an appropriately representative
25 deposition value, which provides adequate protection of ecosystems and can be applied
26 over larger spatial areas. For this reason, EPA Staff proposes evaluating a population of
27 waterbodies that are sensitive to acidification to calculate a benchmark depositional load
28 (called the deposition metric) in which a specified percentage of the population does not
29 exceed their critical load for a selected ANC limit. This approach uses the distribution of
30 critical loads from a population to derive a value that is intended to provide protection
September, 2010 5-25 Draft - Do Not Quote or Cite
-------
1 over a spatial area that is larger than the individual catchment for which a single critical
2 load may be calculated. The deposition metric is expressed by the following equation:
3
4
5 DL%ECO (N + S} = {([Bel - [ANC,m ])Q}%ECO (5)
6
7 The change from the CL term to the DL term reflects the change from a critical load
8 for deposition to a specific catchment to a deposition load that applies across a population
9 of catchments in a broader area. NECO would then be added to DL%Eco to calculate the N
10 and SO2 + SO'2 tradeoff curves (section 5.3.2.8)
11
12 5.3.2.6Deposition metric: What critical load data is available and what populations of
13 catchments are excluded? For any acid sensitive category, how is the distribution
14 of critical loads evaluated in developing the deposition metric?
15
16 Steady state critical loads calculated for ANC that were used in this analysis were
17 either previously published in the literature or calculated using the SSWC method
18 from water quality data collected by national monitoring programs. A map of critical
19 loads is presented in Fig. 5.7.
20
September, 2010 5-26 Draft - Do Not Quote or Cite
-------
1
2
3
4
Figure 5-7 Sites of CLs calculated by SSWC that are used in the Policy
Assessment analysis
To ensure the population of water bodies included in the analysis were those
6 sensitive to acidity caused by atmospheric deposition, several criteria were applied to the
7 CL dataset to remove watersheds in which organic acids, acid mine drainage or naturally
8 low base cation weathering caused acidification.
9 • The pre-industrial ANC was calculated for each site in which there was
10 sufficient data. Certain pre-industrial ANC values, defined as ANC< the target
11 ANC( e.g. 50 /zeq/L), indicate natural acidity due to naturally low BC
12 weathering and were removed from the data set.
13 • All CL <10 meq/m2/yr were removed. This is a second screen to remove
14 catchments that may be naturally acidic for which pre-industrial ANC values
15 could not be calculated.
16 • To remove catchments that are dominated by acidity caused by acid mine
17 drainage, if the concentration of SC>4~2of the water was >400 ueq/L twice or
September, 2010
5-27
Draft - Do Not Quote or Cite
-------
1 more than expected by S deposition the catchment was considered to be
2 dominated by acid-mine drainage and removed from the dataset.
3 • The cause of aquatic acidity was considered to be dominated by organic acids
4 if the concentration of DOC in the water was > lOmg/L, sites exceeding 10
5 mg/L were excluded from the dataset.
6
7 The latitude and longitude of each site was used to determine its membership to
8 an acid-sensitive category. Multiple methods for defining an acid sensitive category are
9 presented in section 5.3.2.7. The distribution of critical loads from a population that
10 represents an acid-sensitivity category was then evaluated. A deposition value that
11 protects a specified percentage of the ecosystems (called the deposition metric) would be
12 selected. Each acid-sensitivity category would have a deposition metric that protects the
13 same percentage of ecosystems (see Section 5.4-5.6 for a comprehensive discussion of
14 the levels). The deposition metric would be a value representing a specified percentile of
15 the distribution, such as the 95th percentile. The distributions of critical loads for many
16 definitions of acid sensitive areas are skewed to the right, which implies that a central
17 estimate using the mean, would not be projected to achieve the target ANC of the
18 majority of acid-sensitive ecosystems; therefore it may be preferable to calculate the
19 deposition metric to target protecting a higher percentage of catchments. For example, if
20 the 75% or 90% of the aquatic ecosystems are targeted for protection, then the deposition
21 metric that represents the critical load for the 75h or 90th percentile of the population
22 would be selected.
23
24 5.3.2.7 Landscape categorization: How is the U.S. landscape categorized based on acid-
25 sensitivity?
26 Two general approaches are suggested. The first approach is to have one
27 population of water bodies represent the entire nation and select a deposition metric from
28 this one large population. In this approach the U.S is not subdivided into acid-sensitivity
29 categories. The second approach is to subdivide the U.S. into acid-sensitivity categories
30 based on criteria that are known to be associated with acid-sensitivity and for which
31 national-scale data sets are available. A population of catchments could then be
September, 2010 5-28 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
identified to represent each acid-sensitive category. Several categorization possibilities
are discussed under the second approach.
Option 1: No subdivision of the U.S.
Option 1 is a simple approach in which the U.S. is not subdivided into acid-
sensitivity categories and the total population of critical loads represents the entirety of
the U.S. The deposition metric would be selected by evaluating the distribution of CL
from the entire population after screening out preindustrial ANC< 50 ^eq/L and CL <10
meq/m2/yr, as described in section 5.3.2.6. Deposition metrics that protect 90%, 75% and
50% of the population are given as examples in Table 5.1. Figure 5.8 shows the
cumulative distribution of critical loads of waterbodies at ANO50 /*eq/L. Note that for
an ANC=50 jieq/L the greatest CL is 4179 meq/m2/yr, while 50% of the waterbodies have
CLs less than 118 meq/m2/yr, thus indicating the skewed distribution of the data set.
Table 5.1. Descriptive statistics of the national CL data set.
AN
C
Meq/
L
50
20
#
CL
749
7
749
7
Mea
n
Nee
0
52
52
After
excluding
sites for
sulfate>40
0, DOC
>10, pre-
industrial
ANC<
target
ANC, then
excluding
sites with
CL<10,
the
resulting
CLin
analysis
n=
5280
5613
CL mean
(meq/m2/y
r)
257
259
C
L
St
er
5
4
Example deposition metrics
DL %90
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Cumulative Distributuion of Critical Loads
(ANC =50)
ll
Q.
8.2
••= E
•ft o
100 f
80
60
II
e 2 20
0 --
0
of the population
500
1000 1500 2000 2500 3000
Deposition N + S (meq/m2/yr)
3500
4000
Figure 5-8 Cumulative distribution of the critical loads for ANC =50 jteq/L
considered in the analysis (n =5281)
The benefit to this approach is that it provides one deposition metric for the entire
country, making it simple and easy to calculate and convey. The main weakness of this
approach is the degree of uniformity in protection across the nation, as that it may over-
protect areas of the nation that are least sensitive to acidification and under protect in
areas that are most sensitive. The extent of possible over or under protected is discussed
at the end of this section.
Option 2: Acid-sensitivity categorization of the landscape
In Option 2, four approaches to categorize the U.S based on criteria that influence
acid-sensitivity are presented: 1) binary categorization (sensitive vs. less-sensitive), 2) 5
categories of sensitivity (based on cluster analysis), 3) one less-sensitive category and
sensitive ecoregions considered individually and 4) all ecoregions considered
individually. The extent to which current critical loads represent the categories as defined
by each approach is discussed. Each approach presented uses the concept of ecoregions
to define spatial boundaries of sensitive ecosystems and ANC to determine acid-
September, 2010
5-30
Draft - Do Not Quote or Cite
-------
1 sensitivity. Numerous options were explored, but we have chosen to focus this
2 discussion on those options recommended by the staff as most likely to be useful in
3 setting the standards. Other parameters for which to base acid-sensitivity categorization
4 that were explored include bedrock geology, soil base cation weathering, dissolved
5 organic carbon and alkalinity. Staff suggests moving forward with ANC because it is the
6 best indicator of acid-sensitivity integrating natural buffering capacity, historic load and
7 current load, and there is good national coverage of ANC measurement from almost
8 every ecoregions (level 3). A brief summary of ecoregions and ANC data that were used
9 in the analysis is presented here, prior to the presentation of the options.
10 Ecoregions
11 Ecoregions are "areas of similarity regarding patterns in the mosaic of biotic,
12 abiotic, aquatic, and terrestrial ecosystem components, with humans being considered
13 part of the biota" (McMahon and others 2001). EPA staff proposed considering
14 Omernik's ecoregions in this analysis of acid-sensitivity (maps may be downloaded at
15 www.epa.gov/wed/pages/ecoregions.htm). These ecoregions are categorized using a
16 holistic; "weight-of-evidence" approach where the importance of factors may vary. The
17 method used to map ecoregions is,
18 "based on the premise that ecological regions can be identified through
19 the analysis of the patterns and the composition of biotic and abiotic phenomena
20 that affect or reflect differences in ecosystem quality and integrity (Wiken 1986;
21 Omernik 1987, 1995). These phenomena include geology, physiography,
22 vegetation, climate, soils, land use, wildlife, and hydrology. The relative
23 importance of each characteristic varies from one ecological region to another
24 regardless of the hierarchical level".
25 To evaluate the effectiveness of ecoregions, they must be tested against measures that
26 relate to their purpose (Omemik 2004). The first publication of the ecoregions based on
27 Omernik's weight-of-evidence approach was published in 1987 (Omernik 1987). Current
28 maps are refinements and revisions of the 1987 publication. Hierarchical levels were
29 developed and a Roman numeral classification scheme was adopted to distinguish coarser
30 (more general) and finer (more detailed) categorization. Level I is the coarsest level,
31 dividing North America into 15 ecoregions (CEC 1997). At Level II, the continent is
32 subdivided into 52 ecoregions (CEC 1997). Level III, is a farther subdivision of Level II,
September, 2010 5-31 Draft - Do Not Quote or Cite
-------
1 and divides N. America into 120+ categories. Level IV is a subdivision of Level III, it is
2 the finest scale of Omernik's ecoregions mapping and development of these maps is
3 currently in progress. Ecoregion level 3 are used here to categorize the landscape.
September, 2010 5-32 Draft - Do Not Quote or Cite
-------
u
t-l
o
0>
4
WJ
g
o
u
.a
S
I
o
en
1
•s
O\
o
Q
I
Q
O
(N
tf
,0
on
-------
1 ANC data
2 Water chemistry data on ANC was collected from several national monitoring networks. The
3 data sources are summarized on Table 5.2. These data sources included approximately 11,000
4 observations. For sites with more than one year of data, the measurements were averaged over
5 the past 5 years. For example, if data for 2008, 2007, 2006, 2005, 2004, and 2003 were
6 available, then the years from 2004-2008 were averaged. If it had data from 1990 and 2002 they
7 were not averaged because the 1990 point was beyond the cutoff date.
Table 5-2 Summary of data sources considered for the evaluation of national ANC
Program
EPA Long Term Monitoring Vermont
(LTM VT)
EPA Eastern Lakes Survey (ELS)
Adirondack Lake Survey (ALS)
EPA Western Lake Survey (WLS)
EPA National Stream Survey (NSS)
VTSSS
EPA Long Term Monitoring Colorado sites
(LTM CO)
EPALong Term Monitoring Midwest Sites
(LTM MW)
VT SSS LTM
EPA Long Term Monitoring_Pennsylvania
sites(LTM PA)
EPA Long Term Monitoring Catskill sites
(LTM CAT)
EPA Long Term Monitoring: Annual average
from 1992-2007
EPA EMAP Northeast Lake Survey
EPA Long Term Monitoring_Maine sites
(LTM ME)
Regional Environmental Monitoring
Program_Maine sites (REMAP _ME)
EPA EMAP Mid Atlantic streams (EPA
EMAP MAIA)
EPA EMAP Mid Atlantic streams (EPA
Dates of
observations
1983-2007
1984
1984-1987
1985
1986
1987 & 2000
1990-1994
1990-2000
1990-2007
1990-2007
1990-2007
1990-2007
1991-1994
1992-2007
1993
1993-1996
1997-1998
Reference
EPA/903/R-00/015
EPA/620/R-93/009
Stoddard.et.al.WRR. 19
96
EPA 620-R-05-005
Stoddaid.et.aLWRR.19
96
EPA 841-B-06-002
Stoddaid.et.aLWRR.19
96
Stoddard.et.al.WRR.19
96
Stoddard.et.al.WRR.19
96
Stoddaid.et.al.WRR. 19
96
EPA 905-R-92-001
EPA/600/4-88/032
Stoddard.et.al.WRR.19
96
regl qa.pdf
Stoddaid.et.aLWRR.19
96
EPA/R-06/XX
EPA-600-388-021a
September, 2010
5-34
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
9
10
11
Table 5-2 Summary of data sources considered for the evaluation of national ANC
Program
EMAP _MAIA )
EPA EMAP Western Stream and River
Survey (EMAP WEST )
EPA National Lakes Survey (NLS)
USGS NAWQA Program
EPA Storet Program
Dates of
observations
2000-2004
2010
Reference
EPA/600/3 -86/054b
EPA 841-F-09-007.
http://water.usgs.gOV/n
awqa/
http : //www. epa. go v/sto
ret/
Option 2a: Categorize the U.S. into 2 categories (sensitive or less—sensitive) based on ANC
and ecoregions
In this approach, ANC, considered the best indicator of acidification, is used to determine
which ecoregions in the U.S. are sensitive to acidification. All ecoregions in the U.S. are then
6 categorized as either sensitive or less-sensitive. A deposition metric is calculated for each
7 category. The data sources used in this analysis are described first, followed by a more detailed
8 description of the analysis.
Defining acid-sensitivity of ecoregions based on ANC
This analysis compared whether mean ANC differed between ecoregions when the entire
12 dataset versus a subset of the dataset containing observations that ranged from -50 to 200 //eq/L
13 was used in the analysis. The sample size and mean ANC value for each ecoregion is presented
14 in table 5.3. In the data subset, the values ranging from -50 to 200 /ueq/L were selected to focus
15 on acid-sensitive areas, all ANC observations that were greater than 200 /ieq/L were excluded
16 from the dataset because ANC above this level indicates that an eco-system is not very sensitive
17 to acidification. All observation less than -50 pieq/L were also deleted from the dataset. This is
18 because ANC values below -50 /ieq/L are typically not the result of acidification resulting from
19 deposition (Kauffman et al. 1992). Both analyses indicate that ecoregions 5, 6, 7 and 8 have the
20 among the lowest mean ANC values compared to the other five ecoregions. After observations
21 outside of the -50 to 200 /ieq/L are removed, ecoregions 12, 13, arelS are represented by five or
22 less observations, indicating there are few sites measured with ANC values within the range of
23 acid-sensitivity due to deposition. Therefore ecoregions 5 through 11 are considered sensitive
24 and all other ecoregions are considered less sensitive.
September, 2010
5-35
Draft - Do Not Quote or Cite
-------
1
2
3
4
Next only those level three subdivisions of the ecoregions with 10 or more observations
between -50 to 200 fieq/L were considered sensitive. Fig 5.10 lists the level three ecoregions
which are indicated to be sensitive by ANC data.
Table 5-3 National ANC data. Observations from 6318 sites remained after filtering
data to remove sites with an average value of ANCX-50 and >200 ^eq/L. Color
shaded boxes indicate ecoregions that are considered sensitive based on the ANC values.
Kruskal-Wallis Test indicated ecoregions has a significant effect on ANC.
Ecoregion
level 1
5
6
7
8
9
10
11
12
13
15
All ANC data
# sites
3519
1952
177
5051
1098
592
209
19
102
12
Mean ANC
212
892
775
605
4694
2677
2663
2496
2869
406
ANC data between -50 and 200 ^eq/L
# sites
2603
754
30
2680
94
112
35
1
3
5
Mean ANC
57
90
68
77
149
134
93
198
146
164
5
6
September, 2010
5-36
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
Figure 5-10 Map of sensitive ecoregions (red) using binary categorization approach
(sensitive vs. less-sensitive) with ANC observations between -50 and 200
Table 5-4. Descriptive statistics of the CL populations that result when the US is divided
into two categories, sensitive and less-sensitive based on ANC data.
ANC
/xeq/L
50
50
20
20
Sensitivity
class
sensitive
less-
sensitive
sensitive
less-
sensitive
CLin
class
n=
4553
121
4881
732
Mean
Neco
55
37
55
37
Neco
Ster
0.4
0.9
0.3
0.9
CLmean
(meq/m2/yr)
219
496
222
508
CL
St
er
4
21
4
21
Example deposition metrics
DL %90
(meq/m2/yr)
26
53
32
63
DL %75
(meq/m2/yr)
51
117
58
122
DL %50
(meq/m2/yr)
106
211
114
291
6
7
September, 2010
5-37
Draft - Do Not Quote or Cite
-------
1 Dividing the U.S. into two categories (sensitive and less-sensitive) provides a mechanism
2 to reduce possible over-protection that may occur in less-sensitive regions via Option 1, which
3 does not distinguish between areas that are sensitive and not-sensitive. In Option 2a, the
4 sensitivity category is considered one area, in which one deposition metric could be calculated
5 by evaluating all the CL data from the sensitive ecoregions. This would result in two deposition
6 metrics for the U.S., one for sensitive areas and one for less-sensitive areas. The ecoregions in
7 the sensitive category could be subdivided (See Option 2c).
8
9 Option 2b: Cluster Analysis
\ 0 This approach uses ANC to categorize the acid-sensitivity of ecoregions based on a
11 quantitative cluster analysis. Additional analysis simultaneously considering multiple criteria
12 including ALK, DOC and soil BC weathering datasets were considered, however they are not
13 suggested mainly due to limitation in the national coverage of these datasets, and in the case of
14 soil BC weathering, uncertainty in the method of calculation.
15 Cluster analysis is a method of grouping objects together based on a chosen similarity
16 metric that is determined from a set of variables of interest. In this analysis ecoregions are
17 grouped based on the distributions of ANC from waterbodies. Numerous algorithms exist for
18 performing cluster analysis. We employed the &-means method, which involves assigning a pre-
19 determined number of clusters («) to the dataset and minimizing the Euclidean distances between
20 the cluster centers and the watersheds. Distance in this case refers to the water quality variables,
21 not geographic distance. The process was repeated for a number of different n values (2 to 40),
22 and a combination of fit statistics and graphical examination of the resulting clusters was used to
23 settle on a final solution. The 5 cluster solution was chosen in this way, which has an R2 of 0.71
24 and is shown in Fig 5.11. In selecting the number of clusters, we balanced goals of a
25 parsimonious number of clusters with ability of the clusters to explain a large portion of the
26 variance in ANC across ecoregions.
27 The deposition metrics based on the CL ANC=50 /xeq/L are presented for each cluster in
28 Table 5.5. In general the sensitivity indicated by the mean ANC level of the clusters matches the
29 deposition metrics, with the exception of Cluster 3 and 4 which have deposition metrics
30 indicating they should switch places in regards to sensitivity indicated by the deposition metrics
31 of the 90th, 75th and 50th percentile. This reflects the skewed nature of the distributions of ANC
32 within ecoregions, where ecoregions may be matched more closely on alternative percentiles of
September, 2010 5-38 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
the distribution than others. There are 5 ecoregions which do not have cluster membership
(Table 5.6); this indicates there was insufficient ANC data in these ecoregions for classification.
There are 8 ecoregions that do not have CL data, but do have cluster membership (Table 5.6.).
These ecoregions would be represented by the deposition metric in their cluster.
Mean ANC
167
142
109
80
68
Fig 5.11. Map of sensitive ecoregions based on clusters of log ANC values
between -50 and 200 /ieq/L
Table 5-5. Descriptive statistics of the CL populations that result when the US is divided
into 5 clusters based on log ANC values between -50 and 200 jieq/L.
Target
ANC
50
50
50
50
50
Sensitivity
class
(mean
ANC)
5 (167)
4 (142)
3 (109)
2 (80)
1 (68)
CL
n=
216
655
784
1113
2432
Mean
Neco
(meq/m2/yr)
51
39
41
52
61
Neco
Ster
2
0.7
0.8
1.0
0.5
CL mean
(meq/m2/yr)
354
393
416
183
197
CL
Ster
20
17
18
6
6
Example deposition metrics
DL %90
(meq/m2/yr)
72
30
48
26
24
DL %75
(meq/m2/yr)
114
64
99
52
46
DL %50
(meq/m2/yr)
269
193
232
111
90
September, 2010
5-39
Draft - Do Not Quote or Cite
-------
Option 2c: One less-sensitivity category and individual sensitive ecoregions
This option builds on option 2a. Like option 2a, the CLs from the ecoregions (level 3)
5 considered less-sensitive would be combined into one population to represent the less-sensitive
6 category. The ecoregions (level 3) considered sensitive using the ANC approach described in
7 option 2a would be represented by the CLs only from the same ecoregions. There are 43
8 ecoregions (level 3) considered sensitive using option 2a. The deposition metrics of the sensitive
9 ecoregions are presented in Table 5.6.
10
Table 5-6. Descriptive statistics of the populations of CL ANC=50 peq/L values that result
for each ecoregion
Sensitivity
class
Cluster
5
4
na
5
3
3
3
4
3
5
5
3
3
__ i
5
5
5
1
4
1
1
4
4
2
4
2
AN
C
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Eco-
region
level 3
10.1.2
10.1.3
10.1.4
10.1.5
10.1.6
10.1.7
10.1.8
10.2.1
10.2.2
10.2.4
11.1.1
11.1.2
11.1.3
12.1.1
13.1.1
15.4.1
5.2.1
5.2.2
5.3.1
5.3.3
6.2.10
6.2.11
6.2.12
6.2.13
6.2.14
Neco
mean
28
37
19
27
36
23
39
19
20
22
14
32
38
45
55
79
35
20
17
37
22
CL
N=
22
41
16
43
5
3
7
2
16
5
19
3
52
469
735
159
202
86
105
78
186
CLmean
(meq/m
/yr)
177
95
244
266
106
74
158
75
705
279
367
88
189
158
177
113
230
956
172
245
145
CL
Ster
40
18
47
53
37
8
58
1
211
97
69
41
21
10
9
9
20
92
24
45
12
Example deposition metrics
DL %90
(meq/m2
/yr)
26
18
34
16
22
61
16
73
62
58
41
45
36
16
36
28
26
138
17
21
20
DL %75
(meq/m2
/yr)
37
27
50
41
23
61
73
73
127
88
72
45
76
26
58
52
50
360
28
33
43
DL %50
(meq/m2/y
r)
122
52
258
107
119
71
134
293
262
337
48
120
56
101
90
112
744
68
59
91
September, 2010
5-40
Draft - Do Not Quote or Cite
-------
Table 5-6. Descriptive statistics of the populations of CL ANC=50 /teq/L values that result
for each ecoregion
Sensitivity
class
Cluster
5
3
3
4
3
2
na
4
2
3
2
3
3
na
4
5
5
1
3
3
4
5
5
5
2
5
3
2
1
2
2
4
1
1
3
2
4
3
4
AN
C
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Eco-
region
level 3
6.2.15
6.2.3
6.2.4
6.2.5
6.2.7
6.2.8
6-2.9_j
7.1.7
7.1.8
7.1.9
8.1.1
8.1.3
8.1.2
8.1.4
8.1.5
8.1.6
8.1.7
8.1.8
8.1.10
8.2.1
8.2.2
8.2.3
8.2.4
8.3.1
8.3.2
8.3.3
8.3.4
8.3.5
8.3.6
8.3.7
8.3.8
8.4.1
8.4.2
8.4.3
8.4.4
8.4.5
8.4.6
8.4.7
Neco
mean
38
47
61
21
21
25
32
16
12
4
72
' 72
65
24
90
72
35
81
77
104
50
82
67
75
63
78
65
69
68
82
85
78
70
73
67
CL
N=
86
73
28
133
88
30
108
29
68
11
43
88
45
9
23
169
143
9
3
9
41
12
17
168
75
6
31
4
510
200
19
379
24
14
23
CL mean
(meq/m2
/yr)
155
353
956
425
376
323
244
423
1098
662
716
255
328
961
731
223
177
935
667
892
306
659
1022
174
182
399
211
261
321
160
346
98
938
102
129
CL
Ster
19
35
92
33
29
44
17
64
100
100
102
20
41
125
49
18
16
152
101
76
33
132
147
7
244
242
35
149
22
16
59
5
96
28
22
Example deposition metrics
DL %90
(meq/m2
/yr)
16
46
102
90
60
73
65
21
273
260
133
104
26
158
381
49
51
351
482
642
113
329
429
74
15
52
60
32
31
24
119
21
243
25
33
DL %75
(meq/m2
/yr)
35
119
653
163
116
127
117
185
368
357
280
144
46
794
617
86
70
542
482
674
138
338
558
100
34
98
104
55
53
36
168
38
642
42
64
DL %50
(meq/m2/y
r)
76
228
990
308
284
260
197
325
945
523
541
216
333
946
703
148
108
1011
687
937
218
466
719
155
76
LJ66_
160
156
98
82
316
68
907
67
94
September, 2010
5-41
Draft - Do Not Quote or Cite
-------
Table 5-6. Descriptive statistics of the populations of CL ANC=50 /teq/L values that result
for each ecoreeion
Sensitivity
class
Cluster
5
2
2
1
3
1
1
2
5
5
5
na
5
1
4
5
4
?
5
5
5
3
4
AN
C
X
X
X
X
X
X
X
X
Eco-
region
level 3
8.4.8
8.4.9
8.5.1
8.5.2
8.5.3
8.5.4
9.2.1
9.2.2
9.2.3
9.2.4
9.3.1
9.3.3
9.3.4
9.4.1
9.4.2
9.4.3
9.4.4
9.4.5
9.4.6
9.4.7
9.5.1
9.6.1
Neco
mean
68
61
44
73
59
64
53
21
72
55
32
38
38
59
30
66
69
52
83
62
CL
N=
25
23
16
13
53
43
2
13
16
2
21
3
11
15
2
6
28
5
2
6
CL mean
(meq/m2
/yr)
141
309
159
942
116
62
122
425
365
211
278
79
30
148
36
419
257
265
263
1096
CL
Ster
28
82
39
159
25
6
24_
40
26
101
73
15
3
18
6
55
21
80
179
386
Example deposition metrics
DL %90
(meq/m2
/yr)
30
53
111
223
13
22
98
249
190
110
78
49
17
58
30
268
118
78
84
258
DL %75
(meq/m2
/yr)
65
101
121
461
25
33
98
337
309
110
106
49
20
102
30
315
175
134
84
378
DL %50
(meq/m2/y
r)
94
167
245
987
52
59
122
392
361
211
139
87
27
138
36
398
237
199
263
838
1
2
3
4
5
6
7
Option 2d: All ecoregions
In this approach, each ecoregion (level 3) is represented by CLs that are located from sites within
the spatial boundaries of the same ecoregion (level 3). Eighty-five ecoregions (level 3) occur in
the conterminous U.S., 76 of those are represented by CLs in this analysis. The remaining 9
ecoregions (level 3) for which CLs are not available could be represented by deposition metrics
that are calculated using all of the CL data as described in option 1. This approach yields 77
8 deposition metrics.
9 Comparison
10 In this section, Option 1 and 2a-d are compared to help inform the decision of which
11 option is most appropriate. Table 5.7 summarizes the number of deposition metrics that would
12 result from each option. Option 1 has only one deposition metric. This approach would be
September, 2010
5-42
Draft - Do Not Quote or Cite
-------
1 easiest to communicate and implement. However there is some concern that this approach may
2 under protect the most sensitive areas and over protect the least sensitive areas. Several analyses
3 have been made to determine if and to what extent this concern may be valid.
Table 5-7 Summary of the number of deposition metrics that would result from th4
options for how to categorized the landscape based on acid-sensitivity ^
Acid-
sensitivity
categorization
approach
Op#l
No subdivision
of the U.S.
Op#2a
Binary
categorization
(sensitivity
based on 1 0 or
more
observations of
ANC between -
50 and 200
/ieq/L)
Op#2b
5 Cluster
(based on
centroid of data
between ANC
between -50
and 20 /zeq/L
0)
Op#2c
1 sensitive
category and
individual eco-
region
categories
Op#2d
All ecoregions
^deposition
metrics for
less-
sensitive
areas
n/a
1
n/a
1
n/a
# of tradeoff curves for sensitive areas
1
All sensitive areas could be represented
by the same population of critical loads
The U.S. is categorized into 5 clusters of
acid sensitivity based on ANC. Each
ecoregions is assigned a cluster id, and the
CLs from sites within each cluster are
pooled to create the population for that
cluster.
43 ecoregions (level 3) are considered
sensitive, there is CL data available for 40
ecoregions at ANC= 50, the 3 ecoregions
which do not have CL data could be
represented by values calculated from the
entire sensitive population population.
85 ecoregions (level 3) occur in the US, of
these there is CL data for 76 therefore
they could be represented by CLs only
from their ecoregions. The 9 ecoregions
without CLs could be represented by the
entire population.
TotalN + S
depositions
metrics -
8
1 9
10
2 11
12
13
14
15
16
17
5 18
19
20
21
22
73
44
24
25
26
7.7
77
o o
28
29
30
31
A
com
pari
son
of
the
valu
es
for
the
90
%-
tile
dep
ositi
on
met
ric
at
AN
C=5
0
L is
give
n
for
September, 2010
5-43
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
option 1 (no-subdivision/all data), 2a (sensitive vs. less-sensitive classification) and 2b (5 clusters
fy
based on log ANC) in Fig 5.12. The deposition metric for option 1 is 27 meq/m /yr. When
compared to the deposition metrics from option 2a, it would under protect the sensitive category
0 9
by 1 meq/m /yr and over-protect the less-sensitive category by 27 meq/m /yr. When option 1 is
compared to option 2b, clusters 1 and 2 would be slightly under protected by 3 and 1 meq/m2/yr,
ry
respectively. Clusters 3, 4 and 5 would be over protected by between 3 to 35 meq/m /yr.
300
250
t200
VI
z150
100
50 •
Option 1
D DL %90
• DL %75 269
D DL %50
less-sensilive cluster 1 (68) clusler 2 (80) ckisler 3(109) cluster4 (142) cluster 5 (167)
Figure 5-12 Deposition metrics for 90%,75% and 50% of the population for options, 1
(all data), 2a (binary classification), and 2b (clusters based on log ANC)
A second method to compare the options 1 and 2a is presented in Fig 5.13. Sites that
would not be protected at a 90%-tile deposition metric at ANC=50 /ieq/L are shown. In both
options 10% of the total population is not going to be protected from exceeding the CL, by
comparing these maps we see that the spatial distribution of sites that are not protected between
the two options is similar. In option 2a, there are 10% of total sites that will not be protected in
both the sensitive and the less-sensitive ecoregions. This means that eventhough a region is
considered less sensitive, 10% of the water bodies would not meet the target ANC because more
deposition would be allowed. Considering the analyses represented by Fig 5.11 and 5.12, how
much benefit is gained by subdividing the landscape into acid sensitivity classes?
September, 2010
5-44
Draft - Do Not Quote or Cite
-------
"
, '-. ,4f
. • "..< t ••:'/
••>- ] #\
. ..••'::•/•• •
y
One untiou
Sites for which N+S criticnl loads are
not protected when one deposition
metric to protect 90% of the
population is selected. Red indicates
sites not protected
1
2
3
4
5
fcSS i '•-.. ,/
A^.;J- i*-: <-tf
\ UUi-^
^ "':- iJ ;
;'- i •'-'•• i y-
" j • ; ' '
•'
!'• -
•
Biliary categorization: sensitive aud
non-sensitive
Sites for which 1V+S critical loads are
not protected when the tiro
deposition metrics are selected to
protect 90% of each population.
Red indicates sites not protected in
the sensitive areas and green
indicates sites that are not protected
in the less-sensitive areas
Figure 5-13 Comparison of sites that are protected if the nation is not subdivided and if
the nation is subdivided into a sensitive and non-sensitive category
Option 2c and 2d provide more spatial detail than the other options. Option 2c uses one
6 deposition metric for all ecoregions categorized as less-sensitive areas according to option 2a,
7 while each ecoregion considered sensitive (using log ANC) is represented only by CLs located
8 within the boundaries of the ecoregion. Option 2d uses only CLs from within an ecoregions to
9 represent that ecoregions. A strength of option 2d is that it provides the finest level of spatial
10 detail consistent with available data. A weakness is that some ecoregions are represented by a
11 small sample size of critical loads, which may not accurately convey the population of critical
12 loads for waterbodies across the entire ecoregion. The cumulative distribution of the 90 %-tile
13 deposition metric of the CLs at ANC=50 /zeq/L of all ecoregions (level 3) is shown in Fig 5.14.
14 This figure helps to visualize how option 2d compares to the other options. The deposition to
15 protect 90% of the population in the most sensitive ecoregions is 13 meq/m2/yr. In comparison
16 to option 1, the deposition to protect 90% of the population is 29 meq/m2/yr. If the national
17 value were chosen, then although 90% of the total population would be protected, only ~76% of
September, 2010
5-45
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
the ecoregions would have 90% of their populations protected. This illustrates that the sample
size and distribution of critical loads varies among ecoregions.
Cumulative Distributuion 90th percentile deposition
metric for ecoregions (level 3)
(ANC =50)
CO
o
'o>
£
o
o
£ t>
*• 0)
08
(U Q.
O)
S
0)
100 f 13,100
\ 17,90
80
en
60
40
20
\ 26, 75
58, 50
\
t16.25
269,10
494,2.5
642,0.0
100
200 300 400 500
Deposition N + S (meq/m2/yr)
600
Figure 5-14. Cumulative distribution of the 90% deposition metric for ecoregions is
shown. The x and y values are given next to each data point.
Neco
Neco is added to the deposition metric to develop the N and S tradeoff curves (See
Section 5.3.2.8). Staff suggests that the spatial boundaries used to determine the acid-sensitivity
categories for calculating deposition metrics are also used to calculate Neco values.
5.3.2.8 Deposition metric: Developing N and S tradeoff curves from the deposition metrics for
the acid sensitive categories
Acid sensitivity categories define the spatial area of a population. Therefore the membership of
an individual catchment to a population that represents and acid sensitivity category will vary
depending on which approach for categorizing acid-sensitivity across the landscape is selected
for the NAAQS. For each category that is established, there will be a single, specified value for
a deposition metric. This deposition metric will be based on the population of catchments in that
September, 2010
5-46
Draft - Do Not Quote or Cite
-------
1 specific category. Therefore both the number and the values of the deposition metrics can
2 change based on how finely the U.S. is subdivided into acid sensitivity categories.
3 The deposition metric for a category will be a single, specified value for the deposition of
4 N and S. That deposition load can be achieved by various combinations of N and S loading, and
5 these combinations can be expressed as an N and S deposition tradeoff curve. Thus for each
6 category a specific deposition load and a resulting N and S tradeoff curve can be developed.
7 Regardless of which option is chosen for spatially dividing the U.S. into acid-sensitivity
8 categories, the number of N and S deposition tradeoff curves is the same as the number of
9 deposition metrics (Table 5.7). For each category, the assigned value for Neco is added to the
10 deposition metric to develop the N and S tradeoff curve. Fig. 5.15 provides an example for two
11 of the categorization options - Option 1, with a tradeoff curve for the single national category,
12 reflecting the entire population of critical loads, and Option 2a, with a tradeoff curve for the
13 sensitive and not sensitive binary landscape categorization.
14
September, 2010 5-47 Draft - Do Not Quote or Cite
-------
.>> 100
CM
<3
1 r
0
A. entire U.S. at ANC=50 (n=5280)
= 52
50
N (rneq/m2/yr)
100
100
B. sensitive areas of the U.S. at ANC=50 (n=4553)
50
N (meq/m2/yr)
100
C. less-sensitive areas of the U.S. at ANO50 (n=727)
1
2
3
4
N (meq/m lyr)
Figure 5-15 Examples of N + (SOa +SO4"2) deposition tradeoff curves for option 1 (A) and
option 2a(B and C).
September, 2010
5-48
Draft - Do Not Quote or Cite
-------
1
2 5.3.9 Deposition metric: How is reduced nitrogen appropriately considered in the deposition
3 metric?
4
5 N deposition is composed of NHx deposition and NOy deposition. Deposition from NHx
6 can contribute to acidification. However, the criteria pollutant listed by EPA pursuant to section
7 108 (a) of the Act is oxides of nitrogen, which does not include NHx. The NAAQS is for oxides
8 of nitrogen and sulfur, hence the form of the NAAQS needs to account for the deposition effects
9 of these oxides. To accomplish this, the loadings of reduced forms of nitrogen derived for a
10 given spatial area would be subtracted from the N +S deposition metric after NECO is added, such
11 that the resultant deposition metric is for sulfur and oxidized nitrogen only. Subtraction of
12 reduced nitrogen from the deposition metric based on nitrogen and sulfur is expressed by
13 equation
14 DL%ECO(NOY +S ) = DL%ECO(N + S) + NEco -Dep™ (6)
15 Staff propose the value of reduced nitrogen is initially set using deposition of NHx
16 modeled using the CMAQ, evaluated for the period 2002-2005. The average values of NHx
17 deposition for the 83 ecoregions (level 3) in the U.S. ranges from 4 -54 meq/m2/yr. This would
18 mean that each acid-sensitivity category would have one N+S deposition metric, one NECO value
19 and one NHx deposition value. Staff is considering the most appropriate spatial averaging
20 extent for NHx. Figure 5.16 shows spatially interpolated values of reduced nitrogen
21 deposition based on 12km CMAQ modeling in the Eastern U.S. It is clear that in some
22 locations, there is significant heterogeneity in NHx deposition within ecoregion 3
23 boundaries. Given this information, two possible approaches to estimating reduced
24 nitrogen values are possible: 1) average reduced nitrogen deposition within an ecoregion,
25 acknowledging that this will lead to uncertainties in the level of protection associated with
26 levels of ambient NOy and SOx, or 2) allow for additional spatial refinement of sensitive
27 areas to reflect the heterogeneity of NHx deposition.
September, 2010 5-49 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 5-16. A U.S. map of NH3 deposition (meq/m2/yr) overlaying ecoregions (level 3)
boundaries
The deposition fromNOy and SOx is converted to atmospheric concentrations of NOy
and SOx by the methods described in section 5.2.3 and an example is given in fig 5.19. This
figure also shows air quality in the Adirondack region relative to the NOy and SOx tradeoff
curves for option 1 and option 2a. In option 1 the Adirondacks air quality is slightly out of
attainment for a 75%-tile deposition metric based on CL at ANC=50. In option 2a the
Adirondack air quality is out of attainment for the curve for the sensitive areas, but in attainment
for the less-sensitive areas.
September, 2010
5-50
Draft - Do Not Quote or Cite
-------
•50
CO
i
o
CO
0 +
A. entire U.S. at ANC=50 (n=5280)
Max(S)=55
N (meq/m /yr)
—•— 90%
—»—75%
- - - Neco
NHxdep = 4
NHx dep = 54
100
•f
O 50
CO
to
04-
B. sensitive areas of the U.S. at ANC=50 (n=4553)
50
N (meq/m2/yr)
100
1
2
3
4
E 100
I
50
0
C. less-sensitive areas Of the U.S. at ANO50 (n=727)
Max(S)=117
50
100
N (meq/m2/yr)
Figure 5-18. Examples of N + (SO2 +SO4"2) deposition tradeoff curves for option 1
(A) and option 2a(B and C). Deposition metrics that correspond to protection 75% or 90%
of ecosystems are given with the high and low value for average NHx deposition in
ecoregions indicated by vertical grey lines.
September, 2010
5-51
Draft - Do Not Quote or Cite
-------
2
3
A. entire U.S.
ANC=50
90% and NHx=5
•^90% and NHx=54
75%&NHx=5
75%&NHx=54
A Adirondack air quality
2 3
NOy (ug/m3)
B. sensitive areas of the U.S.
ANC=50
90% and NHx=5
^—90% and NHx=54
75%&NHx=5
75%&NHx=54
A Adirondack air quality
2 3
NOy (ug/m3)
C. non-sensitive areas of the U.S.
ANC=50
90% and NHx=5
90% and NHx=54
75% & NHx =5
75% & NHx =54
Adirondack air quality
2 3
NOy (ug/m3)
Figure 5-19. Examples of NOy + (SO2 +SO4"2) deposition tradeoff curves for option 1
(A) and option 2a(B and C). Deposition metrics that correspond to protection 75% or
90% of the ecosystems considering both the high and low value for NHx deposition
September, 2010
5-52
Draft - Do Not Quote or Cite
-------
1 5.3.3 Conceptual Design: Linking Deposition to Atmospheric Concentration
2
3 5.3.3.1 Background
4 Atmospheric pollutants deposit onto land and water surfaces through at least two
5 major mechanisms: direct contact with the surface (dry deposition), and transfer into
6 liquid precipitation (wet deposition). A third mechanism involving impaction of fog
7 droplets onto the surface (occult deposition) also exists, but has been shown to be very
8 small relative to wet deposition (Fuhrer 1986; Thalmann 2001) and so is not considered
9 in calculations below. The magnitude of each deposition process is related to the ambient
10 concentration through the time-, location-, process- and chemical species-specific
1 1 deposition velocity (Seinfeld and Pandis, 1 998):
1 ~ 7^ Diy Diy .~, Amb fn\
12 Dep, =v, •Ci (7)
•\ ^ r^. Wet Wet s-i Amb /o\
13 Dep, =v, -C, (8)
1 . , Dry .Wet , , , , , . . , _ Dry , _ Wet , ,
14 where v, and v, are the dry and wet deposition velocities, Dep, and Dep , are the
15 dry and wet deposition fluxes, Cjm is the ambient concentration, and the / subscript
16 indicates the pollutant species under study. The total deposition of each pollutant is
17 DePiT°' =DePiDiy+Dep'Vel (9)
1 8 Substituting Equations 7 and 8 into Equation 9 yields
! r> r> Tot Diy ,-, Amb , Wet ,~, Amb fi r>\
19 Dep, =v, •Ci +v, -C. (10)
20 The total deposition of sulfur or nitrogen would therefore be:
_1 ,-^ Tot
21 Depa =
22 where m. is the molar ratio of the atom of interest (a, which refers to either sulfur or
23 nitrogen atoms in SOX or NOy) to the fth pollutant. Note that only species in SOX and
24 NOy (see Table 1) will be input to these equations; NH3 and NH4 are not currently
25 included as listed pollutants (see Chapter 8 for an expanded discussion of the role of
26 NHx).
27
28 5.3.3.2 Aggregation Issues
29 A relationship for converting sulfur or nitrogen deposition to "equivalent"
30 ambient concentrations, which is part of specifying ambient air quality standards, is
September, 2010 5-53 Draft - Do Not Quote or Cite
-------
1 provided in equation 11. A major issue to consider during such conversion is the
2 treatment of spatial, temporal and chemical resolutions of the deposition data and the
3 resulting standards. Since the objective is to specify allowable levels of total ambient
4 SOX and NOy (as S and N),, and this is also the chemical resolution provided by the
5 ecosystem models, it is convenient to use a relationship with the following form:
*r T-\ Tot rri s~i An\b f-\ r\\
6 Depi =T,-Ct (12)
7 where Tt is what we term the "transference ratio", which can be considered an
8 aggregated, "effective" deposition velocity that relates total deposition of sulfur or
9 nitrogen to the total ambient concentration, and represents an average of the chemical
10 species specific v?°l (= viDry + v^el) values in Equation 11. In other words, the
11 transference ratio is effectively the aggregate deposition velocity that would have been
12 required to obtain a level of deposition given a level of atmospheric concentrations. The
13 SOX and NOy concentrations are the result of applying the m\ values to the Cfmb values
14 in Equation 11. Although T, values are representative of pure nitrogen and sulfur, they
15 are derived from concentrations and depositions of the individual pollutants, so that the
16 resulting value is appropriately weighted by the relative levels of each pollutant actually
17 present in the environment.
18 The deposition velocities (and, by extension, the T, values) provide relationships
19 between depositions and concentrations that are concurrent. Since the deposition critical
20 loads derived from ecosystem models that are in need of conversion to atmospheric
21 concentrations are in terms of annual deposition, the ambient concentrations will also be
22 aggregated to the annual level through averaging. Data used to derive annual 71, values
23 will also need to have the same spatial representativeness as the depositional loads
24 derived from ecosystem models. However, many of the calculations involving
25 relationships between concentrations and depositions will be in terms of the masses of the
26 N and S atoms within NOy and SOx removed from the other atoms in the pollutant
27 species, which we will refer to as "oxidized S" and "oxidized N". This is done both
28 because these quantities are the outputs of the ecosystem models discussed in Section
29 XXX, and it removes potential ambiguities that might be introduced by different species
30 mixtures.
31
September, 2010 5-54 Draft - Do Not Quote or Cite
-------
1 5.3.3,3 Air Quality Simulation Models
2 Ideally, T, values would be derived for each area of interest from concurrently
3 collected sulfur and nitrogen deposition and concentration measurements. However, no
4 monitoring network currently exists that can provide such information. We therefore
5 suggest using output of the Community Multi-scale Air Quality (CMAQ) model (EPA,
6 1999) for initial calculation of I7, values.
1 CMAQ provides both concentrations and depositions for a large suite of pollutant
8 species on an hourly basis for 12 km grids across the continental U.S. Its comprehensive
9 structure is ideal for providing T, values that appropriately address the chemical and
10 temporal aggregation issues discussed above, and weighted spatial averages of the
11 gridded data can be used for areas that span multiple grid cells. The major potential
12 drawback to using CMAQ output is that the data is simulated rather than measured.
13 CMAQ does not directly calculate or use T, values; instead the following
14 procedures are used in the code to model deposition:
15 1) vdry values of gaseous pollutants are calculated in the CMAQ weather module
16 called the Meteorology-Chemistry Interface Processor (MCIP) through a complex
11 function of meteorological parameters (e.g. temperature, relative humidity) and properties
18 of the geographic surface (e.g. leaf area index, surface wetness)
19 2) v^ values for particulate pollutants are calculated in the aerosol module of
20 CMAQ, which, in addition to the parameters needed for the gaseous calculations, also
21 accounts for properties of the aerosol size distribution
22 3) vwet values are not explicitly calculated. Wet deposition is derived from the
23 cloud processing module of CMAQ, which performs simulations of mass transfer into
24 cloud droplets and aqueous chemistry to incorporate pollutants into rainwater, all of
25 which is conceptually contained in the vwet parameter in Equation 8.
26 Due to lack of direct measurements, no performance evaluations of CMAQ's dry
27 deposition calculations can be found; however, the current state of MCIP is the product
28 of research that has been based on peer-reviewed literature from the past two decades
29 (EPA, 1999) and is considered to be EPA's best estimate of dry deposition velocities.
30 Some bias has been found between CMAQ's wet deposition predictions and measured
31 values (Morris et al., 2005); recent analyses suggest that poor simulation of precipitation
September, 2010 5-55 Draft - Do Not Quote or Cite
-------
1 could be responsible for this (Davis and Swall, 2006), which can potentially be dealt with
2 by recalculating wet deposition using precipitation measurements. Although the model is
3 continually undergoing improvement, CMAQ is EPA's state-of-the-science
4 computational framework for calculating deposition velocities, and was therefore the
5 logical first choice as a source for Tt values.
6
7 5.3.3.4 Oxidized Sulfur and Nitrogen Pollutant Species
8 Ideally, all possible NOy and SOx air pollutant species that contribute to
9 ecological adversity would be considered for Tt values. The pollutant list is constrained
10 by the source of TL values, which is currently CMAQ output. The oxidized sulfur and
11 nitrogen species currently available in CMAQ whose data will be used for T, values, are
12 listed in Table 1.
13 One issue that needs explicit consideration is the contributions of particles larger
14 than PM2.5 to sulfur and nitrogen deposition. A recent review of particle deposition
15 measurements (Grantz, Garner, and Johnson, 2003) showed that coarse particles
16 generally deposit far more sulfate and nitrate in forest ecosystems than fine particles.
17 However, CMAQ does not currently provide simulations of coarse particulate sulfate and
18 nitrate. This is an issue that needs to be addressed by developers of either the model or
19 the future SOx/NOy measurement network to set scientifically sound standards. This
20 issue is explored further in the discussion of monitoring in Section XXX.
21 5.4.5 Example Calculations
22 Figure 5.20 shows annual inverse Tt values1 calculated for each 12 km grid in the
23 eastern and western domains for a 2002 CMAQ v4.6 simulation, which is the quantity
24 that would be used for conversion of deposition load tradeoff curves. Figure 5.21 shows
25 an example application of these ratios for a lake in the Adirondacks. Deposition load
26 tradeoff curves for this lake (see Section 5.5 for their calculation) are multiplied by the
27 inverse Tt value from the appropriate gridcell in Fig. 5.22 to convert those depositions to
28 ambient concentrations of sulfur and nitrogen.
1 Inverse T, values represent the multiplier needed to convert deposition levels into atmospheric
concentrations of NOy and SOx.
September, 2010 5-56 Draft - Do Not Quote or Cite
-------
1 A CMAQ v4.7 simulation for multiple years (2002-2005) recently became
2 available, which was used to examine the inter-annual variability of inverse Tt values.
3 The grid-specific coefficients of variation (CV) are shown in Fig. 5.22. Fig. 5.22 shows
4 that CV values are relatively small (< 25%) in the Adirondacks and Shenandoah case
5 study areas. This suggests that a 3-year average of the ratios may be a sufficiently stable
6 representation of deposition velocities for converting the deposition load curves to
7 ambient concentrations.
8
9
10
11
12
September, 2010 5-57 Draft - Do Not Quote or Cite
-------
S: Dep/Conc
(kg/ha/ug/m3)
I12
110
8
N: Dep/Conc
(kg/ha/ug/m3)
f12
10
6
4
2
0
2
3
4
Figure 5-20 2005 Tt values for each grid cell in the eastern U.S. domain. The top
map shows values for sulfur and the bottom is for nitrogen.
September, 2010
5-58
Draft - Do Not Quote or Cite
-------
Deposition
Concentration
1
2
3
4
5
6
7
8
9
' ta so
N deposition (kg/ha-Xi
- ANC10Q
ANCSO
ANC20
Current
• Cgodrfons
(CMAO)
NO-dig/m1)
Figure 5-21. Schematic Diagram illustrating the procedure for converting deposition
tradeoff curves of sulfur and nitrogen to atmospheric concentrations of SOx
and NOy.
September, 2010
5-59
Draft - Do Not Quote or Cite
-------
Coeflicwil ol Varimton ot N Conc/Dep rallo, 2002-2005
CV (%)
rSO.O
37.5
25.0
12.5
0.0
1 a)
Cmlllcttil nl V»iUlloti ol S Cora: D»ri nlk>. W02-30IU
CV (%)
rSO.O
37.5
25.0
12.5
0.0
y
2 b)
3
4 Figure 5-22 Inter-annual coefficients of variation (CV) of a) NOy and b) SOx T, values,
5 based on a series of 2002-2005 CMAQ v4.7 simulation.
6
September, 2010
5-60
Draft - Do Not Quote or Cite
-------
1 5.3.3 AAPI
2
3 In sections 5.3.2 and 5.3.3, the form of the standard is expressed as an approach for
4 linking ecosystems, N and S deposition and ambient air concentrations of NOx and SOx to
5 conceptually identify levels of NOx and SOx that are protective of ecosystems via NOx and SOx
6 tradeoff curves. The tradeoff curves will differ among acid-sensitivity categories because they
7 are derived from a specific percentage of ecosystems protected within an overall sensitive area.
8 In this section, the form of the standard is expressed as the AAPI equation, which is an index, set
9 to one level, which can be applied across the nation to convey the potential of an ecosystem to
10 become acidified from atmospheric deposition. Moreover, the AAPI calculates the target ANC
11 for a percentage of aquatic ecosystems within a particular acid sensitive area when atmospheric
12 concentrations of NOx and SOx are input
13
14 5.3.4.2 Definition and Derivation of the AAPI
15 The definition of the AAPI considered here is:
16
17 Annual Average AAPI: Natural background ANC minus the contribution to acidifying
18 deposition from NHx, minus the acidifying contribution of deposition from NOy and
19 SOx.
20
21 In order to derive the AAPI, we start with the basic framework of critical loads discussed in
22 section 5.3. The approach used to calculate N and S deposition values for a specified ANC at a
23 catchment-scale is expressed in Equation 2 from section 5.3.
24
25 CLANC,m (N + S) = ([BC]0 - [ANC^ })Q + Neco (2)
26
27 To develop an equation that allows us to set a single value for the standard across the U.S.,
28 rearrange equation (2) to solve for ANC (place ANC on the left hand side of the equation):
29
30 Q • ANC,tm = NECO + [BC]0 • Q - \DepTN°tal + DepTsolal] (13)
31
September, 2010 5-61 Draft - Do Not Quote or Cite
-------
(14)
2
3
4
5
6
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Building from equation 14, total nitrogen deposition is split into oxidized and reduced
nitrogen because we need to be able to specify the standards in terms of oxides of nitrogen, and
so the contribution of reduced nitrogen has to be separated.
(15)
>-v L J 'Vi/y -* J j ,-*
Where,
j^'= the depositional load of reduced nitrogen, NHx.
In order to judge whether a catchment meets the ANCnmjt given observed NOy and SOx
levels, the associated depositional loadings of NOy and S can be compared directly against
calculated deposition tradeoff curves, atmospheric concentrations of NOy and SOx can be
compared against the atmospheric concentration tradeoff curves or, loadings of NOy and SOx
can be input into the following equations to obtain the calculated value of ANC, equal to ANC*:
[L(NOy] + L(SOx}] - L(NHx)
(16)
Where,
ANC*= the calculated value of ANC given loadings of N and S for comparison against an
L(NOy+S)= the load of NOy+S anions based on observed atmospheric concentrations of NOy
and SOx
L(NHx) = the load of reduced nitrogen deposition
[Note that L(N) = L(NOy+NHx)]
September, 2010
5-62
Draft - Do Not Quote or Cite
-------
1 In equation 16, the ANC* will vary based on the deposition load inputs of NOx, NHx and
2 S at the site of interest. The deposition loads caused by NOx and S and NHx are inputs, leading
3 to
4
ANC* =
fi
fi
(17)
6 If ANC* < ANQim, then the deposition of N and S exceeds the deposition load to maintain
7 ANCiimit. ANC* is still representative of the calculated ANC based on specific catchment level
8 estimates of [BC]0*, NEco, Q and NHX.
9 The AAPI is equivalent to the equation for calculating ANC* when the catchment
10 specific values are replaced by spatially aggregated values that represent an acid sensitive areas
11 (based on a percentile of water bodies targeted for an ANC level selected by the Administrator),
12 NHX is replaced by average values for aggregate ecosystem areas, and aggregated deposition
13 velocities translate NOy and SOx into deposition:
14
15 AAPI =
(18)
eeco L jv ,—v J. uivnx s-\ i jviyv /VL/V JL/J: ou/A" j ^ '
- %eco *-* *•"
16
17 Where C^and C^ are concentrations of NOy and SOx, respectively, TNOy and T50v are the
18 transfer ratios to convert ambient concentrations to deposition for NOy and SOx, respectively
19 (See section 5.3.3 for further description of calculation of these ratios)
20 Note that while equation (18) is used to calculate the value of AAPI for any observed
21 values of NOx and SOx, the level of the standard for AAPI selected by the administrator should
22 reflect a wide number of factors, including desired level of protection indicated by a target
23 ANCiimit, the specified percentile of waterbodies projected to achieve the target ANC, and the
24 various factors and uncertainties involved in specifying all of the other aspects of the standard,
25 such as the classification of landscape areas, the specification of reduced nitrogen deposition, the
26 methodology to determine deposition of NOx and SOx, and the averaging time. As such the
27 administrator may choose an AAPI level higher or lower than the target ANQimJt to reflect the
28 combined effect of the all of the components of the standard and their related uncertainty, such
September, 2010 5-63 Draft - Do Not Quote or Cite
-------
1 that the chosen AAPI, in the context of the overall standard, reflects her informed judgment as to
2 a standard that is sufficient but not more than necessary to protect against adverse public welfare
3 effects.
4
5 5.3.4.2 How are AAPI parameters determined?
6 Other than ambient levels of NOx and SOx, which would be measured values, EPA
7 would determine and specify all of the values for the AAPI parameters, as discussed below.
8 TaNOy and TaSOx are calculated from CMAQ by dividing the annual average NOy, SOx
9 concentration by the total NOy or SOx deposition, respectively, for each grid cell and then
10 aggregating all grid cells in a level three ecoregion. The TaNOy and TaSOx are spatially variable,
11 and for the purposes of setting the standard, are determined based on the ratios of total sulfur and
12 nitrogen depositions to concentrations from CMAQ model outputs (as described in section 5.3).
13 NHx is spatially variable and determined based on monitored and/or CMAQ modeled
14 outputs. The average NHx deposition across grid cells within an acid sensitive region will be
15 used to represent the depositional load of NHx. (See section 5.3)
16 There will be multiple combinations of concentrations of NOy and SOx that result in a
17 specific value of the AAPI. There will be no single combination of NOy and SOx that solves for
18 a particular value of AAPI in all locations. Measured concentrations of annual average NOy and
19 SOx necessary to meet the standards are thus expressed conditionally by the equality in the
20 AAPI equation (18), and not by fixed quantities.
21 In order to provide a set of values for elements of the form, the TNOy, TSOx and NHx,
22 values for specific areas would be estimated based on the best available monitoring and/or
23 modeling data. Given the limited availability of measured deposition velocities, staff concludes
24 that the calculated deposition ratios based on the CMAQ modeling from 2005 provides the best
25 available source of estimates of TNOy and TSOx. Evaluation of the stability of these estimates of
26 deposition ratios over time suggests that in most acid sensitive areas, deposition ratios are quite
27 stable, with a coefficient of variation less than 25 percent across a four year period. While there
28 are a limited number of sites that directly measure deposition of reduced nitrogen, staff
September, 2010 5-64 Draft - Do Not Quote or Cite
-------
1 concludes that the most widely available and defensible estimates of reduced nitrogen deposition
2 (NHx) are the estimates obtained from the CMAQ modeling from 2005.2
3 For each acid sensitive area the natural background ANC is a calculated value and is
4 determined by the expression
12
13
14
18
19
20
21
22
23
24
25
26
Q
from equation 18. The three components
%eco
5 are: [BC]o , the pre-industrial weathering rate; Neco , which represents the amount of deposited
6 nitrogen that is removed via ecosystem uptake, denirification and immobilization; and Q, the
7 runoff parameter. These components act in combination, so that the 95th percentile of the
8 distribution of values for each variable may not equal the 95th percentile value for the
9 combination of the variables. Therefore the value for the combination of these variables will be
10 determine from the critical load which represents the selected % of protection given the
11 distribution of critical loads in the population.
It is important to note for this form of the standard that the same AAPI can be obtained
with different combinations of ambient NOy and SOX concentrations. The implication of the
15 form of the standard expressed in equation (18) is that there will be a tradeoff curve that reflects
16 the combinations of NOy and SOx that satisfy equation (18) for any specific value of the
17 standard. The shape of the tradeoff curve will depend on the specific values of
Q
, Tso and NHx for a limited number of specific areas classified
%eco
based on acid-sensitivity. As discussed in Section 5.3, all parts of the U.S. would be classified
into areas based on acid-sensitivity. Within each such area, EPA would specify the parameter
values of AAPI, leading to a specific tradeoff curve for each area.
The levels of NOy and SOx that meet an AAPI standard expressed for a given
,TNOy,TSOx and NHx:
Q
%eco
Q
-Dep'°£-Q-AAPI (19)
%eco
2 Note to readers: Maps of CMAQ 2005 estimates of NHx deposition will be included in the second draft policy
assessment, along with an evaluation of the representativeness of the 2005 NHx deposition for characterizing
conditions over a multiyear period.
September, 2010
5-65
Draft - Do Not Quote or Cite
-------
1 Note that \TNOy • CaN"£ + TSOx • C™* ] is essentially the critical load of NOy and S expressed in
2 terms of atmospheric concentrations. The pairs of NOy and SOx that will meet a given AAPI
3 limit are related through the following equations
4
5 tf0;=Cm>0,) (20)
6 SOI = C1MX (SOx)VNOy < Cmm(NOy) (21)
Cm (SOx)
7 SO, = C,JKK (SOx) +
(22)
8 Where,
9 NOy is the coordinate point for NOy
10 SO*X is the coordinate point for SOx
1 1 Cmax (SOx) is the concentration of SOx in the atmosphere consistent with DLmax (S)
1 2 Cmax (NOy) is the concentration of NOy in the atmosphere consistent with DL max (N)
1 3 Cmin (NOy ) is the concentration of NOy in the atmosphere consistent with DL min (N)
14
15 CSOx = -DLm(S) (23)
T
laS
16 Cmin (NOY) = —— DLm.m (N - NHx)MNHx < DLm-n (N) (24)
aNOy
17 =WNHx>DL^(N)
18 C^(NOr) = -±-DL(N) (25)
19
20 Where DLmax(S), DLmax (N), and DLmin(N).are based on the critical load within a sensitive
21 areas that protects a specified percentile (e.g. 95%) of water bodies in the area.
22 Note that Cmin(NOy) is a conditional function determined by the relationship between
23 total nitrogen buffering capacity in an ecosystem and the amount of reduced nitrogen deposition.
24 When reduced nitrogen deposition exceeds the buffering capacity of an ecosystem, then all
25 atmospheric oxidized nitrogen contributes to acidification. When reduced nitrogen deposition is
26 less than the buffering capacity of an ecosystem, then some amount of NOy is buffered (i.e. is
27 reflected in Crain(NOy) but that amount reflects the contribution of NHx to total nitrogen (the
28 amount of buffering capacity used up by reduced nitrogen). In this case, some fraction of the
29 atmospheric oxidized nitrogen may not contribute to acidification.
September, 2010 5-66 Draft - Do Not Quote or Cite
-------
1 Recall that these three variables are conditional on the chosen level of AAPI, and reflect
2 the depositional loadings that are associated with an equivalent level of ANC, e.g. for an AAPI
3 of 50, the DLmaX(S), DLmax(N), and DLmin(N) are associated with an ANC of 50. Also recall than
4 DLmax(S) for a given ANC is a function of the "natural" flux of base cations to a watershed,
5 runoff, and the amount of sulfur retention within a waterbody; DLmin(N) is the minimum amount
6 of deposition of total nitrogen (NHx + NOx) that catchment processes can effectively remove
7 without contributing to the acidic balance; and DLmax(N) for a given ANC is a function of
8 DLmin(N) and the "natural" flux of base cations to a watershed, runoff, and the amount of
9 nitrogen retention within a waterbody, assuming S is zero. In our framework, DL^N) is
10 calculated from the FAB critical load modeling (equation 5 from Attachment A of the REA) or
11 estimated through measured or modeled values of total nitrogen deposition and nitrate leaching.
12 As discussed in sections 5.3, the specific estimation of of
23
Q
T
> Ai
NOy 5
- %eco
13 TSOx and NHx in a specific sensitive area will depend on the spatial scale of the sensitive area.
14 Sensitivity can be assessed at the level of individual catchments, however, this presents practical
15 limitations for establishing meaningful standards, as there are thousands of catchments within the
16 U.S. Binning classes of sensitivity within larger spatial areas can provide a more manageable set
17 of values of
Q
, TNOy , TSOx and NHx. These parameters can be estimated in
- %eco
18 several ways for the larger spatial areas. Mean or median values can be generated across
19 catchments, however, this would lead to parameter estimates that do not reflect conditions in the
20 more sensitive lakes in the region. Alternatively, in order to provide a desired level of protection
21 in these larger defined spatial areas, estimates based on higher percentiles of the distributions of
22 parameters across catchments can be generated, e.g. the 75th or 95th percentile values of of
, TSOx and NHx could be used to provide protection for the more
24 vulnerable aquatic ecosystems, however this would potentially lead to over-protection for less
25 vulnerable ecosystems in the area. The Administrator may consider the balance between
26 protection of particularly sensitive ecosystems and the overall protection for ecosystems in an
27 area as an important element to consider in making decisions about the target level of ANC and
28 the percent of aquatic ecosystems within an area targeted to achieve the specified ANC level.
September, 2010 5-67 Draft - Do Not Quote or Cite
-------
1
2 5.4 Options for specifying the targets for the ecological indicator for aquatic
3 acidification (ANC)
4
5 As discussed earlier in this chapter, ANC is the ecological indicator best suited to reflect
6 the sensitivity of aquatic ecosystems to acidification (5.3). ANC is an indicator of the effects
7 expected to occur given the natural buffering capacity of an ecosystem and the loadings of
8 nitrogen and sulfur resulting from atmospheric deposition. As noted by CASAC in their
9 review of the first draft of this Policy Assessment, "information on levels of ANC
10 protective to fish and other aquatic biota has been well developed and presents probably
11 the lowest level of uncertainty in the entire methodology."
12 A target ANC limit is an important starting point in deciding the appropriate level of
13 AAPI, and is a necessary component in determining the critical loads for N and S from which the
14 tradeoff curves for ambient NOy and SOx are derived. In fact, the target ANC together with the
15 selected percent of waterbodies to be protected to that target ANC (see section 5.4) uniquely
16 determines the set of critical loads of N and S across the set of acid sensitivity categories.
17 In reaching staff conclusions regarding the range of target ANC levels that is
18 appropriate to evaluate further in the context of setting a NOx and SOx standard with the AAPI
19 form, considering effects on public welfare due to aquatic acidification, we use three types of
20 information: 1) direct information on levels of biological impairment related to
21 alternative levels of ANC (and related levels of pH), 2) target ANC values and their
22 rationales as identified by states or regions in setting critical loads to protect regional water
23 bodies, and 3) information on ecosystem service losses associated with alternative ANC
24 values.
25 In addition, we evaluate risk-based information derived from the Risk Assessment,
26 focused on case studies of acid sensitive lakes and streams in the Adirondacks and
27 Shenendoahs. As part of our risk-based considerations, we have considered estimates of
28 risks under current conditions in which current NOx and SOx air quality standards are met
29 in the case study areas.
September, 2010 5-68 Draft - Do Not Quote or Cite
-------
1 It is important to note that the choice of a target ANC level starts with an
2 evaluation of the conditions that exist if a certain ANC occurs in fact. There is a
3 significant body of science to inform that evaluation, as discussed below. However the
4 development of a NAAQS based on the AAPI form does not assume or guarantee that
5 any specific ANC level will in fact occur for any specific body of water. A critical issue for
6 the Administrator to address in setting a NAAQS based on the AAPI is to consider and
7 weigh the varying degrees of uncertainty in establishing the elements of the AAPI. These
8 uncertainties impact the likelihood that a specific AAPI standard would in fact achieve a
9 target ANC level for a specified percentage of a population of water bodies. Thus it is
10 appropriate to first look at target ANC from the perspective of what we understand the
11 impacts would be if a certain ANC level were realized. Hence, the discussion below
12 focuses on evaluation of the ecological impacts at various target ANC levels, assuming the
13 ANC level occurs in fact. Judgments made based on that evaluation then become the
14 foundation for further judgments that would be made in the process of establishing the
15 specific level of an AAPI, taking into account the degree of uncertainty in projecting the
16 actual ANC levels that would result from any specific APPI standard.
17
18 5.4.1 What levels of impairment are related to alternative target levels of ANC?
19 As discussed in Chapters 2, 3, and 4, specific levels of ANC are associated with differing
20 levels of risk of ecosystem impairment, with higher levels of ANC resulting in lower risk of
21 ecosystem impacts, and lower levels resulting in risk of both higher intensity of impacts and a
22 broader set of impacts. While ANC is not the causal agent determining biological effects in
23 aquatic ecosystem3, it is a useful metric for determining the level at which a water body is
24 protected against risks of acidification. There is a direct correlation between ANC and pH
25 levels, which, along with dissolved aluminum, are more closely linked to the biological causes of
26 ecosystem response to acidification.
3 Biological effects are primarily attributable to a combination of low pH and high inorganic Al concentration. Such
conditions occur more frequently during rainfall and snowmelt that cause high flows of water and less commonly
during low-flow conditions, except where chronic acidity conditions are severe. Higher ANC values are generally
associated with lower risk of low pH during rainfall and snowmelt events.
September, 2010 5-69 Draft - Do Not Quote or Cite
-------
1 Because there is a direct correlation between ANC and pH levels, we can inform
2 the selection of target ANC in part through information on effects of pH as well as direct
3 studies of ANC. Figure 5-23 demonstrates the relationship between the measure of risk to
4 the aquatic ecosystems, ANC, and pH, the causal indicator of effect in aquatic ecosystems
5 (Chapter 2). Levels of pH are closely associated with ANC in the pH range of 4.5 to 7.
6 Within this range, higher ANC levels are associated with higher pH levels. At a pH level of
7 4.5, further reductions in ANC appear to be uncorrelated with pH, as pH levels remain at
8 4.5 while ANC values fall substantially. Likewise, at a pH value of 7, ANC values
9 continue to increase with no corresponding increase in pH. As pH is the primary causal
10 indicator of acidification related effects, this suggests that ANC values below -50 /ueq/L
11 (the apparent point on the function where pH reaches a minimum) are not likely to result
12 in further damage, while ANC values above 50 /zeq/L (the apparent point on the function
13 where pH reaches a maximum) are not likely to confer additional protection. As a result,
14 our focus will be on ANC values in the range of -50 to 50 /xeq/L.
15
September, 2010 5-70 Draft - Do Not Quote or Cite
-------
o
360 -,
200-
100
0-
-100
-200-
• measured
* calculated
•* ,
A
-300
1
2
3
4
5
Figure 5-23. Relationship between ANC and pH levels (reproduced from Bi et al, 2001)
Within this ANC range, the role of ANC is to protect against the likelihood of
decreased pH (and associated increases in Al). Gerritson et al (1996) demonstrate that
6 the probability of pH decreasing to a given threshold level is affected by the ANC level. In
7 general, the higher the ANC, the lower the probability of decreasing to the threshold.
8 Figure 5-24, reproduced from Gerritson et al (1996) shows that for two common fish
9 species, the probability of exposure to pH levels below threshold values declines as ANC
10 levels increase. This suggests that greater protection against threshold values of pH that cause
11 biological damage would be achieved by target ANC values set high enough to have a low
12 probability of the system going below a threshold pH level.
13
14
September, 2010
5-71
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
cr
3
3
o
5.4 5.8 6.2
b.ANC = 300 nmolc/L
a. Yellow Perch
1.0-1
0,0+
5.0 5,4 5.8 6,2
c. ANC = 500 fimol{ /L
6,6
0.8
0.6
0-4
0.2
0,0
100
b. River Herring
200
300
Wei Voar
AvM|J§ Vsaf
Dry Y
-------
1 Evidence of effects due to lowpH levels:
2 Significant harm to sensitive aquatic species has been observed at pH levels below 6.
3 Normal stream pH levels with little to no toxicity ranges from 6 to 7 (MacAvoy et al, 1995).
4 Baker et al (1990) observed that "lakes with pH less than approximately 6.0 contain significantly
5 fewer species than lakes with pH levels above 6.0". As noted in Chapter 2, typically at pH <4.5
6 and an ANC <0 /zeq/L, complete to near-complete loss of many taxa of organisms occur,
7 including fish and aquatic insect populations, whereas other taxa are reduced to only acidophilic
8 species.
9 Additional evidence can help refine the understanding of effects occurring at pH levels
10 between 4.5 and 6. When pH levels are below 5.6, relatively lower trout survival rates were
11 observed in the Shenendoah National Park. In field observations, when pH levels dropped to 5,
12 mortality rates went to 100 percent. (Bulger et al, 2000). At pH levels ranging from 5.4 to 5.8,
13 cumulative mortality continues to increase. Several studies have shown that trout exposed to
14 water with varying pH levels and fish larvae showed increasing mortality as pH levels decrease.
15 In one study almost 100 percent mortality was observed at a pH of 4.5 compared to almost 100
16 percent survival at a pH of 6.5. Intermediate pH values (6.0, 5.5) in all cases showed reduced
17 survival compared with the control (6.5), but not by statistically significant amounts (ISA
18 3.2.3.3). One study (Woodward, 1991) concluded that the threshold for effects of acidity on
19 greenback cutthroat trout in the absence of inorganic Al was pH 5.0.
20 One important indicator of acid stress is increased fish mortality. Table 5-9 summarizes
21 the evidence of mortality related to different levels of pH for a number of fish species.
September, 2010 5-73 Draft - Do Not Quote or Cite
-------
Table 5-9. Summary of Fish Mortality Response to pH
Source: EPA, 2008 (ISA)
Mortality
Endpoint
Increased
Mortality
>50% larval
mortality
embryo
survival
Significant
decrease
>50%
embryo
mortality
Substantial
reduction
Authors
Johnson et al.
(1987)
Holtze and
Hutchinson
(1989)
Johansson et al.
(1977)
Swenson et al.
(1989)
Mills etal. (1987)
Buckler et al.
(1987)
Klauda et al.
(1987)
Kane and Rabeni
(1987)
McCormick et al.
(1989)
Holtze and
Hutchinson
(1989)
Baker and
Scho field (1980)
Species
Blacknose dace,
creek chub
Brook trout
Common shiner
Lake whitefish,
white sucker,
walleye
Smallmouth bass
Atlantic salmon
Brown trout
Brook Trout
Black crappie
Rock bass
Yellow perch,
largemouth bass
Fathead minnow
Slimy sculpin
Lake Trout
Pearl dace
White sucker
Striped bass
Blueback herring
Smallmouth bass
Fathead minnow
Common shiner
White sucker
pH
Level
5.9-6.0
4.8-5.1
5.4-6.0
5.1 -5.2
4.8
5.0
4.5-5.0
4.5
5.5
5.0
4.5
5.9
5.6-5.9
5.6
5.1
5.0-5.1
6.5
5.7
5.1
6.0
5.4
5.2
Notes
In situ bioassay with early
life stages in Adirondack
surface waters
Laboratory exposure of
early life stages to pH and
Al.
Laboratory tests with eggs
exposed to low pH, no Al.
Laboratory tests with early
life stages exposed to pH
and Al.
Whole-lake treatment (fish
population recruitment
failure)
Lab bioassay
Lab bioassay
Lab bioassay
Lab bioassay
Lab bioassay
Lab bioassay
September, 2010
5-74
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
The response offish to pH is not uniform across species. A number of synoptic surveys
indicated loss of species diversity and absence of several fish species in the pH range of 5.0 to
5.5. If pH is lower, there is a greater likelihood that more fish species could be lost without
replacement, resulting in decreased richness and diversity. In general, populations of salmonids
are not found at pH levels less than 5.0, and smallmouth bass (Micropterus dolomieu)
populations are usually not found at pH values less than about 5.2 to 5.5. From Table 5-10, only
one study showed significant mortality effects above a pH of 6, while a number of studies
showed significant mortality when pH levels are at or below 5.5.
Table 5-10 Threshold response of increased mortality of fish to low pH listed from least
sensitive to most sensitive
„. . „ . Increased Mortality
StudV SPecies Threshold, pH
Johnson etal (1987) Blacknose dace, creek chub 59-60
Brook trout 4.8-51
Holtze and Hutchmson Common shiner 54-60
(1989)
Lake whitefish white suckei, 5.1 - 5 I
walleye
Smallmouth bass 4 8
Johansson et al. (1977) Atlantic salmon 50
Brown trout 45-50
Brook Trout 45
Swenson etal. (1939) Black crappie 55
Rock bass 5 0
Yellow perch, largemouth 4 5
bass
Mills et al (1 987} Fathead minnow 5 9
Slimy sculptn 5.6-59
Lake Trout 5 6
Pearl dace 5 1
White sucker 50 - 5 t
Source: Baker et al 1990, reproduced from the ISA, Table
Study Conditions
In situ bioaasay with early life stages in
Laboratory exposure of early life stages to pH
Laboratory tests with eggs exposed to low pH,
Laboratory tests with early life stages exposed
Whole-lake treatment (fish population
B-23.
September, 2010
5-75
Draft - Do Not Quote or Cite
-------
1 The highest pH threshold for any of the studies reported in Table 5-10 is 6.0, suggesting that pH
2 above 6.0 is protective against mortality effects for most species. Most thresholds are in the
3 range of pH of 5.0 to 6.0, which suggests that a target pH should be no lower than 5.0.
4 Protection against mortality in some recreationally important species such as lake trout (pH
5 threshold of 5.6) and crappie (pH threshold of 5.5), combined with the evidence of effects on
6 larval and embryo survival suggests that pH levels greater than 5.5 should be targeted to provide
7 protection against mortality effects throughout the lifestages of fish.
8 Non-lethal effects have been observed at pH levels as high as 6. A study in the
9 Shenendoah National Park found that the condition factor, a measure offish health expressed as
10 fish weight/length3 multiplied by a scaling constant, is positively correlated with stream pH
11 levels, and that the condition factor is reduced in streams with a pH of 6.0 (ISA 3.2.3.3).
12 Biodiversity is another indicator of aquatic ecosystem health. As discussed in Chapter 2,
13 a key study in the Adirondacks found that lakes with a pH of 6.0 had only half the potential
14 species offish (27 of 53 potential species). There is often a positive relationship between pH and
15 number of fish species, at least for pH values between about 5.0 and 6.5, or ANC values between
16 about 0 to 100 ^eq/L (Bulger et al., 1999; Cosby et al., 2006; Sullivan et al., 2006). Such
17 observed relationships are complicated, however, by the tendency for smaller lakes and streams,
18 having smaller watersheds, to also support fewer fish species, irrespective of acid-base
19 chemistry. This pattern may be due to a decrease in the number of available niches as stream or
20 lake size decreases. Nevertheless, fish species richness is relatively easily determined and is one
21 of the most useful indicators of biological effects of surface water acidification.
22 In a study of Ontario lakes, Matuszek and Beggs (1988) found that the number offish
23 species is positively correlated with pH, with a clear loss of species starting at pH levels less than
24 or equal to 5.5. This relationship is displayed in Figure 5-25.
September, 2010 5-76 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
+2
CO
s+1
UJ
ft o
UJ
CQ
CO
111
£
<
UJ
c
-5
-7
4,5
50
5.5
6.0
6.5
PH
7.0
7.5
8.0
8.5
Figure 5-25.
Mean residual number of species per lake for lakes in Ontario, by pH
interval. The residual number of species for a lake is the deviation of
the observed number from the number predicted by lake area.
Source: Matuszek and Beggs (1988).
A study in the Adirondacks found that among the studied lakes with fish, there was an
unambiguous relationship between the number offish species and lake pH, ranging from about
one species per lake for lakes having pH less than 4.5 to about six species per lake for lakes
having pH higher than 6.5 (Baker et al., 1990b; Driscoll et al., 2001).
Biological responses at differing pH levels are summarized in Table B-13 of the ISA,
reproduced here in Table 5-11.
September, 2010
5-77
Draft - Do Not Quote or Cite
-------
1 Table 5-11 General summary of biological changes anticipated with surface water
2 acidification, expressed as a decrease in surface water pH.
pH General Biological Effects
Decrease _ ___
6.5 to 6.0 Small decrease in species richness of plankton and benthic invertebrate
communities resulting from the loss of a few highly acid-sensitive species, but no
measurable change in total community abundance or production.
Some adverse effects (decreased reproductive success) may occur for highly
acid-sensitive fish species (e.g., fathead minnow, striped bass).
6.0 to 5.5 Loss of sensitive species of minnows and dace, such as fathead minnow and
blacknose dace; in some waters, decreased reproductive success of lake trout and
walleye, which are important sport fish species in some areas.
Visual accumulation of filamentous green algae in the near-shore zone of many
lakes and in some streams.
Distinct decrease in species richness and change in species composition of
plankton and benthic invertebrate communities, although little if any change in
total community abundance or production.
Loss of some common invertebrate species from zooplankton and benthic
communities, including many species of snails, clams, mayflies, and amphipods,
and some crayfish.
5.5 to 5.0 Loss of several important sport fish species, including lake trout, walleye,
rainbow trout, and smallmouth bass, as well as additional nongame species such
as creek chub.
Further increase in the extent and abundance of filamentous green algae in lake
near-shore areas and streams.
Continued shift in species composition and decline in species richness of
plankton, periphyton, and benthic invertebrate communities; decreases in total
abundance and biomass of benthic invertebrates and zooplankton may occur in
some waters.
Loss of several additional invertebrate species common in surface waters,
including all snails, most species of clams, and many species of mayflies,
stoneflies, and other benthic invertebrates.
Inhibition of nitrification.
September, 2010 5-78 Draft - Do Not Quote or Cite
-------
pH General Biological Effects
Decrease _ _ _ _ _ _____
5.0 to 4.5 Loss of most fish species, including most important sport fish species such as
brook trout and Atlantic salmon. A few fish species are able to survive and
reproduce in water below pH 4.5 (e.g., central mudminnow, yellow perch, and in
some waters, largemouth bass).
Measurable decline in the whole-system rates of decomposition of some forms of
organic matter, potentially resulting in decreased rates of nutrient cycling.
2 Evidence of changes in biological endpoints in response to changes in pH indicates that
3 while the time to full response varies, and the pH level at which recovery occurs varies, there is
4 clear evidence that increasing pH levels leads to improvements in a number of measures of
5 ecosystem health. In a study in Ontario, there was varied response to increases in pH across
6 studied lakes, with little increase in phytoplankton diversity in one lake as pH changed from 5.0
7 to 5.8 but a strong recovery of diversity at pH above 6 (Findlay and Kasian, 1996), while in
8 another lake, profound change began at pH 5.5.
9 Fish populations have recovered in acidified lakes when the pH and ANC have been
10 increased through liming or reduction of acidifying deposition (Beggs and Gunn, 1986; Dillon et
1 1 al., 1986; Gunn et al., 1988; Hultberg and Andersson, 1982; Keller and Pitblado, 1986; Kelso
12 and Jeffries, 1988; Raddum et al., 1986).
13 In considering the range of pH values to evaluate further in the context of impacts on
14 public welfare, the overall weight of evidence regarding the full range of biological effect
1 5 indicators, including mortality, fish health, and biodiversity supports further consideration of a
16 target pH level of no less than 5.5, with some support for pH target levels of 6 to 6.5.
17
1 8 Evidence of effects directly or indirectly related to low ANC levels:
1 9 The number of fish species present in a waterbody has been shown to be positively
20 correlated with the ANC level in the water, with higher values supporting a greater richness and
21 diversity of fish species (Figure 5-26 and 5-27). The diversity and distribution of phyto-
22 zooplankton communities are also positively correlated with ANC.
23 When comparing lower ANC to higher ANC streams, the lower ANC streams
24 demonstrated severe mortality in young fish due to a sharp drop in pH and increase in Al
September, 201 0 5-79 Draft - Do Not Quote or Cite
-------
1 concentrations. The higher ANC stream maintained a pH level greater than 6.6 with
2 correspondingly low Al concentrations (MacAvoy et al 1995). In streams with chronically low
3 ANC, pH levels were in the range of 5.3 to 5.6, and a steady decline in trout populations was
4 observed.
5 Within the range where ANC is the limiting indicator, (e.g. ANC range from -50 to 100
6 /ieq/L) the relationship between ANC and ecosystem impacts ranges from linear to non-linear,
7 with a sigmoidal shape. Evidence shows a linear relationship in streams in the Eastern U.S., and
8 a non-linear relationship in lakes in the Northeastern U.S. (Figures 5-26 and 5-27).
9
10
11
12
13
14
-20
20
-r
40 60 BO 100 120
Minimum ANC (period of- record)
teo
Figure 5-26 Relationship between ANC and number of fish species present in aquatic
freshwater ecosystems in Shenandoah National Park (Source: Arthur Bulger,
University of Virginia, reproduced from NAP AP, 2005.)
September, 2010
5-80
Draft - Do Not Quote or Cite
-------
o
0)
Q.
to
14-
12-
10-
6-
E 2-
-2-
-4-
-200
-100
100
I
200
300
I
400
500
ANC(peq/L)
Source: Sullivan et al. (2006)
Figure 5-27. Number of fish species per lake versus acidity status, expressed as ANC, for
Adirondack lakes. The data are presented as mean (filled circles) and range
(bars) of species richness within 10 ueq/L ANC categories, based on data
collected by the Adirondack Lakes Survey Corporation.
1 On average, the fish species richness is lower by one fish species for every 21 fieq/L
2 decrease in ANC in Shenandoah National Park streams (ISA 3.2.3.4).
3 For freshwater systems, ANC levels can be grouped into five major classes: <0, 0-20,
4 20-50, 50-100, and >100 /zeq/L, with each range representing a probability of ecological
5 damage to the community. The five categories of ANC and expected ecological effects are
6 described Table 2-1 in Chapter 2 and are supported by a large body of research completed
7 throughout the eastern United States (Sullivan et al., 2006).
8 In a study in Maryland, Gerritsen et al. (1996) found that at an ANC less than 40 to 70
9 ^teq/L, Yellow perch had a risk of exposure to its critical pH =5.5 of greater than 50 percent.
10 Figures from that study, reproduced in Figure 5-24 above, show the relationships between ANC
11 and the probability of exceeding benchmark pH levels that have been associated with mortality
12 in the Yellow perch and River herring. These figures demonstrate two important principles, 1)
13 as ANC levels fall, the distribution of pH shifts to the left, with increasing probability of low pH
September, 2010
5-81
Draft - Do Not Quote or Cite
-------
1 levels, and 2) as ANC levels fall, the probability of exceeding mortality thresholds increases, but
2 this relationship varies by fish species.
3 The specific relationship between ANC and the probability of exceeding benchmark pH
4 levels varies by water body and fish species. However, based on Figure 5-23, ANC levels
5 matching the target pH levels of 5.5 to 6.5 discussed in the previous section range from
6 approximately 0 to 50 /ieq/L. In considering the range of ANC values to evaluate further in the
7 context of impacts on public welfare, the overall evidence on effects at lower ANC levels
8 described in Table 2-1, supports further consideration of a range of ANC from 20 to 50 /xeq/L.
9 While Figure 5-23 suggests some probability that ANC levels down to 0 /ieq/L are correlated
10 with pH levels of 5.5, the specific relationship between ANC and fish species diversity shown in
11 Figure 5-27 indicates that at an ANC of 0 ^ieq/L, there is significant damage to ecosystems with
12 almost complete loss offish species, which indicates that an ANC of 0 /ieq/L is not an
13 appropriate target level to protect against significant damage to ecosystems.
14
15 Evidence and Risk Based Conclusions regarding target ANC levels
16 ANC values less than or equal to 0 ^eq/L are chronically acidic and can lead to complete
17 loss of species, and major changes in the ability of waterbodies to support diverse biota,
18 especially in water bodies that are highly sensitive to episodic acidification. Biota are generally
19 not harmed when ANC values are >100 ^eq/L, due to the low probability that pH levels will be
20 below 7. In the Adirondacks, the number of fish species also peaks at ANC values >100 /ieq/L.
21 This suggests that at ANC greater than 100, little risk from acidification exists in many aquatic
22 ecosystems. At ANC levels below 100 //eq/L, overall health of an aquatic community can be
23 maintained; however, fish fitness and community diversity begin to decline. At ANC levels
24 between 100 and 50 /ieq/L, the likelihood that fitness of sensitive species (e.g., brook trout,
25 zooplankton) will begins to decline is increased. When ANC concentrations are <50 ^eq/L, the
26 probability of acidification increases substantially, and negative effects on aquatic biota are
27 observed, including large reductions in diversity offish species, and changes in health offish
28 populations, affecting reproductive ability and fitness. ANC levels below 20 //eq/L are generally
29 associated with high probability of low pH, leading to death or loss of fitness of biota that are
30 sensitive to acidification. (ISA 5.2.2.1 and REA 5.2.1.2). At these levels, during episodes of
31 high acidifying deposition, brook trout populations may experience lethal effects. In addition, the
September, 2010 5-82 Draft - Do Not Quote or Cite
-------
1 diversity and distribution of zooplankton communities decline sharply at ANC levels below 20
2 fJLeq/L. Overall, there is little uncertainty that significant effects on aquatic biota are occurring at
3 ANC levels below 20 fieq/L.
4 Based on the field data from the Adirondacks and Shenendoah case study areas, ANC
•5 levels less than 50 /ieq/L are adverse to ecosystem health, and are likely to lead to reductions in
6 ecosystem services related to recreational fishing. However, the types of effects, specific
7 species, and prevalence of effects across water bodies in the U.S. is more uncertain at ANC
8 levels between 20 and 50 /xeq/L. ANC levels between 50 and 100 /xeq/L are potentially adverse
9 to ecosystem health, and may result in losses in ecosystem services, but the effects are less
10 severe and greater uncertainty exists as to the magnitude of ecosystem service impacts.
11 Consideration of the appropriate levels of ANC to target in the standard to reduce
12 the likelihood of effects from aquatic acidification can be based upon the above presented
13 categories of aquatic status in Table 2-1. Using this information as well as information
14 provided by both the ISA and REA, the lowest two categories (0 and <20 //eq/L) would
15 appear inadequate to protect against catastrophic loss of ecosystem function. While
16 ecological effects occur at ANC levels below 50 ^teq/L, the degree and nature of those
17 effects is less significant than at levels below 20 jueq/L. Levels between 50 and 100 /ieq/L
18 would provide additional protection; however, uncertainties regarding the additional
19 reduction in adverse welfare effects are much larger for target ANC levels above 50 /xeq/L.
20 The target ANC level specified in designing the standard is only one part in
21 determining the overall protectiveness of the standard. The degree of protectiveness is
22 based on all elements of the standard, including the target ANC, the size of the spatial
23 areas over which the standard is applied, the percent of aquatic ecosystems targeted within
24 a spatial area that is selected by the Administrator in light of the selected ANC level, the
i
25 atmospheric indicator, the calculated values for the deposition transformation ratios (TNOx
26 and TSOx), and the calculated value for reduced nitrogen deposition (NHx). There are
27 widely varying degrees of uncertainty associated with all of these elements, some being
September, 2010 5-83 Draft - Do Not Quote or Cite
-------
1 much more certain and others being much less certain. The specified target ANC
2 level is a crucial part in determining the adequacy of protection provided by an AAPI, but
3 it is the overall design and content of the standard that must be considered in judging the
4 adequacy of protection it provides.
5 Consideration of the target ANC should also reflect that an adequate level of ANC
6 should protect against episodic as well as long term effects. Selecting a higher chronic
7 ANC level can provide greater protection against short term peaks in acidification. In
8 addition, selection of ANC values in the lower end of the range of 20 to 50 /ieq/L
9 provides less protection against these short term episodic effects. Selection of target ANC
10 values in the upper end of the range from 20 to 50 /xeq/L provides additional protection
11 against episodic peaks in acidification.
12 When considering a standard to protect against aquatic acidification, it is necessary
13 to take into account both the time period desired for recovery as well as the potential of
14 recovery. Ecosystems become adversely impacted by acidifying deposition over long
15 periods of time and have variable time frames and abilities to recover from such
16 perturbations. Modeling presented in the REA (REA Section 4.2.4) shows the estimated ANC
17 values for Adirondack lakes and Shenandoah streams under pre-acidification conditions and
18 indicates that for a small percentage of lakes and streams, natural ANC levels would have
19 been below 50 /ieq/L. Therefore, for these waterbodies, no reduction in input is likely to
20 achieve an ANC of 50 /ueq/L or greater. Conversely, for some lakes and streams the level
21 of perturbation from long periods of acidifying deposition has resulted in very low ANC
22 values compared to estimated natural conditions. For such waterbodies, the time to
23 recovery would be largely dependent on future inputs of acidifying deposition.
24 The concept of target ANC is based on the long-term response of aquatic
25 ecosystems. The time required for a waterbody to achieve the target ANC given a
26 decrease in emissions such that the critical load for that target ANC is not exceeded is
27 often decades if not centuries. In recognition of the potential desire to achieve the target
28 ANC in a shorter time frame, the concept of target loads had been developed. Target
29 loads represent the depositional loading that is expected to achieve a particular level of the
September, 2010 5-84 Draft - Do Not Quote or Cite
-------
1 ecological indicator by a given time. Similarly, one might also consider specifying a target
2 ANC that will yield an intermediate ANC level within a specified time horizon. For
3 example, if there is a desire to obtain an ANC of 20 fieq/L by 2030, it might be necessary
4 to use the depositional loadings equivalent to obtaining a target ANC of 50 fieq/L which
5 will ultimately be realized many years later. The depositional loading associated with the
6 interim ANC target is called the target load. The implication of this is that an important
7 component of the decision on the level of the AAPI is the timeframe in which the desired
8 ANC value is to be achieved. There is a great deal of heterogeneity in the response time
9 among waterbodies, and as such, specific quantification of the relationship between the
10 critical loads and target loads is not possible.
11
12 5.4.2 Additional information on Target ANC levels
13 A number of regional organizations, states, and international organizations have
14 developed critical loads frameworks to protect against acidification of sensitive aquatic
15 ecosystems. In considering the appropriate range of target ANC levels, it is informative to
16 evaluate the target ANC levels selected by these different organizations, as well as the
17 rationale provided in support of the selected levels. Chapter 3 provides a detailed
18 discussion of how critical loads have been developed and used in other contexts. This
19 section summarizes the specific target ANC values and their rationales.
20 The UNECE has developed critical loads in support of international emissions
21 reduction agreements. As noted in Chapter 3, critical loads were established to protect
22 95 percent of surface waters in Europe from an ANC less than 20/veq/L, based on
23 protection of brown trout. Individual countries have set alternative ANC targets, for
24 example, Norway targets an ANC of 30 /;eq/L, based on protection of Atlantic salmon.
25 Several states have established target ANC or pH values related to protection of
26 lakes and streams from acidification. While recognizing that some lakes in the Adirondacks
27 will have a naturally low pH, the state of New York has established a target pH value of
28 6.5 for lakes that are not naturally below 6.5. As noted above this level is associated with
29 an ANC value that is likely to be between 20 and 50 jjeq/l (or higher). Vermont has set
September, 2010 5-85 Draft - Do Not Quote or Cite
-------
1 an explicit ANC target of 50 //eq/L Tennessee has established site specific target ANC
2 values based on assessments of natural acidity, with a default value of 50 /;eq/L when
3 specific data is not available.
4 Taken together, these policy responses to concerns about ecological effects
5 associated with acidification indicate that target ANC values between 20 and 50 /*eq/L
6 have been selected by states and other nations to provide adequate protection of lakes and
7 streams in some of the more sensitive aquatic ecosystems.
8
9 5.4.3 Adversity of ecological impacts associated with alternative target ANC levels
10 The point at which effects on public welfare become adverse is not specifically defined in
11 the Clean Air Act. In Chapter 3 we explained that characterizing a known or anticipated adverse
12 effect to public welfare is an important component of developing any secondary NAAQS.
13 According to the Clean Air Act, welfare effects include:
14 effects on soils, water, crops, vegetation, manmade materials, animals, wildlife,
15 weather, visibility, and climate, damage to and deterioration of property, and
16 hazards to transportation, as well as effect on economic values and on personal
17 comfort and well-being, whether caused by transformation, conversion, or
18 combination with other air pollutants (CAA, Section 302(h)).
19
20 While the text above lists a number of welfare effects, these effects do not define public
21 welfare in and of themselves. Consideration of adversity in the context of the secondary
22 NOx and SOx standards can be informed by information about losses in ecosystem
23 services associated with acid deposition, and the potential economic value of those losses
24 (Chapter 3).
25 Ecosystem service losses at alternative levels of ANC are difficult to enumerate.
26 However, in general there are categories of ecosystem services (discussed in detail in Chapter 3)
27 that are related to the specific ecosystem damages expected to occur at alternative ANC levels.
28 Losses in fish populations due to very low ANC (below 20 /xeq/L) are likely associated with
29 significant losses in value for recreational and subsistence fishers. Many acid sensitive lakes are
30 located in areas with high levels of recreational fishing activity. For example, in the
31 Northeastern U.S., where nearly 8 percent of lakes are considered acidic, more than 9 percent of
32 adults participate in freshwater fishing, with a value of $5 billion in 2006. This suggests that
September, 2010 5-86 Draft - Do Not Quote or Cite
-------
1 improvements in lake fish populations are likely associated with significant recreational fishing
2 value.
3 Inland surface waters also provide cultural services such as aesthetic and existence value
4 and educational services. To the extent that piscivorous birds and other wildlife are harmed by
5 the absence offish in these waters, hunting and birdwatching activities are likely to be adversely
6 affected.
7 A case study of the value to New York residents of improving the health of lakes in the
8 Adirondacks found significant willingness to pay for those improvements. When scaled to
9 evaluate the improvement in lake health from achieving an ANC of 20 to 50 fieq/L, the study
10 implies benefits to the New York population of $400 to $900 million per year (in constant
11 2007$) for achievement of an ANC of 20 //eq/L, and $300 to $800 million per year for achieving
12 an ANC of 50 /^eq/L. The survey administered in this study recognized that participants were
13 thinking about the full range of services provided by the lakes in question - not just the
14 recreational fishing services. Therefore the estimates of WTP include resident's benefits for
15 potential birdwatching and other ancillary services. These results are just for New York
16 populations. If similar benefits exist for improvements in other acid sensitive lakes, the
17 economic value to U.S. populations could be very substantial, suggesting that, at least by one
18 measure of public welfare, impacts associated with ANC greater than 50 jiieq/L are adverse to
19 public welfare.
20
September, 2010 5-87 Draft - Do Not Quote or Cite
-------
1 5.5 Options for specifying the targets for the deposition metric
2
3 This section outlines the steps necessary to aggregate critical loads to form a depositionl
4 metric (DL%ECO) for a broad geographic area at a desired level of protection. In this chapter,
5 evaluate DL%ECO for several of the options for categorization of landscape sensitivity based on
6 acid-sensitivity that are presented in section 5.3. The options include one DL%ECO for the entire
7 population of lakes and streams at a given level of protection, creating two categories (sensitive
8 and less-sensitive) with a distinct DL%ECO for each, and a cluster approach which creates
9 multiple categories of sensitivity for which DL%ECO are calculated. Regardless of the method of
10 aggregation, the question becomes how to choose a DL%ECO value to represent the population of
11 CL for waterbodies. For the purpose of comparison, we have chosen to calculate the DL%ECO
12 from the 90 percentile CL, 75 percentile CL and 50 percentile CL for each method of
13 aggregation. Using this method, the target DL%ECO becomes that value which represents
14 protection for 90%, 75% or 50% of the population of waterbodies for a desired level of ANC and
15 would not offer the same degree of protection to those waterbodies with CL less than the chosen
16 DL%ECO. While we would not expect the same degree of protection for these waterbodies, they
17 would likely receive some benefit from the reduced deposition necessary to meet the selected
18 DL%ECO. This chapter will therefore also present comparisons between the percentage of
19 waterbodies that would be protected at given DL%ECO from an ANC <50 /ieq/L and those that
20 would likely be protected from an ANC <20 /*eq/L at the same DL%ECO. These comparisons are
21 described below.
22 Table 5.12 shows a comparison between the percent of waterbodies that would be
23 protected from ANC<50 using 90%, 75% and 50% DL%ECO for the one population approach and
24 those that, while not protected from an ANC <50 ^eq/L would be protected from an ANC<20
25 /ieq/L. The selection of the DL%ECO for ANC at 50 /xeq/L representing 90% of the waterbodies
26 would likely protect 97% of all waterbodies from having an ANC<20 /ieq/L. If the 75% DL%ECO
27 for ANC 50 /^eq/L was chosen, 83% of waterbodies would likely be protected from an ANC<20
28 ^eq/L and if the 50% DL%ECO was chosen only 56% of waterbodies would likely be protected
September, 2010 5-88 Draft - Do Not Quote or Cite
-------
1 against an ANC <20 /ieq/L. This is an important distinction as severe degradation is likely to
2 occur in lakes and streams with ANCX20 /ieq/L.
3
4
5
Table 5-12. Comparison of percentage protection from ANC values less than 50 and
less than 20 using DL that result when the US is considered one population.
Percentile
protection
90%
75%
50%
DL
(meq/m2/yr)
27
55
118
Total
number
of Sites
in
Analysis
5280
5280
5280
Total
Number
of Sites
protected
from
ANC <50
4778
3973
2654
Total %
Sites
protected
from
ANC <50
90 _j
75
50
Total
Number
of Sites
protected
from
ANC <20
5145
4394
2947
Total %
Sites
protected
from
ANC<20
97
83
56
Table 5.13 shows a comparison between the percent of waterbodies that would be
6 protected from ANC<50 /ieq/L using 90%, 75% and 50% DL%ECO s for subdividing the U.S. into
7 two acid-sensitivity categories and those that, while not protected from an ANC <50 would
8 likely be protected from an ANC<20 /xeq/L. Sensitive Category: The selection of the DL%ECO
9 for ANC50 /ieq/L representing 90% of the sensitive waterbodies would likely protect 98% of all
10 waterbodies from having an ANC<20 /ieq/L. If the 75% DL%ECO for ANC 50 /zeq/L for sensitive
11 waterbodies was chosen, 84% of waterbodies would likely be protected from an ANC<20 £ieq/L
12 and if the 50% DL%ECO were chosen only 57% of waterbodies would likely be protected against
13 an ANC <20 //eq/L. Less-sensitive Category: The selection of the DL%ECO for ANC50
14 representing 90% of the non-sensitive waterbodies would likely protect 92% of all waterbodies
15 from having an ANC<20 /zeq/L. If the 75% DL%ECO for ANC 50 /zeq/L for sensitive waterbodies
16 was chosen, 77% of waterbodies would likely be protected from an ANC<20 fJLeq/L and if the
17 50% DL%ECO were chosen only 52% of waterbodies would likely be protected against an ANC
18 <20/xeq/L.
19
20
September, 2010
5-89
Draft - Do Not Quote or Cite
-------
Table 5-13. Comparison of percentage protection from ANC values less than 50 and less
than 20 using DL that result when the US is divided into two categories, sensitive and less
sensitive based on ANC data.
Sensitive
Less
sensitive
Percentile
protection
90%
75%
50%
90%
75%
50%
(meq/m2/yr)
26
51
106
53
117
277
Total
number
of Sites
in
Analysis
4553
4553
4553
727
727
111
Total
Number
of Sites
protected
from
ANC<50
4104
3428
2284
655
546
364
Total %
Sites
protected
from
ANC <50
90
75
50
90
75
50
Total
Number
of Sites
protected
from
ANC<20
4451
3841
2575
672
560
377
Total %
Sites
protected
from
ANC <20
98
84
57
92
77
52
1
2
3
4
5
6
7
8
9
10
11
The sensitive category can be further subdivided into ecoregions (Option 2c in section
5.3) and the deposition metrics would be calculated only from sites within the ecoregions. The
DL%ECO values calculated for each ecoregion (level 3) are presented in Table 5.14s. The
DL%ECO s calculated for the sensitive category of waterbodies (the CLs from all ecoregions
considered sensitive are pooled together) are compared against the DL%ECO developed for each
ecoregions within the sensitivity category for both ANC 50 /xeq/L and ANC 20 /xeq/L and for
each percentile (90, 75 and 50).
[Placeholder for additional discussion on percentile values]
September, 2010
5-90
Draft - Do Not Quote or Cite
-------
1 5.6 REFERENCES
3 Kaufmann PR, Herlihy AT and Baker LA. Sources of acidity in lakes and streams of the United
4 States Environmental Pollution 77 (1992) 115-122
5 Sverdrup, H., de Vries, W., Henriksen, A., 1990. Mapping Critical Loads: A Guidance to the
6 Criteria, Calculations, Data Collection and Mapping of the Critical Loads. Milforapport
7 (Environmental Report) 1990: 14. Nordic Council of Ministers, Copenhagen, NORD:
8 1990:98, 124pp.
9 USDA Natural Resources Conservation Service (NRCS), 1995. State Soil Geographic
10 STATSGO) Data Base: Data Use Information. Miscellaneous Publication Number 1492.
11 National Soil Survey Center, 113 pp. Available from:
12 .
13 Miller, D.A., White, R.A., 1998. A conterminous United States multi-layer soil characteristics
14 data set for regional climate and hydrology modeling. Earth Interactions 2. Available
15 from: .
16 Omernik, J.M. 2004. Perspectives on the nature and definition of ecological regions.
17 Environmental Management 34(Supplement I):s27-s38.
18 McMahon, G., S. M. Gregonis, S. W. Walton, J. M. Omernik,!. D. Thorson, J. A. Freeouf, A. H.
19 Rorick, and J. E. Keys.2001. Developing a spatial framework of common
20 ecologicalregions for the conterminous United States. Environmental Management
21 8:293-316.
22 Omernik, J. M. 1987. Ecoregions of the conterminous UnitedStates. Annals of the Association of
23 American Geographers77:118 -125.
24 Omernik, J. M. 1995. Ecoregions: A spatial framework for environmental management W. S.
25 Davis T. P. Simon Biological assessment and criteria: Tools for water resource planning
26 and decision making Lewis Publishing Boca Raton,Florida.
27 Griffith, G. E., J. M. Omernik, and A. J. Woods. 1999. Ecoregions, watersheds, basins, and
28 HUCs: How state and federal agencies frame water quality. Journal of Soil and Water
29 Conservation 54:666-677.
30 Omernik, J. M., S. S. Chapman, R. A. Lillie, and R. T. Dumke. 2000. Ecoregions of Wisconsin.
31 Transactions of the Wisconsin Academy of Sciences, Arts, and Letters 88:77-103.
32 Wiken, E. B. 1986. Terrestrial ecozones of Canada. Environment Canada. Ecological Land
33 Classification Series No. 19, Ottawa, Ontario, 26 pp.
34 Commission for Environmental Cooperation. 1997. Ecological regions of North America: toward
35 a common perspective. Commission for Environmental Cooperation, Montreal, Quebec,
36 Canada. 71pp. Map (scale 1:12,500,000).
37 Baker ]P; Gherini SA; Christensen SW; Driscoll CT; Gallagher J; Munson RK; Newton RM;
38 Reckhow KH; Schofield CL. (1990). Adirondack lakes survey: An interpretive
39 analysis of fish communities and water chemistry, 1984-1987. Ray Brook, NY:
40 Adirondack Lakes Survey Corporation.
41 Baker JP; Bernard DP; Christensen SW; Sale M]. (1990). Biological effects of changes in
42 surface water acid-base chemistry. (State of science / technology report
43 #\ 3).Washington DC: National Acid Precipitation Assessment Program (NAPAP).
44 Beggs GL; Gunn JM. (1986). Response of lake trout (Salvelinus namaycush and brook trout (S.
45 fontinalis) to surface water acidification in Ontario. Water Air Soil Pollut, 30, 711-717.
September, 2010 5-91 Draft - Do Not Quote or Cite
-------
1
2 Bi, S.P., S.Q. An, F. Liu. 2001. A practical application of Driscoll's equation for predicting
3 theacid-neutralizing capacity in acidic natural waters equilibria with the mineral phase
4 gibbsite. Environment International 26 (2001) 327-333.
5 Bulger, A.J., B.J. Cosby, C.A. Dolloff, K.N. Eshleman, J.R. Webb, J.N. Galloway. 2000.
6 SNP:FISH Shenandoah National Park: Fish In Sensitive Habitats Project Final Report,
7 Volume I.
8 Cosby BJ; Webb JR; Galloway JN; Deviney FA. (2006). Acidic deposition impacts on natural
9 resources in Shenandoah National Park. (Technical Report NPS/NER/NRTR—2006/066
10 ) Philadelphia, PA; Northeast Region of the National Park Service; U.S. Department of
11 the Interior.
12 Dillon PJ; Reid RA; Girard R. (1986). Changes in the chemistry of lakes near Sudbury, Ontario,
13 following reduction of SO2 emissions. Water Air Soil Pollut, 31, 59-65.
14 Driscoll CT; Lawrence GB; Bulger AJ; Butler TJ; Cronan CS; Eagar C; Lambert KF; Likens
15 GE; Stoddard JL; Weather KC. (2001). Acidic deposition in the northeastern United
16 States: Sources and inputs, ecosystem effects, and management strategies. Bioscience,
17 51,180-198.
18 Findlay DL; Kasian SEM. (1996). The effect of incremental pH recovery on the Lake 223
19 phytoplankton community. Can J Fish Aquat Sci, 53, 856-864.
20 Gerritson et al (1996)
21 Hultberg H; Andersson I. (1982). Liming of acidified lakes: induced long-term changes. Water
22 Air Soil Pollut, 18, 311-331.
23 Gunn JM; McMurtry MJ; Casselman JM; Keller W; Powell MJ. (1988). Changes in the fish
24 community of a limed lake near sudbury, Ontario: Effects of chemical neutralization or
25 reduced atmospheric deposition of acids? Water Air Soil Pollut, 41, 113-136.
26 Keller W; Pitblado JR. (1986). Water quality changes in Sudbury area lakes: a comparison of
27 synoptic surveys in 1974-1976 and 1981-1983. Water Air Soil Pollut, 29, 285-296.
28 Kelso JRM; Jeffries DS. (1988). Response of headwater lakes to varying atmospheric deposition
29 in north-central Ontario. Can J Fish Aquat Sci, 45, 1905-1911.
30 MacAvoy SW; Bulger A]. (1995). Survival of brook trout (Salvelinus fontinalis) embryos
31 and fry in streams of different acid sensitivity in Shenandoah National Park, USA.
32 Water Air Soil Pollut, 85, 445-450.
33 Matuszek JE; Beggs GL. (1988). Fish species richness in relation to lake area, pH, and other
34 abiotic factors in Ontario lakes. Can J Fish Aquat Sci, 45, 1931 -1941.
35 NAPAP. (2005). National acid precipitation assessment program report to Congress: An
36 integrated assessment, http://www.esrl.noaa.gov/csd/aqrs/reports/napapreport05.pdf.
37 Silver Spring, MD: National Acid Precipitation Assessment Program (NAPAP);
38 Committee on Environment and Natural Resources (CENR) of the National Science and
39 Technology Council (NSTC).
40 Raddum GG; Brettum P; Matzow D; Nilssen JP; Skov A; Svealy T; Wright RF. (1986). Liming
41 the acid lake Howatn, Norway: a whole lake study. Water Air Soil Pollut, 31, 721-763.
42 Sullivan TJ; Fernandez IJ; Herlihy AT; Driscoll CT; McDonnell TC; Nowicki NA; Snyder KU;
43 Sutherland JW. (2006). Acid-base characteristics of soils in the Adirondack Mountains,
44 New York. Soil Sci Soc Am J, 70, 141 -152.
September, 2010 5-92 Draft - Do Not Quote or Cite
-------
1 Sullivan TJ; Driscoll CT; Cosby BJ; Fernandez IJ; Herlihy AT; Zhai J; Stemberger R; Snyder
2 KU; Sutherland JW; Nierzwicki-Bauer SA; Boylen CW; McDonnell TC; Nowicki NA.
3 (2006). Assessment of the extent to which intensively studied lakes are representative of
4 the Adirondack Mountain region. (Final Report no 06-17).Corvallis, OR; prepared by
5 Environmental Chemistry, Inc. for: Albany, NY; Environmental Monitoring Evaluation
6 and Protection Program of the New York State Energy Research and Development
7 Authority (NYSERDA).
8 Woodward DF. (1991). Sensitivity of greenback cutthroat trout to acidic pH and elevated
9 aluminum. Trans Am Fish Soc, 120, 34-42.
10
September, 2010 5-93 Draft - Do Not Quote or Cite
-------
1 6. CO-PROTECTION FOR OTHER EFFECTS USING STANDARDS TO
2 PROTECT AGAINST ACIDIFICATION
3 Discussion in this Policy Assessment on the NOx and SOx secondary standard has, to this
4 point, centered on the level of protection of aquatic ecosystems against acidification from
5 atmospheric deposition of NOx and SOx. This chapter focuses on the co-protection such a
6 standard could achieve for other ecological effects, including terrestrial acidification, terrestrial
7 nutrient enrichment, and estuarine eutrophication.
8 6.1 To What Extent Would A Standard Specifically Defined To Protect Against Aquatic
9 Acidification Likely Provide Protection From Terrestrial Acidification?
10 To understand the level of co-protection a NOx and SOx secondary standard based on
11 aquatic acidification can provide for terrestrial ecosystems, EPA staff conducted an analysis to
12 compare the critical acid loads for aquatic and terrestrial components of watersheds in the eastern
13 United States. Aquatic critical acid loads are an integrated function of the chemistry of runoff
14 from stream and lake waters, and the biogeochemical processes that occur within the aquatic and
15 terrestrial components of the entire watershed. Terrestrial critical acid loads, however, are
16 largely determined by the conditions and processes that occur in the root zone of the soil profile
17 of the terrestrial systems of a watershed. Therefore, it is possible to have different critical acid
18 load values for aquatic and terrestrial ecosystems within the same watershed.
19 For the comparative analysis of aquatic and terrestrial critical acid loads, aquatic critical
20 acid loads were selected to protect for an acid neutralizing capacity (ANC) of 50ueq/L, and were
21 taken directly from the Risk and Exposure Assessment (REA) (2009). The 50[j,eq/L ANC value
22 was one of three values modeled in the REA (2009) for aquatic acidification. The terrestrial
23 critical acid loads in this comparative analysis were selected to protect for either a terrestrial base
24 cation to aluminum molar ratio (Bc:Al) of 1.2 or 10.0. The Bc:Al ratio of 10.0 is more
25 conservative, as it provides greater protection against the impacts of acidification on cation
26 availability and aluminum toxicity in the soil solution. The terrestrial critical loads were
27 calculated using the Simple Mass Balance (8MB) method outlined in the REA (2009) and input
28 values averaged across the area of each watershed.
September, 2010 6-1 Draft - Do Not Quote or Cite
-------
1 Aquatic and terrestrial critical acid loads were compared in 16 watersheds from each of
2 the two aquatic acidification Case Study Areas, the Adirondacks and the Shenandoah, identified
3 in the REA (2009). For each Case Study Area, four watersheds were randomly selected from
4 each of the four aquatic acidification sensitivity classes reported in the REA (2009). Those four
5 sensitivity classes are "highly sensitive", "moderately sensitive", "low sensitivity", and "not
6 sensitive". In order for a watershed to be classified as one of these four classes, it had to contain
7 at least one lake or stream with that sensitivity class designation. The Adirondacks Case Study
8 Area contained watersheds representing all four sensitivity classes, and the 16 watersheds that
9 were selected for the analysis contained a total of 29 lakes. However, in the Shenandoah Case
10 Study Area, there were a limited number of watersheds in the "low" and "not sensitive" classes.
11 Therefore, only one of the 16 randomly selected watersheds contained a "low" and a "not
12 sensitive" stream. In total, there were 20 streams located in the 16 Shenandoah watersheds
13 selected for the comparative analysis. In each of the 32 watersheds (16 Adirondacks plus 16
14 Shenandoah), the terrestrial critical acid loads were calculated as a single value for the entire
15 watershed. These terrestrial critical acid loads were then compared to the aquatic critical acid
16 loads for the lakes and streams within each watershed to determine whether the aquatic or
17 terrestrial critical acid load provided greater protection against acidifying nitrogen and sulfur
18 deposition. Appendix B of this Policy Assessment document provides a full description of the
19 methods and results of this comparative analysis.
20 Results of the comparison between the aquatic critical acid load (ANC = 50 ueq/L) and
21 the terrestrial critical acid loads (Bc:Al 1.2 and 10.0) for the 32 watersheds are presented in
22 Tables 6.1 and 6.2. In the 16 Adirondack watersheds, 13 of the 29 lakes had aquatic critical
23 acid loads that were lower (more protective) than the terrestrial critical acid loads when a Be: Al
24 ratio of 10.0 was used. Based on terrestrial critical acid loads determined with a Bc:Al ratio of
25 1.2, 21 of the 29 lakes in the Adirondacks had aquatic critical acid loads lower than the terrestrial
26 critical acid loads. More importantly, for the terrestrial critical acid loads determined with a
27 Bc:Al ratio of 10.0, 13 of the 16 lakes in the Adirondacks classified as "highly" and
28 "moderately" sensitive to acidification had aquatic critical acid loads lower than the terrestrial
29 critical acid loads, and all 16 lakes in these two sensitivity classes had critical acid loads lower
30 than the terrestrial loads determined with a Be:Al of 1.2 The watersheds within the Shenandoah
31 region showed similar results (Table 6.1).
September, 2010 6-2 Draft - Do Not Quote or Cite
-------
1 Table 6-1. Results of the comparison of lake and stream aquatic critical loads (ANC of 50
2 ueq/L) to terrestrial critical loads (Bc:Al molar ratios of 10.0 in soil solution) calculated for
3 the full watershed in each of the 16 watersheds in the Adirondacks and Shenandoah Case
4 Study Areas. The tabular results show the number of times the aquatic acidification critical load
5 would provide more protection than the terrestrial acidification critical load.
Case Study Area
Adirondacks
Shenandoh
Watershed Sensitivity to Aquatic Acidification
Highly Sensitive
7 of 7
14 of 14
Moderately Sensitive
6 of 9
5 of 5
Low Sensitivity
Oof 7
Oof 1
Not Sensitive
Oof 6
Oof 1
7
8
9
10
11
Table 6-2. Results of the comparison of lake and stream aquatic critical loads (ANC of 50
ueq/L) to terrestrial critical loads (Bc:Al molar ratios of 1.2 in soil solution) calculated for
the full watershed in each of the 16 watersheds in the Adirondacks and Shenandoah Case
Study Areas. The tabular results show the number of times the aquatic acidification critical load
would provide more protection than the terrestrial acidification critical load.
Case Study Area
Adirondacks
Shenandoh
Watershed Sensitivity to Aquatic Acidification
Highly Sensitive
7 of 7
14 of 14
Moderately Sensitive
9 of 9
5 of 5
Low Sensitivity
5 of 7
Oof 1
Not Sensitive
Oof 6
Oof 1
12
13
14
15
16
17
18
19
20
In summary, terrestrial and aquatic critical acid loads were compared for watersheds in
the Adirondack and Shenandoah Case Study Areas. Results indicated that, in general, the
aquatic critical acid loads were lower and therefore offered greater protection to the watershed
than did the terrestrial critical acid loads. In situations where the terrestrial critical acid loads
were lower (i.e., more protective) than the aquatic critical acid loads, the lakes or streams in the
watershed were often rated as having "low sensitivity" or "not sensitive" to acidifying nitrogen
and sulfur deposition. Conversely, when the waterbodies were more sensitive to deposition
("highly sensitive" or "moderately sensitive"), the aquatic critical acid loads generally provided a
September, 2010
6-3
Draft - Do Not Quote or Cite
-------
1 greater level of protection against acidifying nitrogen and sulfur deposition in the watershed. It
2 is uncertain whether these results would be consistent for the rest of the country.
3 6.2 To What Extent Would A Standard Specifically Defined To Protect Against Aquatic
4 Acidification Likely Provide Protection From Terrestrial Nutrient Enrichment?
5
6 The figure below summarizes the terrestrial nutrient enrichment effects from nitrogen
7 discussed in the REA (2009). The REA reported these benchmarks as kg/ha/yr. To convert to
8 meq/m2/yr each benchmark number must be multiplied by 7.14. The range in meq/m2/yr thus
9 becomes 10.7 meq/m2/yr for the low end benchmark of changes to diatom community structure
10 to 321 meq/m2/yr at the high end benchmark associated with nitrate leaching, high foliar
11 nitrogen, and high NO emissions in southern California.
12 For each depositional load that is considered for aquatic acidification, whether it is a
13 national number or a sensitivity based number, it can be compared against the chart in figure 6.1
14 to understand the level of protection offered in individual parts of the country where these
15 studies were conducted. For example referring back to the figure under option 1 in section
16 5.3.2, the depositional load selected to encompass 90% of the critical loads on a national basis to
17 protect for an ANC of 50 is described as a tradeoff curve between sulfur and nitrogen. If sulfur
18 were zero then the maximum nitrogen deposition would be 79 meq/m2/yr or 11 kg/ha/yr.
19 Comparing this maximum nitrogen deposition number to the benchmarks in figure 6.1 shows
20 that the depositional load would provide some protection against leaching in northeast forests,
21 but would have to be lower to protect California coastal sage scrub, lichens in mixed conifer
22 forests, alpine lake communities, and Minnesota grasslands.
September, 2010 6-4 Draft - Do Not Quote or Cite
-------
5.2-
5.3
Rocky Mountain alpine lakes: shift in diatom community dominance (Baron, 2006)
• Southern California: N growth requirement threshold (Wood et al., 2006)
• San Bernardino Mountains and Sierra Nevada Mountains: acidophytic lichen
decline in MCF (Fern et al., 2008)
• Eastern Rocky Mountain Slope: low carbon:nitrogen; low lignin:nitrogen (Baron et
al.,2000)
• Eastern Rocky Mountain Slope: increased foliar nitrogen; increased mineralization
(Baronet al.,2000)
• San Bernardino Mountains and Sierra Nevada Mountains: shift from acidophytic
to neutral or nitrogen-tolerant lichen in MCF (Fenn et al., 2008)
• Minnesota grasslands: decreased plant species (Clark and Tilman, 2008)
7-12
9.8-10.2
Northeast U.S.: NO3 leaching (Aber et al., 2003)
Bay Area, CA: Increased cover of nonnative grasses; decreased native
grasses (Weiss, 1999)
San Bernardino Mountains and Sierra Nevada Mountains: loss of acidophytic
lichen in MCF (Fenn et al., 2008)
Southern California: shift in mycorrhizal species in CSS (Egerton-Warburton
and Allen, 2000)
Southern California: shift from native species to invasive grasses in CSS (Allen,
2008)*
• San Bernardino Mountains: high dissolved organic nitrogen (Meixner and
Fenn, 2004)
• San Bernardino Mountains: nitrogen saturation (Meixner and Fenn, 2004)
Increased nitrogen in lichen (Fenn et al., 2007)
• MCF: NO3 leaching (Fenn et al., 2008)
• MCF: 25% decrease in fine-root biomass (Fenn et al., 2008)
• Southern California: NO3" leaching (Fenn et al., 2003)
• Southern California: high foliar nitrogen (Bytnerowicz and
Fenn, 1996)
• Los Angeles Basin, California: High NO emissions
(Bytnerowicz and Fenn, 1996)
' Fraser Experimental Forest, CO:
increased foliar nitrogen; increased
mineralization (Rueth et al., 2003)**
10
15 20 25 30 35
Nitrogen Deposition, kg/ha/yr
40
45
" Personal communication, 2008. Abo referenced in Bobbink et. al ,2010, Ecological Applic<5tions,20(1):30-S9 and USOS FS, 2010,
http:/^vww.nrs.fa.fed.u&'cl6an_alr_wate^/dean_watar/critlcal_lQads/local-rescurcss,'dQcs/Emp[rlGal_CLS_of_N_1
"Nitrogen deposition levels include ambient and experimental additions.
2
3
Figure 6-1 Benchmarks of atmospheric nitrogen deposition for several ecosystem
indicators.
September, 2010
6-5
Draft - Do Not Quote or Cite
-------
2 6.3 To What Extent Would A Standard Specifically Defined To Protect Against Aquatic
3 Acidification Likely Provide Protection From Aquatic Nutrient Enrichment?
4
5 The REA (2009) found that deposition of reactive nitrogen contributed to eutrophication
6 of estuaries; however, it was also noted that atmospheric deposition of nitrogen is only part of
7 the total nitrogen load to the estuaries. Due to the complications of separating out the effects of
8 atmospheric deposition from the effects of other nitrogen loads, CASAC did not recommend that
9 a secondary NAAQS be set to specifically protect against estuarine eutrophi cation at this time.
10 As described in the REA (2009), the Chesapeake Bay is one national estuary that is
11 suffering from eutrophication. In issuing his Executive Order on the Chesapeake Bay, President
12 Obama recognized that the Bay watershed is one of our nation's greatest treasures and must be
13 protected and restored. To that end, EPA is proposing a nitrogen total maximum daily load
14 (TMDL) for the Chesapeake Bay. The TMDL will contain a specific air allocation for nitrogen
15 deposition. The allocations that were provided to the states included assumptions that air
16 deposition levels of nitrogen would be reduced to 14.9 million pounds per year to the tidal waters
17 and to 323 million pounds to the watershed by the year 2020. According to the Chesapeake Bay
18 Program Office, the tidal waters have a surface area of 4,479 square miles and the watershed is
19 64,216 square miles. This means that in 2020, the TMDL currently calls for nitrogen deposition
20 levels to the combined bay and watershed to be reduced to 337.9 million pounds/68,695 square
21 miles/yr, which is equivalent to 8.6 kg/ha/yr or 61 meq/m2/yr. As in Section 6.2, this number
22 can be compared to the maximum deposit!onal load of 79 meq/m2/yr. If sulfur were zero on the
23 national tradeoff curve from the figure under option 1 of section 5.3.2, then the allowed
24 depositional load of 79 meq/m2/yr would not meet the Chesapeake Bay TMDL as currently
25 envisioned.
26
27
September, 2010 6-6 Draft - Do Not Quote or Cite
-------
1 6.4 REFERENCE
2 US EPA (United States Environmental Protection Agency). 2009. Risk and Exposure Assessment
3 for Review of the Secondary National Ambient Air Quality Standards for Oxides of
4 Nitrogen and Sulfur. Final. U.S. Environmental Protection Agency, Office of Research
5 and Development, National Center for Environmental Assessment, Research Triangle
6 Park, NC. September.
7
September, 2010 6-7 Draft - Do Not Quote or Cite
-------
i 7 EVALUATION OF UNCERTAINTY AND VARIABILITY IN THE
2 CONTEXT OF AN AAPI STANDARD, INCLUDING MODEL
3 EVALUATION, SENSITIVITY ANALYSES, AND ASSESSMENT OF
4 INFORMATION GAPS
5
6 7.1 INTRODUCTION AND PURPOSE
7 This chapter provides discussions of results of analyses and assessments intended to
8 address the relative confidence associated with many of the individual and combined
9 components of the linked atmospheric-ecological effects system described in Chapter 5, as well
10 as important uncertainties in the scientific evidence that should be considered in developing
11 options for the standard. This chapter is intended to integrate a variety of analyses related to the
12 sensitivity of the models and model components to uncertainty and variability, and place the
13 results of those analyses within the context of the conclusions that can be drawn regarding the
14 components of the AAPI. These components include ecosystem effects; dose-response
15 relationships; underlying ecosystem sensitivity to acid deposition, biogeochemical, atmospheric
16 and deposition processes; and characterization of ecosystem services. While several processes
17 are imbedded in the AAPI equation introduced in Chapter 5, the level of the AAPI, as in all
18 NAAQS, is to include consideration of information on uncertainty and variability.
19 Consequently, knowledge of the relative confidence and natural variability in the structural
20 components of the AAPI are considered in staff conclusions on options for ranges of the level of
21 the standard. This chapter is not intended to be a comprehensive treatment of all uncertainties
22 that exist relative to the overall review of the standards, instead, it focused on those that are most
23 relevant in evaluating choices regarding the AAPI form of the standard and options regarding the
24 indicator, averaging time, and ranges of levels of the AAPI-based NOx and SOx standard.
25 Uncertainty and sensitivity analyses are used to inform the relative confidence in the
26 components and models that are used in defining the standard. Assessments of variability in the
27 data used to determine parameters of the standard increases the level of understanding about the
28 likelihood that alternative parameterizations of the standard will achieve targeted levels of
29 protection when applied to sensitive ecosystems across the U.S. Assessments of the sensitivity
30 of the overall AAPI to the components of the equation proposed to calculate the AAPI can help
September, 2010 7-1 Draft - Do Not Quote or Cite
-------
1 demonstrate how important uncertainty and variability in those components are in assessing the
2 protection of ecosystems provided by an AAPI standard. To evaluate the potential interactions
3 between uncertain and/or variable AAPI components, a multifactor sensitivity analysis is also
4 conducted. The ranges of component values evaluated in the multifactor sensitivity assessment
5 are guided by individual variability and uncertainty analyses of specific components. In addition
6 to informing considerations of the AAPI level, an additional objective of these "confidence"
7 related analyses and discussions is to help guide research and data collection efforts intended to
8 reduce uncertainty for future NAAQS reviews and implementation efforts. Spatial and
9 temporal variability analyses of AAPI components are especially useful to inform monitoring
10 network design, the spatial boundaries of acid sensitive regions, and averaging periods relevant
11 to NAAQS implementation.
12 Use of confidence assessment results to inform setting of the AAPI level.
13 The analyses summarized in this chapter largely address parameters related to
14 atmospheric/deposition and biogeochemical characterization, which are incorporated in the
15 AAPI. General uncertainty related discussions are provided for elements in the standard setting
16 process including dose-response relationships, effects, economic valuation metrics, as well as for
17 several specific processes (e.g., organic nitrogen) which are directly represented in the AAPI.
18 Confidence related information is not used to make decisions on inclusion or exclusion of a
19 parameter in the AAPI model. Instead, we develop the overall approach using the best available
20 information and most practical approaches, with an understanding of the level of confidence
21 associated with both individual components and the overall resulting AAPI. The overarching
22 principle in setting the level of the AAPI is consideration of the impact that associated NOx and
23 SOx levels will have on the level of protection from adverse effects on public welfare.
24 Information on uncertainty in different elements of the form of the standard and the scientific
25 evidence informing those elements is used to evaluate our confidence that the standard has a high
26 likelihood of protecting public welfare. One way of considering confidence levels is to
27 identify components of the standard for which we have low confidence and adjust the level of the
28 AAPI to ensure that, given the uncertainty in those components, we have a high likelihood of
29 achieving a target level of protection. This approach to considering confidence levels places
30 greater weight on ensuring a high likelihood of achieving protection, and places less weight on
September, 2010 7-2 Draft - Do Not Quote or Cite
-------
1 concerns about providing more that requisite protection. Conversely, low confidence in
2 components of the standard could also be considered to increase the likelihood that more than
3 requisite protection could result, and consequently, the AAPI could be adjusted downward to
4 account for lower confidence that the level of protection is requisite. To the extent that the
5 available information suggests that a particular process or parameter creates a positive or
6 negative bias (as opposed to broader uncertainty with no clear direction) would lead to
7 recommending correspondingly higher or lower AAPI level. The Clean Air Act (CAA)
8 language requires the secondary standards to be set to provide a requisite level of protection
9 against known or anticipated effects, which can be interpreted to include effects which have
10 large uncertainty but for which the expectation of adverse effects on public welfare exists with
11 reasonable confidence.
12 Significant emphasis is placed on evaluations of CMAQ due to the unique role that
13 atmospheric models hold in the linked AAPI system. The AAPI as currently formulated relies
14 on CMAQ for both the initial characterization of reduced nitrogen deposition, and the deposition
15 transformation ratios (TNOX and TSOX) which characterize the relationships between atmospheric
16 concentrations of NOx and SOx and deposition of N and S. Included are interpretations of
17 model evaluation results from the REA (EPA, 2009) as well as more recent results related to wet
18 deposition and the treatment of ammonia deposition. Comparison of model results to
19 observations provides a general sense of the confidence we have that the models capture the
20 spatial, temporal and compositional texture of the relevant atmospheric and deposition species
21 that drive the linked atmospheric-ecosystem processes. Both model evaluation results and
22 assessments of spatial and temporal variability guide implementation strategies for monitoring
23 network design and emission inventory improvement. Sensitivity of CMAQ derived deposition
24 transformation ratios to changes in emissions, and treatment of chemistry [not yet completed]
25 and variability over time provide insight into the stability of these parameters that are used in a
26 relatively static manner in the AAPI, and into how well proposed averaging times capture the
27 overall spatial and temporal trends in the parameters.
28 We evaluate the sensitivity of critical load modeling components by comparing dynamic
29 (MAGIC) and hybrid steady state model results, looking at terminal results of MAGIC. This
30 approach was viewed as a test of the more reduced form approximations used in steady state
September, 2010 7-3 Draft - Do Not Quote or Cite
-------
1 modeling relative to more sophisticated treatment in MAGIC. The MAGIC critical load
2 simulations also provide information on the temporal trajectory of ANC, including the expected
3 time necessary to reach a desired ANC, which can help inform the level of the AAPI,
4 recognizing that there may be additional consideration given to reaching a target ANC within a
5 specific timeframe, e.g. by 2030 or 2040. [[not yet completed]
6 For the purposes of this discussion, we characterize uncertainty regarding models and
7 their outputs as referring to the lack of knowledge regarding both the actual values of model
8 input variables (parameter uncertainty) and the model characterization of physical systems or
9 relationships (model uncertainly). In any application, uncertainty is, ideally, reduced to the
10 maximum extent possible, but significant uncertainty often remains. It can be reduced by
11 improved measurement and improved model formulation. Model evaluation results provide
12 some insight into the relative uncertainty associated with the ability of models to capture key
13 environmental state characteristics. Confidence regarding the fundamental science supporting
14 causal determinations about the effects of acid deposition, and the translation of those effects
15 into ecosystem services and values is less amenable to quantification. As a result, these
16 uncertainties are more difficult to explicity account for in development of the standards. In the
17 case of the equation describing the AAPI, while the degree of uncertainty in some elements can
18 be characterized, sometimes quantitatively, a formal uncertainty analysis using statistical
19 sampling techniques (e.g., Monte Carlo simulation) to identify the relative and combined
20 influences of parameter uncertainty was not performed. However, we did evaluate the sensitivity
21 of the AAPI to its components using two related assessments: analysis of variance (ANOVA)
22 and elasticity calculations. The results of these assessments are addressed in section 7.5.
23
24 Sensitivity refers to the influence on modeled results due to perturbations in input
25 variables or change of process formulations. Sensitivity analysis can provide a sense of how
26 important different parameters and inputs might be to the outcomes of interest, e.g. the AAPI
27 level, but cannot by themselves indicate how important specific parameters actually are, because
28 they do not incorporate information on the range of parameter values or the likelihood associated
29 with any specific parameter value. Sensitivity results in this PAD are intended to provide
30 insight into the relative stability of the AAPI and associated NOx and SOx tradeoff curves and
31 confidence in modeled parameterizations. Sensitivity analyses are especially useful in the
September, 2010 7-4 Draft - Do Not Quote or Cite
-------
1 absence of observed data to challenge models. For example, the NOy and SOx transference
2 ratios are a model construct that is difficult, if not impossible, to compare to observations. The
3 sensitivity of these ratios to changing meteorology, emissions and chemical mechanism
4 treatments is evaluated in reference to the stability of these ratios under changing conditions.
5 Low sensitivity here implies that the choice to use long-term averages of modeled ratios is
6 justified. Sensitivity analyses also are used to discern the relative influence (on AAPI results) of
7 AAPI parameters. Toward that end, ANOVA and elasticity analyses were applied to determine
8 the relative sensitivity of AAPI results associated with individual and combined AAPI
9 parameters.
10 Variability refers to the heterogeneity in a population or variable of interest that is
11 inherent and cannot be reduced through further data collection and research. In the context of
12 the AAPI and trade-off curves, variability is considered in guiding the design of monitoring and
13 modeling analyses supporting implementation activities.
14
15 7.2 Uncertainty associated with ecosystem effects and dose - response relationships.
16 Chapter 2 provides a brief summary of uncertainties based on the REA and is reproduced
17 here to centralize all uncertainty discussions. There are different levels of uncertainty associated
18 with relationships between deposition, ecological effects and ecological indicators. In Chapter 7
19 of the REA, the case study analyses associated with each targeted effect area were synthesized
20 by identifying the strengths, limitations, and uncertainties associated with the available data,
21 modeling approach, and relationship between the ANC and atmospheric deposition. The key
22 uncertainties were characterized as follows to evaluate the strength of the scientific basis for
23 setting a national standard to protect against a given effect (REA 7.0):
24 • Data Availability: high, medium or low quality. This criterion is based on the
25 availability and robustness of data sets, monitoring networks, availability of data that
26 allows for extrapolation to larger assessment areas, and input parameters for modeling
27 and developing the ecological effect function. The scientific basis for the ecological
28 indicator selected is also incorporated into this criterion.
29 • Modeling Approach: high, fairly high, intermediate, or low confidence. This value is
30 based on the strengths and limitations of the models used in the analysis and how
31 accepted they are by the scientific community for their application in this analysis.
September, 2010 7-5 Draft - Do Not Quote or Cite
-------
1 • Ecological Effect Function: high, fairly high, intermediate, or low confidence. This
2 ranking is based on how well the ecological effect function describes the relationship
3 between atmospheric deposition and the ecological indicator of an effect.
4
5 The REA concludes that the available data are robust and considered high quality. There is
6 high confidence about the use of these data and their value for extrapolating to a larger regional
7 population of lakes. The EPA TIME/LTM network represents a source of long-term,
8 representative sampling. Data on sulfate concentrations, nitrate concentrations and ANC from
9 1990 to 2006 used for this analysis as well as EPA EMAP and REMAP surveys, provide
10 considerable data on surface water trends.
11 There is fairly high confidence associated with modeling and input parameters. Uncertainties
12 in water quality estimates (.i.e., ANC) from MAGIC was derived from multiple site calibrations.
13 The 95% confidence interval for pre-acidification of lakes was an average of 15 ueq/L difference
14 in ANC concentrations or 10% and 8 ueq/L or 5% for streams (REA 7.1.2). The use of the
15 critical load model used to estimate aquatic critical loads is limited by the uncertainties
16 associated with runoff and surface water measurements and in estimating the catchment supply
17 of base cations from the weathering of bedrock and soils (McNulty et al., 2007). To propagate
18 uncertainty in the model parameters, Monte Carlo methods were employed to develop an inverse
19 function of exceedences. There is high confidence associated with the ecological effect function
20 developed for aquatic acidification. In calculating the ANC function, the depositional load for N
21 or S is fixed by the deposition of the other, so deposition for either will never be zero (Figure
22 7.1-6 REA)
23 Chapter 2 also reviews the basic evidence underlying effects on fish mortality, aquatic species
24 diversity and more extended food web disruptions leading to adverse impacts on birds associated
25 with aquatic acidification. There is high confidence associated with causality between
26 acidification and these ecological effects. Also, there is extremely high confidence in the
27 relationship between the ecological indicator, ANC, and the more direct chemical properties
28 (lower pH and increased Al) associated with acidification.
29
30
September, 2010 7-6 Draft - Do Not Quote or Cite
-------
1 7.3 Uncertainty in benefits estimates
2
3 Descriptions of the current provision of ecosystem services presented for each of the
4 effect areas analyzed for this review followed by estimations of the damages incurred to selected
5 services due to nitrogen and sulfur deposition. The current services are presented to give the
6 reader a sense of the magnitude of the benefit the public receives from these ecosystems under
7 current conditions. The data used in these descriptive passages is generally derived from
8 government (either federal or state) sources we are reasonably certain to be of the highest
9 quality. Where monetary values are placed on the these services we have generally used widely
10 cited studies, particularly meta analyses that provide an average value that smoothes the variation
11 in WTP estimates. These estimates underestimate the total value of these services as they use
12 benefit estimates for a marginal increase in these services. It is likely that the total benefits of
13 these services are greater because their marginal value likely is lower than the average value.
14 While reductions in sulfur and nitrogen emissions would increase the size of the benefits from
15 these services, for many of them it is unknown how significant the increase will be.
16 The analyses of damages incurred are more uncertain and are limited to those areas where
17 data and tools were available. Only some services were analyzed which in some cases meant
18 that the results were limited to one or two services and in the case of terrestrial nutrient
19 enrichment no services had sufficient data available to attempt an estimate of damage. This
20 means that the estimates presented are a very small part of the total damage incurred due to
21 deposition.
22 Aquatic Acidification
23 Recreational Fishing Model.
24 The analysis of recreational fishing damages presented in Chapter 3 is subject to the
25 assumptions necessary to perform the analysis. The original analysis performed for the REA
26 was based on projecting future benefits of increased recreational fishing based on a complete
27 cessation of all nitrogen and sulfur emissions. These decisions under or over estimate the current
28 damages to public welfare incurred from nitrogen and sulfur deposition. The magnitude of the
29 bias in results is unknown in either direction however the majority of the assumptions influence
September, 2010 7-7 Draft - Do Not Quote or Cite
-------
1 the estimates downward. These include the use of emissions estimates that include projected
2 decreases due to implementation of Title IV regulations in 2020. These emissions estimates are
3 lower than current emissions and therefore lead to underestimation of damages. Because the
4 models only value this improvement for New York residents (without accounting for out-of-state
5 visitors) the damages are underestimates of the benefits of these improvements in the
6 Adirondacks region.
7 The use of projected population in the REA analyses contributes to an overestimate of
8 current damages since current population is smaller than future population. Further, these
9 estimates are extrapolated from a 44 lake subset and applied to all Adirondack lakes. The
10 representativeness of this sample is unknown. This analysis also does not account for any
11 change in fishing demand (possible overestimate) and income (possible underestimate).
12 Banzhaf, et al Benefits Transfer.
13 The approach using the WTP estimates from the Banzhaf study is subject to the same
14 uncertainties described above and some additional considerations. Specifically there is some
15 uncertainty regarding which types of ecosystem services are reflected in the study's estimates of
16 the improvements in ecosystem services of reducing acidification, particularly provisioning and
17 regulating services. The values likely include recreational fishing services, which means they
18 cannot be added to the recreational fishing model results, and other cultural services including
19 other recreation and nonuse services. The inclusion in the survey of other ecosystem changes
20 (birds, trees, etc.) leads to an overestimation of WTP for remediation of lake acidification alone.
21 Finally, assumptions were required to align the Banzhaf survey scenarios to the likely results of
22 complete removal of all nitrogen and sulfur emissions. These are reasonably close but not exact
23 and may not be applicable to another baseline.
24 Conclusion.
25 While these estimates are subject to uncertainty we are reasonably confident that they
26 represent a good first-order approximation of the damages to recreational fishing due to nitrogen
27 and sulfur deposition. Additionally it should be noted that the Banzhaf survey results represent a
28 broader picture (though by no means complete) of the damages to ecosystem services in the
29 Adirondacks. Finally, we would again like to emphasize that these estimates represent only a
September, 2010 7-8 Draft - Do Not Quote or Cite
-------
1 small sample of the damages incurred to a broad range of ecosystem services affected and the
2 areas of the nation where acidic deposition is an ongoing issue.
3 7.4 CMAQ Application and Evaluation
4 7.4.1 Overview of CMAQ model application
5 The CMAQ model is a comprehensive, peer-reviewed (Aiyyer et al., 2007), three-
6 dimensional grid-based Eulerian air quality model designed to simulate the formation and fate of
7 gaseous and particle (i.e., particulate matter or PM) species, including ozone, oxidant precursors,
8 and primary and secondary PM concentrations and deposition over urban, regional, and larger
9 spatial scales (Dennis et al., 1996; U.S. EPA, 1999; Bryun and Schere, 2006). CMAQ is run for
10 user-defined input sets of meteorological conditions and emissions. For this analysis, we are
11 using predictions from several existing CMAQ runs. These runs include annual simulations for
12 2002 using CMAQv4.6 and annual simulations for each of the years 2002 through 2005 using
13 CMAQv4.7 (Foley et al., 2010). CMAQv4.6 was released by the U.S. Environmental Protection
14 Agency's (EPA's) Office of Research and Development (ORD) in October 2007. CMAQv4.7
15 along with an updated version of CMAQ's meteorological preprocessor (MCIPv3.4, Otte and
16 Pleim, 2010)1 were released in October 20082. The 2002 simulation with CMAQv4.6 was
17 performed for both the Eastern and Western domains. The horizontal spatial resolution of the
18 CMAQ grid cells in these domains is 12 x 12 km. The 2002 through 2005 simulations with
19 CMAQv4.7 were performed for the eastern 12-km domain and for the continental United States
20 domain, which has a grid resolution of 36 x 36 km. The CMAQv4.6 and v4.7 annual simulations
21 feature year-specific meteorology, as well as year-specific emissions inventories for key source
22 sectors, such as utilities, on-road vehicles, nonroad vehicles, wild fires, and natural biogenic
23 sources. Emissions for other sectors of the inventory for each of the years modeled rely on
24 inventories for 2002. Details on the development of emissions, meteorology, and other inputs to
25 the 2002 CMAQv4.6 runs can be found in a separate report (U.S. EPA, 2008). Inputs for the
26 CMAQv4.7 runs for 2002 through 2005 were derived using procedures similar to those for the
27 CMAQv4.6 2002 runs.
1 The scientific updates in CMAQ v4.7 and MCIP v3.4 can be found at the following web links:
http://www.cmascenter.org/help/model_docs/cmaq/4.77RELEASE_NOTES.txt
http://www.cmascenter.org/help/model_docs/mcip/3.4/ReleaseNotes
2 The differences in nitrogen and sulfur deposition in the case study areas between CMAQ v4.6 and v4.7 for 2002
are small, as described in Chapter 3.
September, 2010 7-9 Draft - Do Not Quote or Cite
-------
1 Additional details of the modeling domain, emissions and meteorological inputs are
2 provided in EPA (2009; REA Appendices).
3
4 7.4.2 CMAQ Evaluation, Sensitivity and Variability Analyses
5 Past results. A variety of comparisons of modeled estimates to observations were
6 included in the REA (EPA, 2009), and some of the highlights are summarized here in addition to
7 new work on ammonia characterization and wet deposition. Readers are encouraged to review
8 the earlier report. Ambient air concentrations and wet deposition observations are paired against
9 modeled estimates. In contrast, dry deposition is always a modeled value, either derived from
10 ambient or modeled ambient concentrations. Given the interest in relevant nitrogen and sulfur
11 species, CASTNET observations were used extensively. Comparisons of modeled annual
12 average total nitrate (sum of nitric acid and particulate nitrate), ammonium, sulfate, and sulfur
13 dioxide to observations for the 2002 base year are provided in Figures 7-1 through 7-4.
14 Normalized mean bias statistics for 2002-2005 base years are provided in Table 7-1.
15 CMAQ overpredicts 862 and underpredicts 864. Although model performance is good
16 for total SOX, the inclusion of co-located 862 and sulfate measurements required for future
17 secondary NOX/SOX NAAQS comparisons will help diagnose issues with the model's ability to
18 partition these two species. CMAQ generally overpredicts total nitrate and slightly
19 underpredicts ammonium and the model captures the monthly temporal patterns of sulfate, total
20 nitrate and ammonium when all sites are aggregated (Figures 7-5 to 7-7). There are some basic
21 incommensurabilities between model estimates and observations that complicate interpretation
22 of model to observation comparisons, most notably the representation of space as a model
23 represents a volume average of roughly 144 km2, which depends on the time varying vertical
24 depth of the lowest modeled layer. Most surface based observations rely on point sampling and
25 the extent to which a point is representative of broader volume space varies with meteorology,
26 distribution of emissions and surface characteristics.
27
September, 2010 7-10 Draft - Do Not Quote or Cite
-------
1 Table 7-1. Normalized Mean Bias Statistics for Predicted and Observed Pollutant
2 Concentration
Pollutant
Concentrations
SO2
SO42"
TNO3
NH4+
2002
45%
-13%
22%
4%
2003
39%
-9%
26%
11%
2004
47%
-13%
22%
7%
2005
41%
-17%
24%
2%
3 Ammonia.
4 Characterizing ammonia deposition is of increasing importance for assessing ecosystem
5 responses to nitrogen deposition. Nitrogen deposition includes both deposition of NOy and
6 deposition of reduced nitrogen (ammonia + ammonium). Ecosystem effects are due to total
7 nitrogenand, as such, the recommended AAPI form of the standard directly includes reduced
8 nitrogen to account for its role in using up an ecosystems ability to absorb nitrogen. The
9 proposed form of the standard uses CMAQ modeled deposition of reduced nitrogen.
10 Because of a shortage of routinely available ammonia observations, model evaluation
11 studies have relied on ammonium measurements from CASTNET as the only routine source of
12 reduced nitrogen observations. Clearly, the lack of ammonia observations must be addressed in
13 future implementation scenarios as there is relatively greater uncertainty in characterizing
14 ammonia relative to SOX and NOy. Recently, Dennis et al. (2011) explored the sensitivity of
15 CMAQ (addressed below) to varying treatments affecting ammonia deposition velocity. As part
16 of that study, CMAQ ammonia estimates were compared with ammonia observations at two sites
17 in North Carolina (Figure 7-8).
18 Wet deposition.
19 Modeled wet deposition in CMAQ is a function of the volume of predicted precipitation
20 within a grid cell and the pollutant concentrations scavenged from the atmosphere during
21 precipitation events. As a result, errors in modeled precipitation and in emission inputs can lead
22 to significant bias and error in the wet deposition predictions compared to observed values. EPA
September, 2010
7-11
Draft - Do Not Quote or Cite
-------
1 (Dennis and Foley, 2010) has corrected CMAQ wet deposition predictions by scaling the model
2 output based on observation-based gridded precipitation data generated by the Parameter-
3 elevation Regressions on Independent Slopes Model (PRISM, 2004). The precipitation adjusted
4 deposition fields are more highly correlated with observed values for all wet deposited nitrogen
5 and sulfur species compared to the base model output (Figures 7-9). In addition, the adjusted
6 fields are better able to capture the spatial heterogeneity of accumulated wet deposition due to
7 orographic effects on precipitation amounts.
8 Adjusting the wet deposition values to account for over-predictions in the model
9 precipitation inputs revealed compensating errors for nitrate and ammonium. The negative bias
10 seen in these species after the precipitation adjustment is believed to be due to missing emissions
11 sources. A second bias adjustment was performed for nitrate and ammonium based on observed
12 levels at the NADP/NTN sites (Figure 7-10). The final adjusted spatial fields of annual total wet
13 deposition values are more consistent with observed wet deposition values. Ongoing studies
14 suggest that much of this bias can be reduced in the Eastern half of the US by including nitrogen
15 oxide produced by lightning and accounting for the bi-directional flux of ammonia. Once these
16 model improvements are incorporated in CMAQ a second bias adjustment may not be needed in
17 the East.
18 7.4.3 Variability and sensitivity of CMAQ generated components.
19 7.4.3.1 Ambient Concentration to Deposition Transformation ratios.
20 The deposition transference ratios^ introduced in Chapter 5 are referenced as TSox and
21 TNOY, to distinguish these parameters from an exact linkage to deposition velocity, which is
22 uniquely associated with individual atmospheric species. Deposition transference ratios are
23 defined as the annual wet and dry deposition of all oxidized species (NOy for TNOY, SO2 plus SO4
24 for TSOX) divided by the average annual concentration of NOy, for TNOy, or SO4 plus SO2, for
25 TSOX. The units for TNoy and TSox are distance/time. Deposition transference ratios provide a
26 mechanism to associate ambient concentrations to deposition loads and to determine if an area's
3 In the first draft of the Policy Assessment, the deposition transformation ratios were labeled VNOx and VSOx. For
this draft, based on recommendations from CASAC, we have renamed these ratios TNOx and TSOx.
September, 2010 7-12 Draft - Do Not Quote or Cite
-------
1 air concentrations of NOy and SOX meet a NAAQS level using the AAPI form. A deposition
2 transformation ratio is an aggregate representation of the deposition process generated through
3 modeling which does not lend itself to a traditional analysis relating observations and
4 predictions. Furthermore, there is an implicit assumption that the response of deposition
5 transformation ratios to changes in meteorology and emissions is relatively stiff, as these ratios
6 are an attribute of the system that channels ambient air response associated with decreases in
7 emissions of NOX and SOX to changes in deposition. The stiffness of the deposition transference
8 ratios would suggest that the relationship between ambient concentrations and deposition is
9 strictly a constant proportion, not impacted by the mixture and level of emissions or by changes
10 in meteorology. To better understand the implications of this assumption, we investigated the
11 relative variability of the modeled deposition transformation ratios across time and space, and the
12 stability of the ratios relative to emissions and meteorological inputs was conducted to guide
13 EPA in determining how uncertainties in this parameter may eventually impact AAPI related
14 calculations.
15 7.4.3.2 Spatial and Interannual Variation of Ts and TN.
16 Generally small spatial and inter-annual variability exist in the deposition transformation
17 ratios for the 2002 -2005 model years (Figure 7-11). The inter-annual variability, calculated at
18 the grid cell level, as measured by the median coefficient of variation is around 10% and the
19 absolute values of the ratios remain stable (Figure 7-12), suggesting that year to year changes in
20 meteorology have minimal impact on the ratios. Spatial homogeneity of deposition
21 transformation ratios within the two acid sensitive areas we evaluated in the REA (Adirondacks
22 and Shenendoah) (Figure 7-11) is consistent with a relatively homogeneous ambient
23 concentration environment overlaid upon a landscape of similar vegetation and surface
24 conditions. Such spatial homogeneity within case study areas provides confidence that an area
25 wide application AAPI will not be strongly dependent on the exact boundaries chosen to define
26 an acid sensitive area.
27
28
29
September, 2010 7-13 Draft - Do Not Quote or Cite
-------
1 7.4.3.3 TSOX and TNoy Sensitivity to emission changes.
2 The response of TSox and TNOy to emission changes was explored by analyzing available
3 base case 2005 and 2030 CMAQ simulations. The 2030 case reflected expected changes in
4 emissions associated with simulated implementation of a variety of national rules and represents
5 Eastern U.S. domain wide NOx and SOx emission reductions of 48% and 40%, respectively.
6 Median changes in deposition transference ratios tended to be around zero (figure 7-12), with the
7 Adirondack region exhibiting slightly higher response than the Shenandoah region and
8 remainder of the Eastern U.S. domain.
9 7.4.3.4 TSOX and TNOy sensitivity to different chemical mechanisms.
10 [Intended to be added in the final PA]
September, 2010 7-14 Draft - Do Not Quote or Cite
-------
1 7.4.3.5 Ammonia sensitivity.
2 The role of NHX deposition is incorporated in the AAPI expression as a parameter that
3 influences the level of allowable concentrations of NOy and SOX, due to its role as part of the
4 total reactive nitrogen budget which affects acidification.. Characterizing ammonia deposition
5 is challenging due to the variety of surface and vegetation types that influence ammonia dry
6 deposition velocities as well the potential for bi-directional flux of ammonia. In addition,
7 ammonia emission estimates remain relatively more uncertain than emissions of NOX and 862
8 given the complexity of meteorology and agricultural practices that influence the spatial and
9 temporal patterns of ammonia releases. An exploration of the sensitivity of ammonia to three
10 different treatments of deposition processes in CMAQ was performed by EPA (Dennis et al.,
11 2010) to test the inclusion of a bi-directional NHa flux algorithm and elucidate the relative
12 importance associated with advection, deposition and chemical transformation on ammonia
13 patterns. These treatments included a (1) base case of current CMAQ treatment using existing
14 ammonia deposition velocity schemes and uni-directional deposition, (2) modified the base case
15 by replacing ammonia deposition velocity calculations with 862 deposition velocities (862
16 interacts with surfaces and vegetation similarly to NH3, but with reduced velocity) as a lower
17 bound and (3) introducing a bi-directional flux algorithm to the base case (retaining NH3
18 deposition velocities). Based on modeled process analysis that delineates the effects of
19 deposition, chemical transformation and advection (horizontal and vertical) on emitted ammonia,
20 the results (Figure 7-13) suggest that ammonia patterns, especially when a bi-directional flux
21 process is incorporated, are more indicative of a transported pollutant where emissions influence
22 can span hundreds of kilometers, markedly different from some earlier perspectives where
23 ammonia often was thought of as near source phenomenon due to high deposition velocities.
24 The process analysis illustrates the importance of vertical advection which enables the movement
25 of ammonia into traditional mesoscale flow patterns. The effect is enhanced by the
26 reintroduction of deposited ammonia through bi-directional flux into the ambient environment.
27 From a monitoring perspective, a design that addresses the regional characterization of
28 NOy and SOX would be consistent with characterizing NHX. Not only would ammonia and
29 ammonium measurements be useful for estimating dry deposition through deposition modeling
30 approaches such as those used in CASTNET, but they would serve as important diagnostic data
September, 2010 7-15 Draft - Do Not Quote or Cite
-------
1 to continually assess the effectiveness of NHX deposition processes in models like CMAQ. This
2 is especially important as we recognize a large uncertainty in the bi-directional formulation
3 associated with the estimation of F, the emissions potential due to the existence of compensation
4 points. Nonetheless, we can learn much about the NHa budget in spite of these uncertainties.
5 High priority research is ongoing to improve the bi-directional parameterization and the
6 estimates of the leaf and soil gammas across different cropping regions and throughout the year.
7 We are developing a software tool to estimate the soil F associated with fertilizer application.
8 When we have a spatially and temporally varying Fg, we will investigate the emissions budgets
9 for fertilized fields, as well as reexamine the animal operation emission budgets, as this will be
10 of interest. Work to examine the seasonality of single cell budgets and their range of influence is
11 continuing. Current and future CMAQ applications to ecosystem deposition will incorporate bi-
12 directional flux treatment of ammonia.
13 7.4.3.6 CMAQ uncertainties and the AAPI
14 The AAPI relies on CMAQ for the sulfur and nitrogen transference ratios and NHX
15 deposition. The model evaluation results, including the ammonia and wet precipitation
16 treatments, reflect a continual process of model improvement designed to ingest the latest
17 science within a framework links a myriad of atmospheric and surface processes across multiple
18 pollutant species. While this document focuses in on the more direct processes affecting N and
19 S deposition, these modifications are incorporated with the philosophy that the best science is
20 being adopted and they in turn support the overall improvement of the models' treatment of all
21 processes. The inclusion of better chemistry and physics of a particular process acts as an
22 internal diagnostic tool for other processes that are linked throughout the model framework
23 through basic conservation of mass principles. With respect to the AAPI, the CMAQ model
24 must be relied on to provide the spatial flexibility attendant with a national standard. As the
25 model continually adopts the best science, confidence in relevant CMAQ generated AAPI
26 parameters is raised for both near term and future scenarios.
September, 2010 7-16 Draft - Do Not Quote or Cite
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
I
0
2002ac met2v33 12kmESO2 for 20020101 to 20021231
o .
CD -
O -I
n CASTNet(2002ac_met2v33_12kmE)
D
(ug/m3)
IA
RMSE =
RMSEs =
RMSEu =
MB
ME
MdnB =
MdnE =
NMB =
NME =
NMdnB =
NMdnE =
FB =
FE =
Period Average
SO2 ( ug/m3)
0
2
4
6
8
Observation
10
12
14
Figure 7-1 2002 CMAQv4.6 annual average SOi predicted concentrations versus
observations at CASTNet sites in the eastern domain (note, units are in
actual mass for SOi, including oxygen)
September, 2010
7-17
Draft - Do Not Quote or Cite
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2002ac met2v33 12kmE SO4 for 20020101 to 20021231
O
O
n CASTNet 2002ac_met2v33_12kmE
IA
RMSE =
RMSEs =
RMSEu =
MB
ME
NMB =
NME =
NMdnB =
NMdnE =
FB =
FE =
MdnB = -0.41
Period Average
SO4 ( ug/m3)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Observation
Figure 7-2
\2-
2002 CMAQv4.6 annual average SO "4 predicted concentrations versus
observations at CASTNet sites in the eastern domain (note, units are in
actual mass for SC>4, including oxygen).
September, 2010
7-18
Draft - Do Not Quote or Cite
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
2002ac met2v33 12kmE TNO3 for 20020101 to 20021231
o
q
CD
in
q
uri
in
in
co
q
co
m
c\i
q
c\i
LO
m
ci
n CASTNet (2002ac_met2v33_12kmE)
(ug/m3)
IA
RMSE =
RMSEs =
RMSEu =
MB
ME
MdnB =
MdnE =
0.95
0.58
0.31
0.49
0.30
0.45
0.20
0.35
NMB
NME
NMdnB =
NMdnE =
FB
FE
Period Average
TNO3 ( ug/m3)
o i i i i i i i i i i i i r
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Observation
Figure 7-3 2002 CMAQv4.6 annual average TNOs predicted concentrations versus
observations at CASTNet sites in the eastern domain (note, units are in
actual mass for NO3, including oxygen).
September, 2010
7-19
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
2002ac met2v33 12kmE NH4 for 20020101 to 20021231
un
c\i
n CASTNet (2002ac_met2v33_12kmE)
(ug/m3)
IA
RMSE =
RMSEs =
RMSEu =
MB
ME
MdnB =
MdnE =
NMB
NME
NMdnB =
NMdnE =
FB
FE
Period Average
NH4 ( ug/m3
0.0
0.5
1.0 1.5
Observation
2.0
2.5
Figure 7-4. 2002 CMAQv4.6 annual average NH4+ predicted concentrations versus
observations at CASTNet sites in the eastern domain (note, units are in
actual mass for NH4, including hydrogen).
September, 2010
7-20
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
CDC_PHASE
7.0 -
6.5 -
6.0 -
5.5 -
5.0 -
^ 4.5 -
co
ro 4-° -
^ 3-5 -
D
3.0 -
2.5 -
2.0 -
1.5 -
1.0 -
0.5 -
=S= CASTNET
-*- CMAQ
A
/\
\
\
\
A
A \
x\
/ ?
I /
x
1
A4
//"
AX
/ /
AA X
X
X
RUNS CASTNET SO4 for 20020101 to 20051231; State: All; Site: All
1
\\ A
A \
$jt
x \\
V
\
s.
t
2002
i i i i i i i i i
13579
A
fl
A
Inl
1
\
/ 1
4
/t \
ix k
/\ // 1
/ \ J \
//X> \
l \
I \
1 4\
x xV
/ x^
\
\
1
2003
1 1 1 1 1 1 1 1 1 1 1
A"X\
/ / \ '
// \A
/x \ \
A/ *\
4 \
1 , \\
// fe
// f\
X \\
H \
L, ^, v \\
/xxx L
/ \ /
f U
I
2004
1 1 1 1 1 1 1 1 1 1 1
AA
AX
A'
— 1
1 / \
/ / X
x
|X
~i
K
//
// ^
/ ' \
/ \X V
/ X X
V \ A
/ »
/ "x
' X
2005
1 1 1 1 1 1 1 11
11 1 3579 11 1 3579 11 1 3579 11
Months
Figure 7-5. 2002-2005 Domain-wide average SC>42 predicted concentrations and
observations by month at CASTNet Sites in the eastern domain(note, units
are in actual mass for SC>4, including oxygen) .
September, 2010
7-21
Draft - Do Not Quote or Cite
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
CDC_PHASE_RUNS CASTNET TNO3 for 20020101 to 20051231 ; State: All; Site: All
5.5 -
5.0 -
4.5 -
4.0 -
E
D) 3 5 -
«•**
CO
D
Z 3.0 -
2.5 -
2.0 -
1.5 -
-A- CASTNET
-*- CMAQ
x
x x A
V \
Ax | /
, \ xx,
\ / r<
\ \
XAx / /
W I
V A /
A \ A /
XA
2002
X
1
1 1
/
/ X
XA |\
//I \
/A U \
/A I1 I
X
11
1
1
' r \ 1
\\ \ I
\\ 1 U
^1 X x/
W ?
V A \l A
* X /
A i-i
2003
i i i i i i i i i i i i i i i i i i i i i i i i
1 3579 11 1 3579 11 1
X
/\
/ \
' / \
1 / X
\ v -
\ , '
A\ / i
\ 1
\ 1 1
r\x-xX X
A f\
'A /
'A
2004
X
X
// A
tj
I
1
( \
I I 1 1 1 1 1 1 I I
3 5 7 9 11 1 3
X
/
XX
il
A
1
t 1 I
\
\ '
A \ X
'A\ X
V /A
AX y
\A /\
AA
2005
1 1 I I I 1 I I I
5 7 9 11
Months
Figure 7-6. 2002-2005 Domain-wide average TNOs predicted concentrations and
observations by month at CASTNet sites in the eastern domain (note, units
are in actual mass for NOs, including oxygen).
September, 2010
7-22
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
CDC_PHASE_RUNS CASTNET NH4 for 20020101 to 20051231; State: All; Site: All
CO
E
2.0 -
1.8 -
1.6 -
1.4 -
1.2 -
1.0 -
0.8 -
0.6 -
CASTNET
CMAQ
: i i ; i ' ' i ! i . i > • i i i : i i i
1 3579 11 1 3579 11 1 3579 11 1 3579 11
Months
Figure 7-7 2002-2005 Domain-wide average NH4+ predicted concentrations and
observations by month at CASTNet sites in the eastern domain (note, units
are in actual mass for NH4, including hydrogen).
September, 2010
7-23
Draft - Do Not Quote or Cite
-------
1
2
3
4
Kenansville Ammonia July 2004
12-hour Averages: Sam-Spin
184 186 188 190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
-Observations
Millbrook (Raleigh) Ammonia July 2004
12-hour Averages: 6am-6pm
183.5 185.5 187.5 189.5 191.5 193.5 195.5 197.5 199.5 201.5 203.5 205.5 207.5 209.5 211.5 213.5
Julian Day: TickMark at Midnight (July 2004)
-Observations
6
7
8
9
10
11
Kenansville Ammonia August 2004
12 hour averages: Sam-Spin
214 216 218 220 222 224 226 228 230 232 234 236 238 240 242 244
Julian Day: TickMark at Midnight (August 2004)
-Observations
Millbrook (Raleigh) Ammonia August 2004
12 hour averages: 6am-6pm
214 216 218 220 222 224 226 228 230 232 234 236 238 240 242 244
Julian Day: TickMark at Midnight (August 2004)
Figure 7-8. Comparison of CMAQ predictions and measurements for 12-hour (6am-
6pm) average NHs concentrations, with a monitoring cycle of 4 days on and
4days off, at a high emission site (Kenansville) and a low emission urban site
(Raleigh) in North Carolina compared to CMAQ for July 2004 (top) and
August 2004 (bottom), from Dennis et al., 2010 (note, units are in actual mass
for NHa, including hydrogen).
September, 2010
7-24
Draft - Do Not Quote or Cite
-------
! -
s '
i
MB = 2.4 kgfha
NMB = 20%
RMSEa =2.9 kg/ha
RMSEu = 4.2 kg/ha
RMSE = 5.1 kg/ha
NT.IL - .
K
„••
observed SO4 wel deposition (kg'hoj
I
MB = 0.7 kg'ha
NMB = 6 %
RMSE = 29 kg.'ha
NME = 18%
RMSEs = 0.8 kg/ha
RMSEu = 2.B kg/ha | 51 V« decrease )
'• *.*
.•'.<
.<^
observed SO4 wet deposition Ikg'ha)
2 Figure 7-9 Unadjusted (left) and PRISM (right) adjusted CMAQ annual wet deposited
3 sulfate for 2002 (note, units are in actual mass for SC>4, including oxygen).
5
September, 2010
7-25
Draft - Do Not Quote or Cite
-------
MB = -0.6 kgma
NMB = -6 1i
RMSEs = 0.7 kg/ha
RMSEu = 2.E kg'tia
RMSE= !.
HME = 21 %
MB = 0.1 kg 'I-.!
MMB = I %
RMSE = 2.4 kgfhl
RMSEa = 0.1 kg/ha
RMSEu=2.4kgfia (17* decrease)
•*."*••
>*
~
Stwilll
observed N03 wet deposlicn (
wei deposition (kg.'ha
2
3
4
5
MB = -0.1 kg/ha RMEE = O.S Kg(ha
RMSEs = 0.2 kg/hs
RMSEu = 0.8 MM
•*. * 1-
1 . . 1
2 4 6 S
obser/ed NK4 vetdapostion (kg/ha)
MB = 0.1 kg •!•," RUSE = 0.7 kglh*
RHSEs = 0.2 fcgfha
RMSEu = 0.7 hg/ha ( 17 % decrease
2 4 6
observed NH4 wet deposition (
Figure 7-10 Unadjusted (left) and PRISM and bias (right) adjusted CMAQ annual wet
deposition of nitrate (top) and ammonium (bottom) (note, units are in actual
mass for NH4 and NH3, including hydrogen).
September, 2010
7-26
Draft - Do Not Quote or Cite
-------
Canc:Dep Hallos In ADH outline
I
2004 2003 Kftlkp Kd&p 2002 2003 2004 2003 2005ce 2030W
Cone-Bap Ralios inShan Lak*
I
11111
2002 2003 200« 2005 2003o> 20Xa> 2002 20tt 20» 2005
Sulfur
Nitrogen
Sulfur
Nitrogen
4 Figure 7-11 Spatial and interannual variability of inverse deposition transference ratios,
5 1/Tsox and l/TNOy, for Adirondack (top) and Shenandoah case study areas.
September, 2010
7-27
Draft - Do Not Quote or Cite
-------
Inter-year CV of Sulfur Conc:Dep Ratio
Inter-year CV of Nitrogen Conc:Dep Ratio
>
o
o
03
O
in
o _
ADR
Other
o
i
Shen
o
If)
o _
S
s
ADR
Other
Shen
Changes in Sulfur Conc:Dep
Ratio from Emissions
ADR
I
Qher
Changes in Nitrogen Conc:Dep
Ratio from Emissions
-2005)/2005
20 40
i l
o
CO
o
CM
CO ,-.
c °
(0
O
6?
o
oj -
o
o
e
!
1
,
t
_j
i
I
1
" i i
,_ 8
; o
-2005)72005
40 60 80
l i i
§ 8-
CM
O) °
CO
-^ o
O CM -
o _
1
0
o
1
1
• 1
1
_ 1
I
C±3
r-rH
r T
Shen
ADR
Other
Shen
2 Figure 7-12. Summary of inter-annual and emissions sensitivity variability of sulfur and
3 nitrogen deposition transference ratios.
September, 2010
7-28
Draft - Do Not Quote or Cite
-------
7
8
LU
n—
O
c
o
'o
Base
— Advection
— Dry Deposition
•Wet Deposition
VclSO2 Bi-directional
Advection -o— Advection
Dry Deposition -o—Dry Deposition
Wet Deposition -o—Wet Deposition
oo ooooo oo oooo oo oo
100
200 300
Distance (km)
400
500
3 Figure 7-13. Cumulative regional NH3 budget of advection, wet- and dry deposition,
4 calculated for an expanding box starting at the high-emitting Sampson
5 County NC cell (from Dennis et al, 2010)
September, 2010
7-29
Draft - Do Not Quote or Cite
-------
1 7.5. Sensitivity of AAPI to component parameters
2 An elasticity analysis was applied to investigate sensitivity of the AAPI to its components
3 (Appendix A). The means, medians and quartiles of the AAPI component variables were based
4 on the range variable values across ecoregions that overlapped with the CMAQ domains.
5 Elasticities measure the percent change in the AAPI for a 1% change in the AAPI parameters: Q,
6 Neco, NHX, BCl , TNOy, TSOx, NOy, and (SO2 + SO4).
8 The elasticity results identified significance for all the AAPI parameters. Detailed
9 results are provided in Appendix A. Base cation weathering, BC* , and hydraulic flow rate, Q,
10 exerted strong influence on AAPI, an expected result given the explicit dependency evident in
11 the AAPI expression. The transference rations for NOy (TNoy) and SOx (Tsox) exhibited
12 relatively less influence on AAPI calculations than all other parameters when evaluated at means
13 of the variables. However, in some locations, when evaluated at other values of the variables,
14 AAPI can be more sensitive to the deposition transformation ratios.
15 These results suggest focusing on the uncertainties in the non-atmospheric inputs,
16 including base cation weathering and runoff rates, and the implications of those uncertainties in
17 setting an AAPI that will have a high likelihood of providing the targeted level of protection.
18 An analysis of variance (ANOVA) analysis of the AAPI parameters will be added to the final
19 PA.
20
21 7.6 Uncertainty in Critical Load and ANC modeling
22 7.6.1 MAGIC modeling
23 An extensive uncertainty analysis of the MAGIC model was conducted as part of the
24 REA, and documented in Appendix 4 of the REA. This uncertainty analysis included
25 comparison of MAGIC outputs with observed water chemistry and ANC values. The
26 uncertainty analysis also included an approach for generating confidence intervals for predicted
27 ANC, using ensembles of model results based on alternative model calibration methods.
September, 2010 7-30 Draft - Do Not Quote or Cite
-------
1 The model performance comparisons documented in Appendix 4 of the REA show close
2 correspondence between simulated and observed annual average surface water 864, NOs, and
3 ANC during the model calibration period for 44 lakes in the Adirondacks Case Study Area and
4 60 streams in the Shenendoah Case Study Area. These comparisons are reproduced in Figures 7-
5 14 and 7-15. Comparisons in the ability of MAGIC to reproduce the temporal pattern of ANC
6 for individual lakes was also assessed, and the model does reasonably well at matching the
7 pattern of ANC, although the fit is not as good as during the model calibration period.
8 The estimated confidence bounds on predicted ANC suggest that the 95 percent upper
9 confidence bound is on average 10 percent higher in lakes, and 5 percent higher in streams. This
10 suggests relatively low uncertainty introduced by the MAGIC modeling assumptions. MAGIC
11 modeling is used in developing the estimates of base cation weathering for comparison to the F-
12 factor approach described in Chapter 5.
13 7.6.2 SSWC modeling
14 As stated in Appendix 4 of the REA, uncertainties in some elements of the SSWC
15 modeling are not well understood. The version of the SSWC model used here uses the F-factor
16 approach to estimate the preindustrial base cation supply for a given catchment. While this
17 approach has been widely applied in Canada and Europe, it has only been used in a few cases
18 within the United States and its assumptions and parameters have not been fully evaluated for
19 aquatic systems. The natural or preindustrial catchment supply of base cations (i.e. weathering
20 rates) has the most influence on the critical load calculation and also has the largest uncertainty
21 (Li and McNulty, 2007). The uncertainty and ability to accurately estimate this parameter has
22 not fully been evaluated and its uncertainty is unknown. It is important to note that for the
23 United States, there is only one study for surface waters critical loads that compared steady-state
24 and dynamic models and different steady-state approaches (MAGIC and F-factor) (Holdren et al.
25 1992) other than what is presented in Chapter 5. Holdren et al. 1992 compared critical loads
26 calculated by the steady-state MAGIC and the SSWC F-factor model for lakes in the Northeast.
27 In this study, steady-state MAGIC model yielded critical load values that show the same general
28 trend and on average were 14 kg/(ha-yr) SC>4 higher than those from the SSWC F-factor
29 approach, which is consistent with results, presented in Chapter 5. The two models converge at
September, 2010 7-31 Draft - Do Not Quote or Cite
-------
1 low critical, but diverge as the buffering potential for watersheds increase, as indicated by
2 increasing critical loads.
3 The REA conducted an uncertainty assessment using Monte Carlo simulation methods to
4 characterize the uncertainty in estimated critical loads using the SSWC, varying a number of
5 important inputs including runoff rates, water chemistry variables, and acid deposition. The
6 coefficients of variation (CV) for the estimated critical loads (standard deviation divided by the
7 mean) were calculated for each lake in the study as a measure of relative uncertainty.
8 The results of this uncertainty analysis show that the coefficients of variation are on
9 average very low for target ANC values within the range we are recommending (20 to 50 |ieq/L).
10 The CVs for critical loads are only 5% and 9% for critical load limits of 20 and 50 [j,eq/L,
11 respectively. Although the average CV is relatively small for the population of sites modeled,
12 individual site CV can vary from 1% to 45%. This difference is due to the high degree of
13 uncertainty in site specific parameters for particular sites.
14 These analyses suggest that uncertainties introduced in the AAPI directly by the SSWC
15 Factor model are likely to be moderate. Additional uncertainties are introduced by the
16 generalization of the F-factor approach to estimate critical loads in locations where F-factors
17 have not been developed.
September, 2010 7-32 Draft - Do Not Quote or Cite
-------
150!
100-
O
3 50
300
£• 200 -
'£
o>
i
CO
100-1
0 -
-100
SO4 (jeq/L
50 100
Observed Chemistry
150
ANC
-100 0 100 200
Observed Chemistry
300
o
5 10 15
Observed Chemistry
20
8 i
7 -
6 -
E
to
567
Observed Chemistry
2 Figure 7-14 Simulated versus observed annual average surface water SO42-, NO3-, ANC,
3 and pH during the model calibration period for each of the 44 lakes in the
4 Adirondacks Case Study Area. The black line is the 1:1 line. (Source:
5 reproduced from REA, Appendix 4, Figure 1.1-1)
September, 2010
7-33
Draft - Do Not Quote or Cite
-------
150 -i
| 100
E
o
"i
S04 |jeq/L
50 100
Observed Chemistry
150
10
20 30 40
Observed Chemistry
50
300
0
E
to
200 -
100 -
0 -
-100
ANC [jeq/L
7 -
6 -
E
to
-100 0 100 200
Observed Chemistry
300
pH [jeq/L
567
Observed Chemistry
2 Figure 7-15. Simulated versus observed annual average surface water SO42-, NO3-, ANC,
3 and pH during the model calibration period for each of the 60 streams in the
4 Shenandoah Case Study Area. The black line is the 1:1 line. (Source:
5 reproduced from REA, Appendix 4, Figure 1.1-2)
6 7.7. Modeling and Data Gaps (To be expanded in final PA)
7 Atmospheric and deposition processes. The previous section introduced two important
8 enhancements regarding the treatment of wet precipitation and the bi-directional flux of
9 ammonia. The interest in deposition of sulfur and nitrogen raises the potential importance of
10 occult (cloud and fog related processes) deposition associated with mists and clouds, which may
11 be particularly relevant for aquatic acidification of high elevation watersheds. Occult deposition
12 currently is in the early stages of development within the CMAQ framework.
September, 2010
7-34
Draft - Do Not Quote or Cite
-------
1 Lightning generated NOx emissions have been an active area of research over the last
2 decade and approaches that incorporate lightning count data and estimated NOx generation based
3 on satellite measurements and aircraft campaigns have been tested in modern air quality models,
4 including CMAQ. Lightning NOx is hypothesized to increase upper tropospheric ozone levels
5 and wet nitrogen precipitation, with relatively negligible impact on near surface ambient nitrogen
6 patterns. It is anticipated that CMAQ will incorporate lightning NOx for EPA assessments in the
7 2012timeframe.
8 Interest in organic bound nitrogen has increased based on NADP measurements
9 suggesting that organic nitrogen contributes as much as 30% of the total nitrogen in precipitation
10 samples. Significant uncertainties regarding the origin and composition of organic nitrogen
11 (Altieri et al., 2009) suggest a need for research to improve our understanding of organic
12 nitrogen prior to developing parameterizations in air quality models. Questions regarding the
13 the relative contribution of anthropogenic or natural sources as well as the effects of re-
14 entrainment from the surface require attention.
15 Atmospheric Observations. Chapter 4 addresses the current state of atmospheric
16 observations relative to the NOx/SOx secondary standard and Chapter 8 addresses preliminary
17 recommendations for monitoring methods. This new standard poses measurement resource
18 challenges as the current networks, with the exception of CASTNET and some National Park
19 Service (NFS) efforts, are deficient in spatial coverage relevant to anticipated acid sensitive areas
20 and the specific measurement needs related to NOy, speciated NOy and ammonia and
21 ammonium.
22 Source emissions. Anthropogenic emissions of nitrogen oxides (NO and NO2) and
23 sulfur dioxide generally are believed to be well characterized as the major contributors of NOx
24 and SO2 from energy generation and transportation sectors have a history of continuous
25 improvements of emissions modeling as well as direct emission measurements for major power
26 generating units. Greater uncertainty resides in natural emissions of NOx from lightning
27 processes (discussed above) and soil and agricultural related phenomena. Both NOx and
28 ammonia emissions are subject to re-emission after deposition as part of the complex cycling of
29 nitrogen in soils and biota. Characterizing the variety of agricultural practices that impact both
30 ammonia and NOx is complicated by the dispersed nature of agriculture processes as well as the
September, 2010 7-35 Draft - Do Not Quote or Cite
-------
1 influence of various meteorological factors on relevant biogeochemical processes controlling
2 transformation and removal of nitrogen species.
3 Ecosystem processes and surface water observations. [To be completed in final PA]
4 The critical load modeling approaches that produced the N/S deposition tradeoff curves require a
5 variety of input data depending on the approach chosen. In general terms, the availability of
6 watershed related deposition, soil and vegetation characteristics and surface water chemistry
7 determine the approach taken. There is a relatively extensive source of data for critical load
8 modeling in the Eastern U.S., as illustrated by the frequent reliance on the Adirondack and
9 Shenandoah Case studies. For this PA, critical load estimates were developed for national level
10 coverage, largely through SSWC modeling relying on surface water data. Several ecoregions
11 included an extremely small sample size of critical load estimates that challenged the
12 development of a national scale assessment of acid sensitive areas. A more thorough
13 characterization of nitrogen retention, dissolved organic carbon, soil chemistry in all acid
14 sensitive areas would lead to reduced uncertainties in applying the AAPI as well as future
15 considerations for standards that incorporate terrestrial acidification and nutrient enrichment
16 effects.
17 7.8. Summary and Conclusions
18 Uncertainty and natural variability exist in all of the components of the structure of the
19 NOx and SOx standard introduced in this PA, and should be considered in establishing the level
20 of the AAPI. A summary of the relative uncertainties of these components is provided in table
21 7-1 (To be added). On balance, the confidence level in the information and processes associated
22 with the linkages from ecological effects to atmospheric conditions through deposition and
23 ecosystem modeling is very high. The considerable body of evidence is conclusive with regard
24 to causality between aquatic acidification and biological and ecological effects. Confidence in
25 the linkage associating aquatic acidification and ANC is extremely high, as the aquatic chemistry
26 describing this relationship, while nonlinear, is relatively simple with regard to chemical species
27 and reactions. The relationships between deposition and ANC, while complicated by a variety
28 of biogeochemical and hydrological processes and data requirements within watersheds, are well
29 established and the critical load models have been thoroughly vetted through the scientific
30 community with a demonstrated level of successful evaluation. The linkages between ambient
September, 2010 7-36 Draft - Do Not Quote or Cite
-------
1 concentrations of relevant species and deposition is best handled through air quality modeling
2 systems like CMAQ. The relationship between concentrations and deposition loads is well
3 characterized by these models, which are constrained by mass balance principles. While much
4 of the physical and chemical processing that determine concentrations and consequent deposition
5 is interwoven with numerous fundamental processes characterizing mass transport and
6 atmospheric chemical oxidation, the science is relatively mature with years of applications and
7 continued evolution of the models. The specific processes guiding nitrogen and sulfur
8 chemistry and deposition are relatively simple. More challenging is the ability to parameterize
9 processes at the air-surface interface which guide the estimation of deposition velocities and the
10 re-emission of certain species, as well as many of the area wide natural processes and
11 agricultural practices which influence emissions of oxidized and reduced forms of nitrogen.
12 The variety of uncertainty, variability and sensitivity analyses included in this chapter
13 have been conducted under the assumption that the basic model construct is solid, as discussed
14 immediately above, and are used to inform conclusions regarding the level of the AAPI that
15 incorporate consideration of uncertainty. These analyses are also useful in guiding
16 implementation efforts related to future monitoring, emissions and model process improvements.
17 The influence of uncertainty on the level of the AAPI can be thought of as reducing or
18 increasing relative stringency of the level to increase the likelihood that requisite protection of
19 public welfare is provided. Throughout these discussions there is no apparent directional bias
20 in the uncertainty regarding the biological, chemical and physical processes incorporated in the
21 AAPI. From the perspective of valuation of ecosystem services, the estimates generally are
22 believed to be biased low, meaning the values of reaching a target level of protection are
23 underestimated. However, quantification of these values is perhaps the most uncertain of all
24 aspects considered. Consequently, the level of the AAPI should be relatively high in a buffering
25 context to account for the existence of uncertainties in several components. In addition to, but
26 related to these uncertainties discussions, are considerations of time lag to reach a target level
27 ANC due to ecosystem response dynamics, as well the uncertainties in the severity and
28 prevalence of episodic events. Both of these considerations suggest support for an AAPI that is
29 somewhat higher than the target ANC supported by the specific evidence and risk information.
30
September, 2010 7-37 Draft - Do Not Quote or Cite
-------
1 Table 7-2. Summary (Incomplete) of Qualitative Uncertainty Analysis of Key Modeling Elements in the NOx/SOx AAPL
Source
Description
Potential influence of
uncertainty on risk
estimates
Direction Magnitude
Knowledge-
Base
uncertainty*
Comments
(KB: knowledge base, INF: influence of uncertainty on AAPI
estimates)
Major elements (and sub-models) of the ecological effects to ambient concentration framework
Biological/ecosystem
response to
acidification
Clear associations between
aquatic acidification (pH,
elevated Al) and adverse
ecosystem effects (fish
mortality, decreased species
diversity)
Both
Low
Low
Linkage between
direct acidification
species to ecological
indicator(ANC)
The relationships across ANC,
pH and dissolved Al are
controlled by well defined
aquatic equilibrium chemistry
Both
Low
Low
ANC is the preferred ecosystem indicator as it has a direct
relationship with pH and the deposition species relevant to
the NOx/SOx standard.
Linkage between
ecological indictor
and adverse
ecological effects
Direct associations between
ANC and fish mortality and
species diversity
Both
Low-medium
Low
Although the pH dependency on ANC is nonlinear, it is always
directionally consistent. In extremely low and high ANC
environments the relationship is of minimal value as
catchments are in relatively "less sensitive" regimes due to
natural conditions or extreme anthropogenic influence (i.e.,
acid mine drainage). In sensitive areas of concern the
relationship essentially is similar to the relationships between
direct acidification species and adverse effects.
Deposition to ANC
linkage through
Critical Load
modeling
Both MAGIC and Steady State
critical load models are
applied to determine critical
load models. The Steady
State critical load model
formulation is used as the
foundation for deriving the
Both
Low
Low
The model formulation is well conceived and based on a
substantial amount of research and applications available in
the peer reviewed literature. There is greater uncertainty
associated with the availability of data to support certain
model components.
September, 2010
7-38
Draft - Do Not Quote or Cite
-------
Source
Atmospheric
concentrations to
Deposition
Description
AAPI equation.
Deposition is a direct function
of ambient concentration,
influenced by several
processes, and handled in this
PA through air quality
modeling.
Potential influence of
uncertainty on risk
estimates
Direction
Both
Magnitude
Low
Knowledge-
Base
uncertainty*
Low
Comments
(KB: knowledge base, INF: influence of uncertainty on AAPI
estimates)
The model design is appropriate given the spatial and
temporal complexities that influence deposition velocity, as
well as the variety of atmospheric species that generally are
not measured. Greater uncertainty resides in the information
driving (e,g,, ammonia emissions) these calculations and
availability of observations to evaluate model behavior.
Sub-components and data of individual models
Deposition
Transference Ratios
CMAQ derived ratio of total
oxidized deposition to
concentration averaged over
one year
both
low
unknown
Transference ratios enable the connection between
deposition and the policy relevant ambient air indicators, NOy
and (SO2 + SO4). They are strictly a model construct and
cannot be evaluated in a traditional model to observation
context. The low sensitivity of these ratios to emission
changes and inter annual meteorology combined with low
spatial variability indicate that these ratios are necessarily
stable.
Additional elements to be added (.BCo, NECO, Q, DOC, benefits, ambient obs, surface water obs, emissions, )
September, 2010
7-39
Draft - Do Not Quote or Cite
-------
1 7.9 REFERENCES:
2 Altieri, K.E., BJ. Turpin and S.B. Seitzinger, 2009, Composition of dissolved organic nitrogen
3 in continental precipitation investigated by bu ultra-high resolution FT_ICR mass
4 spectrometry, Environmental Science and Technology, 43, 18, 6950 - 6955.
5 Foley, K.M., Roselle, SJ, Appel, K.W., Phave, P.V., Pleim, I.E., Otte, T.L., Mathur, R., Sarwar,
6 G., Young, J.O., Gilliam, R.C., Nolte, C.G., Kelly, J.T., Gilliland, A.B., Bash, J.O. (2010)
7 Incremental testing of the Community Multiscale Air Quality (CMAQ) modeling system
8 version 4.7, Geosci. Model Dev., 3, 205-226.
9 Otte, T.L. and Pleim, I.E. (2010) The Meteorology-Chemistry Interface Processor (MCIP) for
10 the CMAQ modeling system, Geosci. Model Dev., 3, 243-256.
11 PRISM Climate Group, Oregon State University, http://www.prismclimate.org, created 4 Feb
12 2004
13 Dennis, R. and K. Foley, 2009, Adapting CMAQ deposition fields for critical loads analyses,
14 presented at 2009 NADP Confernce; manuscript in preparation
15 Dennis, R., R. Mathur, I.E. Pleim and J.T. Walker, 2010, Fate of Ammonia Emissions at the
16 Local to Regional Scale as Simulated by the Community Multiscale Air Quality Model,
17 accepted by Atmospheric Pollution Research
18 Aiyyer, A, Cohan, D., Russell, A., Stockwell, W., Tanrikulu, S., Vizuete, W., and Wilczak, J.
19 2007. Final Report: Third Peer Review of the CMAQ Model.
20 Byun, D.W., and Schere, K.L. 2006. Review of the Governing Equations, Computational
21 Algorithms, and Other Components of the Models-3 Community Multiscale Air Quality
22 (CMAQ) Modeling System. J. Applied Mechanics Reviews, 59 (2), 51-77.
23
September, 2010 7-40 Draft - Do Not Quote or Cite
-------
1 8 AMBIENT AIR MONITORING
2
3 Ambient air measurements of nitrogen and sulfur species support implementation of this
4 proposed NAAQS secondary standard and improve the information basis for subsequent reviews.
5 These uses extend beyond the basic need to measure the proposed NAAQS indicators, NOy, 862
6 and SO4, in approximate priority order:
7
8 • Direct NAAQS comparisons.
9 • Reduced nitrogen, NH3 and NH4, to evaluate CMAQ and other air quality models
10 ability to characterize NHx deposition, a component of the AAPI expression.
11 • Model and process improvement - In combination with NAAQS indicators and NHx,
12 additional speciated NOy components including HNOs, PAN, NC>2, and NO to
13 continually assess air quality model behavior and associated deposition processes.
14 • Subsequent NAAQS reviews of the secondary NOx/SOx standard as well as related
15 primary and secondary standards reviews (ozone, NC>2, SC>2 and PM).
16 • Accountability - Assessing the effectiveness of implemented programs addressing
17 emission strategies to meet attainment of the proposed NAAQS using all noted
18 measurements.
19
20 8.1 What Are The Appropriate Ambient Air Indicators To Consider In Developing The
21 Standards?
22
23 The recommendation of NOY, SO2 and SO4 as the ambient air indicators for the proposed
24 NOx/SOx standard was introduced in the first draft of the PAD and endorsed by CAS AC
25 (CASAC, 2010). Essentially, NOy, which is an aggregate of all reactive oxidized nitrogen
26 compounds and the two sulfur species represent the oxidized ambient air species of relevance to
27 the criteria pollutants NOx and SOx with potential to adversely affect acid-base balance in
28 aquatic systems. Contributions of reduced nitrogen, which would not be part of the indicator for
29 the standard under the approach suggested in this PAD, are provided by CMAQ.
30
September, 2010 8-1 Draft - Do Not Quote or Cite
-------
1 Why not use each individual species as indicators?
2 One could consider using all NOy species as NAAQS indicators, requiring, for example,
3 measurements of the dominant species: HNO3, particulate nitrate, true NO2, NO, and PAN.
4 Conceptually, each species would be paired with a species - specific deposition velocity in the
5 AAPI expression to transfer individual deposition values to ambient values. The attraction of
6 using individual species would be the reliance on actual deposition velocities that have more
7 physical meaning in comparison to model constructed transference ratios which aggregate dry
8 and wet deposition and all nitrogen species. The transference ratio approach does retain the
9 necessary conservation of mass which underlies virtually all parameterization schemes, but loses
10 some degree of physical relevance due to use of modeled outputs - an admittedly unique
11 construct. The major drawback of using individual species as NAAQS indicators is the lack of
12 routinely available measurement techniques and an associated resource burden even if adequate
13 techniques were available. Currently, technology for measuring true NO2, HNOs, and PAN
14 generally is not available for routine network applications. In addition to this practical
15 consideration, there is another important reason for using the aggregated transference ratios.
16 Because the standard, and the explicit Clean Air Act authority, is based on ambient air there must
17 be an effective link connecting deposition and concentrations of the criteria pollutants in the
18 ambient air. The use of individual species conceptually allows for a more physically meaningful
19 approach to characterize and calculate deposition. However, the transference ratios also enable
20 incorporation of the contributions of wet precipitation in the ambient air indicators. There is no
21 practical alternative that allows for the disentangling of wet deposition as a function of ambient
22 air concentrations as that relationship is best addressed through the coupling of numerous
23 meteorological and chemical processes imbedded in the air quality modeling platform. One
24 could consider wet precipitation as a separate parameter and isolate dry precipitation in the AAPI
25 equation. But doing so would lose the important connection between wet precipitation of
26 nitrogen and the same emission sources responsible for dry deposition.
27
28 Finally, one might advocate for direct measurements of dry deposition of individual NOY
29 species. Again, technologies simply are not ready for consideration in routine network
30 applications. And, it is arguably practical to model dry deposition even if direct dry deposition
31 measurement technologies were available. One reason for this is that there is significant spatial
September, 2010 8-2 Draft - Do Not Quote or Cite
-------
1 heterogeneity in the factors (vegetation and surface type, micrometeorology) that define
2 deposition velocity. Consequently, direct dry deposition measurements would have limited
3 spatial representativeness in comparison to ambient air observations. The model conceptually
4 accounts for the spatial variance, at the level of horizontal grid cell resolution, of the factors
5 defining deposition velocity on a species by species basis. However, one also could reason that a
6 well placed direct measurement of dry deposition is more realistic than a modeled result that
7 relies on numerous assumptions. The development of technologies to measure direct dry
8 deposition will benefit the diagnosis and improvement of process formulations in models.
9
10 Consideration has been given to change the atmospheric indicator from NOy to total
11 nitrate (the sum of nitric acid and particulate nitrate). The rationale for that approach is that a
12 larger fraction of the deposited NOy is accounted for by total nitrate, which currently is
13 measured in CASTNET with high confidence. One can reason adequately that nitrate may
14 correlate well with total oxidized nitrogen deposition relative to NOy (as discussed in Chapter 4),
15 given the inherent noise associated with variable contributions of low deposition velocity species
16 of ambient level significance (e.g., NCh). The disadvantages of using total nitrate as an
17 indicator are that significant ambient mass with the potential for deposition is not captured, and
18 NOy is a preferred measurement for model evaluation and accountability purposes.
19 Accountability refers to assessing if emissions reductions have the intended consequences on
20 ambient air and deposition levels in the context of, "Aare our emissions reductions strategies
21 working as planned?" All three of these concerns relate to the benefit of closing mass balances in
22 whatever environmental medium is being characterized. In addition, the use of nitrate alone
23 would create an increased distancing from the listed criteria pollutant oxides of nitrogen as NOy
24 does include NO and NO2.
25
26 8.2 Reactive Oxidized Nitrogen and Sulfur Species.
27
28 NOy species. Air quality models and deposition models that use direct observations calculate
29 'deposition on a species by species basis to account for differences in deposition velocities.
30 Consequently, the relative fractional contributions of individual NOy or SOx species to
31 deposition or concentration is influenced by the differences in species deposition velocities. For
September, 2010 8-3 Draft - Do Not Quote or Cite
-------
1 example, nitric acid with a high deposition velocity would exhibit a larger relative contribution
2 to overall deposition compared to ambient concentrations in a particular area (Figure 8-1). The
3 dominant ambient air NOY species are NO, NO2, HNO3, P-NO3 and PAN. Near source urban
4 environments typically have a relatively higher fraction of NOx compared to the products of
5 NOx reactions, nitrates and PAN, which are relatively more dominant in rural locations (Figures
6 8-2 - 8-5).
7
8 The differences in the relative patterns between ambient air and deposition on a species
9 by species basis illustrate a number of challenges and considerations in developing a monitoring
10 strategy. It is clear in the Adirondacks and Shenandoah areas that nitric acid is the most
11 dominant contributing species from a deposition perspective (Figure 8-1), with significant
12 contributions from particulate nitrate, PAN and NO2. The original source of emissions (NO
13 accounts for 90-95% of all emitted NOx) provides very small (< 5%) contributions to ambient air
14 and virtually nothing to deposition. The combination of nitric acid and particulate nitrate
15 consistently contribute greater than 50% of the oxidized nitrogen dry deposition load, whereas
16 PAN and NO2 contribute roughly 15-25% of the deposition load. These broad summary
17 statements speak to some of the monitoring considerations addressed earlier, particularly the case
18 for monitoring for total nitrate. However, caution should be exercised when considering not
19 measuring a considerable fraction of the ambient NOy burden reflected in NO2 and PAN.
20 Characterization of NO2 deposition is an area requiring further refinement especially considering
21 that NO2 is a significant component of total oxidized nitrogen. Zhang et al. (2005) suggest that
22 NO2 contributes up to 36% of dry NOy deposition in rural Eastern Canadian locations, and
23 suggest, based on observational evidence (Figure 8-2), that in some locations NO2 deposition
24 may be similar to nitric acid contributions.
25
26 A sampling of co-located NOy species observations in rural Eastern Canada (Figure 8-2),
27 particularly in Egbert, Ontario, illustrates the concern of the general assumption that NO2 may
28 not contribute significantly to NOY deposition in rural locations. While it may be true that in
29 general NO2 is of less of concern in rural areas relative to urban areas, that does not dismiss the
30 potential for significant misrepresentation of total nitrogen budget in certain rural locations.
31
September, 2010 8-4 Draft - Do Not Quote or Cite
-------
1 The results also raise the question of potential efficiencies gained from not cycling
2 between NO and NOy analyses, which is the standard configuration in commercial NOy
3 instruments, acknowledging the limited use of NO data in rural, acid sensitive environments
4 (note that NO present in the air would still be captured in the NOy measurement).
5
6 These examples are used to support the rationalization of using NOy as an appropriate
7 atmospheric indicator in applying the AAPI. However, while it may be required to measure
8 NOy for explicit AAPI calculations to determine compliance with a NOx/SOx standard, there
9 should be additional measurements of true NO2, HNOs, p-NOs and PAN to allow for diagnostic
10 evaluations of both air quality models and the NOy measurement itself. This recommendation
11 would leverage existing CASTNET filter pack (FP) observations necessary to capture particulate
12 sulfate (discussed below) and therefore require the addition of true NO2 measurements and
13 periodic sampling for PAN.
14
15 Sulfur Species. Although sulfur dioxide and particulate sulfate contribute approximately 60 and
16 40 %, respectively, to ambient SOx concentrations, sulfur dioxide is the dominant contributor to
17 SOx deposition (Figure 8-6), which is consistent with CASTNET observational studies (Sickles
18 and Shadwick, 2007). Measurement technology issues are not as complex for SOx as they are
19 for NOy and individual NOy species. Issues related to particle size fraction and averaging
20 period for SOx are discussed below.
21
22 Reduced Nitrogen.
23
24 The AAPI does not include reduced nitrogen (ammonia gas and ammonium ion) as an
25 ambient air indicator. However, reduced nitrogen deposition is an explicit AAPI component
26 which is estimated through air quality modeling. As discussed in the Chapter 7, characterization
27 of reduced nitrogen deposition processes is an active developmental area which would benefit
28 markedly from NHX measurements in order to assess modeled predictions of ambient patterns of
29 ammonia and ammonium. This need for monitoring ammonia in rural environments is further
30 supported by emerging evidence that ammonia acts as a regionally dispersed species based on
31 the inclusion of ammonia bi-directional flux in CMAQ simulations. (Dennis et al., 2010).
September, 2010 8-5 Draft - Do Not Quote or Cite
-------
1 Monitoring method approaches under consideration for routine application typically are limited
2 to time averaged filter and denuder technologies, including passive sampling approaches.
O
4 8.3 What measurements would be used to characterize NOy and SOx ambient air
5 concentrations for the purposes of the AAPI based standard?
6
7 Ambient NOy, SC>2 and particulate sulfate (804) concentrations would be used as the
8 indicators in determining compliance with the standard. All of these indicators are measured in
9 different places within the current routine monitoring networks (section 3.2). However, there are
10 issues requiring resolution associated with Federal Reference or Equivalency Measurement
11 (FRM/FEM) status of measurement techniques that to date have served as supplemental
12 information, which will require resolution. A FRM for SC>2 exists, but not for NOyor SO/t. Only
13 recently have NOy measurements, which historically were viewed as research venue
14 measurements, been incorporated as "routine" observations, partly as a result of the NCore
15 program. Acquiring FRM status may require better characterization of the conversion
16 efficiencies, mass loss and clear guidance on operating and siting procedures. Particulate sulfate
17 has been measured for several years in the IMPROVE, CASTNET and EPA CSN networks. The
18 nation has over 500 24-hour average, every third day sulfate measurements produced by the
19 PM2.5 speciation networks (IMPROVE and EPA CSN) and nearly 80 CASTNET sites that
20 provide continuous weekly average samples of sulfate with an open inlet accommodating all
21 particle sizes. With minor exceptions, the PM2.5 fraction generally accounts for over 80% of the
22 ambient sulfate mass. Unfortunately, as particle size diameters increase beyond 2.5 \i,
23 gravitational settling imparts greater influence resulting in substantially enhanced deposition
24 velocities. Consequently, the sulfate mass in size fractions greater than 2.5 |i potentially
25 provides correspondingly greater contribution to as much as 50% of dry sulfate deposition in
26 certain locations (Grantz et al., 2003).
27
28 Sample collection period is not an issue for gaseous measurements of NOY and SO2 that
29 operate continuously. However, consideration should be given to using the CASTNET FP for
30 SO2 measurements as a resource saving option, assuming the FPs will be used for particulate
31 sulfate. However, the availability of highly time resolved data will support the continual
September, 2010 8-6 Draft - Do Not Quote or Cite
-------
1 evaluation of 862 and sulfate balance in air quality modeling systems which is a critical
2 underpinning for both human and ecosystem health assessments. Some concerns have been
3 raised about the possibility of exclusion of coarse particles from NOY samplers operating at low
4 flow conditions as well as potential difficulties of reducing organically bound and mineralized
5 nitrate. Insight into conversion and capture efficiency characteristics will be advanced both by
6 research catalyzed by the need to support this standard and through ongoing and future network
7 operations.
8
9 8.4 What additional complementary measurements are recommended?
10
11 We recommend that there be 3-4 locations nationally, in airsheds with different
12 atmospheric chemistries, that sample not only for the NAAQS indicator NOy but for the suite of
13 major NOy species as well; HNOs, p-NOs, PAN, true NC>2, and NO as discussed earlier. Not
14 only is this important from a modeling and process diagnosis perspective, but it is especially
15 useful in the introduction of new measurements that have a limited track record to provide
16 insight into instrument performance. In the case of NOy, it is even more relevant since there
17 effectively are no standards that explicitly challenge instrument accuracy given the highly
18 variable nature of NOy species distribution and the instability associated with mixing NOy gases.
19 This quality assurance issue is analogous to PM2.5 where aerosol standards are not available and
20 measurement accuracy is judged against periodic challenges relative to a "gold standard"
21 instrument. Reduced nitrogen measurements of ammonia and ammonium ion are recommended
22 at all locations with FRM/FEM instruments based on the need to support the AAPI as discussed
23 above.
24
25 8.5 What sampling frequency would be required?
26
27 The averaging time for the standard is likely to be an annual average, perhaps based on 3-
28 5 years of data collection to minimize the influence of interannual varaibility in meteorology,
29 especially precipitation.. Conceptually, extended sampling periods no longer than one year
30 would be adequate for the specific purposes of comparison to a standard. However, there are
31 significant peripheral benefits relevant to improving the scientific foundation for subsequent
September, 2010 8-7 Draft - Do Not Quote or Cite
-------
1 reviews and a variety of related air quality and deposition assessments to be gleaned from more
2 highly time resolved data. In particular, the critical role of air quality models in deposition
3 assessments implies value to be derived from measurements that support model evaluation and
4 improvement. Many of the monitoring approaches that are used throughout the nation sample
5 (or at least report out) on daily (PM2.5 chemical speciation), weekly (CASTNET) and hourly (all
6 inorganic gases) periods. There is a tradeoff to consider in sampling period design. For
7 example, the weekly CASTNET collection scheme covers all time periods throughout a year, but
8 only provides weekly resolution that misses key temporal and episodic features valuable for
9 diagnosing model behavior. The every third day, 24-hour sampling scheme used in IMPROVE
10 and EPA speciation monitoring does provide more information for a specific day of interest yet
11 misses 2/3 of all sampling periods. The missing sampling period generally is not a concern when
12 aggregating upward to a longer term average value as the sample number adequately represents
13 an aggregated mean value. Additionally, there is a benefit to leveraging existing networks which
14 should be considered in sampling frequency recommendations. A possible starting point would
15 be to assume gaseous oxidized species, NOy and SC>2, are run continually all year reporting
16 values every hour, consistent with current routine network operations. Sulfate sampling periods
17 should coincide with either the chemical speciation network schedules or with CASTNET.
18 There are advantages to coordinating with either network. Ammonia gas and ammonium ion
19 present challenges in that they are not routinely sampled and analyzed for, and the combined
20 quantity, NHX is of interest. Because NHX is of interest, some of the problems of volatile
21 ammonia loss from filters may be mitigated. However, for model diagnostic purposes,
22 delineation of both species at the highest temporal resolution is preferred.
23
24 8.6 What are the spatial scale issues associated with monitoring for compliance, and
25 how should these be addressed?
26
27 The current observation network for NOy, NHX and SOx is very modest and includes a
28 monitoring network infrastructure that is largely population oriented with the exception of
29 CASTNET and IMPROVE. While there is platform and access infrastructure support provided
30 by CASTNET, NADP and IMPROVE, those locations by themselves are not likely to provide
31 the needed spatial coverage to address acid sensitive watersheds across the United States.
September, 2010 8-8 Draft - Do Not Quote or Cite
-------
1 Ambient monitoring at every watershed will not be required given the reality of resource
2 constraints and the relative spatial homogeneity of air concentrations that are averaged over
3 annual time periods and within 'acid sensitive" areas. The spatial monitoring requirements will
4 be associated with the determination of acid sensitive areas, which is discussed at length in
5 Chapter 5. The number of sites per area will be addressed in rule development and general
6 guidance based on an understanding of the spatial variability of NOy, NHx, sulfate and SC>2
7 combined with resource allocations will help inform those decisions.
8
9 Critical load models applied for the purposes of this standard would be based on annual
10 averages, which effectively serves to dampen much of the spatial variability. Furthermore, the
11 development of an area-wide depositional load tradeoff curve implies focus on region wide
12 characterization. Toward that end, CMAQ concentration fields will provide insight into the
13 likely spatial representativeness of monitors leading to efficient application of monitoring
14 resources. For example, the CMAQ based spatial coefficient of variation (standard
15 deviation/mean) of oxidized nitrogen in the Adirondacks was 1.46%. Improved dry deposition
16 estimates will result from enhancements of ambient monitoring addressing the N/S secondary
17 standards as each additional location could serves a similar role that existing CASTNET sites
18 provide in estimating dry deposition.
19
20 8.7 What specific monitoring methods would be used?
21
22 Federal reference and/or equivalent methods (FRM/FEM) are presently available only for
23 SC>2. Particulate SC>4 is measured at over 500 sites nationally, and there is a general consensus
24 that methods available are reliable and provide consistent data. NOy measurement is in a
25 transition period from largely being viewed as a research level measurement to now being
26 deployed as a routine measurement in EPA's national 75 site NCORE network. The general
27 consensus on NOy measurement is that the methodology is sound and applicable for
28 routine/regulatory use, but there does not exist a well defined understanding of the quality of
29 NOy data. Inorganic dry nitrate (nitric acid and particulate nitrate) is measured routinely in the
30 CASTNET network with filter packs (FP).
31
September, 2010 8-9 Draft - Do Not Quote or Cite
-------
1 One of the challenges associated with specifying performance attributes for p-SC>4 and
2 NOy is the lack of specific challenge standards. For example, instruments measuring discrete
3 gases such as ozone or nitrogen oxide can be challenged by comparing an instrument's reading
4 when measuring known concentrations of gases which are readily provided for single gas
5 concentrations. Particle standards are not available. NOy performance typically is challenged by
6 known mixtures of NO2, and occasionally with N-propyl nitrate, which only addresses part of the
7 spectrum of nitrogen species in an NOy mix. Consequently, instrument performance in EPA's
8 national networks for aerosol mass is quantified in terms of bias and precision relative to a co-
9 located "performance evaluation" instrument. There is no comparable program in place for p-
10 SO4orNOY.
11
12 p-SC>4. The routinely operating methodology for particulate sulfate (p-SO4) is based on
13 an integrated (i.e., time averaged over several hours or days) sample collection on a Teflon filter
14 followed by ion chromatography (1C) detection in the laboratory. Two major variations of this
15 approach are applied in the PM2 5 speciation (exclusion of particles larger than 2.5 ji and 24-hour
16 collection typically every third day) and CASTNET (weekly average integrated sampling all year
17 with an open inlet to include all size fractions). There are additional variations related to inlet
18 design and flow characteristics of PM2.5 speciation samplers in which two designs are prevalent
19 in the networks: (IMPROVE and EPA CSN SASS samplers). These variations are considered
20 minor as sulfate species (dominated by ammonium sulfate) typically are not subject to major
21 sampling artifacts associated with volatilization or condensation. The difference in inlets (open
22 vs. 2.5 ji) is perceived by some as not an issue of concern as 80 - 90 % of the PM sulfate mass is
23 distributed in size fractions less than 2.5 \i. However, the higher deposition velocities associated
24 with larger diameter particles argue for including all size fractions as discussed above.
25 Continuously operating in-situ sulfate instruments that allow for hourly, or less, data reporting
26 are available. However, the limited deployment (less than 20 sites nationally) of these
27 instruments combined with the 2.5 |i inlet cutoff configuration preclude consideration at this
28 time.
29
30 The CASTNET FP offers three important attributes: a history of high quality data,
31 existing infrastructure and network to build on and an open inlet to capture the full range of
September, 2010 8-10 Draft - Do Not Quote or Cite
-------
1 particle diameters. EPA intends to develop FRM status for this method. A significant
2 additional advantages of using the FP method will be the availability of important co-measured
3 species (e.g., SO2, total nitrate, ammonium). While EPA will expedite the certification process
4 for the CASTNET FP, in the future consideration should be given to other available methods to
5 more efficiently leverage network assets. For example, the SASS sampler potentially would
6 accommodate ammonia gas and ammonium ion measurements, as well as other standard
7 chemical speciation parameters depending on the configuration of this multi channel system.
8 Continuous sulfate measurements would be extremely useful for model evaluation, especially
9 considering the availability of continuous 862 data that would be required as part of the NAAQS
10 indicators. A performance based approach to meet equivalency requirements, given the variety
11 of sulfate measurement approaches and well vetted and accurate analytical procedures.
12
13 SO2. Sulfur dioxide is a NAAQS pollutant and a FRM is available. See 75 FR at 35554-
14 56 and 35593-95 (June 22, 2010) (adopting a second FRM for SO2). As part of the NCore
15 network development effort, trace gas SC>2 analyzers capable of sub ppb resolution became
16 commercially available and are the preferred instruments for implementation in rural locations.
17 As discussed above, the near continuous data output of gaseous analyzers is desired for
18 peripheral support of model evaluation. Nevertheless, the convenience and resource savings
19 associated with the CASTNET FP suggest that Federal Equivalency Method (FEM) status should
20 be incorporated in concert with the sulfate certification process.
21
22 NOy. In principle, measured NOy is based on catalytic conversion of all oxidized
23 species to NO followed by chemiluminescence NO detection. While there are caveats associated
24 with instrument conversion efficiency and possible inlet losses, the technique is considered
25 adequate and routinely operational. Approximately 25 sites (out of a planned 75) in EPA's
26 NCORE network are operating NOy instruments, and an additional 5-110 sites are operated in
27 SEARCH, CASTNET and other programs. NOy measurements are nearly continuous, reporting
28 at hourly intervals providing far greater temporal information compared to filter or denuder
29 based methods. FRM certificaton for NOy presents more considerable challenges given the
30 limited history of routinely operating instruments. The process EPA is pursuing for certification
31 status for NOy will be addressed in the final PA.
September, 2010 8-11 Draft - Do Not Quote or Cite
-------
1 8.8 REFERENCES
2
3 CAS AC, 2010, Letter to Administrator Jackson on Review of the Policy Assessment of the
4 Secondary National Ambient Air Quality Standard for NOx and SOx : First Draft.,
5 Grantz D.A., J.H.B. Gardner and D.W. Johnson, 2003, Ecological effects of particulate matter,
6 Environment International, 29 (2003) 213 -239.
7 NARSTO, 2010, Technical Challenges of Multipollutant Air Quality Management; Hidy, G.M.,
8 J.R. Brook, K.L. Demerjian, L.T. Molina, W.T. Pennell, and R.D. Scheffe (eds).
9 Springer, Dordrecht, 2010.
10 Sickles, J.E. and D.S. Shadwick, 2007, Changes in air quality and atmospheric deposition in the
11 Eastern United States: 1990- 2004, JGR, 112, D17301
12 Zhang, L., J.R. Broo, R. Vet, C. Mihele, M. Straw, J.M. O'Brien and S. Iqbal, 2005, Estimation
13 of contributions of NO2 and PAN to total atmospheric deposition of oxidized nitrogen
14 across Eastern Canada. Atmospheric Environment 39 (2005) 7030-7043.
September, 2010 8-12 Draft - Do Not Quote or Cite
-------
Mean 3pec;cs Breakdowns ol N Concern re
an Species Breakdowns ol N DC DO si tic
PAN
PANX
NTR
NO
NO3
N02
MONO
HNO3
DQEP_NQ3
DDEP_PANT
DDEPJJTR
DDEP.HN03
DDEF
DDEP.N02
DDEPJJO
2 Figure 8-1 Annual 2002 - 2004 CMAQ derived annual average fraction of ambient
3 concentrations (above) and deposition (below) of individual NOy species
4 delineated by the Adirondack and Shenandoah case study areas and the
5 remainder of the Eastern U.S. domain.
September, 2010
8-13
Draft - Do Not Quote or Cite
-------
Jan Fed Mar Apr May Jun Jul Aug se-p Get Nov Oec
Note Oct. and Nov. NO. vakws ata
i :•:>
Ft® Mar (iff May Jun Jii! *ug Sep OO Ni>« Dec.
Jan and Feb PA.M values are
3 Figure 8-2 Examples of the Relative Abundance of Several NOy Species Measured at Two
4 Rural Southeastern Canadian Sites as a Fraction of the Total Measured NOy
5 Concentration — Kejimkujik, NS, (top) and Egbert, ON, (bottom) during 2003.
6 Although both sites are in rural locations, the Kejimkujik, NS site represents more
7 aged air masses as it lies considerably further downwind from major sources of
8 NOx relative to the Egbert site. (Source: NARSTO, 2010)
September, 2010
8-14
Draft - Do Not Quote or Cite
-------
Layer 1 NO2[1]/ NOY[1]
[1 ]=cctrn_N1 a_2005af_05b_v3.4rjeta3.12EUS1 .yearly.avgaconc.sumdepl
231
221 •
211
201 •
191 •
181
171 •
161 •
1 51
141 •
131 '
121
111 •
1 01
91 •
81 '
71
61
51 •
41
31 •
21
11
1
32
63
94 125 156
April 5, 3 00:00:00 UTC
Min (5, 10) = 0.093, Max (10,34) - 0.819
187 218
2 4 5
0 a 1 9
0.723
H
0.637
0.547
10.456
0 365
0.2754
0.184^
3 Figure 8-3
4
Annual average fraction of NOy ambient air contributed by NOi based on
2005 CMAQ Eastern U.S. simulation at 12 km grid cell resolution.
September, 2010
8-15
Draft - Do Not Quote or Cite
-------
Layer 1 HNO3[1]/ NOY[1]
[1 ]=cctm_N1 a_2005af_05b_v3.4t)eta3.12EUS1 .yearly.avgaconc.sumdepl
231
221 •
211
201 •
191 •
181
171 •
161 •
1 51
141 •
131 '
121
111 •
1 01
91 •
81 '
71
61
51 •
41
31 •
21
11
1
1 32 63 94 125 156
April 5,3 00:00:00 UTC
Mil. (2, 235) = 0.005, Max (279, 24) = 0.376
187 218
24'J
0 376
0.330
H
0.237
10.190
0.051 -
3 Figure 8-4
4
Annual average fraction of NOy ambient air contributed by HNOs based on
2005 CMAQ Eastern U.S. simulation at 12 km grid cell resolution.
September, 2010
8-16
Draft - Do Not Quote or Cite
-------
Layer 1 PAN[1] / NOY[1]
[1 ]=cctm_N1 a_2005af_05b_v3.4tieta3.12EUS1 .yearly.avgaconc.sumdepl
231
221 •
211
201 •
191 •
181
171 •
161 •
1 51
141 •
131 '
121
111 •
1 01
91
81 '
71
61
51 •
41
31 •
21 •
11
1
V
1 32 63 94 125 156 187
April 5, 3 00:00:00 UTC
Min (136,45) = 0.003, Max (4,176) = 0.285
218 249
0.285
0.249
0.214
H
=- 0.144
0 IMS
0.073J
0 U ;::):-
3 Figure 8-5
4
Annual average fraction of NOy ambient air contributed by PAN based on
2005 CMAQ Eastern U.S. simulation at 12 km grid cell resolution.
September, 2010
8-17
Draft - Do Not Quote or Cite
-------
Mean Species Breakdowns of S Can-can I rang
SO2
Moon SpSSies Breakdowns oJ S Dop
DDEP_S02
2 Figure 8-6 Annual 2002 - 2004 CMAQ derived annual average fraction of ambient
3 concentrations (above) and deposition (below) of individual SOx species
4 delineated by the Adirondack and Shenandoah case study areas and the
5 remainder of the Eastern U.S. domain.
September, 2010
8-18
Draft - Do Not Quote or Cite
-------
1 9 INITIAL CONCLUSIONS
2
3 Staff initial conclusions on the elements of the secondary NOx and SOx standards for the
4 Administrator's consideration in making decisions on the secondary NOx and SOx standards are
5 summarized below, together with supporting conclusions from previous chapters. We recognize
6 that selecting from among alternative policy options will necessarily reflect consideration of
7 qualitative and quantitative uncertainties inherent in the relevant evidence and in the assumptions
8 of the quantitative exposure and risk assessments. Any such standard should protect public
9 welfare from any known or anticipated adverse effects associated with the presence of the
10 pollutant(s) in the ambient air, "whether caused by transformation, conversion, or combination
11 with other air pollutants." CAA § 302(h). . In providing these options for consideration, we are
12 mindful that the Act requires standards that, in the judgment of the Administrator, are requisite to
13 protect public welfare. The standards are to be neither more nor less stringent than necessary.
14 Our focus in this review on ecosystems that are both sensitive to acidification and responsive to
15 atmospheric acid deposition is intended to ensure that the resulting standards are appropriately
16 protective and not more protective than necessary in ecosystems that are not adversely affected
17 by acid deposition.
18 To evaluate whether the current secondary NAAQS is adequate or whether consideration
19 of revisions is appropriate, the conclusions and options for the Administrator to consider in this
20 review are based on effects-, exposure- and risk-based considerations. The exposure and risk
21 assessments reflect the availability of new tools, assessment methods, and a larger and more
22 diverse body of evidence than was available in the last reviews. We have taken a weight of
23 evidence approach that evaluates information across the variety of research areas described in the
24 ISA and in addition includes assessments of air quality, exposures, and qualitative and
25 quantitative risks associated with alternative air quality scenarios.
26 Staff notes that since the last review, additional policy-relevant developments have
27 occurred that may also warrant consideration by the Administrator when making decisions about
28 what is requisite to protect public welfare. The NRC report (described in Chapter 5) states:
29 "Whatever the reason that led EPA to use identical primary and secondary NAAQS in the past, it
30 is becoming increasingly evident that a new approach will be needed in the future. There is
September, 2010 9-1 Draft -Do Not Quote or Cite
-------
1 growing evidence that the current forms of the NAAQS are not providing adequate protection to
2 sensitive ecosystems and crops" (NRC, 2004).
3 The last review raised the following key issues as a rationale for not setting a separate
4 standard for NOx to protect against acidification and nutrient enrichment effects in sensitive
5 ecosystems:
6 1) Lack of enough consistent information to support a revision of the current secondary
7 standard to protect these aquatic systems.
8 2) Lack of adequate quantitative evidence on the relationship between deposition rates
9 and environmental impacts
10 3) Significant uncertainties with regard to the long-term role of nitrogen deposition in
11 surface water acidity and with regard to the quantification of the magnitude and timing of the
12 relationship between atmospheric deposition and the appearance of nitrogen in surface water.
13 In this current review, staff concludes that important new information has become
14 available since the last review that supports revising the current NOx and SOx standards.
15 Specifically, the ISA has concluded that there are causal relationships between NOx and SOx
16 acidifying deposition and effects on aquatic and terrestrial ecosystems, and the ISA and REA
17 provide substantial quantitative evidence of effects occurring in locations that meet the current
18 NO2 and SO2 standards. In addition, substantial new information, based on observational data
19 and rigorous atmospheric modeling, has become available regarding the role of both nitrogen and
20 sulfur deposition in acidification of sensitive water bodies. This information is sufficient to
21 inform the development of revised secondary standards for NOx and SOx to protect against the
22 effects of acidification. While there is also new information available on the role of nitrogen
23 deposition on nutrient enrichment effects in terrestrial and aquatic ecosystems, and the ISA
24 concludes there is a causal relationship between NOx and nutrient enrichment effects, for this
25 draft policy assessment, staff have focused on aquatic acidification effects due to the
26 substantially greater amount of information available to inform the development of secondary
27 standards for those effects. This is consistent with the available science and data, and also with
28 the recommendations of CAS AC, which stated that "EPA Staff is advised to focus on an AAPI
29 standard driven by aquatic effects concerns." There is not sufficient information at this time to
30 develop secondary standards directly focused on protection of sensitive terrestrial ecosystems
31 from acidification or sensitive terrestrial and aquatic ecosystems from adverse effects from
September, 2010 9-2 Draft -Do Not Quote or Cite
-------
1 nutrient enrichment. Establishing standards that are multipollutant and ecologically relevant is
2 an inherently complex process. We note that the aquatic acidification based standards are an
3 important step in providing additional protections for sensitive ecosystems. While they do not
4 provide complete protection for all sensitive ecosystems against all adverse effects, they will
5 likely result in reductions in NOy and SO2 across broad regions of the U.S., resulting in
6 decreased deposition and related effects for both terrestrial and aquatic ecosystems across all
7 types of effects (see Chapter 6 for a broader discussion of this issue).
8 Staff highlights the progress made in considering the j oint nature of ecosystem responses
9 to acidifying deposition of NOx and SOx, and notes that the ability to consider revisions to the
10 NOx and SOx secondary standards has been enhanced by our ability to consider a joint standard
11 for NOX and SOX to protect against aquatic acidification effects. The development of an
12 appropriate form of the standard linked to a common indicator of aquatic acidification, ANC, is
13 also a significant step forward, as it allows for development of a standard for aquatic
14 acidification designed to provide generally the same degree of protection across the country,
15 while still reflecting the underlying variability in ecosystem sensitivity to acidifying NOx and
16 SOx deposition.
17
18 9.1 CONSIDERATION OF ALTERNATIVE STANDARDS
19
20 We begin by noting that the existing evidence continues to support existing NO2 and SO2
21 standards to protect against adverse effects associated with direct exposure of vegetation to gas
22 phase NOx and SOx. The ISA concluded that there was sufficient evidence to infer a causal
23 relationship between exposure to SO2, NO, NO2 and PAN and injury to vegetation. Additional
24 research on acute foliar injury has been limited and there is no evidence to suggest foliar injury
25 below the levels of the current secondary standards for SOx and NOx. There is sufficient
26 evidence to suggest that the levels of the current standards are likely adequate to protect against
27 direct phytotoxic effects. As such, staff concludes that retaining the existing NO2 and SO2
28 standards to continue protection against these effects is appropriate. However, as discussed in
29 Chapter 4, we also conclude that the existing secondary NOx and SOx standards are not adequate
30 to provide protection of aquatic and terrestrial ecosystems against the effects of acidifying
31 deposition. In response to this conclusion, staff considers a second overarching question:
September, 2010 9-3 Draft -Do Not Quote or Cite
-------
1
2 What additional NOx and SOx standards are supported by the currently available scientific
3 evidence and risk-based information, as reflected in the ISA and REA ?
4
5 To inform the answer to this overarching question, we have posed a series of more specific
6 questions to aid in considering how the current NOx and SOx standards might be revised to
7 provide requisite public welfare protection.
8 Chapter 5 provided a conceptual framework for a secondary standard that is designed to
9 provide protection of ecosystems against the effects associated with deposition of ambient
10 concentrations of NOx and SOx. Chapter 5 also provided a discussion of potential options for the
11 elements of a standard based on that conceptual framework, with a focus on the form of the
12 standards. While we recognize the potential for significant impacts of current levels of NOx and
13 SOx on terrestrial ecosystems and the effect of current levels of NOx on nutrient enrichment in
14 sensitive aquatic ecosystems, we conclude that the currently available information is insufficient
15 to develop either individual or joint standards to protect against these effects. We note that
16 development of a standard for protection against terrestrial acidification may be appropriate
17 using the same structure as we are proposing for aquatic acidification, using the Bc:Al ratio as
18 the ecological indicator. However, the data needed to parameterize the form of such a standard
19 is not sufficient at this time. As a result, we conclude that the current state of knowledge
20 supports a standard to protect against the adverse effects associated with acidification of aquatic
21 ecosystems. Such a standard is likely to provide some level of protection against other endpoints
22 associated with deposition of N and S, but is not likely to adequately protect all sensitive
23 terrestrial ecosystems or all N nutrient sensitive aquatic ecosystems.
24 Building on the options discussed in Chapter 5, this section offers staff conclusions
25 regarding the elements of the standard, including the indicators for NOx and SOx, the form of
26 the standard, the averaging times, and presents for consideration options for target ANC levels
27 associated with protection against specific ecological effects in aquatic ecosystems. Ultimately,
28 the levels of AAPI considered by the Administrator should incorporate consideration of target
29 levels of ANC, percent of waterbodies protected within defined spatial areas, and trajectories for
30 ecosystem recovery, as well as uncertainties in the components of the AAPI. Associated with
31 these elements of the AAPI are sets of NOy/(SO2+SO4) tradeoff curves which determine the
September, 2010 9-4 Draft -Do Not Quote or Cite
-------
1 levels of ambient NOy and SO2+SO4 which will satisfy the level of the standard. The
2 expression of these trade-off curves embodies the deposit onal load for a specified spatial area
3 that is equivalent to the critical load for the waterbody in that area that represents a selected
4 percentile (e.g. 95th percentile) of critical loads across waterbodies in the area, such that the
5 selected percent of waterbodies in that area are expected to achieve an ANC at the target level.
6 If a target load for a specific temporal period (e.g. by 2030) is evaluated instead of the critical
7 load, then the deposit!onal load represents the amount of deposition that is expected to achieve a
8 target ANC value by a specific year for the selected percent of waterbodies in the area. The
9 equivalent AAPI for an area can be calculated by inputting the values for each parameter of the
10 AAPI equation for the selected percentile waterbody and the observed values of NOy and
11 SO2+SO4.
12 These elements will be considered collectively in evaluating the protection from welfare
13 effects associated with aquatic acidification afforded by alternative standards under
14 consideration. In considering the currently available scientific and technical information, we
15 consider both the information available in the last review and information that is newly available
16 since the last review as assessed and presented in the ISA and RA prepared for this review (US
17 EPA, 2008; US EPA, 2009).
18
19 9.1.1. Indicators
20
21 Staff concludes that the appropriate indicators for NOx and SOx, as described in detail in
22 Chapter 5, are total NOy and the sum of SO2 and SO4, respectively. Total NOy includes all
23 nitrogen oxides, including e.g. total reactive oxidized atmospheric nitrogen, defined as NOx (NO
24 and NO2) and all oxidized NOX products: NOy = NO2 + NO + HNO3 + PAN +2N2O5 +
25 HONO+ NO3 + organic nitrates + particulate NO3. The sum of SO2 and SO4 constitutes
26 virtually the entire ambient air sulfur budget and SO2 and SO4 are measured routinely in
27 existing monitoring networks.
28
29 9.1.2. Averaging times
30
September, 2010 9-5 Draft -Do Not Quote or Cite
-------
1 As noted in Chapter 2 and Chapter 5, episodes of acidification and chronic acidification
2 levels are associated with deposition over longer periods of time due to the storage and release of
3 deposited nitrogen and sulfur in soils, snow, and ice. As a result, while episodic acidification
4 may occur on much shorter timeframes, e.g. days to weeks, the cause of these episodes is largely
5 due to shifts in hydrological flow paths, and the impact of these episodes is still determined by
6 long term deposition of NOy and SO2+SO4 and associated long-term ANC, and thus the
7 appropriate averaging time for ambient NOy and SO2+SO4 will be longer term. The averaging
8 time for ambient concentrations of NOy and SO2+SO4 should be reflective of the long-term
9 cumulative nature of deposition. CAS AC supports using a three to five year averaging period
10 "to help smooth out the year-to-year climatic variation in air concentration and deposition
11 estimates." (CASAC, 2010)
12
13 9.1.3. Form
14
15 The "form" of a standard defines the air quality statistic that is to be compared to the
16 level of the standard in determining whether an area attains the standard. As discussed
17 extensively in chapter 5, staff concludes that the current forms of the NOx and SOx secondary
18 standards are not appropriate for addressing ecosystem acidification effects, and also concludes
19 that a form that combines NOx and SOx levels with information on ecosystem sensitivity and
20 nitrogen retention and uptake is most appropriate to maximize the likelihood of protecting
21 sensitive ecosystems from the effects of acidification, without requiring standards that are more
22 than requisite to provide that protection. Specifically, staff concludes that the Atmospheric
23 Acidification Protection Index form as described in Chapter 5 is best suited to provide for
24 protection against adverse effects due to acidifying deposition related to NOx and SOx.
25 Within the AAPI, it is also appropriate to consider the specification of values of non-air
26 quality parameters of the AAPI, including pre-industrial base cation weathering, nitrogen
27 retention and uptake, runoff, and levels of reduced nitrogen deposition. As discussed in Chapter
28 5, staff is proposing that the pre-industrial base cation weathering, nitrogen retention and uptake,
29 and runoff values be determined by assessing the critical loads associated with a specific
30 percentile of the waterbodies within defined spatial boundaries (as noted in Chapter 5,
31 consideration is being given to a number of different methods for defining spatial boundaries).
September, 2010 9-6 Draft -Do Not Quote or Cite
-------
1 The values of pre-industrial base cation weathering, nitrogen retention and uptake, and runoff
2 values for a selected percentile waterbody are then used as the values in the form of the standard
3 as realized for the specific ecoregion.
4 The value of reduced nitrogen is initially set using deposition of NHx modeled using the
5 CMAQ, evaluated for the period 2002-2005. Staff is considering the most appropriate spatial
6 averaging extent for reduced nitrogen. Figure 9-1 shows spatially interpolated values of reduced
7 nitrogen deposition based on 12km CMAQ modeling in the Eastern U.S. It is clear that in some
8 locations, there is significant heterogeneity in reduced nitrogen deposition within ecoregion III
9 boundaries. Given this information, two possible approaches to estimating reduced nitrogen
10 values in the AAPI algorithm are 1) average reduced nitrogen deposition within an ecoregion,
11 acknowledging that this will lead to uncertainties in the level of protection associated with levels
12 of ambient NOy and SOx, or 2) allow for additional spatial refinement of sensitive areas to
13 reflect the heterogeneity of reduced nitrogen deposition. This will result in multiple
14 parameterizations of the AAPI within a single ecoregion.
September, 2010 9-7 Draft -Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Figure 9-1. Spatially interpolated CMAQ estimates of deposition of reduced nitrogen
(2002-2004 average)
Unlike other parameters in the AAPI, reduced nitrogen is expected to change significantly over
time because of the largely anthropogenic sources of reduced nitrogen deposition. In order to
address this potential, we are exploring methods for specifying the standards in a way that would
provide for updating_the values of reduced nitrogen in the AAPI based on new modeling of
reduced nitrogen deposition, following requirements for modeling established by EPA.
September, 2010
9-8
Draft -Do Not Quote or Cite
-------
1 9.1.4. Considerations in Defining Options for the AAPI Standard
2 Conclusions regarding an appropriate range of levels of the AAPI standard will be
3 informed by considerations of the levels of the target ecological indicator ANC (related to levels
4 of protection from effects of chronic and episodic acidification), relationships between target
5 ANC levels and trajectories of recovery over time, and uncertainties in the various elements of
6 the AAPI that affect the likelihood that the level of protection intended by a particular target
7 ANC will be realized when atmospheric concentrations of NOy and SO2+SO4 fall below the
8 tradeoff curve for that target ANC. Chapter 5 provides an assessment of the information related
9 to selection of a target ANC, and the rationale for focusing additional considerations on specific
10 target ANC levels, including 20 and 50 |ieq/L.
11 The "levels" of the ambient indicators are determined by the selection of a level for the
12 AAPI, as they represent the quantities of the ambient indicators that will result in the specified
13 level of the AAPI. Those levels will vary for different locations depending on the non-
14 atmospheric related characteristics of the ecosystem, the level of reduced nitrogen in the
15 ecosystem, and the atmospheric transformation ratios (TNOy and T(So2+so4))-
16 The secondary NAAQS will reflect the public welfare policy judgments of the
17 Administrator, based on the science, as to the level of air quality which is requisite to protect the
18 public welfare from any known or anticipated adverse effects associated with the pollutant in the
19 ambient air. The exposure and risk assessment provide information regarding the effects
20 associated with a number of different welfare endpoints at different levels of air quality,
21 expressed in terms of the joint multiyear mean concentrations of NOy and SO2+SO4 determined
22 such that specific levels of ecosystem protection (for example, ANC greater than 50 jieq/L) are
23 met. Staff also recognizes that in certain naturally acidic ecosystems, even though the ecological
24 benchmarks are exceeded, e.g. ANC may be quite low; NOy and SO2+SO4 are not contributing
25 to effects because those systems have chronic natural acidity and will not benefit from reductions
26 in atmospheric deposition. The secondary NAAQS are not intended to provide protection in
27 these types of naturally acidic systems. As a result, in our determination of appropriate lakes and
28 streams to include in the populations of critical loads that determine protective NOy and
29 SO2+SO4 levels, we apply filters to remove lakes and streams that are naturally acidic or
30 acidified due to mine drainage. The secondary NAAQS are focused on providing protection in
September, 2010 9-9 Draft -Do Not Quote or Cite
-------
1 areas where ambient NOy and SO2+SO4 are resulting in effects in ecosystems with low natural
2 levels of acidification that are highly sensitive to additional inputs of acid deposition.
3 Staff concludes that ecosystem effects of NOy and SO2+SO4 deposition in aquatic
4 ecosystems are an important public welfare effect of concern, based on the types and extent of
5 ecosystem services likely to be affected by deposition, as well as the location of some aquatic
6 ecosystems within state and national protected lands, including Class I national parks and
7 wilderness areas (Chapter 3).
8
9 9.1.4.1 Target ANC Level
10 In reaching staff conclusions regarding target ANC levels that are appropriate to consider for an
11 AAPI-based standard, staff take into account the currently available scientific information
12 including: evidence from field and laboratory studies, including evidence of effects in highly
13 sensitive ecosystems and estimates of risk reductions associated with alternative annual standard
14 levels, as well as the related limitations and uncertainties associated with this information as
15 presented and discussed more fully in the ISA and RA (US EPA, 2008; US EPA, 2009). In
16 developing conclusions regarding the target ANC, we evaluated both evidence and risk based
17 information. In addition, we consider information on target ANC and pH levels used by other
18 organizations that have established critical loads for protection of aquatic ecosystems from
19 effects of acidification.
20 We conclude that it is appropriate to define NOX and SOX standards that will provide
21 generally consistent protection for acid sensitive lakes and streams across the country. In order
22 to do so, we focus attention on considering target ANC levels that will lead to a level of AAPI
23 that provides the same protection for sensitive aquatic ecosystems throughout the U.S. The
24 result of focusing on a nationally protective level of AAPI is a set of varying NOy/(SO2+SO4)
25 tradeoff curves across the U.S. reflecting that NOy and SO2+SO4 affect acidification in different
26 ways depending on underlying ecosystem characteristics and levels of reduced nitrogen, such
27 that the same AAPI is calculated with differing levels of NOy and SO2+SO4. This approach
28 recognizes that changes in air quality to meet the standards may reflect differing combinations of
29 NOy and SO2+SO4 leading to the same level of the AAPI.
30 Based on our analyses of risks of impacts on aquatic species diversity and fitness and on
31 the basis of the scientific effects literature, we conclude that achieving a target ANC of 50 jieq/L
September, 2010 9-10 Draft -Do Not Quote or Cite
-------
1 would substantially decrease the effects of acidification due to NOy and SO2+SO4 on aquatic
2 ecosystems, decreasing the risk of losses in biodiversity and mortality in fish and other aquatic
3 organisms, and improving the overall health of aquatic ecosystems. Additionally, it is
4 anticipated that achieving a target ANC of 50 jieq/L would provide increased protection from
5 NOy and SO2+SO4 in areas with higher levels of variability in ecosystem sensitivity due to
6 variability in meteorology, bedrock geology, topography, land use characteristics, or reduced
7 nitrogen deposition.
8 It is recognized, however, that a achieving a target ANC of 50 |ieq/L would likely not
9 protect the most sensitive aquatic ecosystems or species within those ecosystems from the effects
10 of NOy and SO2+SO4. At ANC levels below 100 |ieq/L, while overall health of an aquatic
11 community can be maintained, ANC levels are expected to be such that fish fitness and
12 community diversity begin to decline. At ANC levels between 100 and 50 |ieq/L, ANC levels
13 are expected to be such that the fitness of sensitive species (e.g., brook trout, zooplankton) also
14 begins to decline.
15
16 9.1.4.2 Target Percent of Waterbodies to Meet a Target ANC Level
17 The appropriate range of levels of AAPI is also informed by the selection of a target
18 percent of protection for waterbodies within particular acid sensitivity classes or ecoregions (see
19 Chapter 5). More specifically, the greater percentage of waterbodies that are to be protected
20 generally indicates a greater likelihood that sensitive waterbodies will achieve a target ANC
21 level, and as such, in setting the level of the standard, there will be less need to reflect
22 uncertainty in the likelihood that those sensitive waterbodies will be protected.
23 It is also important to consider that while the target ANC level will not be met in all
24 sensitive waterbodies if a target percent less than 100 is selected, all waterbodies will realize
25 some level of protection due to decreases in NOy and SO2+SO4 to meet the target ANC in
26 targeted waterbodies.
27 Additional analyses of the implications of alternative target percentages of waterbodies
28 are underway and are expected to be completed for inclusion in the final PA.
29
30 9.1.4.2 Additional Considerations Related to Developing Options for the AAPI Form
31 of the Standard
September, 2010 9-11 Draft -Do Not Quote or Cite
-------
1 In developing options for the standard, consideration should of the degree to which any
2 specific AAPI would lead to achieving the desired ANC level, and a judgment as to the degree of
3 protection of public welfare that is warranted. These considerations should incorporate a wide
4 number of factors, including the percent of water bodies within acid sensitive areas that the
5 Administrator determines should be protected at the targeted ANC level, as well as consideration
6 of achieving desired levels of protection within generational timeframes, e.g., 20 to 40 years,
7 concerns about protection against episodic acidification events, and uncertainties in the modeling
8 of critical loads, nitrogen uptake and retention, reduced nitrogen deposition, and relationships of
9 atmospheric concentrations to deposition. In considering options for the standard that would
10 reflect consideration for providing requisite welfare protection against know or anticipated
11 effects, we believe that while the available information is insufficient to set separate standards
12 for terrestrial acidification and aquatic and terrestrial nutrient enrichment effects, it is also
13 appropriate to consider the evidence of those effects and the likelihood for co-protection
14 provided by standards targeted at protection against effects of aquatic acidification (Chapter 6).
15 Chapter 7 provided a summary and synthesis of critical uncertainties and implications for
16 the standards. While many uncertainties cannot be quantitatively assessed, and as such cannot be
17 used to recommend specific quantitative changes to the AAPI, there are several uncertainty
18 analyses which give some insight into the magnitude and direction of the uncertainty. For
19 example, uncertainty analysis of the MAGIC model of critical loads indicates that modeled pre-
20 industrial ANC (which informs the distribution of critical loads on which NOy and SO2+SO4
21 levels are based by establishing the natural ability of an aquatic ecosystem to neutralize acid
22 inputs) has 95 percent confidence intervals that are 10 percent higher (or lower) than the mean
23 estimate for lakes, and 5 percent higher (or lower) than the mean estimate for streams. Similar
24 uncertainty exists regarding overall uncertainty in the models used to generate critical loads for
25 determining the NOy and SO2+SO4 tradeoff curves.
26 The deposition of reduced nitrogen is a critical input to the AAPI and has a large
27 expected uncertainty due to the use of CMAQ modeling which relies on uncertain chemistry and
28 uncertain inventories of ammonia emissions. Much of the uncertainty introduced by reduced
29 nitrogen deposition can be decreased by improvements in measurements of reduced nitrogen
30 deposition and improvements in the emissions inventories of ammonia and characterization of
31 ammonia chemistry within the CMAQ modeling. These improvements are underway, and our
September, 2010 9-12 Draft -Do Not Quote or Cite
-------
1 form of the standard is designed to allow for dynamic updates to reduced nitrogen deposition
2 parameters.
O
4
5 9.1.5 Additional protections for ecosystems against the effects of terrestrial acidification
6 and terrestrial and aquatic nutrient enrichment
7
8 While we are not basing the elements of this standard primarily on consideration of
9 effects other than aquatic acidification, our approach recognizes that some level of protection
10 against effects of acidification in terrestrial ecosystems and effects of nutrient enrichment in
11 terrestrial and aquatic ecosystems is likely to be realized through changes in air quality to meet
12 the AAPI standard. We recognize that an annual standard focused on aquatic acidification
13 cannot be expected to offer protection against all of the welfare effects from NOy and SO2+SO4,
14 especially in areas that are sensitive to nutrient enrichment but are not acid sensitive. However,
15 based on the information available in the ISA and REA, we conclude that the available
16 information is not sufficient to set a complementary standard to provide protection against
17 additional effects of NOy through nutrient enrichment, and that additional research is necessary
18 to support the setting of such a standard, especially in the areas of identifying the specific
19 impacts of decreases in atmospheric nitrogen deposition (see chapter 7). CASAC has noted that
20 our current framework, with appropriate modifications, should be applicable to developing
21 standards to provide protection against acidification effects in sensitive terrestrial ecosystems.
22 However, staff has concluded that the current data is not sufficient to develop a separate
23 terrestrial acidification based standard at this time. A primary limitation is the identification of
24 specific levels of harm associated with the BC/A1 ecological indicator.
25 Sensitive terrestrial ecosystems that are located in watersheds with acid sensitive water
26 bodies are likely to receive the most protection under an aquatic acidification targeted standard.
27 Terrestrial ecosystems outside of these watersheds are likely to see some level of protection, but
28 will not realize targeted changes in ambient NOy and SO2+SO4 (Chapter 6).
29 Chapter 6 evaluated the relative protection for terrestrial ecosystems in areas from
30 meeting a target ANC of 50 jieq/L compared to meeting a target Bc:Al ratio of 10. Critical
31 loads for N and S were compared to determine which of the targets would result in a more
September, 2010 9-13 Draft -Do Not Quote or Cite
-------
1 stringent critical load. Over half of the watersheds had a lower critical load to meet the ANC
2 target compared to the critical load to meet the Be: Al ratio target. As a result, those watersheds
3 are likely to be protected from both terrestrial and aquatic impacts when the ANC target is met.
4 When the water bodies are more sensitive to deposition ("highly sensitive" or "moderately
5 sensitive"), the aquatic critical acid loads generally provide a greater level of protection against
6 acidifying nitrogen and sulfur deposition in the watershed.
7 The tradeoff curves for NOy and SO2+SO4 that are associated with protection against
8 aquatic acidification also provide bounding conditions for nitrogen that can be compared against
9 benchmarks of effects associated with nitrogen deposition in sensitive terrestrial and aquatic
10 ecosystems. Achieving an ANC level of 50 jieq/L for 90 percent of lakes and streams
11 nationwide would provide some protection against leaching in northeast forests, but would need
12 to be lower to protect California coastal sage scrub, lichens in mixed conifer forests, alpine lake
13 communities, and Minnesota grasslands (Section 6.2). In the case of aquatic nutrient
14 enrichment, comparison of maximum allowable NOx levels on the tradeoff curves with
15 deposition requirements to meet the Chesapeake Bay TMDL shows that without further
16 restrictions on NOy concentrations, standards set to protect against aquatic acidification will not
17 be protective against effects of aquatic nutrient enrichment in the Chesapeake Bay.
18
19 9.1.6 Summary of options
20
21 To facilitate evaluation of the elements of the standard and staff conclusions regarding
22 those elements, we have constructed a summary table showing the elements of the AAPI, options
23 for each element, and staff conclusions where appropriate.
24
25 [Table 9-1 to be provided]
26
27 9.2 CONCLUSIONS
28 The following secondary NAAQS conclusions encompass the breadth of policy-relevant
29 considerations described in this policy assessment. We note that staff conclusions to be
30 presented in the final PA will consider input received from CASAC and the public on this second
31 draft PA. We recognize that selecting from among alternative standards will necessarily reflect
September, 2010 9-14 Draft -Do Not Quote or Cite
-------
1 consideration of the qualitative and quantitative uncertainties inherent in the relevant evidence
2 and in the assumptions that underlie the quantitative risk assessment. In identifying these
3 alternative secondary standards and ranges of levels for consideration, we are mindful that the
4 Clean Air Act requires standards to be set that are requisite to protect public from known or
5 anticipated adverse effects, such that the standards are to be neither more nor less stringent than
6 necessary. Thus, the Act does not require that the NAAQS be set at no effect levels, but rather at
7 levels that avoid adverse effects on public welfare:
8 (1) Based on the policy-relevant findings from the ISA described in Chapter 2, and while
9 recognizing that important uncertainties and research questions remain, staff conclude that great
10 progress has been made since the last reviews of the secondary standards for NOx and SOx. We
11 generally find support in the available effects-based evidence for consideration of NOx and SOx
12 standards that are at least as protective as the current standard and do not find support for
13 consideration of NOx and SOx standards that are less protective than the current standard. The
14 staff also concludes that consideration of joint standards for NOx and SOx is appropriate given
15 the common atmospheric processes governing the deposition of NOx and SOx to sensitive
16 ecosystems, and given the combined effects of N and S deposition on acidification of soil and
17 water.
18 (2) Staff concludes that ambient NOx is a significant component of atmospheric nitrogen
19 deposition, even in areas with relatively high rates of deposition of reduced nitrogen. Staff make
20 this conclusion based on the analysis in Chapter 3 of the REA, which provides a thorough
21 assessment of the contribution of NOx to nitrogen deposition throughout the U.S., and the
22 relative contributions of ambient NOx and reduced forms of nitrogen.
23 (3) Staff concludes based on the case study results provided in the REA, that current
24 levels of NOx and SOx are associated with deposition that leads to ANC values below
25 benchmark values that cause ecological harm and losses in ecosystem services. Staff concludes
26 that the evidence and risk assessment support strongly a relationship between atmospheric
27 deposition of NOx and SOx and ANC, and that ANC is an excellent indicator of aquatic
28 acidification. Staff also concludes that at levels of deposition associated with NOX and SOX
29 concentrations at or below the current standards, ANC levels are expected to be below
30 benchmark values that are associated with significant losses in fish species richness, which is
31 associated with reductions in recreational fishing services. Although there are many other
September, 2010 9-15 Draft -Do Not Quote or Cite
-------
1 ecosystem services potentially affected by reductions in ANC, including subsistence fishing,
2 natural habitat provision, and biological control, confidence in the specific translation of ANC
3 values to these additional ecosystem services is much lower.
4 (4) Losses in aquatic resources associated with ANC levels below 50 jieq/L are clearly
5 associated with significant losses in economic value. Based on the best available data, just in
6 New York., increasing ANC levels to 50 in the Adirondacks is estimated to result in $300 to
7 $800 million in annual benefits in 2006 dollars. This estimate represents only a fraction of the
8 total economic value of ecosystem damages as many impacted resources are not amenable to
9 economic valuation methods. In addition, economic damages are also likely to occur in other
10 areas affected by acidification, including New England, the Appalachian Mountains (northern
11 Appalachian Plateau and Ridge/Blue Ridge region), and the Upper Midwest. Staff concludes
12 that reducing acidifying deposition of NOx and SOx will result in improvements in public
13 welfare by increasing the quantity and quality of ecosystem services, including recreational
14 fishing and other services associated with improved water quality.
15 (5) Staff initially concludes based on the case study results that current levels of ambient
16 NOx and SOx are associated with deposition that leads to Bc:Al values below benchmark values
17 that cause ecological harm and losses in ecosystem services. Staff concludes that the evidence
18 and risk assessment support strongly a relationship between atmospheric deposition of NOx and
19 SOx and Bc:Al, and that Be: Al is a good indicator of terrestrial acidification. Staff also
20 concludes that at levels of deposition associated with NOX and SOX concentrations at or below
21 the current standards, Bc:Al levels are expected to be below benchmark values that are
22 associated with significant losses in tree health and growth, which are associated with reductions
23 in timber production. While there are many other ecosystem services, including maple syrup
24 production, natural habitat provision, and regulation of water, climate, and erosion, potentially
25 affected by reductions in Bc:Al, confidence in the specific translation of Bc:Al values to these
26 additional ecosystem services is much lower.
27 (6) On the basis of the acidification and nutrient enrichment effects that have been
28 observed to still occur under current ambient conditions and those predicted to occur under the
29 scenario of just meeting the current secondary NAAQS, staff concludes that the current
30 secondary NAAQS are inadequate to protect the public welfare from known and anticipated
31 adverse welfare effects from aquatic and terrestrial acidification associated with deposition of
September, 2010 9-16 Draft -Do Not Quote or Cite
-------
1 NOx and SOx.. As discussed above, this conclusion derives from several lines of evidence. Staff
2 also concludes that the current NOx and SOx secondary standards are adequate to protect against
3 direct gas-phase effects on vegetation, and as such, should be retained to preserve protection
4 against these welfare effects.
5 (7) Staff has concluded, based on the completeness of the available evidence and
6 quantitative risk information, that effects due to aquatic acidification are most suitable for
7 defining additional secondary standards for NOx and SOx. Staff notes that in developing a
8 standard designed to protect against the effects of aquatic acidification due to deposition of NOx
9 and SOx, the resulting standards may not provide protection against known effects associated
10 with terrestrial acidification and with nutrient enrichment in sensitive aquatic and terrestrial
11 ecosystems.
12 (8) It is appropriate to consider using indicators other than NO2 and SO2 as the indicators
13 for an additional standard that is intended to address the ecological effects associated with
14 deposition of NOx and SOx to sensitive ecosystems. Given the reasons discussed in Chapters 2,
15 4, and 5 of this policy assessment, staff concludes that NOx (oxides of nitrogen, the definition in
16 section 302 (v) of the CAA), is best represented by the atmospheric indicator NOy, defined as
17 NO2 + NO + HNO3 + PAN +2N2O5 + HONO+ NO3 + organic nitrates + paniculate NO3 is the
18 more appropriate indicator of oxides of nitrogen, and that SO2+SO4 is the more appropriate
19 indicator of oxides of sulfur for purposes of a secondary standard addressing aquatic
20 acidification.
21 (9) It is appropriate to use multi-year averages of concentrations of NOy and SO2+SO4 as
22 the averaging times for the secondary standards, based on the chronic nature of acidification, and
23 the protection against episodic acidification provided by a standard based on annual average
24 concentrations. Averaging periods in the range of 3 to 5 years are most appropriate.
25 (10) It is appropriate to consider adding a new standard with a different form for NOy
26 and SO2+SO4 as the current form does not take into account the linkages between NOx and SOx
27 in the causation of effects associated with acidification of aquatic ecosystems. Based on the
28 causal linkages between NOX and SOX, deposition of N and S, and the indicator of acidification,
29 ANC, staff concludes that a standard with a form specified as an atmospheric acidification
30 protection index (AAPI) should be added. A standard with this form would reflect the important
September, 2010 9-17 Draft -Do Not Quote or Cite
-------
1 roles of underlying ecosystem characteristics, determinants of deposition, and deposition of
2 reduced nitrogen in determining the potential effects from deposition of NOx and SOx.
3 (11) Staff has concluded, based on the evidence and risk based information, and
4 consideration of information related to definitions of adversity, that
5 a) a target level of ANC of 20 jieq/L will protect against significant losses in
6 fish mortality in many sensitive lakes, but will place less weight on protection against
7 losses in aquatic biodiversity, and will be less protective against potential acidification
8 episodes,
9 b) a target level of ANC of 50 |ieq/L will protect against significant mortality
10 in aquatic organisms and loss of fish health and biodiversity in sensitive lakes and
11 streams, and will give weight to considerations of uncertainties in the time to recovery of
12 aquatic ecosystems,
13 c) target levels of ANC above 50 jieq/L may provide additional protection
14 against declines in fitness of sensitive species (e.g., brook trout, zooplankton), however,
15 overall health of aquatic communities may not be impacted.
16
September, 2010 9-18 Draft -Do Not Quote or Cite
-------
1 9.3 REFERENCES
2
3 National Research Council. 2004. Air Quality Management in the United States. National
4 Academies Press, Washington, D.C.
5 CAS AC (2010). Review of the Policy Assessment for the Review of the Secondary National
6 Ambient Air Quality Standards for NOx and SOx: First Draft (March 2010). EPA-
7 CASAC-10-014. June 22.
8 US EPA (2008) U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen and
9 Sulfur Ecological Criteria (Final Report). U.S. Environmental Protection Agency,
10 Washington, D.C., EPA/600/R-08/082F
11 US EPA (2009) Risk and Exposure Assessment for Review of the Secondary National Ambient
12 Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur-Main Content - Final
13 Report. U.S. Environmental Protection Agency, Washington, D.C., EPA-452/R-09-008a
14
September, 2010 9-19 Draft -Do Not Quote or Cite
-------
1
2
3
4
5
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
APPENDIX A
ADDITIONAL INFORMATION FOR CHAPTER 5
A.lConceptual Design of the Standard
This is supplemental information to support the discussion of the conceptual
design of the standard that is presented in Chapter 5 of the Policy Assessment Document.
The aquatic acidification analyses developed in the REA used a number of different
models and calculation techniques that are important for the development of the standard.
The goal of this Appendix is to summarize information from the REA analysis that is
most relevant to the Policy Assessment. A brief summary of the REA analyses are
presented in section A. 1.1. In section A. 1.2 there is a general summary and technical
discussion of the critical loads modeling approaches that were used in the REA, followed
by a brief description of MAGIC model data requirements.
A.1.1 Technical summary of methods used in the REA Aquatic Acidification
analysis
The aquatic acidification analysis is presented in Chapter 4 and Appendix 4 of
the REA. The analysis uses multiple techniques to show the relationship between ANC
and NOx and SOx deposition, as well as determine the current level of risk to water
bodies that occur in sensitive areas. A brief summary of the techniques and objectives of
the REA analysis is given in Table A-l.
Table A-l. Brief summary of objects and methods used in the REA Aquatic Acidification
analysis.
Technique
Time-series
graphs of
current
conditions
MAGIC
Critical Loads
modeling
Objectives
1
1
2
3
1
Data from monitoring networks collected from 1990 to 2006 were
plotted to show trends in concentrations of pollutants, deposition and
acidification for each case study site. The data included surface water
concentration of nitrate, sulfate and ANC; deposition of sulfate and
nitrate; as well as air concentration of SOx, NOx and NH4
Used to estimate the relationship between ANC values and
anthropogenic NOx and SOx emission from the past (preacidification
-I860), present (2002 and 2006) and projected into the future (2020
and 2050). Analysis included 44 lakes from Adirondacks and 60
streams from Shenandoah.
Used to develop input parameters for critical loads modeling (i.e.
weathering rates)
Used for uncertainty analysis
SSWC and FAB models used to calculate critical loads for critical
limits of ANC = 0, 20, 50, 100
September, 2010
A-l
Draft - Do Not Quote or Cite
-------
Table A-l. Brief summary of objects and methods used in the REA Aquatic Acidification
analysis.
Regional
Extrapolation
2
3
1
2
Critical loads for ANC critical limits calculated for 169 lakes in the
Adirondacks and 60 streams in the Shenandoah using water quality
data from monitoring sites collected in 2006
Critical loads exceedences calculated by comparing the critical loads
that were calculated by SSWC with deposition data from NADP for
wet deposition and CMAQ for dry deposition, both for the year 2002
117 of the critical loads calculated for the Adirondacks were
extrapolated to lakes defined by the New England EMAP probability
survey, representing 1842 lakes, to infer the # of lakes that exceeded
their critical load
69 of the critical loads calculated for the Shenandoah were
extrapolated to 330 streams based on bed rock geology classification.
1
2
3
4
5
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
A.1.2 Technical summary of critical loads modeling in the REA
The critical load of acidity for lakes or streams was derived from present-day
water chemistry using a combination of steady-state models. Both the Steady-State Water
Chemistry (SSWC) model and First-order Acidity Balance model (FAB) is based on the
principle that excess base-cation production within a catchment area should be equal to or
greater than the acid anion input, thereby maintaining the ANC above a preselected level
(Reynolds and Norris, 2001; Posch et al. 1997). These models assume steady-state
conditions and assume that all SC>42 in runoff originates from sea salt spray and
anthropogenic deposition. Given a critical ANC protection level, the critical load of
acidity is simply the input flux of acid anions from atmospheric deposition (i.e., natural
and anthropogenic) subtracted from the natural (i.e., preindustrial) inputs of base cations
in the surface water. Final Risk and Exposure Assessment September 2009 Appendix 4,
Attachment A - 15 Aquatic Acidification Case Study Atmospheric deposition of NOx and
SOx contributes to acidification in aquatic ecosystems through the input of acid anions,
such as NO3- and SC>42 The acid balance of headwater lakes and streams is controlled by
the level of this acidifying deposition of NOs- and SC>42 and a series of biogeochemical
processes that produce and consume acidity in watersheds. The biotic integrity of
freshwater ecosystems is then a function of the acid-base balance, and the resulting
acidity-related stress on the biota that occupy the water. The calculated ANC of the
surface waters is a measure of the acid-base balance:
ANC = [BC]* - [AN]* (1)
September, 2010
A-2
Draft - Do Not Quote or Cite
-------
1 where [BC]* and [AN]* are the sum of base cations and acid anions (NCV and SC>42 ),
2 respectively. Equation (1) forms the basis of the linkage between deposition and surface
3 water acidic condition and the modeling approach used. Given some "target" ANC
4 concentration [ANClimit]) that protects biological integrity, the amount of deposition of
5 acid anions (AN) or depositional load of acidity CL(A) is simply the input flux of acid
6 anions from atmospheric deposition that result in a surface water ANC concentration
7 equal to the [ANClimit] when balanced by the sustainable flux of base cations input and
8 the sinks of nitrogen and sulfur in the lake and watershed catchment.
9
10 Critical loads for nitrogen and sulfur (CL(N) + CL(S) ) or critical load of acidity
11 CL(A) were calculated for each waterbody from the principle that the acid load should
12 not exceed the nonmarine, nonanthropogenic base cation input and sources and sinks in
13 the catchment minus a neutralizing to protect selected biota from being damaged:
14
15 CL(N) + CL(S) or CL(A) = BC*dep + BCw - Ecu - AN - ANClimit (2)
16
17 Where,
18 BC*dep = (BC*=Ca*+Mg*+K*+Na*), nonanthropogenic deposition flux of base cations
19 BCw = the average weathering flux, producing base cations
20 Ecu (Bc=Ca*+Mg*+K*) = the net long-term average uptake flux of base cations in the
21 biomass (i.e., the annual average removal of base cations due to harvesting)
22 AN = the net long-term average uptake, denitrification, and immobilization of nitrogen
23 anions (e.g. NO3") and uptake of SO42
24 ANClimit = the lowest ANC-flux that protects the biological communities.
25
26 Since the average flux of base cations weathered in a catchment and reaching the
27 lake or streams is difficult to measure or compute from available information, the average
28 flux of base cations and the resulting critical load estimation were derived from water
29 quality data (Henriksen and Posch, 2001; Henriksen et al., 1992; Sverdrup et al., 1990).
30 Weighted annual mean water chemistry values were used to estimate average base cation
31 fluxes, which were calculated from water chemistry data collected from the Temporally
32 Integrated Monitoring of Ecosystems (TIME)/Long-Term Monitoring (LTM) monitoring
September, 2010 A-3 Draft - Do Not Quote or Cite
-------
1 networks, that include Adirondack Longterm Monitoring (ALTM), Virginia Trout Stream
2 Sensitivity Study (VTSSS), and the Shenandoah Watershed Study (SWAS), and
3 Environmental Monitoring and Assessment Program (EMAP) (see Section 4.1.2.1 of
4 Chapter 4).
5
6 The preacidification nonmarine flux of base cations for each lake or stream, BC*0, is
7
8 BC*0 = BC*dep + BCw - Ecu (3)
9
10 Thus, critical load for acidity can be rewritten as
11
12 CL(N) + CL(S) = BC*0 - AN - ANClimit = Q.([BC*]0 - [AN] - [ANC]limit), (4)
13
14 where the second identity expresses the critical load for acidity in terms of catchment
15 runoff (Q) m/yr and concentration ([x] = X/Q). The sink of nitrogen in the watershed is
16 equal to the uptake (Nupt), immobilization (Nimm), and denitrification (Nden) of
17 nitrogen in the catchment. Thus, critical load for acidity can be rewritten as
18
19 CL(N) + CL(S) = (fNupt + (1 - r)(Nimm + Nden)} + ( [BC]0* - [ANClimit])Q (5)
20
21 where f and r are dimensionless parameters that define the fraction of forest cover in the
22 catchment and the lake/catchment ratio. The in-lake retention of nitrogen and sulfur was
23 assumed to be negligible. Equation 5 described the FAB model that was applied when
24 sufficient data was available to estimate the uptake, immobilization, and denitrification of
25 nitrogen and the neutralization of acid anions (e.g. NO3-) in the catchment. In the case
26 were data was not available, the contribution of nitrogen anions to acidification was
27 assumed to be equal to the nitrogen leaching rate (Nleach) into the surface water. The
28 flux of acid anions in the surface water is assumed to represent the amount of nitrogen
29 that is not retained by the catchment, which is determined from the sum of measured
30 concentration of NO3- and ammonia in the stream chemistry. This case describes the
31 SSWC model and the critical load for acidity is
32
33 CL(A) = Q.([BC*]0 - [ANC]limit) (6)
34
35 where the contribution of acid anions is considered as part of the exceedances calculation
36 (see Section 1.2.5, below). For the assessment of current condition in both case study
September, 2010 A-4 Draft - Do Not Quote or Cite
-------
1 areas, the critical load calculation described in Equation 6 was used for most lakes and
2 streams. The lack of sufficient data for quantifying nitrogen denitrification and
3 immobilization prohibited the wide use of the FAB model. In addition, given the
4 uncertainty in quantifying nitrogen denitrification and immobilization, the flux of
5 nitrogen anions in the surface water was assumed to more accurately reflect the
6 contribution of NO3- to acidification. Several major assumptions are made: (1) steady-
7 state conditions exist, (2) the effect of nutrient cycling between plants and soil is
8 negligible, (3) there are no significant nitrogen inputs from sources other than
9 atmospheric deposition, (4) ammonium leaching is negligible because any inputs are
10 either taken up by biota or adsorbed onto soils or nitrate compounds, and (5) longterm
11 sinks of sulfate in the catchment soils are negligible.
12
13 A. 1.2.1 Preindustrial Base Cation Concentration
14 Present-day surface water concentrations of base cations are elevated above their
15 steady state preindustrial concentrations because of base cation leaching through ion
16 exchange in the soil due to anthropogenic inputs of SC>42 to the watershed. For this
17 reason, present-day surface water base cation concentrations are higher than natural or
18 preindustrial levels, which, if not corrected for, would result in critical load values not in
19 steady-state condition. To estimate the preacidification flux of base cations, the present
20 flux of base cations was estimated,
21
22 BC*t, given by BC*t = BC*dep + BCw - Ecu +BCexc, (7)
23
24 Where BCexc = the release of base cations due to ion-exchange processes. Assuming
25 that deposition, weathering rate, and net uptake have not changed over time, BCexc can
26 be obtained by subtracting Equation 5 from Equation 7:
27
28 BCexc = BC*t-BC*0(8)
29
30 This present-day excess production of base cations in the catchment was related to the
31 long-term changes in inputs of nonmarine acid anions (ASO*2 + ANO3) by the F-factor
32 (see below):
September, 2010 A-5 Draft - Do Not Quote or Cite
-------
1
2 BCexc = F (ASO*2 + ANO3) (9)
3
4 For the preacidification base cation flux, solving Equation 5 for BC*0 and then
5 substituting Equation 8 for BCexc and explicitly describing the long-term changes in
6 nonmarine acid ion inputs:
7
8 BC*0 = BC*t - F (SO*4,t - SO*4,0 + NO*3,t - NO*3,0) (10)
9
10 The preacidification NO3- concentration, NO*3,0, was assumed to be zero.
11
12 A.l.2.2 F-factor
13 An F-factor was used to correct the concentrations and estimate preindustrial base
14 concentrations for lakes in the Adirondack Case Study Area. In the case of streams in the
15 Shenandoah Case Study Area, the preindustrial base concentrations were derived from
16 the MAGIC model as the base cation supply in 1860 (hindcast) because the F-factor
17 approach is untested in this region. An F-factor is a ratio of the change in nonmarine base
18 cation concentration due to changes in strong acid anion concentrations (Henriksen,
19 1984; Brakkeetal., 1990):
20
21 F =([BC*]t - [BC*]0)/([SO4*]t - [SO4*]0 + [NO3*]t - [NO3*]0), (12)
22
23 where the subscripts t and 0 refer to present and preacidification conditions, respectively.
24 If F=l, all incoming protons are neutralized in the catchment (only soil acidification); at
25 F=0, none of the incoming protons are neutralized in the catchment (only water
26 acidification). The F-factor was estimated empirically to be in the range 0.2 to 0.4, based
27 on the analysis of historical data from Norway, Sweden, the United States, and Canada
28 (Henriksen, 1984). Brakke et al. (1990) later suggested that the F-factor should be a
29 function of the base cation concentration:
30
31 F = sin (Ti/2 Q[BC*]t/[S]) (13)
September, 2010 A-6 Draft - Do Not Quote or Cite
-------
1
2 where
3 Q = the annual runoff (m/yr). [S] = the base cation concentration at which F=l; and for
4 [BC*]t>[S] F is set to 1. For Norway [S] has been set to 400 milliequivalents per cubic
5 meter (meq/m3)(circa.8 mg Ca/L) (Brakke et al., 1990). The preacidification SO42-
6 concentration in lakes, [SO4*]0, is assumed to consist of a constant atmospheric
7 contribution and a geologic contribution proportional to the concentration of base cations
8 (Brakke et al., 1989). The preacidification SO42- concentration in lakes, [SO4*]0 was
9 estimated from the relationship between [SO42-]o* and [BC]t* based on work completed
10 by Henriksen et al., 2002 as described by the following equation:
11
12 [SO42-]o* = 15 + 0.16* [BC]t* (14)
13
14
September, 2010 A-7 Draft - Do Not Quote or Cite
-------
1
2
Table A-2 Illustrates SSWC Approach - Environmental Variables
CL(A) = BC*dep + BCW - Bcu - ANClimit
CL(A) = Q'([BC*]o - [ANCliMt)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Variable
Code
BC dep
BCW
Bcu
ANClimit
Ca*
Mg*
Na*
K*
SO/
CL
SO/
N03*
Q
[BC*]0
[S0/]o
[N03*]0
F
Description
Sum (Ca*+Mg*+K*+Na*), nonanthropogenic
deposition flux of base cations
Average weathering flux of base cations
Sum (Ca+Mg+K), the net long-term average
uptake flux of base cations in the biomass
Lowest ANC-flux that protects the biological
communities
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Ca - (CL x
0.0213))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Mg - (CL
x 0.0669))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Na - (CL x
0.557))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (K - (CL x
0.0.0206))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (SO4 - (CL x
0.14))
Surface water concentration (ueq/L) growing
season average.
Surface water concentration (ueq/L) growing
season average.
Surface water concentration (ueq/L) growing
season average.
The annual runoff (m/yr)
Preindustrial flux of base cations in surface
water, corrected for sea salts
Preindustrial flux of sulfate in surface water,
corrected for sea salts
Preindustrial flux of nitrate, corrected for sea
salts
Calculated factor
Source
Wet NADP and Dry
CASTNET
Calculated (5-17)
USFS-FIA data
Set
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
USGS
Calculated from water
quality data
Estimated
Equal to 0
Fix values
September, 2010
A-8
Draft - Do Not Quote or Cite
-------
Table A-2 FAB Approach - Environmental Variables
DL(N) + DL(S) = |fNupt + (1 - r)(Nimm + Nden) + (Nret + Sret)| + ( [BC]0* - [ANClimit])Q
1
2
3
4
5
6
7
8
9
10
11
12
13
14
14
15
16
17
18
19
20
Variable
Code
Ndepo
ANClimit
[BC*]0
Ca*
Mg*
Na*
K*
SO/
CL
SO/
NO3*
Q
f
r
Nret
Sret
Nupt
J^imm
Nden
Lake Size
WSH
Description
Total N deposition
Lowest ANC-flux that protects the biological
communities
Preindustrial flux of base cations in surface
water, corrected for sea salt
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Ca - (CL x
0.0213))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Mg - (CL x
0.0669))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (Na - (CL x
0.557))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (K - (CL x
0.0.0206))
Sea Salt corrected Surface water concentration
(ueq/L) growing season average. (SO4 - (CL x
0.14))
Surface water concentration (ueq/L) growing
season average.
Surface water concentration (ueq/L) growing
season average.
Surface water concentration (ueq/L) growing
season average.
The annual runoff (m/yr)
f is a dimensionless parameter that define the
fraction of forest cover in the catchment
r is a dimensionless parameter that define the
lake/catchment ratio
The in-lake retention of nitrogen
The in-lake retention of sulfur
The net long-term average uptake flux of N in
the biomass
Immobilization of N in the soils
Denitrification
Lake size (ha)
Watershed area (ha)
Source
NADP/CMAQ
Set
Calculated from
water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
Water quality data
USGS
Estimated
Estimated
USFS-FIA data
Estimated fix value
Estimated fix value
DLMs
Calculated
1
2
September, 2010
A-9
Draft - Do Not Quote or Cite
-------
1
2 Data requirements for MAGIC
3 The MAGIC model (Cosby et al., 1985a; 1985b; 1985c) is a mathematical model
4 (a lumped-parameter model) of soil and surface water acidification in response to
5 atmospheric deposition based on process-level information about acidification. A process
6 model, such as MAGIC, characterizes acidification into (l)a section in which the
7 concentrations of major ions are assumed to be governed by simultaneous reactions
8 involving SC>42" adsorption, cation exchange, dissolution-precipitation- speciation of
9 aluminum, and dissolution-speciation of inorganic carbon; and (2) a mass balance section
10 in which the flux of major ions to and from the soil is assumed to be controlled by
11 atmospheric inputs, chemical weathering, net uptake and loss in biomass and losses to
12 runoff. At the heart of MAGIC is the size of the pool of exchangeable base cations in the
13 soil. As the fluxes to and from this pool change over time owing to changes in
14 atmospheric deposition, the chemical equilibria between soil and soil solution shift to
15 give changes in surface water chemistry. The degree and rate of change of surface water
16 acidity thus depend both on flux factors and the inherent characteristics of the affected
17 soils.
18 There are numerous input data required to run MAGIC making it rather data
19 intensive. Atmospheric deposition fluxes for the base cations and strong acid anions are
20 required as inputs to the model. These inputs are generally assumed to be uniform over
21 the catchment. The volume discharge for the catchment must also be provided to the
22 model. In general, the model is implemented using average hydrologic conditions and
23 meteorological conditions in annual simulations, i.e., mean annual deposition,
24 precipitation and lake discharge are used to drive the model. Values for soil and surface
25 water temperature, partial pressure of carbon dioxide and organic acid concentrations
26 must also be provided at the appropriate temporal resolution.
27 The aggregated nature of the model requires that it be calibrated to observed data
28 from a system before it can be used to examine potential system response. Calibrations
29 are based on volume weighted mean annual or seasonal fluxes for a given period of
30 observation. The length of the period of observation used for calibration is not arbitrary.
31 Model output will be more reliable if the annual flux estimates used in calibration are
September, 2010 A-10 Draft - Do Not Quote or Cite
-------
1 based on a number of years rather than just one year. There is a lot of year-to-year
2 variability in atmospheric deposition and catchment runoff. Averaging over a number of
3 years reduces the likelihood that an "outlier" year (very dry, etc.) is used to specify the
4 primary data on which model forecasts are based. On the other hand, averaging over too
5 long a period may remove important trends in the data that need to be simulated by the
6 model.
7 The calibration procedure requires that stream water quality, soil chemical and
8 physical characteristics, and atmospheric deposition data be available for each catchment.
9 The water quality data needed for calibration are the concentrations of the individual base
10 cations (Ca, Mg, Na, and K) and acid anions (Cl, SC>42", and N(V) and the pH. The soil
11 data used in the model include soil depth and bulk density, soil pH, soil cation-exchange
12 capacity, and exchangeable bases in the soil (Ca, Mg, Na, and K). The atmospheric
13 deposition inputs to the model must be estimates of total deposition, not just wet
14 deposition. In some instances, direct measurements of either atmospheric deposition or
15 soil properties may not be available for a given site with stream water data. In these
16 cases, the required data can often be estimated by: (a) assigning soil properties based on
17 some landscape classification of the catchment; and (b) assigning deposition using model
18 extrapolations from some national or regional atmospheric deposition monitoring
19 network. Soil data for model calibration are usually derived as aerially averaged values
20 of soil parameters within a catchment. If soils data for a given location are vertically
21 stratified, the soils data for the individual soil horizons at that sampling site can be
22 aggregated based on horizon, depth, and bulk density to obtain single vertically
23 aggregated values for the site, or the stratified data can be used directly in the model.
24
25
26 A.1.3 Example of the two ways to calculate NECO from the first draft NOx and SOx
27 secondary NAAQS Policy Assessment Document
28
29 The steady-state critical load model suggest for use in the NAAQS by the PA
30 could be constrained by a quantity of N which would be taken up, immobilized or
31 denitrified by ecosystems and used to adjust the quantity of deposition required to meet a
32 specified critical load. This term is abbreviated by Neco, and could be derived multiple
September, 2010 A-ll Draft - Do Not Quote or Cite
-------
1 ways. The first is by taking the mean value calculated to represent the long-term amount
2 of N an ecosystem can immobilize and denitrify before leaching (i.e. N saturation) that is
3 derived from the FAB model. This approach requires the input of multiple ecosystem
4 parameters. Its components are expressed by eq 1.
5
6 Neco = JNupt + Nret + (1 - rlNmm +Nden) (1)
7
8 Where,
9 Nupt= nitirogen uptake by the catchment
10 Nimm= nitrogen immobilization by the catchment soil
11 Nden=denitrification of nitrogen in the catchment,
12 Nret = in-lake retention of nitrogen
13 f =forest cover in the catchment (dimensionless parameter)
14 r = fraction lake/catchment ratio (dimensionless parameter)
15
16 The second approach for estimating Neco is to take the difference between N
17 deposition and measured N leaching in a catchment as expressed by eq 2.
18
19 Neco=DL(N)-Nleach (2)
20 The site specific values of critical loads can be used to derive such a deposition
21 loading, here called the deposition metric, which represents a group or percentage of
22 water bodies that reach a specified ANC (or higher) in a given spatial area. For example,
23 if it is desired that all water bodies reach a specified ANC, the allowable amount of
24 deposition for all water bodies is equal to the lowest critical load the population of water
25 bodies. Because the deposition metric represents a percentage of individual catchments
26 from a population of water bodies, and not an individual catchment, the deposition metric
27 is noted by the follow abbreviation DLo/oECo.
28 As an example of the above approach, we evaluated the population of 169
29 waterbodies in the Adirondacks used in the REA analysis. For each individual waterbody
30 in the population the critical load at ANCiim = 50 jieq/L was calculated using the two
31 equations for deriving the NECO term (eq 2 and 3). The distribution of deposition loads
September, 2010 A-12 Draft - Do Not Quote or Cite
-------
1 for the population was assessed and Table A.l.3-1 shows example for the percentage of
2 protection. The mean value for DLo/oECo for the 169 water bodies is presented, as well as
3 the values for which 50, 75, 85, 95 and 100% of the water bodies in the population will
4 not exceed their critical load at ANC = 50 ^eq/L. Note, only 32% of water bodies would
5 not exceed their critical load at ANC = 50 jieq/L for the mean value DLo/oECO because
6 variability is high in the data set, therefore the mean can be problematic for areas with
7 high variability.
8
9
Table A.l.3-1.
target ANC
levels.
Example
This example
calculations for determining
the percent of water
is based the population of DLANCiim for and ANC=50
the Adirondacks. These catchments occur across three
the DLANciim
values into sensitivity
categories
categories (if info is available)
of geologic sensitivity
bodies achieving
for 169 catchments
in
. We could separate
and do the analysis for each category or
calculate one DLANClim for combined geologic categories. Units are in meq/m2/yr.
Mean
Stdev
Ster
Rank
%tile
50%
75%
85%
95%
100%
10 The
NHx
dep
20.40
3.22
0.25
values
Neco DL./.ECO (S+N) +Neco
(eq2) (eq2)
19.19
3.03
0.23
162.36
162.92
13.04
99.33
65.62
54.89
45.12
30.22
for the maximum deposition N
1 1 using the two approaches for NECO and protective
Neco
(eq3)
63.95
11.15
0.86
DL%ECO(S+N)
+Neco (eq 3)
207.55
165.42
13.24
139.22
110.37
95.53
83.99
59.07
% of lakes within the
population that have
ANC > 50neq/L
31.7%
50%
75%
85%
95%
100%
and S based on DLo/oECo at ANC=50
of 95 and
50% of the population of
12 water bodies, are given in Table A.l.3-2.
13
14
September, 2010
A-13
Draft - Do Not Quote or Cite
-------
Table A-3 Values for N and S deposition tradeoff curves for ANC = 50, protecting
95 and 50% of the population, in Adirondacks case study area as illustrated on Fig
5.1 and Fig 5.2. Units are in meq/m2/yr unless noted otherwise.
% protection
95
50
95
50
Eq2
Eq2
Eq3
Eq3
NHxdep
20.4
20.4
20.4
20.4
Neco
19.19
19.19
63.75
63.75
DLo/oECO (max N)
45.12
99.33
83.99
139.22
DL%ECO (max S)
25.93
80.14
20.04
75.27
DL./.ECO (max NOY)
141.96
78.9.3
63.59
118.82
2
3
4
6 A.1.4 Additional Analysis of Bedrock Geology as a basis for defining acid-sensitivity
7 classes across the landscape
8
9 APPROACH
10 This analysis applied a methodology developed by Sullivan et al. (2007) for their
11 exercise to test the hypotheses on watershed sensitivity to acidic deposition. Sullivan et
12 al.'s exercise focused on streams in the Southern Appalachian Mountains region. The
13 classification scheme was based on lithologic maps and the stream chemistry for more
14 than 900 sites. Using logistic regression to model the presence of acid-sensitive
15 waterbodies (expressed as ANC) with respect to lithology class, the authors found "four
16 variables were highly significant in predicting the probability of occurrence of acid-
17 sensitive sites (defined for this analysis as having ANC < 20 |ieq/L) in the Southern
18 Appalachians using logistic regressions: % siliceous bedrock in watershed, % forested
19 watershed, elevation, and watershed area". (Sullivan et al., 2007)
20
21 Using stream and lake ANC data from the Shenandoah Mountains and the
22 Adirondack Mountains, along with lithologic classifications, linked forested area,
23 elevation and watershed area, this NOx SOx exercise used logistic regression to
24 determine if a correlation exists between lithology and ANC that is similar to the
25 correlation found by Sullivan et al. in the Southern Appalachians. If similar to Sullivan
26 et al.'s findings, then the findings of this NOx SOx exercise will be extrapolated spatially
27 to identify areas of the U.S. potentially sensitive to aquatic acidification.
September, 2010
A-14
Draft - Do Not Quote or Cite
-------
1
2 Phase I of this analysis applied the Sullivan et al. methodology separately to the two 2009 NOx-
3 SOx Risk and Exposure Assessment (Shenandoah Mountains and the Adirondack Mountains
4 Case Study Areas) and consists of 4 steps:
5
6 1. Site selection (i.e., lakes and streams)
7 2. Stream chemistry data acquisition and conversion to a spatial data layer
8 3. Data acquisition for % lithology classification of watershed (e.g., % siliceous bedrock),
9 elevation, % forested area of watershed, and watershed area
10 4. Logistic regression modeling.
11
12 The modeling results are compared to Sullivan's SAMI analysis and, if consistent,
13 will be extrapolated to the region surrounding the case study area (Phase II) and to other
14 regions of the U.S. (Phase III) upon EPA directive.
15
16 The output desired for this Phase I exercise was a set of constants and coefficients
17 that predict the probability that the acid-sensitive watersheds (having stream ANC either
18 < 0 jieq/L or between 0 and 20 jieq/L) in the selected study areas based upon the input
19 variables (percent lithology classification in watershed, elevation, watershed size, and
20 percent forest cover in watershed).
21
22 PHASE I
23
24 Step 1 - Site Selection: Streams and lakes to be assessed
25
26 The Adirondack Case Study Area and the Shenandoah Case Study Area provide ideal
27 areas to assess the risk to aquatic ecosystems from NOx and SOx acidifying deposition. Four
28 main reasons support the selection of these two areas. First, both regions fall within the areas of
29 the United States known to be sensitive to acidifying deposition because of a host of
30 environmental factors that make these regions predisposed to acidification. Second, these areas
31 are representative of other areas sensitive to acidification. Third, these regions have in the past
32 and continue to experience substantial exposure to NOx and SOx air pollution. Fourth, these
September, 2010 A-15 Draft - Do Not Quote or Cite
-------
1 areas have been extensively studied ... over the last 3 decades (see Section 4 of the ISA Report
2 (US EPA 2008). (US EPA, 2009, REA, Appendix 4, Section 3.1)
3
4 Freshwater surveys and monitoring in the eastern United States have been
5 conducted by many programs since the mid-1980s, including the National Lake/Stream
6 Surveys (NSWS), EPA's Environmental Monitoring and Assessment Program (EMAP),
7 the Temporally Integrated Monitoring of Ecosystems (TIME) monitoring program
8 (Stoddard, 1990), and Long-Term Monitoring (LTM) project (Ford et al., 1993; Stoddard
9 et al, 1998). The purpose of these programs is to determine the current state and
10 document the trends over time in surface water chemistry for regional populations of
11 lakes or streams impacted by acidifying deposition. Based on extensive surveys and
12 surface water data from these programs, it was determined that the most sensitive lakes
13 and streams (i.e., ANC less than about 50 |ieq/l) in the eastern US are found in New
14 England, the Adirondack Mountains, the Appalachian Mountains (northern Appalachian
15 Plateau and Ridge/Blue Ridge region), northern Florida, and the Upper Midwest. These
16 areas are estimated to contain 95% of the lakes and 84% of the streams in the United
17 States that have been anthropogenically acidified through deposition (see Annex 4.3.3.2
18 of the ISA, US EPA 2008). (US EPA, 2009, REA, Appendix 4, Section 3.1)
19
20 ANC in surface water from 50 lakes in the Adirondack Case Study Area were monitored
21 through the Adirondack LTM program and 38 lakes from the TIME program (Figure A.l.4-3).
22
23 For the Adirondack Case Study Area, the regional EMAP probability survey of 117 lakes
24 (i.e., weighting factors) were used to infer the number of lakes and percentage of lakes that
25 receive acidifying deposition above their critical load of a target population of 1,842 lakes. The
26 117 lakes... represent a subset of 344 sampled lakes throughout New England from 1991 through
27 1994. (ME, NH, VT, RI, MA, CT, NY, NJ). (U.S. EPA, 2009, Appendix 4, Section 4.3.1.1)
28
29 Shenandoah Case Study Area ANC monitoring occurred as part of the
30 Shenandoah National Park Surface Water Acidification Study (SWAS), Virginia Trout
31 Stream Sensitivity Survey (VTSSS), and LTM programs. There are a significant number
32 of the 67 streams in SWAS-VTSSS and LTM programs that currently have ANC of
September, 2010 A-16 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
about 50 jieq/L based on the observed annual average ANC concentrations (Figure
A.l.4-4)
The total number of brook trout streams represented by the SWAS-VTSSS LTM
quarterly monitored sites is approximately 310 out of 440 mountain headwater streams known to
support reproducing brook trout in the Shenandoah Case study Area. ... The SWAS-VTSSS LTM
programs began in Spring 1987, when water samples were collected form 440 streams known to
have brook trout. Following the 1987 survey , a representative subset of 69 streams was selected
for long-term quarterly monitoring of water quality, mostly located on National Forest lands or
within the Shenandoah National Park Case Study Areas (14 SWAS and 55 VTSSS streams).
(U.S. EPA, Appendix 4, Section 4.3.1.2).
TIME/LTM: 2005-2006
ANC
Source: TIME/ALTM 2009
Figure A. 1.4-3. Current yearly average for 2005 to 2006 ANC (ueq/L) in
surface waters from 88 monitored lakes in the Temporally Integrated
Monitoring of Ecosystems (38 Lakes) and Adirondacks Long-Term
Monitoring (50 Lakes) networks in the Adirondack Case Study Area.
(U.S. EPA, 2009)
September, 2010
A-17
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
SWAS-VTSSS/LTM: 2005 - 2006
ANC
Source- SWAS-VTSSS/LTM 2009
Figure A. 1.4-4. Current yearly average for 2005 to 2006 ANC (ueq/L) in surface
waters from 67 monitored streams in the Surface Water Acidification Study,
Virginia Trout Stream Sensitivity, and Long-Term Monitoring network in the
Shenandoah Case Study Area.
[In Phase I, efforts focused on the Shenandoah and Adirondack Mountains case study
areas.]
Step 2: Stream Chemistry Data Acquisition - Water chemistry for those streams and lakes
There are numerous chemical constituents in surface water that can be use to indicate the
acidification condition of lakes and streams and to assess the effects of acidifying deposition on
ecosystem components. These include surface water pH and concentrations of SO42", NO3", A13+,
and Ca2+; the sum of base cations; the recently developed base cation surplus; and the acid
neutralizing capacity (ANC). Each of these chemical indicators provides direct links to the health
of individual biota and the overall health and integrity of aquatic ecosystems as a result of surface
water acidification.
Although ANC does not directly affect the health of biotic communities, it is calculated
(or measured) based on the concentrations of chemical constituents that directly contribute to or
ameliorate acidity-related stress, in particular, pH, Ca2+, and dissolved inorganic aluminum.
Furthermore, numerical models of surface water acidification can more accurately estimate ANC
than all of the individual constituents that comprise it. Consequently, for the purpose of the
performing the case studies reported in the Risk and Exposure Assessment (US EPA, 2009),
September, 2010
A-18
Draft - Do Not Quote or Cite
-------
1 annual average ANC of surface waters was used as the primary metric to quantify the current
2 acidic conditions and biological impacts for a subset of waterbodies in the study areas. (U.S.
3 EPA, 2009 (REA, Appendix 4, Section 2.2)).
4
5 All lake and stream monitoring data within the Adirondack and Shenandoah study
6 areas were used.. Although Sullivan et al. used only Spring data because those values
7 represented the time of year when ANC is at its lowest, and is also a time when sensitive
8 life stages (eggs and young) are present for many fish species (Kaufman, Herlihy, Mitch,
9 Messer,& Overton, 1991) this study used data throughout the calendar year. This was
10 done because of the considerable variation within and between the study areas in terms of
11 the time of lowest seasonal ANC, and the life cycles of the dependent biota. ANC
12 values were then averaged by sampling station. Those stations in the Adirondacks and
13 Shenandoah regions that fell within a 12 digit HUC that crossed into either of the study
14 area boundarieswere extracted to separate GIS data layers. These two collections of 12
15 digit HUCs were then used to extract and average the other data parameters.
16
17 Step 3: Data acquisition for percent lithology classification, elevation, percent forested area
18 of watershed, and watershed size
19
20 Bedrock and surficial lithology - Information on lithology for the eastern United States
21 was been acquired from the states' Geologic Surveys (USGS) and processed by EPA. This
22 information was needed to determine if Sullivan et al.'s geologic classification system for the
23 Southern Appalachian Mountains as applied to the eastern United States is a significant predictor
24 of acid sensitivity in the Shenandoah Mountains and the Adirondack Mountains. Determining
25 which acid class a given rock lithology falls into can be a somewhat subjective process. Upon
26 initial inspection of the data, it appeared that there were some differences found between adjacent
27 states (Figure A.-l). Following discussion with EPA, it was determined that the Adirondack
28 data would benefit from re-evaluation of the classifications assigned. The acid sensitivity classes
29 were reviewed for consistency and correctness and found to be accurate in all but a handful of
30 cases. These exceptions were modified to match their lithologic descriptions. The Shenandoah
31 class assignments were deemed acceptable. The percent of each of the five classes of acid
32 sensitivity (siliceous, argillaceous, felsic, mafic, and carbonate) were tabulated over each of the
33 12 digit HUCs that fell within the study areas.
September, 2010 A-19 Draft - Do Not Quote or Cite
-------
1 Figure A.I Distribution of geology acid classes in the eastern United States and the
2 location of the ELSDS4 sample locations.
7
3
4
5
9
10
11
12
13
14
15
16
Geology Acid Class
| Siliceous - Most Sensitive
Argillaceous - More Sensitive
Felsic - Sensitive
Mafic - Less Sensitive
| Carbonate - Least Sensitive
CTher
Wfeter
Elevation, feet (30m Digital Elevation Maps) The water quality sample locations for the
Adirondack and Shenandoah Mountains were extracted to a single shapefile, which was then
overlaid with a national 30 meter digital elevation model (DEM). The elevation was determined
at each location and extracted as an attribute.
Watershed area, km2 (HUC12) The HUC 12 identifier was extracted using GIS in a
manner similar to the extraction of elevation information for each acid-sensitive sample location.
The area (km2) of the HUC 12 watershed was extracted to the water sampling location shapefile
as an attribute.
Percent forested area by watershed has been estimated using 1992 US Forest Service
vegetative databases and the Advanced Very High Resolution Radiometer (AVHRR) satellite
September, 2010
A-20
Draft - Do Not Quote or Cite
-------
1 data (1,000 m cell size) to produce a raster forest type image with a 250 m cell size. This forest
2 type image was processed using ArcGIS's Spatial Analyst by HUC 12 to determine the number
3 of forested versus non-forested cells in each watershed. The results of this analysis were then
4 used to calculate the percent of forested area in each watershed.
5
6 Step 4: Logistic Regression Modeling
7
8 Following Sullivan et al. (2007) a logistic regression approach with stepwise
9 variable selection procedure was used to model binary variables derived from ANC. A
10 significance level of oc= 0.20 was used as the criterion for variable entry, and oc= 0.25
11 was selected as the criterion for remaining in the model (Hosmer and Lemeshow (1989),
12 p. 108). The following continuous predictor variables were considered in this modeling
13 exercise : Elevation (ft), Area (km2), % Felsic, %_Carbonate, %Mafic, %_Siliceous, %
14 Argillaceous, and %_FOREST. SAS PROC Logistic (SAS (version 9.2)) was used to
15 determine the best model.
16
17 Three binary responses were considered in the logistic modeling:
18 1-. ANC<=20 then outcome=l; else outcome=0
19 2- ANC<=50 then outcome=l; else outcome=0
20 3- ANC<= 100 then outcome=l; else outcome=0.
21
22
23 ADDITIONAL INFORMATION ABOUT THE METHODOLOGY
24
25 The model was run for all sites' elevations and then followed by subsequent
26 model runs that eliminate sites below a selected elevation and above a selected elevation
27 to discern any improvement in correlation sensitive to elevation. The lower and upper
28 elevations were selected based on the distribution of locations provided in Step 1.
29
30 EXTRAPOLATION - PHASES II AND III
31
32 The overall goal of this exercise was to develop constants and coefficients that
33 predict the acid-sensitivity of watersheds with a high degree of probability. A different
34 set of coefficients were developed for each distinct geographic region for which there is
September, 2010 A-21 Draft - Do Not Quote or Cite
-------
1 water quality monitoring data. In Phase II, the acid-sensitivity probability was calculated
2 and mapped using GIS for all watersheds within each distinct geographic region of the
3 case study areas, using the coefficients developed within the region.
4
5 In Phase III, the acid-sensitive probability for watersheds in other regions outside
6 the sampled regions would be calculated and mapped using GIS by applying the set of
7 coefficients from the sampled region which is most similar, and for which all 4 types of
8 input data are available (i.e., percent lithology classification of watershed, percent
9 forested watershed, elevation, and watershed area). Classified lithology data are
10 currently only compiled for the eastern half of the United States, so extrapolation will
11 only be possible for this portion of the country at this time.
12
13 RESULTS
14
15 GIS Mapping of Geologic Classifications and Measured ANC
16
17 Figures A-2andA-3 show the distribution of water sample locations in the Adirondack
18 and Shenandoah Mountains, respectively. The average ANC values are symbolized
19 using 20, 50, and 100 microequilavlents/litre as the break points. The five geology acid
20 classes are also shown.
21
22
September, 2010 A-22 Draft - Do Not Quote or Cite
-------
Average ANC
-134.00-20.00
20.01 - 50 00
50 01 - 100.00
10001 -2671.00
Adirondack Park
HUC12 boundary
Geology Acid Class
AcidClass
| Siliceous - Most Sensitive
Argillaceous - More Sensitive
Felsic - Sensitive
Mafic - Less Sensitive
| Carbonate - Least Sensitive
Other
Water
1
2
Figure A-2. Average ANC and Geology Classification: Adirondack Region
September, 2010
A-23
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
1
8
9
10
11
12
13
14
15
16
17
18
19
20
Average ANC (lien)
• -24.00 - 20.00
20.01 - 50.00
5001 - 100.00
100.01 - S431 65
HUC12 boundary
Shenandoah Region
Acid Class
j^^f Siliceous - Most Sensitive
Argillaceous - More Sensitive
Felsic - Sensitive
Mafic - Less Sensitive
| Carbonate - Least Sensitive
Other
Water
Figure A-3. Average ANC and Geology Classification: Shenandoah Region
Logistic Regression Analysis Results
Before running any modeling technique, the correlation between the predictors and ANC
was explored (See Figure A.-4) A cell in the interception of a row and column represents
the correlation between the variables at the end of the row and column. The correlation
coefficient, R2 , ranges from -70 to 30%, suggesting a low to moderate correlation
between predictors and outcome. Negative correlations indicate that as the predictor
value increases, the value of ANC decreases. Positive correlations suggest that as the
value of the predictor increases so does the value of ANC. The correlation matrix also
serves to explore the multi-co-linearity issue (presence of highly correlated predictors).
In the Adirondacks region, Elevation is the variable with higher correlation (-30% to -
50%), suggesting that this variable will likely be included as a predictor in the model.
In the Shenandoah region, % Carbonate (R2 between 40% and 70%) and % Forest (R2
between -40% to -70%) show a modest to high correlation with ANC. (See Figure A-5.)
Also, % Forest is moderately correlated with % Siliceous (R2 between 40% to 70%) and
September, 2010
A-24
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
% Carbonate (R2 between -40% to -70%), suggesting that it is likely that one of these
three variables will be in the model. Similarly, % Felsic is moderately correlated with %
Carbonate and % Argillaceous (R2 between -40% to -70%) suggesting that if one of these
variables is in the model, then the other two won't be in the model.
Figure A-.4. Correlation matrix for Adirondacks Region.
9
10
11
12
13
14
PCT Mafic
PCT Felsic
PCT_Argillaceous
PCT Siliceous
PCT Carbonate
AVG ANC
PCT FOREST
AREA SQKM
ELEV FT
-0.1
September, 2010
A-25
Draft - Do Not Quote or Cite
-------
1
2 Figure A-5. Correlation matrix for Shenandoah Region.
3
PCT Mafic
4
5
9
10
11
12
13
14
15
PCT Felsic
PCT_Argillaceous
PCT Siliceous
PCT Carbonate
PCT FOREST
AREA SQKM
ELEV FT
AVG ANC
-0.7
Among the measurements of model assessment considered were C statistic (a C value of
0.5 means the model predictions are equivalent to a random guess; >0.7 denotes
worthwhile to the model); the Akaike information Criteria (AIC) (Hosmer and
Lemeshow (1989), p. 184); and the Schwarz information criterion (SC) (Kass, R. E. &
Wasserman, L. (1995)); and the -2*maximized log-likelihood (-2ML). Smaller values of
all three statistics suggest a better model. To evaluate the goodness of fit of the model,
the Hosmer and Lemeshow goodness of fit test (Hosmer and Lemeshow (1989)) was
used (larger p-values suggest no difference between observed and model-predicted
values, implying that the model's estimates fit the data at an acceptable level).
September, 2010
A-26
Draft - Do Not Quote or Cite
-------
1 The final model for the Adirondacks Region binary variables is the following:
2
fl ANC<20
3 For the binary variable 79n = <
y 20 [0 ANC>20
4
5 the logistic model is the following:
logitO) = -3.3933 + 0.00305 xElevation - 0.0142 x (%Forest) - 0.0216 x (%Carbonate)
6 - 0.0366 x (%Siliceous) - 0.0172 x (%Mafic)
7
fl ANC<50
8 For the binary variable Kn = 4
y 50 [0 ANC>50
9
10 the logistic model is the following:
logitO) = -4.2271 + 0.00306Elevation - 0.0197(%Carbonate) - 0.0463 (%Siliceous) - 0.0155(%Mafic)
12
13
fl
14 For the binary variable Km H
100 [0
15
16 the logistic model is the following:
17
18
19 The final model for the Shenandoah region binary variables is the following:
20
fl ANC<20
21 For the binary variable Y^=<
20 [0 ANC>20
22
23 the logistic model is the following:
logitO) = -5.8725 + 0.0358(%Carbonate)+0.0632(%Siliceous)
fl ANC<50
25 For the binary variable Kn = <
y 50 [0 ANC>50
26
27 the logistic model is the following:
logitO) = -6.0745 + 0.0233(%Carbonate)+0.0750(%Siliceous)
28
29
fl
30 For the binary variable Y,m=<
10° [0
September, 2010 A-27 Draft - Do Not Quote or Cite
-------
Intercept
ELEV_FT
PCT_FOREST
PCT Carbonate
PCT_Argillaceous
7 . 9493
-0.0015
-0.0085
0.0021
0.0533
0.3694
0.0001
0.0039
0.0012
0.0208
7.2254
-0.0016
-0.0162
-0.0003
0.0125
8.6733
-0.0013
-0.0008
0.0045
0.0941
1
2 the logistic model is the following:
logitO) = -3.9068 + 0.000699(%Carbonate)+0.047(%Siliceous)
4
5 Betas, standard error for betas, odds ratio and relevant statistics can be found in Appendix
6 2.
7
8 Weibull modeling using Proc Reliability (SAS) was also used to model ANC. LCL and
9 UCL denote the lower and upper confidence limits at 95% confidence level. If the value 0
10 is contained between the confidence limits then the parameter is not statistically
11 significant at 5% level.
12
13 For the Adirondacks Region, the final predictors and estimates are shown below:
14
15 Parameter Estimate SE 95% LCL
16 95%UCL
17
18
19
20
21
22
23 According to this model, ANC decreases with each unit increase in Elevation and Forest
24 (negative estimates) and increases with each unit increase in % Carbonate and %
25 Argillaceous (positive estimates).
26
27 Model assessment statistics (R2, AIC, etc.) were not found in the documentation. It is
28 possible to calculate them to evaluate the model fit to the data.
29
30 For the Shenandoah Region, the final predictors and estimates are shown below:
31
32
33
34
35
36
37
38 According to this model, ANC decreases with each unit increase in Elevation and %
39 Felsic (negative estimates) and increases with each unit increase in % Siliceous and %
40 Argillaceous (positive estimates).
41
42 Appendix 2 shows the odds ratio for each predictor. The larger the Odds ratio, the
43 higher the chances that ANC <=20 for every change in unit of the predictor.
44
45
September, 2010 A-28 Draft - Do Not Quote or Cite
Intercept
ELEV_FT
PCT Siliceous
PCT_Argillaceous
PCT Felsic
7.3437
-0.0015
0.0167
0.0564
-0.0041
0.1585
0.0001
0.0054
0.0209
0.0011
7.0330
-0.0016
0.0061
0.0154
-0.0062
7.6543
-0.0013
0.0273
0.0973
-0.0020
-------
1 For the Adirondacks region and ANC<=20
2 When elevation increases one unit, the odds that ANC<=20 increase by a factor of 1.03,
3 when other variables are controlled (meaning the same level of % Fforest, % Carbonate,
4 % Siliceous and % Mafic). When % Forest increases one unit, then the odds that
5 ANC<=20 decreases by a factor of 0.972 when other predictors are held constant. Similar
6 interpretations exist for the rest of the models.
7
8
9 For the Shenandoah region and ANC<=20:
10
11 When % Carbonate increases one unit, the odds that ANC<=20 increases by a
12 factor of 1.036 when other predictors are held constant. Similarly,when % Siliceous
13 increases one unit, the odds that ANC<=20 increases by a factor of 1.065 when other
14 variables are held constant.
15
16 UNCERTAINTIES
17
18 Lithology classification assignments. RTI received a geodatabase of geology data
19 from EPA to be used in the ANC GIS comparative analysis. The geodatabase includes a
20 list of lithologies from the most recent USGS (Eastern Mineral Resources Team) national
21 geologic map data for the continental United States. Each lithology polygon from the
22 USGS map data set is defined by two lithology types labeled'ROCKTYPEl' and
23 'ROCKTYPE2'. According to the metadata for the geologic map cover, ROCKTYPE1
24 characterizes more than 50% of the area in the associated polygon. EPA has categorized
25 each lithology record into one of five geologic sensitivity classes (i.e., Siliceous,
26 Argillaceous, Felsic, Mafic, and Carbonate), after Sullivan, et al (2006). The geologic
27 sensitivity classification is included for each lithology record of the geodatabase in a field
28 indentified as'AcidClassDesc'.
29 A cursory review was performed on EPA's assigned classifications suggesting—
30 • Not all rock types in the geodatabase are included in Sullivan (Table 1), but in
31 most instances (except where noted) there was agreement between Sullivan's
32 classification and the EPA assigned classes for rock types included in both the
33 technical document and the geodatabase;
September, 2010 A-29 Draft - Do Not Quote or Cite
-------
1 • Potential discrepancies in class assignments were identified such as those
2 associated with non-descriptive lithogic description (e.g., sedimentary rock),
3 others with typical rock chemistry and GSC (e.g., Carbonate GSC was assigned to
4 beach sand), also classification assigned to sedimentary environments (e.g.,
5 olistostrome), or terms describing non-mineralogy specific rock texture (e.g.,
6 hornfel). The majority of these fell outside the Adirondack or Shenandoah study
7 areas.
8
9 After discussion with EPA and a more detailed review, only a few revisions were
10 required in the Adirondack datasets and those revisions were assigned less acid-sensitive
11 classes. No revisions were required in the Shenandoah dataset.
12
September, 2010 A-30 Draft - Do Not Quote or Cite
-------
1 A.1.5 Additional techniques explored for which to base subdivision of the U.S. based based
2 on acid sensitivity.
3 Alkalinity
4 We also considered dividing the U.S. based on the Alkalinity from the 1984
5 Omernick Alkalinity map. We do not recommend using alkalinity because it includes
6 more ecoregions as sensitive than the ANC approach, which causes the deposition loads
7 for sensitive and less-sensitive areas to be more similar.
Total Alkalinity of Surface Waters
Fig A. 1.5-1. Omernick's alkalinity map (Omernick 1984)
September, 2010
A-31
Draft - Do Not Quote or Cite
-------
Table A. 1.5-1. Descriptive statistics of the CL populations that result when the U.S. is divided
into two categories, sensitive and non-sensitive based on Alk
Omernick sensitivity classification
CL
for
ANC
50
50
Sensiti
vity
class
sensit
ive
Non-
sensit
ive
CL in class
n=
5446
346
Mean
Neco
48
37
Neco St
er
0.4
1
CL mean
(meq/m2/yr)
274
343
CL
Ster
5
15
Example deposition metrics
DL %90
(meq/m2/yr)
28
50
DL %75
(meq/m2/yr)
57
109
DL %50
(meq/m2/yr)
123
265
3
4
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Cluster Analysis
This approach uses ANC, ALK, DOC and soil BCw to categorize the acid-
sensitivity of ecoregions based on a quantitative cluster analysis. The ANC dataset is
described in Chapter 5, the ALK dataset is described in the previous section, and the
DOC and soil BC weathering datasets are described prior to the explanation of the cluster
analysis
DOC
Water chemistry data on DOC was collected from several national monitoring
networks. The data sources are the same as those for ANC and summarized in Chapter
Sand include approximately 11,000 observations.
Bed rock geology
Bedrock geology is the parent material to soils and can be related to acid-
sensitivity. In the first draft PA the EPA staff proposed using bedrock geology to inform
acid sensitivity categorization of the landscape. This approach entailed defining
categories of ecosystem sensitivity to acidification based on bedrock geology/ lithology.
The approach is supported by conclusions from the ISA in which geology is determined
to be the governing factor that drives ecosystem sensitivity to acidification (ISA 3.2.4.1).
A map was developed in this policy assessment to capture the heterogeneity of geologic
bedrock that occurs across the eastern U.S. and link it to ecosystem acid-sensitivity (Fig
A. 1.5-2). The method is based on Sullivan et al. (2007) in which 70+ primary lithologies
September, 2010
A-32
Draft - Do Not Quote or Cite
-------
1 are grouped into 5 categories of acid-sensitivity, using ANC as the ecosystem indicator
2 upon which acid-sensitivity is based. Sullivan et al. (2007) evaluated multiple features of
3 the landscape and found that geology is the landscape parameter that governs ecosystem
4 sensitivity to acidic deposition. The analysis in Sullivan et al. 2007 was conducted in the
5 Southern Appalachian Mountains region, which included sites from the states of GA, TN,
6 NC, KT, VA and WV. EPA conducted additional analyses to further test the concept that
7 lithology correlates to acid sensitivity based on ANC, using data from the Adirondacks
8 and Shenandoahs. Results from this analysis indicated that North of the glaciation line,
9 bedrock geology was not a good predictor of acid sensitivity. For example, the
10 Adirondacks which are known to be one of the most sensitive areas to acidification is
11 dominated by types of bedrock geology that are considered moderately to less sensitive in
12 areas South of the glaciation line.
13
September, 2010 A-33 Draft - Do Not Quote or Cite
-------
2
3
4
5
6
7
Acid-Sensitive Areas of the Eastern United States
A Classification based on Bedrock Geology
US Environmental Protection
September 2010
0 140 280 560 840 1,120
Kilometers
n Alters Equal Area Conic fUSCS)
False Eaarog- 0.000000
False_Mor1»al_M«ri*an -96.000000
Slandard_Para1lflM 29.500000
Siardard Parallel 2 45.SOO&00
UWude Of Origm- 33.000000
Linear Unit: Meier
Datum D North. American 1S63
About the Classification
TNs map ts based on 1:250.0OO bedrock geotogy data
from the US Geological Survey, tl has been classified
basket on a method In Sultoan el al |2006>, with ihe
eKcepteon of the "olher" calegon/. This eaiegoty
tepresents meta morphic or olrwr rock lyp*s wfwse
composition requires ruilh*! study twfoft twmg classified
State Boundaries
Updated Acid Sensitivity
AcidClass
| Carbonate - Least Sensitive
Mafic - Less Sensitive
Felsic - Sensitive
Argillaceous - More Sensitive
H Siliceous - Most Sensitive
Fig A. 1.5-2. A map of acid sensitive areas of the Eastern U.S.
developed from a lithology-based five-unit geologic classification
system after methods in Sullivan et al. (2007).
Soil base-cation weathering
As previously noted, the use of bedrock geology for the classification of acid sensitive
catchments is problematic in areas north of the glaciation line because glacial activity has
caused the surface till to become spatially disconnected from the parent geologic material.
Surficial geology is therefore important to the evaluation of the spatial variability of acid-
September, 2010
A-34
Draft - Do Not Quote or Cite
-------
1 sensitivity. For that reason a national map for soil base-cation weathering (McNulty et al.
2 2007) was considered in the analysis of acid-sensitivity (Fig A. 1.5-3).
3
4
5
6
7
8
9
10
11
13
14
15
eq^ha/yr
| 42 - 500
| 500-1,000
| 1,000-2,000
| 2,000 - 3,000
| 3,000 - 3,934
0 250 500
1,000 1,500
2,000
^H Kilometers
Fig A. 1.5-3. A map of average annual forest soil base cation
-i -i
weathering expressed in eq ha" yr" for the conterminous US for the
years 1994 to 2000 at a 1-km2 spatial resolution from McNulty et al.
McNulty et al. 2007 estimated the base cation weathering rate using the clay correlation-
substrate method (Sverdrup et al., 1990). This method used a combination of parent
12 material and clay percent to determine the weathering rate. Base cation weathering rate
-1 -K
(eq ha" yr" ). A temperature correction can be applied to this method, but this correction
is more suitable for northern climates and was not used in the model. Clay fraction was
derived from a weighted average of soil fraction per map unit.
September, 2010
A-35
Draft - Do Not Quote or Cite
-------
1
2
Table A.I.5-3. Data input to the U.S. national soil base cation weathering map
developed by McNulty et al. 2007
Data source
Description ofdataset
State Soil
Geographic Data
Base (STATSGO)
Compilation by the Natural Resource Conservation Service (NRCS) of geology,
topography, vegetation, climate, Landsat Thematic Mapper (TM) satellite imagery,
and detailed, county level soil survey data. Each soil map unit in STATSGO includes
multiple components and data layers (USDA NRCS, 1995). STATSGO is organized
by state and consists of one geospatial vector representing the soil map units for that
state and 15 tables describing characteristics of those map units. Multiple soil layers
are associated with each map unit. Soil layer sampling depth is not consistent within a
state or between states
CONUS-SOIL
developed by the
Earth System
Science Center
(ESSC) at
Pennsylvania
State University
This is a 1-km multi-layer soil data set based on the STATSGO. ESSC converted the
vector map unit layer in STATSGO to a 1-km2 grid, remapped many of the original
STATSGO attribute layers, and defined 11 standard soil layers (Miller and White,
1998). These data layers and tables linked the standardized data set to the original
STATSGO data set distributed by ESSC as 1-km2 soil map unit grids for the
conterminous US. The CONUS-SOIL was much better suited for national-scale
modeling than the original STATSGO attribute layers and was therefore used
McNulty et al. 2007. Key soil data inputs included CONUS-SOIL map units and clay
fraction (Miller and White, 1998).
forest soil percent
organic matter
(OM) layer
Was created by McNulty et al. 2007 by first averaging the maximum and minimum
recorded values for OM were averaged for each layer in the STATSGO data set.
Second, the average OM layers were remapped into the 11 standard CONUS-SOIL
layers using a weighted average to redistribute the average OM STATSGO layers into
the CONUS-SOIL layers. If a STATSGO layer was completely contained in a
CONUS-SOIL layer, then the average OM was multiplied by the component percent
to determine the average OM contribution to the standard layer. If the STATSGO
layer overlapped more than one CONUS-SOIL layer, then the proportion of overlap
was multiplied by the average OM and the component percent, where the component
percent was the proportion of the soil map unit comprised of that soil component.
Once the conversion from STATSGO to CONUS-SOIL layer was complete, the 11
standard layers were summed by layer and divided by the sum of component percent.
Finally, the weighted average was calculated according to equation 8.
parent material
class
was derived from the STATSGO map unit component (comp) and taxonomic (tax)
classification tables (USDA NRCS, 1995). The dominant mineralogy for each soil
map unit was determined from the comp and tax tables. Each unit was classified into
parent material class based on the mineralogical description (USDA NRCS, 2006)
Soil depth
in meters was obtained from the CONUS-SOIL depth to bedrock layer. This layer
identified map units with bedrock less than 1.52 m below the soil surface (i.e., map
units coded with a depth of 1.52 m did not encounter bedrock) (Miller and White,
1998).
3
4
5
6
Results
This approach uses ANC, ALK, DOC and soil BCw to categorize the acid-sensitivity of
ecoregions based on a quantitative cluster analysis. Cluster analysis is a method to sort a
set of observations (in this case CL from watersheds) into subgroups so that the degree of
association between observations is maximal if they belong to the same group and
September, 2010
A-36
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
9
10
11
12
13
14
15
16
17
minimal otherwise. The similarity of each site was determined using ANC, ALK, DOC
and BCw. In the cluster analysis, the &-means algorithm assigned each point to the cluster
whose center (also called centroid) is nearest. The center is the average of all the points in
the cluster — that is, its coordinates are the arithmetic mean for each dimension
separately over all the points in the cluster. The ecoregions was then assigned
membership to a cluster. The weakness of this approach is that each of the datesets for
the criteria (ANC, soilBCw, Alk and DOC), have varying levels of national coverage. If
not all four criteria are available for the analysis the ecoregions drops out and is not
assigned a cluster identification. The majority of ecoregions in the U.S. were not
represented in the datasets for all four criteria and were not assigned a cluster
identification (Fig A. 1.5-4). Therefore the ulitility of this approach is limited.
Mean ANC
September, 2010
A-37
Draft - Do Not Quote or Cite
-------
1 REFERENCES
2
3 Kaufmann, P.E., A.T. Herlihy, M.E. Mitch, JJ. Messer, and W.S. Overton, 1991.
4 Chemical characteristics of streams in the Eastern United States: I. Synoptic
5 survey design, acid-base status and regional chemical patterns. Water Resources
6 Research, 27, 611-627.
7 NAPAP (National Acid Precipitation Assessment Program), 2005. National Acid
8 Precipitation Assessment Program Report to Congress: An Integrated
9 Assessment. National Acid Precipitation Assessment Program, Washington, DC.
10 Stoddard, 1, J.S. Kahl, F.A. Deviney, D.R. DeWalle, C.T. Driscoll, A.T. Herlihy, J.H.
11 Kellogg, P.S. Murdoch, J.R. Webb, and K.E. Webster, 2003. Response of
12 Surface water Chemistry to the Clean Air Act Amendments of 1990. EPA 620/R-
13 03-001. U.S. Environmental Protection Agency, Office of Research and
14 Development, National Health and Environmental Effects Research Laboratory,
15 Re search Tri angl e Park, NC.
16 Sullivan, T.J., C.T. Driscoll, B.J. Cosby, I.J. Fernandez, A.T. Herlihy, J. Zhai, R.
17 Stemberger, K.U. Snyder, J.W. Sutherland, S.A. Nierzwicki-Bauer, C.W. Boylen,
18 T.C. McDonnell, and N.A. Nowicki. 2006. Assessment of the Extent to which
19 Intensively-Studied Lakes are Representative of the Adirondack Mountain
20 Region. Final report. New York State Energy Research and Development
21 Authority (NYSERDA), Albany, NY. Available at
22 http://nvsl.nvsed.gOv/uhtbin/cgisirsi/Qcwd6NzFbv/NYSL/138650099/8/4298474
23 (accessed November 1, 2007).
24 Sullivan, T. J., J. R. Webb, K.U. Snyder, A.T. Herlihy, B.J. Cosby, 2007"Spatial
25 Distribution of Acid-sensitive and Acid-impacted Streams in Relation to
26 Watershed Features in the Southern Appalachian Mountains". Water Air Soil
27 Pollution (2007) 182:57-71
28 U.S. EPA, 2009. Risk and Exposure Assessment for Review of the Secondary National
29 Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur.
30 EPA-452/R-09-008b. Final. U.S. Environmental Protection Agency, Office of
31 Air Quality Planning and Standards. Research triangle Park, NC.
32 U.S. EPA, 2008. Integrated Science Assessment (ISA) for Oxides of Nitrogen and
33 Sulfur-Ecological Criteria. Final Report. EPA/600/R-08/082F. U.S.
34 Environmental Protection Agency, National Center for Environmental
35 Assessment - RTF Division, Office of Research and Development, Research
36 Triangle Park, NC.
37 Davis, J. M., Swall, J. L., 2006. An examination of the CMAQ simulations of the wet
38 deposition of ammonium from a Bayesian perspective. Atmospheric Environment
39 40,4562-4573.
40 EPA, 1999. Science Algorithms of the EPA Models-3 Community Multiscale Air Quality
41 (CMAQ) Modeling System. Tech. Rep. EPA/600/R-99/030, U.S. Environmental
42 Protection Agency, Washington DC.
43 Grantz, D., Garner, J., Johnson, D., 2003. Ecological effects of particulate matter.
44 Environment International 29, 213-239.
September, 2010 A-3 8 Draft - Do Not Quote or Cite
-------
1 Lien L; Raddum GG; Fjellheim A. (1992). Critical loads for surface waters: invertebrates
2 and fish. (Acid rain research report no 21). Oslo, Norway: Norwegian Institute for
3 Water Research
4 Morris, R. E., McNally, D. E., Tesche, T. W., Tonnesen, G., Boylan, J. W., Brewer, P.,
5 2005. Preliminary evaluation of the community multiscale air quality model for 2002
6 over the Southeastern United States. Journal of the Air and Waste Management
7 Association 55, 1694-1708.
8 Seinfeld, J., Pandis, S., 1998. Atmospheric Chemistry and Physics. John Wiley and Sons,
9 Inc., New York.
10 Sullivan TJ; Webb JR; Snyder KU; Herlihy AT; Cosby BJ. (2007). Spatial distribution of
11 acid-sensitive and acid-impacted streams in relation to watershed features in the
12 southern Appalachian mountains. Water Air Soil Pollut, 182, 57-71.
13 Sullivan TJ; Fernandez IJ; Herlihy AT; Driscoll CT; McDonnell TC; Nowicki NA;
14 Snyder KU; Sutherland JW. (2006). Acid-base characteristics of soils in the
15 Adirondack Mountains, New York. Soil Sci Soc Am J, 70, 141-152.
16
September, 2010 A-39 Draft - Do Not Quote or Cite
-------
1 APPENDIX B
2 Analysis of Critical Loads, Comparing Aquatic and Terrestrial Acidification
3
4 Background
5 Critical load is defined as, "a quantitative estimate of ecosystem exposure to one or more
6 pollutants below which significant harmful effects on specified sensitive elements of the
7 environment do not occur, according to present knowledge" (McNulty et al., 2007), and critical
8 loads can be estimated for aquatic and terrestrial ecosystems. Within the Risk and Exposure
9 Assessment for Review of the Secondary National Ambient Air Quality Standards for Oxides of
10 Nitrogen and Sulfur (hereafter referred to as REA Report) (US EPA, 2009), critical loads of
11 acidification for aquatic systems were determined by relating specific amounts of acidifying
12 nitrogen and sulfur deposition to selected Acid Neutralizing Capacities (ANC) within freshwater
13 lakes or streams. The presence and abundance offish species served as the biological indicator
14 of the impacts of the exceedance of critical acid loads by nitrogen and sulfur deposition.
15 Estimation of critical acid loads for terrestrial systems within the REA Report (US EPA, 2009)
16 related acidifying nitrogen and sulfur deposition to the base cation to aluminum (Bc/Al) ratio in
17 the soil solution, and the health of sugar maple and red spruce in forest ecosystems served as the
18 biological indicator of the impacts of critical acid load exceedance. A main distinction between
19 these two critical loads is that aquatic critical loads are largely an integrated function of the
20 chemistry of run-off waters that feed the lake or stream within a watershed, while terrestrial
21 critical acid loads are determined by the rooting zone section of the soil profile in a forest
22 ecosystem. Therefore, it is possible to have different critical load values for aquatic and
23 terrestrial ecosystems within the same watershed.
24
25 The goal of this Task was to determine the relative degree of protection offered by
26 aquatic versus terrestrial critical acid loads within a landscape. Critical acid loads for lakes and
27 streams within watersheds of the Adirondacks and Shenandoah Valley were compared against
28 terrestrial critical loads calculated for same watersheds to determine which estimate had the
29 lowest, most protective critical load for acidifying nitrogen and sulfur deposition.
30
31
September, 2010 B-1 Draft - Do Not Quote or Cite
-------
1 Methods
2 For the REA Report (US EPA, 2009), critical acid loads were determined for 169 lakes
3 and 60 streams in the Adirondacks and Shenandoah Case Study Areas, respectively. These
4 critical loads were calculated using four different ANCs, 0, 20, 50 and 100 |ieq/L, that ranged in
5 the level of protection offered to fish species abundance and diversity, and the resulting critical
6 acid loads were classified into four "current condition of acidity and sensitivity to acidification"
7 categories. "Highly Sensitive" water bodies had critical loads less than or equal to 50
8 meq/m2/yr, "Moderately Sensitive" systems had critical loads ranging from 51 to 100 meq/m2/yr,
9 "Low Sensitivity" lakes and streams had critical loads that ranged from 101 to 200 meq/m2/yr,
10 and "Not Sensitive" systems had critical acid loads greater than 201 meq/m2/yr.
11
12 For the purposes of this Task, aquatic critical acid loads corresponding to an ANC of 50
13 meq/m2/yr were selected, and the locations of the lakes and streams in the Adirondacks and
14 Shenandoah Case Study Areas were mapped by HUC12 watersheds. Availability of data for
15 terrestrial acidification estimates was determined for each HUC, and only HUCs that had
16 sufficient data were mapped. Data from the U. S. Department of Agriculture- Natural Resources
17 Conservation Service (USDA-NRCS) SSURGO soils database (USDA-NRCS, 2008) had the
18 poorest coverage. This data restriction limited the number of water bodies that could be included
19 in the analysis to 62 and 35 for the Adirondacks and Shenandoah Case Study Areas, respectively.
20
21 To examine a representative selection of water bodies in each Case Study Area, four
22 watersheds containing lakes or streams from each of the four "current condition of acidity and
23 sensitivity to acidification" categories were randomly selected. Therefore, a total of 16
24 watersheds were chosen for each Case Study Area. All four "current condition of acidity and
25 sensitivity to acidification" categories were evenly represented for the Adirondacks Case Study
26 Area (four watersheds for each of the four categories). However, due to the limited number of
27 watersheds in the Shenandoah Area and a lower proportion of lakes with low sensitivities to
28 acidifying nitrogen and sulfur deposition ("Low Sensitivity" and "Not Sensitive"), it was not
29 possible to have equal representation of all "current condition of acidity and sensitivity to
30 acidification" categories. Therefore, there was a larger representation of streams that were more
31 sensitive to acidification ("Highly Sensitive" and "Moderately Sensitive"). All water bodies that
September, 2010 B-2 Draft - Do Not Quote or Cite
-------
1 were located in each of the selected HUCs were included in the analyses. In many cases, these
2 water bodies ranged in sensitivity to acidification. In total, 29 lakes and 20 streams were
3 analyzed in the Adirondacks and Shenandoah Case Study Areas, respectively (Table 1 and 2).
September, 2010 B-3 Draft - Do Not Quote or Cite
-------
1 Table 1. Watersheds (HUC 12) and fresh water lakes in the Adirondacks Case Study Area that were used in the comparison of
2 aquatic and terrestrial critical acid loads. Lake IDs and associated aquatic critical acid loads (CL) in meq/m2/yr, based on an ANC of
3 50 |ieq/L, are indicated in each cell and are from the REA REPORT (US EPA, 2009).
HUC
020100010103
020100040203
020100080304
020100081602
020200020101
020200020704
020200040805
041501011001
041503020801
041503040102
041503040204
041503050103
041503050104
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive (CL
< 50 meq/m2/yr)
1 A2-078O (CL = 33)
NY029L (CL = 39)
NY284L (CL = 23)
NY285L (CL = 42)
1A1-089O(CL = 43)
NY290L (CL = 30)
NY289L (CL = 50)
Moderately Sensitive (CL
= 51-100 meq/m2/yr)
NY013L(CL = 64)
NY536L (CL = 69)
NY278L (CL = 57)
050215AO(CL = 74)
NY793L (CL = 97)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
1A2-028O(CL= 106)
NY310L(CL= 147)
NY292L(CL= 117)
Not Sensitive
(CL > 201 meq/m2/yr)
NY534L(CL= 1043)
NY308L (CL = 485)
NY312L(CL = 588)
NY313L(CL = 598)
NY500L(CL = 610)
NY783L (CL = 455)
September, 2010
B-4
Draft - Do Not Quote or Cite
-------
HUC
041503050302
041503050407
041503050601
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive (CL
< 50 meq/m2/yr)
Moderately Sensitive (CL
= 51-100 meq/m2/yr)
NY767L(CL = 51)
NY529L (CL = 73)
NY528L (CL = 82)
NY769L (CL = 99)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
NY008L (CL = 146)
NY007L(CL= 165)
NY768L(CL= 114)
NY004L(CL= 168)
Not Sensitive
(CL > 201 meq/m2/yr)
September, 2010
B-5
Draft - Do Not Quote or Cite
-------
1
2
3
4
Table 2. Watersheds (HUC 12) and streams in the Shenandoah Case Study Area that were used in the comparison of aquatic and
terrestrial critical acid loads. Stream IDs and associated aquatic critical acid loads (CL) in meq/m2/yr, based on an ANC of 50 |ieq/L,
are indicated in each cell and are from the REA REPORT (US EPA, 2009).
HUC
020700050401
020700050502
020700050703
020700050705
020700050801
020700050803
020700060101
020801030301
020801030402
020802010702
020802010703
020802010801
020802010803
020802020102
020802020401
020802030601
CURRENT CONDITION OF ACIDITY AND SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive (CL
< 50 meq/m2/yr)
VT37 (CL = 26)
VT57 (CL = 39)
VT40(CL= 13)
VT35(CL = 37)
VT36 (CL = 24)
DR01 (CL = 33)
WOR1 (CL = 43)
VT53 (CL = 40)
VT10(CL= 15)
VT11 (CL= 14)
VT14(CL= 14)
VT15(CL= 13)
VT16(CL = 20)
VT41 (CL= 15)
Moderately Sensitive (CL
= 51-100 meq/m2/yr)
VT54 (CL = 69)
VT62 (CL = 68)
VT12 (C = 75)
VT38(CL = 66)
VT46 (CL = 52)
Low Sensitivity (CL
= 101-200 meq/m2/yr)
VT60(CL= 198)
Not Sensitive
(CL > 201 meq/m2/yr)
VT61 (CL = 231)
September, 2010
Draft - Do Not Quote or Cite
-------
1 Terrestrial critical acid loads were calculated for each of the 16 watersheds using the
2 simple mass balance method (UNECE, 2004) and data sources outlined in the REA Report (US
3 EPA, 2009), and Bc/Al soil solution indicator values of 1.2 and 10.0. Briefly, average values for
4 base cation deposition (calcium, potassium, magnesium and sodium), chloride deposition, and
5 annual runoff (m3/ha/yr) were determined for each watershed (Table 3). The Kgibb constant
6 (m6eq2) was determined by the average percent organic matter in the soil (Table 4), and N
7 immobilization in the soil was set to the constant value of 42.86 eq/ha/yr (McNulty et al., 2007).
8 It was assumed that active harvesting did not occur in each of the watersheds. Therefore base
9 cation (calcium, magnesium and potassium) and nitrogen uptake were 0 eq/ha/yr (UNECE,
10 2004). Similarly, it was assumed that the majority of each watershed consisted of upland sites.
11 Therefore, denitrification losses were assumed to be 0 eq/ha/yr (McNulty et al., 2007). Base
12 cation weathering was estimated using the clay substrate model (equations 1-3) (McNulty et al.,
13 2007).
14 Acid Substrate: BCe = (56.7 x %clay)- (o.32 x (%clay)2) (1)
15 Intermediate Substrate: BCe = 500 + (53.6 x %clay)- (o. 18 x (%clay)2) (2)
16 Basic Substrate: BCe = 500 + (59.2 x %clay) (3)
17 where
18 BCe = empirical soil base cation (Ca2+ + K+ + Mg2+ + Na+) weathering rate
19 (eq/ha/yr)
20 % clay = the percentage of clay within the top 50cm of the soil.
21
22 The U.S. Department of Agriculture- Natural Resources Conservation Service (USDA-NRCS)
23 SSURGO soils database (USDA-NRCS, 2008)) and state-level geology (U.S. Geological Survey
24 (USGS) state-level integrated map database for the United States (USGS, 2009)) were used to
25 determine parent material acidity classification. Parent material acidity was determined for each
26 SSURGO polygon within each watershed using the criteria outlined in the REA Report (US
27 EPA, 2009), and the contributions of base cations from the weathering of acid, intermediate and
28 basic substrates (eq/ha/yr) were determined by a weighted average based on the proportion of
29 area occupied by each parent material acidity class. Rooting depth was assumed to be 50 cm and
30 masses of calcium, magnesium, potassium, sodium and nitrogen were converted to eq/ha/yr units
September, 2010 B-7 Draft - Do Not Quote or Cite
-------
1 based on molar charge equivalents. Unless indicated otherwise, the units used in the calculation
2 of critical acid loads were eq/ha/yr. The estimated terrestrial critical loads for the 16 watersheds
3 in the Adirondacks and Shenandoah Case Study Areas are presented in Table 5.
4
5
6
7
8
Table 3. Name, type and source of data used in the simple mass balance estimates of terrestrial
critical acid loads for the watersheds in the Adirondacks and Shenandoah Case Study Areas.
DATA
Base cation (Ca2+,
Mg2+,Na+,K+)
deposition — wet
Chloride (Cl')
deposition — wet
Runoff
Soil horizon depth
Percentage of clay
by soil horizon
Percentage of
organic matter by
soil horizon
Soil parent
material
State-level
bedrock geology
NAME
CMAQ/
NADP
NADP
Annual run-off
(1:7,500,000
scale)
SSURGO
SSURGO
SSURGO
SSURGO
State
Geological
Map
Compilation
TYPE
GIS
datalayers
GIS
datalayer
GIS
datalayer
GIS
datalayer
GIS
datalayer
GIS
datalayer
GIS
datalayer
GIS
datalayer
SOURCE
Provided by U.S. Environmental
Protection Agency (EPA)/NADP,
2003a,c, d, e
NADP, 2003b
Gebertetal., 1987
USDA-NRCS, 2008
USDA-NRCS, 2008
USDA-NRCS, 2008
USDA-NRCS, 2008
USGS, 2009
Note: CMAQ = Community Multiscale Air Quality Model; NADP = National Atmospheric
Deposition Program; GIS = Geographic Information System; SSURGO = Soil Survey
Geographic Database
10
September, 2010
Draft - Do Not Quote or Cite
-------
1 Table 4. Gibbsite equilibrium (Kgibb) constant determined by percentage of soil organic matter
2 (modified from McNulty et al. 2007).
Soil Type Layer
Mineral soils: C layer
Soils with low organic matter: B/C layers
Soils with some organic material: A/E
layers
Peaty and organic soils: organic layers
Organic
Matter %
<5
5tol5
15 to 70
>70
Kgibb (m6/eq2)
950
300
100
9.5
September, 2010
B-9
Draft - Do Not Quote or Cite
-------
1
2
3
Table 5. Terrestrial critical acid loads (in eq/ha/yr) for the watersheds in the Adirondacks and
Shenandoah Case Study Areas.
Case Study
Area
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Adirondacks
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
Shenandoah
HUC12
020100010103
020100040203
020100080304
020100081602
020200020101
020200020704
020200040805
041501011001
041503020801
041503040102
041503040204
041503050103
041503050104
041503050302
041503050407
041503050601
020700050401
020700050502
020700050703
020700050705
020700050801
020700050803
020700060101
020801030301
020801030402
020802010702
020802010703
020802010801
020802010803
020802020102
020802020401
020802030601
Terrestrial Critical Acid Load
(eq/ha/yr)
Bc/Al = 1.2
2045
1316
1329
1670
1484
1707
1770
1770
1664
1627
1436
1774
1794
1754
1447
1203
1440
1560
1762
1852
1799
1975
1638
1511
1393
1603
1642
1635
1573
1519
1264
1660
Bc/Al = 10.0
1134
712
731
922
819
935
951
955
912
880
786
957
968
947
789
656
802
871
979
1032
1003
1102
914
843
776
890
912
909
876
845
703
918
September, 2010
B-10
Draft - Do Not Quote or Cite
-------
1 Maps were generated to compare the aquatic and terrestrial critical acid loads in each
2 watershed to determine which estimate provided the greatest protection against acidifying
3 nitrogen and sulfur deposition. In each watershed, the terrestrial critical load estimate was
4 compared against each aquatic critical load, and the load with the lowest value was set to
5 represent the most sensitive component in the watershed. All critical load estimates were
6 converted to eq/ha/yr for the comparisons.
7
8 Results
9 Maps indicating and comparing the sensitivities of the terrestrial and aquatic critical loads
10 to nitrogen and sulfur deposition in each watershed of the Adirondacks and Shenandoah Case
11 Study Areas are presented in Figures 1-4 and Tables 6-9.
12
13 In the Adirondacks Case Study Area, 7 of the 16 watersheds had terrestrial critical acid
14 loads (based on a Bc/Al of 10.0) that were lower and therefore more sensitive to acidification
15 than all the lakes in the watershed. However, when the terrestrial critical loads were calculated
16 with a Bc/Al soil solution ratio of 1.2, only 5 of the 16 watersheds were protected by a terrestrial
17 critical load that was lower than the aquatic critical loads of the lakes. Three watersheds in the
18 Adriondacks Case Study Area had terrestrial critical loads (based on a Bc/Al of 10.0) that were
19 lower and higher than the critical loads for the lakes in the watershed, and one watershed had a
20 similar mixture of aquatic versus terrestrial acid load protections for terrestrial critical loads
21 estimated with a Bc/Al of 1.2. In general, a main trend in the Adirondacks Case Study Area was
22 that watersheds with "Highly Sensitive" and "Moderately Sensitive" lakes were more protected
23 by aquatic than terrestrial critical acid loads, while the watersheds with "Low Sensitivity" and
24 "Not Sensitive" lakes were more protected by terrestrial critical acid loads.
25
26 Similar trends were found in the Shenandoah Case Study Area. However, there was little
27 distinction between terrestrial acid loads that were calculated with a Bc/Al of 10.0 versus 1.2.
28 Terrestrial critical acid loads offered a higher level of protection than did the stream aquatic
29 critical loads in only one watershed. The two lakes in this watershed had "Low Sensitivity" or
30 were "Not Sensitive" to acidifying nitrogen and sulfur deposition. The 15 watersheds that had
31 streams with aquatic critical loads lower and more protective than the terrestrial critical loads,
September, 2010 B-11 Draft - Do Not Quote or Cite
-------
1 were all "Highly Sensitive" or "Moderately Sensitive" to acidifying nitrogen and sulfur
2 deposition.
3
4 In summary, a comparison of the terrestrial and aquatic critical acid loads for watersheds
5 in the Adirondacks and Shenandoah Case Study Areas indicated that, in general, the aquatic
6 critical acid loads offered greater protection to the watersheds than did the terrestrial critical
7 loads. In situations where the terrestrial loads were more protective, the lakes or streams in the
8 watershed were rated as having "Low Sensitivity" or "Not Sensitive" to acidifying nitrogen and
9 sulfur deposition. Conversely, when the water bodies were more sensitive to deposition
10 ("Highly Sensitive" or "Moderately Sensitive"), the aquatic critical acid loads consistently
11 provided a greater level of protection against acidifying nitrogen and sulfur deposition in the
12 watershed.
13
14
15
September, 2010 B-12 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
Figure 1. Comparison of aquatic and terrestrial critical loads of acidification (in eq/ha/yr) in the
16 watersheds of the Adirondacks Case Study Area, based on an ANC of 50 eq/L for the aquatic
loads and a Bc/Al of 10.0 for the terrestrial loads. Colored circles indicate the locations of the
waters bodies within each watershed. Green circles indicate lakes with critical load values less
than the terrestrial critical load for the same watershed. Red circles indicate a condition where
the terrestrial critical load is lower than the lake critical load.
Aquatic more sensitive
Terrestrial more sensitive
Adirondacks State Park
HUC 12 Watersheds
September, 2010
B-13
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
Figure 2. Comparison of aquatic and terrestrial critical loads of acidification (in eq/ha/yr) in the
16 watersheds of the Adirondacks Case Study Area, based on an ANC of 50 eq/L for the aquatic
loads and a Bc/Al of 1.2 for the terrestrial loads. Colored circles indicate the locations of the
waters bodies within each watershed. Green circles indicate lakes with critical load values less
than the terrestrial critical load for the same watershed. Red circles indicate a condition where
the terrestrial critical load is lower than the lake critical load.
Aquatic more sensitive
Terrestrial more sensitive
Adirondacks State Park
HUG 12 Watersheds
September, 2010
B-14
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
Figure 3. Comparison of aquatic and terrestrial critical loads of acidification (in eq/ha/yr) in the
16 watersheds of the Shenandoah Case Study Area, based on an ANC of 50 eq/L for the aquatic
loads and a Bc/Al of 10.0 for the terrestrial loads. Colored circles indicate the locations of the
waters bodies within each watershed. Green circles indicate streams with critical load values
less than the terrestrial critical load for the same watershed. Red circles indicate a condition
where the terrestrial critical load is lower than the stream critical load.
700050401
020802020102 VIRGINIA
• Aquatic more sensitive
• Terrestrial more sensitive
] HUC 12 Watersheds
National Forests
September, 2010
B-15
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
Figure 4. Comparison of aquatic and terrestrial critical loads of acidification (in eq/ha/yr) in the
16 watersheds of the Shenandoah Case Study Area, based on an ANC of 50 eq/L for the aquatic
loads and a Bc/Al of 1.2 for the terrestrial loads. Colored circles indicate the locations of the
waters bodies within each watershed. Green circles indicate streams with critical load values
less than the terrestrial critical load for the same watershed. Red circles indicate a condition
where the terrestrial critical load is lower than the stream critical load.
WEST VIRGINIA
020700050803
7^~\Jt/
020700050801
O
020700050401
•
102 VIRGINIA
Legend
Aquatic more sensitive
Terrestrial more sensitive
HUC 12 Watersheds
National Forests
September, 2010
B-16
Draft - Do Not Quote or Cite
-------
Table 6. Relative sensitivities of aquatic versus terrestrial critical loads in the 29 lakes and 16 watersheds of the Adirondack Case
Study Area (based on an ANC of 50 jieq/L for the aquatic loads and a Bc/Al of 10.0 for the terrestrial critical loads and common unit
of eq/ha/yr) to acidifying nitrogen and sulfur deposition. Lake IDs are indicated in each cell and are from the REA REPORT (US
EPA, 2009). Green text indicates lakes where the aquatic critical load was less than the terrestrial critical load value for the
watershed. Red text indicates a condition where the terrestrial critical load for the watershed was lower than the aquatic critical load
for the lake within the same watershed.
HUC
020100010103
020100040203
020100080304
020100081602
020200020101
020200020704
020200040805
041501011001
041503020801
041503040102
041503040204
041503050103
041503050104
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
] " •
' '' j
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
NY793L
Low Sensitivity
(CL = 101-200
meq/m2/yr)
1A2-028O
NY310L
NY292L
Not Sensitive
(CL > 201
meq/m2/yr)
NY534L
NY308L
NY312L
NY313L
NY500L
NY783L
September, 2010
B-17
Draft - Do Not Quote or Cite
-------
HUC
041503050302
041503050407
041503050601
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
' • ' . . ' i
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
NY528L
NY769L
Low Sensitivity
(CL = 101-200
meq/m2/yr)
NY008L
NY007L
NY768L
NY004L
Not Sensitive
(CL > 201
meq/m2/yr)
September, 2010
B-18
Draft - Do Not Quote or Cite
-------
Table 7. Relative sensitivities of aquatic versus terrestrial critical loads in the 29 lakes and 16 watersheds of the Adirondack Case
Study Area (based on an ANC of 50 jieq/L for the aquatic loads and a Bc/Al of 1.2 for the terrestrial critical loads and common unit of
eq/ha/yr) to acidifying nitrogen and sulfur deposition. Lake IDs are indicated in each cell and are from the REA REPORT (US EPA,
2009). Green text indicates lakes where the aquatic critical load was less than the terrestrial critical load value for the watershed. Red
text indicates a condition where the terrestrial critical load for the watershed was lower than the aquatic critical load for the lake within
the same watershed.
HUC
020100010103
020100040203
020100080304
020100081602
020200020101
020200020704
020200040805
041501011001
041503020801
041503040102
041503040204
041503050103
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
NY310L
Not Sensitive
(CL > 201
meq/m2/yr)
NY534L
NY308L
NY312L
NY313L
NY500L
NY783L
September, 2010
B-19
Draft - Do Not Quote or Cite
-------
HUC
041503050104
041503050302
041503050407
041503050601
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
M t
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
i M 1 '
NY004L
Not Sensitive
(CL > 201
meq/m2/yr)
September, 2010
B-20
Draft - Do Not Quote or Cite
-------
Table 8. Relative sensitivities of aquatic versus terrestrial critical loads in the 20 streams and 16 watersheds of the Shenandoah Case
Study Area (based on an ANC of 50 jieq/L for the aquatic loads and a Bc/Al of 10.0 for the terrestrial critical loads and common unit
of eq/ha/yr) to acidifying nitrogen and sulfur deposition. Stream IDs are indicated in each cell and are from the REA REPORT (US
EPA, 2009). Green text indicates streams where the aquatic critical load was less than the terrestrial critical load value for the
watershed. Red text indicates a condition where the terrestrial critical load for the watershed was lower than the aquatic critical load
for the stream within the same watershed.
HUC
020700050401
020700050502
020700050703
020700050705
020700050801
020700050803
020700060101
020801030301
020801030402
020802010702
020802010703
020802010801
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
iii
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
VT60
Not Sensitive
(CL > 201
meq/m2/yr)
VT61
September, 2010
B-21
Draft - Do Not Quote or Cite
-------
HUC
020802010803
020802020102
020802020401
020802030601
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
Not Sensitive
(CL > 201
meq/m2/yr)
September, 2010
B-22
Draft - Do Not Quote or Cite
-------
Table 9. Relative sensitivities of aquatic versus terrestrial critical loads in the 20 streams and 16 watersheds of the Shenandoah Case
Study Area (based on an ANC of 50 jieq/L for the aquatic loads and a Bc/Al of 1.2 for the terrestrial critical loads and common unit of
eq/ha/yr) to acidifying nitrogen and sulfur deposition. Stream IDs are indicated in each cell and are from the REA REPORT (US
EPA, 2009). Green text indicates streams where aquatic critical loads were less than the terrestrial critical load value for the
watershed. Red text indicates a condition where the terrestrial critical load for the watershed was lower than the aquatic critical load
for the stream within the same watershed.
HUC
020700050401
020700050502
020700050703
020700050705
020700050801
020700050803
020700060101
020801030301
020801030402
020802010702
020802010703
020802010801
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
. •' i
iii
•. i l
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
VT60
Not Sensitive
(CL > 201
meq/m2/yr)
VT61
September, 2010
B-23
Draft - Do Not Quote or Cite
-------
HUC
020802010803
020802020102
020802020401
020802030601
CURRENT CONDITION OF ACIDITY AND
SENSITIVITY TO ACIDIFICATION CATEGORY
Highly Sensitive
(CL < 50
meq/m2/yr)
Moderately Sensitive
(CL = 51-100
meq/m2/yr)
Low Sensitivity
(CL = 101-200
meq/m2/yr)
Not Sensitive
(CL > 201
meq/m2/yr)
September, 2010
B-24
Draft - Do Not Quote or Cite
-------
REFERENCES
Gebert, W.A., DJ. Graczyk, and W.R. Krug, 1987. Average Annual Runoff in the United States,
1951-80: U.S. Geological Survey Hydrologic Investigations Atlas HA-710, Scale
1:7,500,000. GIS datalayer. U.S. Department of Interior, U.S. Geological Survey,
Madison, WI. Available at: http://water.usgs.gov/GIS/dsdl/runoff.eOO.gz (accessed
September 9, 2009).
McNulty, S.G., B.C. Cohen, H. Li, and J.A. Moore-Myers. 2007. Estimates of critical acid loads
and exceedences for forest soils across the conterminous United States. Environmental
Pollution 149:281-292.
NADP (National Atmospheric Deposition Program). 2003 a. Annual Calcium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003b. Annual Chloride Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003c. Annual Magnesium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003d. Annual Potassium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
Survey, Champaign, IL. Available at http://nadp.sws.uiuc.edu/maps/2002.
NADP (National Atmospheric Deposition Program). 2003e. Annual Sodium Wet Deposition,
2002. GIS datalayer. National Atmospheric Deposition Program, Illinois State Water
UNECE (United Nations Economic Commission for Europe). 2004. Manual on Methodologies
and Criteria for Modeling and Mapping Critical Loads and Levels and Air Pollution
Effects, Risks, and Trends. Convention on Long-Range Transboundary Air Pollution,
Geneva Switzerland. Available at http://www.icpmapping.org (accessed August 16,
2006).
USDA-NRCS (United States Department of Agriculture-Natural Resources Conservation
Service). 2008. Soil Survey Geographic (SSURGO) Database. GIS datalayer. U.S.
Department of Agriculture, Natural Resources Conservation Service, Washington, DC.
Available at http://datagateway.nrcs.usda.gov.
US EPA (United States Environmental Protection Agency). 2009. Risk and Exposure Assessment
for Review of the Secondary National Ambient Air Quality Standards for Oxides of
Nitrogen and Sulfur. Final. U.S. Environmental Protection Agency, Office of Research
and Development, National Center for Environmental Assessment, Research Triangle
Park, NC. September.
USGS (U.S. Geological Survey). 2009. State Geological Map Compilation. U.S. Department of
the Interior, U.S. Geological Survey, Reston, VA. Available at:
http://tin.er.usgs.gov/geology/state (accessed January 28, 2009).
September, 2010 B-25 Draft - Do Not Quote or Cite
-------
1
2 APPENDIX C: Elasticity AAPI sensitivity analyses
3 Elasticity. One metric for determining sensitivity of the AAPI to its component parameters is
4 elasticity. Elasticity measures the percent change in AAPI for a one percent change in the
5 component. In general, the formula for an elasticity is:
R
AAPI _ dAAPI
Xi AAPI
9 Where Ex is the elasticity of AAPI with respect to component XJ; and j is the number of components.
10
11 So, for AAPI defined as
12
13 AAPI = ±Neco
14
15 The set of relevant elasticities are:
16
17 For runoff, Q:
18
19 EAQAP1 = --^\Neco - NHx - TNOyNOy - TSOxSOx]x-JL_, which can be rewritten as
20
21 Efpl = - — \AAPI -BC:]x-Q—, or
2 0L °J AAPI
22
September, 2010 C-1 Draft - Do Not Quote or Cite
-------
BC;
AAPI
3 For BC*
AAPI
1 For Neco,
8
9 pAAPI -_L.
Neco ~ f)
10
11 For NHx,
12
-'
13 E
AAPI _ _ J_ NHx
NHx
Q AAPI
14
15 For TNOy,
16
17
* NOy
18
19 For TSOx,
20
21
Q AAPI
September, 2010
C-2 Draft - Do Not Quote or Cite
-------
SO> Q
2
3 For NOy,
4
1 T NOy
Q y AAPI
6
7 For SOx,
8
T-^AAPI 1 rf,
Q "ux AAPI
10
11 These elasticities can be evaluated at various points along the ranges of each component, as well as along
12 ranges of the AAPI. We evaluate the elasticities at the sample means, medians, first quartiles, and third
13 quartiles. Elasticities are evaluated only for ecoregions that overlap the CMAQ modeling domain which
14 provides values for reduced nitrogen and the transformation ratios (TNOx and TSOx). This will provide a
15 reasonable assessment of the sensitivity of the AAPI to input components. Table 1 provides the estimated
16 elasticities. Elasticities are summarized across ecoregions using means, medians, minimums, and
17 maximums.
18 Note that elasticities can be either positive or negative. A negative elasticity means that the calculated
19 AAPI will decrease as a component increases. The magnitude of the elasticity depends on the values of
20 the components and the starting value of AAPI.
21 Based on the calculated elasticities, AAPI is most responsive to changes in Q, BCO, and Neco with some
22 responsiveness to reduced N. Note that for some components, such as Q, the elasticities switch signs
23 depending on the values of the variables for which the elasticity is evaluated. This suggests potentially
24 important interactions. AAPI is not responsive to the transformation ratios, TNOx and TSOx at mean
25 values of the AAPI components. However, when the elasticities for TNOx and TSOx are evaluated at the
26 first quartiles of the data, some locations in the Eastern U.S. show higher responsiveness to changes in
27 TNOx and TSOx, with elasticities as high as 2.
28
29
September, 2010 C-3 Draft - Do Not Quote or Cite
-------
1 Table 1. Elasticity* of AAPI to Component Variables
AAPI
Component
Runoff (Q)
Base Cation
Weathering (BCO)
o
o
(L>
fc
Reduced Nitrogen
*+J
C3
^•1-
& 1* i
§ £ S fc,
Metric for
Which
Elasticities
are
Evaluated
Mean
Median
1st Quartile
3rd Quartile
Mean
Median
1st Quartile
3rd Quartile
Mean
Median
1st Quartile
3rd Quartile
Mean
Median
1st Quartile
3rd Quartile
Mean
Median
1st Quartile
3rd Quartile
Mean
Elasticity
Across
Ecoregions
-0.1047
0.2561
-0.9283
0.2988
0.8953
1.2561
0.0717
1.2988
0.0179
0.3376
0.3543
0.3016
-0.0409
-0.1407
-0.0308
-0.1128
0.0089
-0.0061
-0.0191
-0.0154
Median
Elasticity
Across
Ecoregions
0.1221
0.1426
0.1303
0.1110
1.1221
1.1426
1.1303
1.1110
0.1464
0.2563
0.2596
0.1440
-0.0702
-0.1031
-0.1190
-0.0615
-0.0053
-0.0064
-0.0071
-0.0044
Minimum
Elasticity
Across
Ecoregions
-20.4572
-6.2481
-135.2544
-0.1810
-19.4572
-5.2481
-134.2544
0.8190
-13.8545
-4.7203
-115.1112
0.0137
-0.4708
-0.7957
-33.9063
-1.9167
-0.0597
-0.0506
-2.7061
-0.5028
Maximum
Elasticity
Across
Ecoregions
1.6005
2.4578
88.1063
7.5684
2.6005
3.4578
89.1063
8.5684
1.6051
2.5044
137.0526
6.5565
3.9332
1.0061
35.3783
-0.0050
1.0598
0.2751
1.7749
-0.0002
September, 2010
C-4
Draft - Do Not Quote or Cite
-------
AAPI
Component
SOx Transformation
Ratio (TSOx)
Metric for
Which
Elasticities
are
Evaluated
Mean
Median
1st Quartile
3rd Quartile
Mean
Elasticity
Across
Ecoregions
0.0019
-0.0045
0.0040
-0.0058
Median
Elasticity
Across
Ecoregions
-0.0024
-0.0036
-0.0032
-0.0021
Minimum
Elasticity
Across
Ecoregions
-0.0185
-0.0413
-1.8023
-0.1534
Maximum
Elasticity
Across
Ecoregions
0.3608
0.1091
1.9860
-0.0002
2 *Elasticity is the percent change in AAPI for a one percent change in the component variable.
3 For example, when evaluated at the means of all component variables, the mean elasticity of
4 AAPI to the runoff variable Q is -0.1047, which means that for each 1 percent increase in Q, the
5 AAPI is reduced by 0.1047 percent.
September, 2010
C-5
Draft - Do Not Quote or Cite
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/P-10-008
Environmental Protection Health and Environmental Impacts Division September, 2010
Agency Research Triangle Park, NC
------- |