EPA/600/R-11/068 June 2011 www.epa.gov/ord
vvEPA
United States
Environmental Protection
Agency
Summary Report of the
Atmospheric Modeling and
Analysis Division's
Research Activities for 2010
Office of Research and Development
National Exposure Research Laboratory
-------
-------
EPA/600/R-11/068 June 2011 www.epa.gov/ord
Summary Report of the
Atmospheric Modeling
and Analysis Division's
Research Activities for 2010
S.T. Rao, Jesse Bash, Sherry Brown, Robert Gilliam, David Heist, David Mobley,
Sergey Napelenok, Chris Nolte, and Tom Pierce
Atmospheric Modeling and Analysis Division
National Exposure Research Laboratory
Office of Research and Development
United States Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
Disclaimer
The information in this document has been funded by the U.S. Environmental Protection Agency. It has been subjected to the
Agency's peer and administrative review and has been approved for publication as an EPA document. Mention of trade names
or commercial products docs not constitute endorsement or recommendation for use.
ii
-------
Abstract
The research presented here was performed by the Atmospheric Modeling and Analysis Division (AM AD) of the National
Exposure Research Laboratory in the U.S. Environmental Protection Agency's (EPA's) Office of Research and Development
in Research Triangle Park. NC. The Division leads the development and evaluation of predictive atmospheric models on all
spatial and temporal scales for assessing changes in air quality and air pollutant exposures, as affected by changes in ecosystem
management and regulatory decisions, and for forecasting the Nation's air quality and reducing exposures to sensitive
populations and ecosystems. A MAD is responsible for providing a sound scientific and technical basis for regulatory policies
to improve ambient air quality. The models developed by A MAD arc being used by EPA and the air pollution community in
understanding and forecasting the magnitude of the air pollution problem and also in developing emission control policies and
regulations for air quality improvements. AMAD applies air quality models to support key integrated, interdisciplinary science
research. This includes linking air quality models to other models in the sourcc-to-outcome continuum framework to effectively
address issues involving human health and ecosystem exposure science. The Community Multiscale Air Quality Model is the
flagship model of the Division. This report summarizes the research and operational activities of AMAD for calendar year
2010.
iii
-------
-------
Table of Contents
Disclaimer
Abstract
List of Figures
1.0 Introduction
2.0 Summary of Accomplishments for the Division
2.1 Division-Wide Accomplishments
2.2 Model Development and Diagnostic Testing
2.3 Air Quality Model Evaluation
2.4 Air Quality-Global Climate Change
2.5 Linking Air Quality to Human Exposure
2.6 Linking Air Quality and Ecosystems
3.0 Model Development and Diagnostic Testing
3.1 Overview of Air Quality Model Development
3.2 CMAQ Aerosol Module
3.3 CMAQ Gas-Phase Chemistry
3.4 Planetary Boundary Layer Modeling
3.5 Meteorology Modeling for Air Quality
3.6 Integrated Meteorology-Chemistry Modeling
3.7 Mercury Modeling
3.8 Air Toxics
3.9 Nanoparticles Modeling
3.10 Emissions Modeling Research
4.0 Air Quality Model Evaluation
4.1 Introduction
4.2 Atmosphepheric Model Evaluation Tool
4.3 Air Quality Model Evaluation Initiative
4.4 Diagnostic Evaluation
4.5 Dynamic Evaluation
4.6 Probabilistic Model Evaluation
5.0 Climate and Air Quality Interactions
5.1 Introduction
5.2 Regional Climate Modeling: Dynamical Downscaling
-------
5.3 Regional Climate Modeling: Statistical Downscaling 33
5.4 Development of Coupled Regional Climate and Chemistry Modeling System 34
5.5 Decision Support Tools for Managing Air Quality and Mitigating Climate Change 36
5.6 Biosphere-Atmosphere Interactions: Improving the Treatment of Isoprene Oxidation 37
6.0 Linking Air Quality to Human Health 39
6.1 Introduction 39
6.2. Research Description 39
6.3. Accomplishments 40
6.4. Next Steps 42
7.0 Linking Air Quality and Ecosystem 45
7.1 Introduction 45
7.2 Air Deposition and Ecosystem Services Assessments 46
7.2.1 Introduction 46
7.2.2 Research Direction 46
7.2.3 Accomplishments 47
7.2.4 Next Steps 47
7.3 Air-Ecosystem Linkage Studies 48
7.3.1 Introduction 48
7.3.2 Research Direction 48
7.3.3 Accomplishments 48
7.3.4 Next Steps 48
7.4 Model and Tool Development 48
7.4.1 Introduction 48
7.4.2 Research Direction 49
7.4.3 Accomplishments 49
7.4.4 Next Steps 51
References 53
APPENDIX A: Atmospheric Modeling and Analysis Division Staff Roster
(as of December 31, 2010) 57
APPENDIX B: Division and Branch Descriptions 58
APPENDIX C: 2010 Awards and Recognition 59
APPENDIX D: 2010 Publications (Division authors are in bold.) 60
APPENDIX E: Acronyms and Abbreviations 63
-------
List of Figures
Figure 1-1. AMAD's role in the source-exposure-dose-effects continuum from the atmospheric
science perspective. (Adapted from "A Conceptional Framework for U.S. EPA's National
Exposure Research Laboratory," EPA/600/R-09/003, January 2009) 1
Figure 1-2. AMAD's structure and organization 2
Figure 1-3. AMAD's integrated interdisciplinary modeling research 2
Figure 4-1. Examples of two types of model evaluation techniques: (a) probabilistic—time series
of daily maximum 8-h 03 concentrations from a 200-member CMAQ model ensemble at a
monitoring site in an urban location; and (b) diagnostic—percent contribution of individual
aerosol species comprising the total average regional PM25 mass concentrations predicted by
CMAQ and measured by the Speciated Trends Network (STN) sites 25
Figure 4-2. Modeled (gray) and observed (rose) wind speed profiles averaged over the nocturnal
periods of August 11-15, 2002, at Ft. Meade, MD. Boxes span the 25th to 75th percentiles,
and whiskers extend from the 10th to 90th percentiles 27
Figure 4-3. Box/whisker plot of the percentage change in modeled (gray) and observed (green)
maximum 8-h 03 concentrations (>95th percentile) for each summer period relative to a 5-year
mean at the CASTNET monitoring sites in the eastern United States. The boxes span the 25th
to 75th percentiles and whiskers extend from the 10th to 90th percentiles 28
Figure 4-4. Change in weekday 3-h average morning NOx concentrations. Model results (gray) and
observations (white) are based on 42 urban sites. Each box/whisker plot shows the median
values (line inside boxes). Boxes span the 25th to 75th percentiles, and whiskers extend from
the 10th to the 90th percentiles of the concentration distributions 28
Figure 4-5. Base model, observational, and model ensemble empirical cumulative distributions of
maximum daily average 8-h 03 concentrations for 2002 and 2005 at two AQS sites: (a) Terre
Haute, IN; and (b) Detroit, Ml. All ensembles were constructed based on ±50% uncertainty in
emissions of area and mobile NOx and ±3% uncertainty in emissions of point NOx. The wide
spread of the ensemble at the Terre Haute site indicates greater sensitivity to NOx emissions in
comparison with the site in Detroit 30
Figure 5-1. Example global telescoping grid structure 35
Figure 5-2. Schematic describing model linkages 36
Figure 5-3: Change in simulated mean surface-level 03 concentration because of updated
chemical mechanism 38
Figure 6-1. Wind tunnel model of Las Vegas field study area built at 1:200 scale 40
Figure 6-2. Aerial photograph of study location in Raleigh, NC 41
vii
-------
Figure 6-3. Photograph of bag samplers and anemometers downwind of a simulated noise barrier
at the Idaho Falls field site 42
Figure 6-4. Example of a controlled burn of spilt oil in the Gulf of Mexico 43
Figure 6-5. Graphic on left shows high NOx emissions in Midwest (July 1997) and graphic on right
shows transport of pollution from Midwest into New York State. Red squares are locations
from which back-trajectories were initiated 43
Figure 7-1. Range of influence of NH3 emissions from a single, isolated Sampson County, NC cell
in 2002 CMAQ simulations using unidirectional, the deposition velocity as a surrogate, and
bidirectional NH3 surface exchange parameterizations, Dennis et al. (2010) 46
Figure 7-2. Monthly NH4 wet deposition bias when compared to NADP deposition for an annual
2002 simulation for a base case unidirectional exchange and a bidi case with bidirectional
NH3 exchange. The red diamond in the box plot represents the mean monthly deposition
bias. Monthly adjustments to precipitation biases have been applied where precipitation
errors significantly correlate with deposition error; a star above the box plot indicates that this
correction was not applied 47
Figure 7-3, Rainfed grain corn (A) example management schedule, (B) second fertilizer
application date, and (C) second first fertilizer application rate 50
Figure 7-4. Watershed deposition tool output showing the average (per unit area) annual total
(wet+dry) oxidized nitrogen deposition (kg-H/ha) estimated for each 12-digit HUC in the
Albemarle-Pamlico Basin for 2002 51
-------
1.0
Introduction
The research presented here was performed by the
Atmospheric Modeling and Analysis Division (AMAD)
of the National Exposure Research Laboratory (NERL)
in the U.S. Environmental Protection Agency's (EPA's)
Office of Research and Development (ORD) in Research
Triangle Park, NC. This report summarizes the research
and operational activities of the Division for calendar
year 2010.
The Division structure includes the following four research
branches.
1. Atmospheric Model Development Branch (AMDB)
2. Emissions and Model Evaluation Branch
(EMEB)
3. Atmospheric Exposure Integration Branch (AEIB)
4. Applied Modeling Branch (AMB)
Included in this report are a list of Division employees
(Appendix A), missions of the Division and its branches
(Appendix B), awards earned by Division personnel
(Appendix C), citations for Division publications
(Appendix D). and a list of acronyms and abbreviations
used herein (Appendix E).
The Division's role within the EPA NERL's "Exposure
Framework" and ORD's source-to-outcome continuum
is to conduct research that improves the Agency's
understanding of the linkages from source to exposure
(see Figure 1-1). Through its four research branches, the
Division provides atmospheric sciences expertise, air
quality forecasting support, and technical guidance on the
meteorological and atmospheric pollution aspects of air
quality management to various EPA offices (including the
Office of Air Quality Planning and Standards [OAQPS]
and regional offices), other Federal agencies, and State and
local pollution control agencies.
The Division provides this technical support and expertise
using an interdisciplinary approach that emphasizes
integration and partnership with EPA and public and
private research communities. Specific research and
development activities are conducted in-house and
externally via external funding.
The Division's research activities were subjected to a
comprehensive peer review in January 2009. (Additional
information from the peer review is available on the
Division's Web site [www.epa.gov/amad/].) To present
materials and programs for the peer review, the Division's
activities were summarized with focuses on five outcome-
oriented theme areas:
1. model development and diagnostic testing,
2. air quality model evaluation,
3. climate and air quality interactions,
4. linking air quality to human health, and
5. linking air quality and ecosystem health.
Source-to-Outcome
Adverse
Outcome
Source / Origin
Division s
Role
Fate and
Transport
Early Signs of
Effects
Exposure
Ambient
~
X
Concentrations
Figure 1-1. AMAD's role in the source-exposure-dose-effects continuum from the atmospheric science perspective.
(Adapted from "A Conceptional Framework for U.S. EPA's National Exposure Research Laboratory,"
EPA/600/R-09/003, January 2009)
-------
Sound Science tor Environmental Decisions
Model Development and Diagnostic Testing
Model Evaluation: Establishing Model s Credibility
Climate Change and Air Quality Interactions
Linking Air Quality and Human Health
Linking Air Quality and Ecosystem Health
~o co
m
AIMAD Structure and Organization
Figure 1-2. AMAD's structure and organization.
Research tasks were developed within each theme area by
considering the following questions.
• Over the next 2 to 3 years, who are the major clients,
and what are their needs?
• What research investments are needed to further the
science in ways that help the clients? How will we
lead or influence the science in this area?
• What personnel expertise, resources, and partners are
needed to do this work?
• Does the proposed work fall within the current
scope and plans of existing projects, or would
personnel resources need to be shifted from other
projects to make this happen?
The result is a research strategy for meeting user needs that
is built around the above-mentioned five major theme areas
and supported by the four branches of the Division as
depicted in Figure 1-2.
This report summarizes the research and operational
activities of the Division for calendaryear 2010. The
integration of research activities is illustrated in Figure
1-3. This report includes descriptions of research and
operational efforts in air pollution meteorology, air
quality model development, and model evaluation and
applications. Chapters 2 through 6 of this report are
organized according to the five major program themes
listed above (and shown in Figure 1-2).
AMAD's Intregrated Interdisciplinary Modeling Research
Climate
Change Air Quality
Interactions
Extreme Weather
(heat waves, forest
fires, floods,droughts)
Climate Model Downscaling
(clouds, aerosols precip)
Impacts On
Human Health
Impacts On
Ecosystems
Hydrological modeling,
precip prediction, atoms.deposition
Air quality,
i exposure
I modeling
Projected future
air quality-climate
(change effects
Projected air qualityt
conditions
Integrated Atmospheric
Modeling Systems
• Meteorology, air quality modeling
• Data analysis
• CMAQ, AERMOD, other models
\
Weather, Climate,
Air Quality
Observations
Regulatory Policies
Evaluation v
Accountability
of Regulations ™
Figure 1-3. AMAD's integrated interdisciplinary modeling research.
-------
2.0
Summary of Accomplis
As a summary of and introduction to the annual report
for 2010, the following Division accomplishments arc
highlighted.
2.1 Division-Wide Accomplishments
1. Issue: Strategic thinking regarding integrated
transdisciplinary approaches to solving
environmental problems.
Accomplishment: Coordination of various articles
for the November 2010 issue of the Air and Waste
Management Association's Environmental Manager
(EM) on Integrated Transdisciplinary Research
Findings: To properly address complex
environmental problems, there is growing consensus
among the research community that we must move
away from the current single-pollutant, single-
medium. and single-discipline approach to problem
solving with a more integrated transdisciplinary
approach. The articles in this issue address the need
for integrated transdisciplinary research to solve the
environmental problems of tomorrow.
Impact: This EM special issue provides a thought
-provoking set of articles for managers to consider
integrated transdisciplinary research for problem
solving—with examples of complex environmental
problems confronting us now (e.g., climate change,
transportation fuels in the 21st century, human
exposure near roadways and in urban environments).
2. Issue: Developing AMAD's strategic research plan
Accomplishment: A comprehensive Research
Strategy was developed outlining the Division's
core research in atmospheric model development
and evaluation, as well as three major application
areas in which air quality models arc used for human
exposure, ecosystems exposure, and climate change
air quality assessments.
Findings: AMAD's research directions were
articulated in three white papers on emerging
application areas. These white papers were reviewed
by an external panel of experts. The panel's report
was complimentary of the Division's proposed
research and included constructive comments and
suggestions.
Impact: AMAD's Strategic Plan was made
publicly available to promote transparency in the
Division's research planning process and facilitate
collaborations both within the Laboratory and
Agency as well as with external partners. The
Strategic Plan positioned AMAD to adapt to
emerging research directions within ORD.
its for the Division
3. Issue: A key uncertainty in quantification of aerosol
radiative effects and their impacts on climate
change is the verification of the spatial and temporal
variability in its magnitude and directionality and.
consequently, its cumulative effect on the radiation
balance of the earth-atmosphere system.
Accomplishment: A research proposal was
developed for a comprehensive observational-
modeling study to investigate and verify aerosol
radiative effects for past air quality conditions (e.g.,
changes in sulfate burden arising from sulfur dioxide
| SO, | reductions as a result of Title IV of the Clean
Air Act [CAA]).
Findings: Based on a rigorous peer review, the
proposal was recommended for funding as a
Department of Energy-EPA Interagency Agreement
(IAG). This project will support three post doctoral
fellows.
Impact: The successful attainment of the project
goals will help quantify the impacts of
troposplieric aerosol loading on atmospheric
radiation budgets, and also help build confidence in
the use of such modeling tools (e.g., the two-way
WRF-CMAQ) for climate projection scenarios and
emission management options.
4. Issue: Systematic intcrcomparisons and evaluations
arc needed for regional air quality models over
different continental regions.
Accomplishment: The second Air Quality Model
Evaluation International Initiative (AQMEII)
Workshop was held September 26 and 27, 2010, in
Turin, Italy. The meeting provided a venue for the 50
participants to share and discuss progress on model
simulations and data analysis related to this modcl-
intcrcomparison study.
Findings: The workshop provided a venue to discuss
and finali/c activities of the first phase of model
intcrcomparison and evaluation. All participating
groups made a commitment to submit their model
data to a central site by December. Participants also
developed plans for joint publications for a special
issue in Atmospheric Environment, manuscripts arc
due in May 2011.
Impact: A model intcrcomparison exercise has been
initiated for U.S., Canadian, and European air quality
modeling systems to be applied on each continent for
full-year simulations for operational and diagnostic
evaluations. This is the first of its kind international
collaborative effort in air quality modeling.
3
-------
5. Issue: Understanding global air pollution issues to
develop sustainable solutions for air quality and
climate.
Accomplishment: Developed a collaborative
proposal with N.C. State University (NCSU) and the
University of North Carolina-Chapel Hill (UNC-CH)
to host a U.S.-India workshop on air quality; the
proposal was funded by the National Science
Foundation (NSF).
Findings: The rapid growth of the Indian economy,
spurred by industrial and urban expansion, has
been accompanied by environmental stresses,
particularly in air quality. The current capacity in
India in research and operational aspects of air
quality forecasting and regulatory management
needs further enhancement. A workshop, to be held
March 14-24, 2011, with partial support from the
NSF and the U.S. Department of Energy, will bring
together several experts from the United States and
India with a common vision for identifying priority
areas of research and development in the air quality
and climate area, and a commitment for long-term
collaboration.
Impact: To engage participation by emerging
scientists in the field, the workshop will include
three days of invited lectures and presentations,
followed by a 7-day hands-on training segment.
This portion of the workshop will provide training
to participants on a publicly available suite of
numerical models (i.e., YVRF, CMAQ) that are
used world wide in a variety of air quality
applications, from basic research to local- and
regional-level planning and management. The
workshop will strive to improve understanding of the
emission sources and meteorological conditions that
contribute to regional-to-urban-scale air quality and
climate issues of relevance to protecting public health
and the environment.
6. Issue: Supporting international cooperation on
air pollution modeling under the North Atlantic
Treaty Organization (NATO).
Accomplishment: Contributed substantially to the
organization and program development for the
September 27 to October 1, 2010, 31st NATO/
Science for Peace and Security (SPS) International
Technical Meeting (ITM) on Air Pollution Modeling
and its Applications, in Turin, Italy.
Findings: The ITM has been broadened
(Topic 7) from air quality and human health to cover
ecosystems and economy (including air quality
trends, cost-benefit analysis of regulatory programs
and their effectiveness, and integrated modeling
approaches).
Impact: More than 130 participants from 35
countries attended the NATO/SPS meeting,
presenting papers on a wide variety of air pollution
modeling topics ranging from local- to global-scale
applications. The meeting provided an important
forum for synthesizing progress on air quality
modeling programs around the world. All papers arc
subject to peer review by the scientific community
and will be published in a book under the NATO
banner.
7. Issue: Promoting collaboration between U.S. and
U.K. scientists on air pollution exposure research.
Accomplishment: Nearly 50 international experts,
including NERL scientists and others from the U.K.
Department for Environment. Food, and Rural
Affairs, and the Environmental Agency for England
and Wales, met during December 6-10, 2010, in
London to discuss progress in collaborations to
improve and apply air quality models.
Findings: NERL/AMAD scientists made several
presentations at the workshop to promote
collaborations in areas of air pollution modeling
and its use to guide policy development, local-scale
air pollution modeling and analysis, air pollution
climate interactions, and air quality human health
linkage studies.
Impact: This collaboration is proposed in recognition
of the common concerns for serious health, welfare,
and economic impacts of atmospheric pollution. This
collaboration recognizes the benefit of defining a
research enterprise that involves joint interests and
lays the foundation for a co-sponsored program that
would improve efficiencies, and avoid redundant
research activities and products through the
leverage of resources and capabilities within each
organization.
2.2 Model Development and
Diagnostic Testing
1. Issue: Scientifically credible air quality models arc
needed both for research purposes as well as
regulatory applications in the implementation of the
National Ambient Air Quality Standards.
Accomplishment: An interim version of the
Community Multiscale Air Quality (CMAQ)
modeling system (CMAQv4.7.1) was released
publicly.
Findings: CMAQv4.7.1 incorporated scientific
advancements in the representation of 3D advcctivc
transport, in numerical solution of the aqueous-
phase chemistry, incorporation of tools to assist
in diagnostic analysis, and several code fixes
recommended by the CMAQ user community.
-------
Impact: CMAQv4.7.1 is being used by the OITicc of
Air Quality Planning and Standards (OAQPS) in the
regulatory impact analysis for the Transport Rule.
Utility MACT Rule, and the Greenhouse Gas Rule.
2. Issue: As U.S. air quality improves, global
background pollutant concentrations play an
increasingly more important role in determining
compliance with U.S. ambient air standards.
Accomplishment: The CMAQ model was extended
to hemispheric scales. Annual simulations for the
year 2006 were completed, and model predictions
were compared with a variety of measurements from
surface, aircraft, and satellite-based platforms.
Findings: Air quality modeling results for O,. PM,
Hg, and other pollutants over the U nited States
arc sensitive to the specification of boundary
concentrations. Modeling over the hemispheric
domain enables robust examination of modeled
processes from an atmospheric budget perspective.
Impact: CMAQ modeling capability has now been
extended to the full Northern Hemisphere, enabling
consistent specification of North American boundary
concentrations and helping understand how the
intercontinental transport of pollution a fleets air
quality over the United States.
3. Issue: The combustion of engineered nanoniatcrials
in diesel fuel can significantly alter ambient
exposures to O,. fine PM, and hazardous air
pollutants, with potential health consequences.
Accomplishment: Completed Annual Performance
Measure (APM) 366: Model the local-scale fate and
transport of a combusted nanoniatcrial and its effect
on regional-scale air quality.
Findings: If nanoccriuin additives receive
widespread use. cerium emissions from the United
States will increase by a factor of 25. Ambient air
concentrations of cerium could exceed 20 ng/m3 in
major cities, representing a 100-fold increase from
their current levels.
Impact: Regulators in the Office of Transportation
and Air Quality can now estimate the large-scale
consequences of approving versus disapproving
these additives in the E21 lb registration process.
In addition. EPA health scientists, who arc seeking
to develop realistic scenarios for assessing the
health impacts of nanoccriuin additives, now have
reasonable estimates of the level to which humans
would be exposed in the future if the additives arc
registered and widely used.
2.3 Air Quality Model Evaluation
1. Issue: Despite a reliance on regional air quality
models by the international community, methods
for evaluating the integrity of these models remain
primitive.
Accomplishment: Under the Air Quality
Model Intcrnational Initiative (AQMEII), an
international consortium of model evaluation
experts representing 22 countries began to evaluate
their models for the North American and European
continents. AQMEII accepted an invitation to publish
a scries of papers in a special issue of Atmospheric
Environment.
Findings: Preliminary results on regional air
quality model evaluations across North America and
Europe were presented at an AQMEII data analysis
workshop in Turin. Italy, during September 2010. The
importance of boundary conditions on the accuracy
of model performance was stressed.
Impact: AQMEII is facilitating the sharing of
improved model evaluation techniques, which
will lead to improvements in regional air quality
model simulations and characterization of model
uncertainties for communication to environmental
decisionmakers.
2. Issue: Implementing effective emission control
strategics to manage the cfleets of niultipollutant
mixtures requires air quality models tliat can
predict the response of pollutant levels to changes in
emissions and meteorology.
Accomplishment: Several peer-reviewed articles
documenting the dynamic response of the CMAQ
model in simulating the weekly and niultiannual
variations of O, precursor emissions were
published in the scientific literature.
Findings: Although the CMAQ mode ling system
showed skill at replicating the overall response of
O, and Oxides of Nitrogen (NOx) concentrations
because of cliangcs in emissions and meteorology.
potential weaknesses were identified with the
characterization of O, boundary conditions and
the characterization of Volatile Organic Compound
(VOC) and NOx emissions.
Impact: The research findings arc informing the
program office regarding the uncertainty of
regional air quality models for O, attainment
demo nst rat ions and arc helping shape priorities
for model development, particularly in the areas
of O, boundary conditions and weekly cycles in O,
precursor emissions.
3. Issue: The quantification of uncertainty in air quality
modeling results has been an important but elusive
goal.
5
-------
Accomplishment: The Direct Decoupled Method
(DDM) was used to provide a probabilistic-based
dynamic evaluation of the CMAQ model, and the
results were summarized in a manuscript submitted
to Atmospheric Environment.
Findings: In addition to uncertainties in NOx
emissions, uncertainties in VOC emissions and
boundary conditions were identified as areas needing
further quantification.
Impact: Characterizing the dynamic response of
O, to changes in emissions suggests that
uncertainties in key inputs, such as emissions and
boundary conditions, should be explicitly considered
in regional air quality modeling simulations.
4. Issue: The public release of CMAQvS.O needs to be
evaluated.
Accomplishment: The evaluation team began
working closely with the development team and has
developed a schedule and protocol for model
evaluations.
Findings: Improved communication between
CMAQ model developers and evaluators will be
accomplished via an internal Wiki site.
Impact: It is anticipated that a robust model
evaluation will be accomplished in an efficient and
timely manner to support the release of CMAQvS.O
to the air quality modeling community.
2.4 Air Quality-Global Climate Change
1. Issue: Traditional techniques for dynamical
downscaling of global model results to the regional
scale have relied only on specification of boundary
conditions for the regional model. However, this
specification in itself is insufficient to constrain the
regional model. New techniques are needed to assure
better consistency between global and regional model
results.
Accomplishment: Completed APM 284: Test
model linkage approaches for downscaling global to
regional climate with WRF. WRF downscaling has
been tested using both global reanalysis data and
output from the Goddard Institute for Space Studies
(GISS) Model E Global Climate Model (GCM). Most
of the testing thus far has focused on spectral and
analysis nudging.
Findings: Initial testing of dynamical downscaling
from GISS Model E to WRF using various nudging
techniques suggests that WRF is able to provide
the needed regional texture to a simulated climate,
while maintaining fidelity at larger scales to the
driving fields. Results arc sensitive to the nudging
parameters; thus, more testing is needed to determine
best configuration for refined climate modeling.
Impact: AMAD's experiments with data assimilation
in the process of downscaling from global to regional
climate models have shown much promise in moving
this discipline forward. Initial results presented at
recent conferences have generated much discussion
and interest in the scientific community.
2. Issue: The WRF model originally was designed
for short-term numerical weather prediction. It
was not designed as a regional climate model.
Certain aspects of the model, particularly its radiation
budget and its treatment of soil heat and moisture
fluxes, must be examined and possibly improved for
long-term simulations.
Accomplishment: Two 20-year continuous WRF
simulations have been completed using different
radiation parameterizations available in WRF. Tliese
simulations did not use any nudging.
Findings: The top-of-atmosphcre and surface
radiation fields, a.s well as cloud cover, from these
simulations will be compared with 5 years of
satellite observations.
Impact: This work will add confidence to future
regional climate simulations conducted by WRF
and will facilitate its use as a tool to understand
climate-chemistry interactions.
3. Issue: Adjoint methods can be used to link the
impacts of pollution to sources. Building the adjoint
of a complex model like CMAQ requires a wide
range of expertise and a community approach.
Accomplishment: Convened a workshop with
international participation on the development of
CMAQ-adjoint methods for advanced sensitivity
analysis.
Findings: The CM AQ-adjoint workshop brought
together model developers for a intense development
session, where the components of the CM AQ-adjoint
were assembled together for the first time. Second,
a panel discussion was convened. Participants from
OAQPS and acadeniia outlined the use of CMAQ-
adjoint to address air pollution, ecosystem impact,
and climate change mitigation needs.
Impact: This meeting fostered the community of
model developers who arc continuing to develop
the model to better address air pollution, ecosystem
impact, and climate change mitigation needs.
4. Issue: Thus far. AMAD's air quality-climate research
has focused on the potential impacts of future global
climate change on air quality . The reverse process
(i.e., the impacts of local and regional air quality on
climate) is also of intense scientific interest.
Accomplishment: The WRF-CMAQ coupled
meteorology-chemistry model has been tested,
including direct aerosol feedback on shortwave
-------
(SW) radiation and O, feedback on longwave (LW)
radiation. Indirect feedback is under development.
Findings: A prototype two-way coupled atmospheric
modeling system, based on the WRF and CMAQ
models, has been developed and further tested. Direct
feedback of aerosols on SW radiation has been
successfully implemented. Initial testing suggests
that these effects can be large in regions with
significant aerosol loading. Comparisons with limited
measurements show improvements in simulation
skill for SW radiation and 2-m temperature. Inclusion
of direct effects leads to suppression of simulated
Planetary Boundary Layer (PBL) heights which can
then impact simulated air quality.
Impact: The two-way coupled WRF-CMAQ system
provides a framework to properly characterize the
spatial heterogeneity in radiative forcing associated
with short-lived aerosol and gases and. consequently,
to better understand their aggregate influence on
the earth's radiation budgets. This evolving system
is expected to play a critical role in the Agency "s
evolving research and regulatory applications
exploring air quality-climate interactions. The
flexible design of the system facilitates coupling
meteorological and chemical calculations at
finer temporal resolutions, which enables more
consistent applications at fine spatial scales to better
characterize variability in air quality and its linkage
with health studies.
2.5 Linking Air Quality to Human Exposure
1. Issue: Cohort studies designed to estimate human
health effects of exposures to urban pollutants require
accurate determination of ambient concentrations
to minimize exposure misclassification errors.
However, it is often difficult to collect concentration
information at each study subject location. In the
absence of complete subject-specific measurements,
land-use regression (LUR) models frequently
have been used for estimating individual levels of
exposures to ambient air pollution. The LUR models,
however, have several limitations mainly dealing
with extensive monitoring data needs and challenges
involved in their broader applicability to other
locations.
Accomplishment: Critically evaluated the LUR
model being used by the community.
Findings: A 2010 publication in Atmospheric
Environment by M. Johnson, V. Isakov, J. Touma.
S. Mukerjee, and H. O/kaynak presents the results
of evaluation of land use regression models used to
predict air quality concentrations in an urban area.
Modeled hybrid air quality concentrations of PM, 5,
NOx, and benzene in New Haven. CT, were used
as pseudo-observations to develop and evaluate the
different LUR models. LUR models appeared to
perform well in the training datasets. However, when
these LUR models were tested against independent
hold-out (test) datasets. their performance diminished
considerably. This indicated that that complex
emissions and atmospheric processes resulting from
meteorological, transport, and diffusion and chemical
mechanisms can severely limit the predictive power
of most LUR-based modeling applications. The
paper also provided recommendations on future
research to examine best ways to augment basic LUR
models with site-specific source-receptor information
generated from air quality models and to test how one
might transfer LUR model results across different
geographical locations or even countries.
Impact: AMAD's results confirm the challenges
facing the LUR community in attempting to
fit empirical response surfaces to spatially and
temporally varying pollution levels using LUR
techniques that arc site dependent. These results
also illustrate the potential benefits of enhancing
basic LUR models by utilizing air quality modeling
tools or concepts to improve the models' reliability
or transferability. The information derived from
this study will be used by EPA as a resource for
developing appropriate tools in support of exposure
assessments. Through this effort, the Divison has
helped advance exposure science.
2. Issue: A principal route of human exposure to
pollutants occurs for those living and working
within several hundred meters of roadways. A better
understanding of the mechanisms for near-road
exposures is needed.
Accomplishment: For the near-road research
program, developed wind tunnel and field study
databases to improve model algorithms for urban
roadways in support of human exposure and health
assessments.
Findings: A 2009 Journal of the Air & Waste
Management Association paper by V. Isakov and
co-investigators presents an innovative methodology
to link regional- and local-scale air quality models
with human exposure models. It shows the presence
of strong spatial gradients in exposures near
roadways and industrial facilities that can vary
by almost a factor of two across the urban area
and much higher at the high end of the exposure
distribution.
Impact: A principal route of human exposure to
pollutants occurs for those living and working
within several hundred meters of roadways. A better
understanding of the mechanisms of such exposure is
needed. The importance of this work was recognized
by EPA and external stakeholders by its inclusion
-------
in the proposed nitrogen dioxide (NO,)NAAQS for
near-road monitoring requirements. Results from
this work also were used by the FHA in addressing
near-road monitoring needs associated with their
settlement agreement in the Sierra Club litigation.
The Federal Highway Administration requested
EPA's guidance and expertise in implementing
their near road research requirements as part of
this litigation, and an LAG has been established to
that end. In addition to regulatory applications, the
nominated papers have been cited in numerous other
peer-reviewed journal articles related to near-road
and local-scale dispersion topics. Through this work,
the Division has helped to advance exposure science.
3. Issue: General consensus exists that populations
spending significant amounts of time near major
roads face increased risks for several adverse health
effects. State and local agencies are interested in
considering a roadway design that includes the
presence of roadside vegetation as a means of
mitigating air pollutant concentrations near roads.
However, there are potential advantages and
disadvantages of implementing vegetation to mitigate
near-road air quality impacts. Vegetation in urban
settings can provide numerous benefits beyond air
quality improvements, such as ecosystem services,
associated with improved phy sical and mental health
and community vitality . Potential disbenefits include
pollen production, water demand, channeling of
invasive pests and fire into the urban environment,
and exacerbation of sprawl by distancing buildings
and other land use activities from roadways. Trees
also may obstruct visibility on the road and reduce
wind speed, cause damage or injury by falling, and
create slippery conditions from dropped debris.
Ideally, a large suite of costs and benefits should be
evaluated in concert to optimize the use of urban
vegetation to protect human health and promote
sustainable communities. Clearly, more research is
needed to assess the role of vegetation in mitigating
air quality impacts from traffic emissions.
Accomplishment: As a first step in evaluating
this concept, in April 2010, A MAD scientists in
collaboration with scientists from NERL, National
Risk Management Research Laboratory (NRMRL),
and National Health and Environmental Effects
Research Laboratory (NHEERL) organized a
workshop on the Role of Vegetation in Mitigating
Air Quality Impacts from Traffic Emissions. The
workshop included representatives from government
agencies, academia. State and local agencies, and
nongovernmental environmental organizations with
expertise in air quality , urban forestry, ecosystem
services, and environmental policy . The participants
reviewed the current science and identified future
activities in evaluating the potential role of vegetation
in mitigating near-road air pollutant concentrations.
Findings: A niultidisciplinary group of researchers
and policy makers met to discuss the state of the
science regarding the potential of roadside vegetation
to mitigate near-road air quality impacts. The results
are summarized in a journal article in the January
2011 issue of EM by Richard Baldauf. Laura Jackson.
Gayle Hagler. Vlad Isakov. Greg McPherson, David
Nowak, Thomas Caliill. Max Zhang. Rich Cook.
Chad Bailey, and Pcriann Wood.
Impact: This workshop was a first step in evaluating
the potential role of vegetation in mitigating near-
road air pollutant concentrations. A niultidisciplinary
group of researchers and policymakers from
government agencies, academia. State and local
agencies, and nongovernmental environmental
organizations with expertise in air quality, urban
forestry, ecosystem services, and environmental
policy reviewed the current science and identified
future activities to further assess the role of
vegetation in mitigating air quality impacts from
traffic emissions.
4. Issue: In its mission to protect human health and
the environment. EPA implemented the NOx Budget
Trading Program (NBP) to reduce the emissions of
NOx and the secondarily formed 0,. These pollutants
and their precursors can be transported downwind,
contributing to pollutant levels at locations far from
the emission sources, potentially impacting human
health in downwind areas. This study investigated
the liealth impacts in New York State from exposure
to polluted air parcels transported from the Midwest.
Accomplishment: This study developed and applied
a methodology to identify and target the transport of
polluted air parcels and demonstrated that the risk for
hospital admission resulting from respiratory-related
illness was increased in New York State on those
days that the air parcel originated over the polluted
Ohio River Valley .
Findings: The results of this analy sis indicate tliat
the risk of being hospitalized for respiratory-related
illness in New York State is greater on those days
when air is transported from the Midwest into
New York State as compared to days when air is
transported from the North. Using a refined method
to examine air parcels moving through a boundary
drawn around high-emitting power plants in the
midwestern United States resulted in stronger
associations across more regions (significant odds
ratios ranging from 1.06 to 1.16 for the entire study
time period for six of the eight New York State
regions). An assessment of temperature and its
impact on the odds ratio calculation in the New York
City metropolitan region indicates that temperature
-------
alone does not explain the increased association
between air pollution and respiratory-related hospital
admissions.
Impact: This technique is unique in that it examines
the health impact of air pollution that can travel
hundreds of kilometers downwind of its source
region. It is one of the first studies to associate
transported air pollution with a health end point in
this manner. The approach developed by AMAD can
be used as an indicator of exposure from transported
air pollution.
2.6 Linking Air Quality and Ecosystems
1. Issue: Full implementation of the bi-directional flux
approach for ammonia (NH3) in CMAQ requires
grid-level information regarding fertiliser application
timing, depth, and amount for each agricultural crop
in the model domain. This information needs to be
produced in a nationally consistent, process-based
manner that can respond to changes in weather/
climate, atmospheric deposition, and land use/land
cover conditions.
Accomplishment: The Fertilizer Emission
Scenario Tool for CMAQ (FEST-C) and associated
agricultural fertilizer modeling system (AFMS)
were developed to produce the input information
needed for implementation of bi-directional CMAQ
(APM 372). FEST-C is the interface that enables
the user to identify the combination of preexisting
weather, soil, land use/land cover, and agricultural
management files that best describe the emissions
setting of interest and to execute the programs that
estimate the information needed by CMAQ. The
tool produces a CMAQ-rcady input file for use in
bi-directional CMAQ, as well as other supporting
variables of interest to Ecological Services Research
Program (ESRP) partners that can be read, analyzed,
and displayed using the Visualization Environment
for Rich Data Interpretation (VERDI) visualization
and analysis tool.
Findings: FEST-C was developed successfully, and
a beta version was released in September 2010. The
test client set indicated that the current capabilities
appear to be adequate. However, the management
input files need further improvement and evaluation
to fully meet the needs of our ESRP partners and to
reduce current emission uncertainties in the CMAQ
NH, bi-directional model.
Impact: Development of the FEST-C/AFMS
system informed the design and execution of the
CMAQ bi-directional NH, pilot study completed in
FY 10. Completion of the beta version of FEST-C/
AFMS has laid the foundation for completing the
full implementation of bi-directional CMAQ for
NH, in 2011. Completion of the tool increased the
visibility of AMAD within the ESRP. and requests
for fertilizer and supplemental variable information
have been received from scientists working with
the National Atlas. Future Midwest Landscapes,
and Albemarle-Pamlico Watershed Study teams
to support their nitrogen budget, watershed, and
water quality modeling work. A summary of the
FEST-C/AFMS system was presented at the 31st
International Meeting on Air Pollution Modeling and
its Application. September 27 to October 1, 2010, in
Turin, Italy.
2. Issue: Annual Mercury wet deposition from CMAQ
4.7 niultipollutant was biased high when compared
with the Mercury Deposition Network (MDN)
observations, and. when the CMAQ 4.7 was run
at hcniisplieric scale, mercury was depleted during
the period of the run. indicating that the mercury
sinks were not balanced by emissions. Oxidized
mercury (Hg[ 111) is soluble and reactive and removed
from the atmosphere much more rapidly than
elemental mercury (Hg[0]). The biases in mercury
wet deposition and depletion of mercury in the
hcniisplieric runs were the result of the parameterized
oxidation rates of Hg|0|.
Accomplishment: Hg|0| has an atmospheric
lifetime ranging from 0.5 to 1 year, and Hg[II| has
an atmospheric lifetime of days. The relatively slow
oxidation rates of Hg|0| to Hg|ll| and differences in
atmospheric lifetimes required the simulation and
evaluation of this chemical nicclianisni on two scales.
A model sensitivity study of mercury oxidation rates
was run at the lieniispheric and continental United
States (CONUS) scales. At the hcniisplieric scale, the
budget of mercury sources and sinks and the model's
ability to sustain ambient concentrations near the
global background were used to evaluate changes
to the mercury oxidation rates. Identical runs were
conducted at the CONUS scale and evaluated against
MDN and Ambient Mercury Network (AMNet)
observations.
Findings: A controversial gas phase reaction was
found to be driving both the depletion of mercury
at the hcniisplieric scale and model deposition
biases at the regional scale. The reaction rates in
the mercury chemical nicclianisni in CMAQ and all
other chemical transport models remain uncertain and
high-quality measurements of mercury oxidation and
reduction rates from chamber studies and estimates
using computational cliemistry arc needed to further
reduce uncertainty in the model results.
Impact: This refined chemical mechanism reduced
the biases in the modeled wet deposition, improved
the hcniisplieric mercury budget, and better captured
9
-------
the seasonal variation in mercury wet deposition. The
mercury chemical mechanism was refined based on
the sensitivity study that improved the hemispheric
mercury budget and reduced model deposition biases
to be included in CMAQ 5.0 and for use in model
simulations in support of the EPA's toxics rule.
3. Issue: FY 10 research published by AMAD
scientists suggests that full implementation of the
bi-directional flux approach for NH, in CMAQ
would alter the temporal and spatial simulation of
ammonia deposition contributing to excess nutrient
deposition to aquatic and terrestrial ecosystems
and the formation of fine particulates. The research
conclusions were based on a 1 -mo simulation,
however. Full implementation of this approach into a
research version of CMAQ will not be complete until
2011. An eastern U.S. pilot study was developed
and executed for an annual simulation to inform
the implementation process and to provide initial
estimates of the potential value of the investment to
incorporate bi-directional exchange of ammonia in
CMAQ.
Accomplishment: A pilot study was designed and
executed for the eastern U.S. 12-kni CMAQ grid.
Preliminary algorithms published in 2010 were added
to CMAQ. Parameter time series were estimated
offline for the full CMAQ domain and were then
combined with these algorithms within CMAQ to
estimate ammonia emissions from agricultural soils
amended with inorganic nitrogen, canopy uptake, and
net flux from the surface.
Findings: The pilot study suggests that the
bi-directional NH, flux CMAQ will reduce current
unidirectional estimations of nitrogen dry deposition
by a factor of two at background sites and by a
factor of three for the model domain as a whole.
Partitioning of nitrogen to the aerosol phase and wet
deposition is increased. Transport of reduced nitrogen
out of the modeling domain is increased by about
10%. Precipitation corrected wet deposition and
ambient aerosol estimates arc improved relative to
observations when this approach is used.
Impact: Numerous "lessons learned" were
transferred to the full model implementation process.
Findings of the pilot were presented to O AQPS.
which now is considering their implications for
ongoing rulemaking and the possible contributions
of bi-directional CMAQ results to the development
of new secondary standards. Four manuscripts arc
in development, and numerous presentations at
professional meetings have been made to begin
documenting the new algorithms.
4. Issue: To meet total maximum daily load (TMDL)
targets to restore water quality, it is necessary to
understand the sources of the deposition-driven
loading of nitrogen to water bodies, including state
and sector-level sources of atmospheric deposition.
This requires a sensitivity analysis with an air quality
model. "Brute force" sensitivity approaches do not
work for pollutants with complex interactions and
feedbacks, such as NH,. The capability to dcfensiblv
estimate source attribution for NH, deposition has
been lacking.
Accomplishment: Adaptation of the Decoupled
Direct Method-3D (DDM-3D) sensitivity technique
for atmospheric deposition in CMAQ4.7.1 for
reduced nitrogen stemming from NH, emissions.
Findings: The NH, deposition from CMAQ could
be partitioned successfully to state and sector-level
sources of NH, emissions. However, the sensitivity
algorithms do not consider all NH, nonlincarities.
so the model sensitivity must be carefully set up to
achieve an internally consistent interpretation of the
budget partitioning.
Impact: The foundation has been laid for source
attribution analyses that will provide needed
NH, atmospheric deposition source responsibility
estimates to the Cliesapeake Bay TMDL process
to provide guidance for management to consider
to further reduce atmospheric deposition to the
watershed and Bay.
10
-------
3.0
Model Development and
3.1 Overview of Air Quality Model
Development
Introduction
AMAD has a primary mission of developing air quality
models in a "one-atmosphere" multipollutant. paradigm that
are based on the current state of the science and directly
applicable to EPA's air quality management and policy
needs. The first release of the AMAD CMAQ model was
in 1998, and AMAD has ongoing responsibilities to further
develop, test, and evaluate the CMAQ model in an effort
to identify model improvements and advance the science
within CMAQ. The AMAD Model Development and
Model Evaluation Program elements work closely to carry
out these responsibilities.
Within the AMAD Modeling Program, the Model
Development Program element is responsible for the
continued development, testing, and refinement of the
CMAQ model for a variety of spatial (urban through
hemispheric) and temporal (days to years) scales and for a
variety of pollutants (03, PM, air toxics, visibility, and acid
deposition). The one-atmosphere model concept enables
the interaction of these pollutant regimes within one
modeling construct. Through detailed treatment of physical
and chemical processes affecting the fate of atmospheric
pollutants, these modeling systems provide scientifically
sound tools to understand the relationships between
sources of air pollution and ambient concentrations over
spatial scales ranging from urban to hemispheric and
temporal scales ranging from hourly to annual. This is
accomplished through an integrated niultidisciplinary
approach involving physical, chemical, numerical, and
computational science to develop a "numerical laboratory."
wherein atmospheric physico-chemical interactions can be
simulated effectively to guide development of air pollution
abatement strategies. Through synthesis of laboratory and
field measurements in paranictcri/ations included in the
model and diagnostic testing against measurements over
wider spatial and temporal scales, the models provide
a framework to test and refine hypotheses and process
formulations based on limited and controlled data, thereby
improving our understanding of key processes regulating
the atmospheric fate of pollutants. This comprehensive
approach permits the testing of emissions control strategy
impacts on the target pollutant, as well as collateral impacts
on other pollutants. Model sensitivity and uncertainty
tests arc conducted to understand the areas of the CMAQ
model system that arc most in need of developmental
focus and where the model response will be greatest. New
experimental and theoretical knowledge of important
chemical and physical processes in the atmosphere arc
monitored, analyzed. and incorporated into the model when
iagnostic Testing
appropriate. New CMAQ model versions arc released for
public access on an as needed basis. The Model Evaluation
Program conducts evaluations to assess how well the
CMAQ model is performing and to better understand the
role of the model inputs and model processes in the air
quality predictions. This requires comparisons against
observational data from a variety of perspectives, where
analyses consider different spatial and temporal scales
to assess model performance. Interrelationships among
different chemical species must also be considered, as
well as the influence of uncertainties in meteorological
predictions and emission estimates. Model evaluation
serves dual purposes: (1) to characterize the accuracy of
model predictions and (2) to identify needed improvements
in modeled processes within the air quality model or model
inputs.
Research Activities and Accomplishments
Several significant scientific and structural updates to
the CMAQ modeling system were completed in FY08,
resulting in the public release of CMAQv4.7. CMAQ v4.7
features a new aerosol module, that contains substantial
scientific improvements over the aerosol modules released
in previous versions of CMAQ.
New pathways for secondary organic aerosol (SO A)
formation from precursors, including isoprenc.
sesquiterpenes, benzene, glyoxal, and nicthylglyoxal. were
incorporated to improve the model's ability to represent
the contribution of organic carbon (OC) to airborne fine
PM. The model was updated to represent sea-salt emissions
from wave-breaking in the coastal surf zone. A new
parameterization to represent the heterogeneous dinitrogen
pentoxide (N205) hydrolysis on particle surfaces, which
includes dependence on temperature, relative humidity,
inorganic PM composition, and phase state, was developed
and incorporated in CMAQ. Modifications to the treatment
of gas-phase chemistry in CMAQv4.7 were directed at
improving the partitioning of airborne oxidi/cd nitrogen, as
well as providing a consistent treatment for multipollutant
applications (03, PM, air toxics, and mercury).
In FY 10, an interim version of CMAQ (v4.7.1) was
released publicly and included the following features:
• Instrumented versions of CMAQ (including the
sulfur tracking model, the primary carbon source
apportionment model, and the direct decoupled
sensitivity analysis method) to aid in diagnostic
investigations
• Updates to the vertical advcction scheme to reduce
numerical diffusion associated with the original scheme
11
-------
• Updates to the numerical solver for aqueous chemistry
to improve robustness and mass conservation
characteristics
• An on-line photolysis rate module that incorporates
the radiative impacts of aerosol loading simulated
by the model has been included; this enhancement
enables investigation of potentially important impacts
of scattering and absorbing aerosols in modulating
photolysis rates and atmospheric photochemistry
regulating the formation of secondary air pollutants.
A prototype of a new modeling system that couples the
meteorological model (based on the WRF) and chemistry -
transport (CMAQ) calculations within a single executable,
following a two-way coupled modeling paradigm, was
developed successfully.
Efforts to expand the applicability of the CMAQ
modeling system to hemispheric scales were pursued.
An annual simulation for 2006 and numerous sensitivity
runs were conducted over a domain encompassing the
Northern Hemisphere. Model predictions of O,. PM25 and
constituents, and precursor species were compared with a
variety of measure incuts from surface, aloft, and remote
sensing platforms.
Anew chemical mechanism. SAPRC-07TB, was
developed and incorporated into CMAQ. This version was
customized for EPA to include an explicit description of
reactive H APs and high-emissions O, and PM precursors.
In this mechanism, CMAQ uses an efficient operator
technique to reduce the number of equations representing
low-NO chemistry, which speeds up the run time while
allowing more detail in the chemistry.
Many chemical reactions in the atmosphere arc initiated
by the photo-dissociation of trace gases. These reactions
arc responsible for most of the smog buildup, which is
detrimental to humans, animals, plant life, and materials.
To accurately model and predict the effects of air pollution,
good photo-dissociation reaction rate (or photolysis rate)
estimates must be made. The CMAQv4.7 release included
an optional inline photolysis rate module. This module
allows for feedbacks of modeled atmospheric pollutants in
the radiative transfer calculations. Several refinements arc
planned for the inline photolysis module in preparation for
the FY 11 CMAQ release, including: (1) specification of the
O, column, (2) cloud attenuation effects, (3) incorporation
of temperature (and possibly pressure) dependencies on
the absorption cross sections/quantum yields, and (4)
improvement to the surface albedo (including snow albedo
and sea ice effects). In addition to these refinements, the
module may be refined structurally to accomplish the
following: (1) move cross-sections and quantum yields for
photolysis rates from the module's source code to an input
file. (2) use a preprocessor program to create the input
file, and (3) enable photolysis rates to be shared with other
CMAQ modules (e.g., the cloud module).
Next Steps
Two APMs related to CMAQ development arc required for
FY 11.
1. Improved CMAQ mode ling system for use in urban-
scale residual nonattainnicnt
2. An operational two-way coupled WRF-CMAQ
mode ling system will be publicly released.
Collectively, these require exploration of novel modeling
methodologies to extend the CMAQ mode ling system
to address emerging environmental problems at scales
different from traditional CMAQ applications. Model
applications to date have demonstrated clearly the
continued need to account for and to improve the
representation of interactions of atmospheric processes
occurring at the various spatial and temporal scales.
Model simulations over annual cycles have pointed to
the need for more robust methods for specifying lateral
boundary conditions. Although linking with global scale
atinosplieric chemistry models has been pursued and
will continue to be investigated, it is also recognized that
biases in the global model can propagate and influence
regional CMAQ calculations and often confound the
interpretation of regional model results. The specifications
of the lateral boundary conditions, to a large extent, dictate
the simulated variability in the free troposphere, which in
turn can impact the simulated surface-level background
values for a variety of trace species. The tightening of
the NAAQS to lower threshold values (e.g., the recent
revisions to the O, NAAQS) places additional requirements
on the ability of atinosplieric chemistry transport models
to accurately represent the entire spectrum of ambient
concentrations, including the background values.
Expansion of the mode ling system to hemispheric scales
provides opportunities to consistently represent processes
at all scales and will improve the characterization of
long-lived pollutants (e.g., mercury). The extension also
supports future applications to study the linkage and
interactions between global climate and air quality. On
the other end of the spectrum, emerging Agency problems
focusing on air quality-human exposure linkage will
require application of the model at significantly finer
resolutions to capture variability in ambient concentrations
of a number of pollutants (03, PM, and air toxics) and
resultant human exposure. These extensions will require
further development and cnlianccnicnt of various physical
and chemical process modules/algorithms included in the
mode ling system.
The two-way coupled WRF-CMAQ system will be
essential for high resolution modeling because of the very
high-frequency of meteorology data needed for air quality
mode ling at urban scale resolutions. Many other aspects
of mode ling science and techniques also need further
development for urban-scale applications. Although both
meteorology and air quality models have been applied
at grid cell si/cs down to 1 km and even less, t lie re
arc science issues at these fine scales that need further
-------
development and analysis. For example, meteorology
simulations at these scales have been shown to better
capture local dynamical effects caused by complex terrain
and coastal circulations; however, current nicsoscale
models need considerable modifications to realistically
capture the effects of urban environments on local wind
fields, temperature and humidity, boundary layer dynamics,
and local atmospheric dispersion. Also, there arc scale
limitations for Eulcrian grid models. For example, the
subgrid scale assumption of PBL models begins to break
down as grid cell si/c is reduced toward 1 km because
si/c of the largest eddies in the convcctivc boundary layer
arc on the order of the PBL depth, which approaches
or exceeds the 1-km grid scale. Wind field modeling in
urban areas also presents a scale issue where the uiban
buildings, which arc treated as roughness elements at larger
grid resolutions, become obstacles to the flow at finer
resolutions. What happens to pollutants emitted in such an
environment is the key question.
An alternate photolysis module is planned for the FY 11
release. This module will use a radiance module to
calculate actinic flux at 27 wavelength bins (from 290nm
to 800nm) at all altitudes within the troposphere. The
radiance model will incorporate the independent variables
of solar zenith angle, earth-sun distance as a function of
Julian day. surface elevation, surface reflectance, cloud
attenuation and reflectance, integrated aerosol optical
depth, modulation of stratospheric aerosol depth, aerosol
single scattering albedo and scattering phase function, and
a latitude and season-specific model of atniosplieric gases
and season-specific dependence for aerosol vertical profile.
The radiance model is expected to reproduce the accuracy
of a 64 stream calculation with a computational speed
tliat is faster than a two-stream calculation. The radiance
model will be coupled with a photochemistry model that
will incorporate temperature and pressure dependencies
in the cross-sections and quantum yields of currently
modeled photochemical reactions. The coupled radiance
and photochemistry models then will comprise a complete
photolysis module.
Clouds cover about 60% of the earth's surface and provide
a means to efficiently transport constituents from the
polluted boundary layer to the free troposphere, with
substantial implications for long-range pollution transport
and climate. Although CMAQ has hundreds of explicit
or semi-explicit kinetic expressions to describe gas-pha.se
chemistry, the aqueous chemical mechanism is limited
to only a few (about seven) oxidation reactions. A more
accurate and flexible solver for aqueous chemistry will
be developed for the CMAQ model, which will enable
clianging the reactions within the aqueous mechanism.
The new solver will be bcnchniarkcd against the original
by using box model and CMAQ simulations. Through
collaborations with investigators at Rutgers University,
new nitrogen-related chemistry will be added to the
aqueous module. The new clieniistry will be evaluated with
nitrogen deposition data collected during special studies
(e.g., Bay Regional Atmospheric Chemistry Experiment
|BRACE]) prior to public release.
3.2 CMAQ Aerosol Module
Introduction
The treatment of aerosol chemistry and physics is a
critical component of the CMAQ modeling system.
Ambient particles contain a mixture of numerous chemical
species that originate from primary particle emissions and
from secondary format ion pathways linked to gaseous
precursors. Representation of this chemical diversity
is an ongoing modeling challenge that requires the
simulation of mass transfer between the gas and particle
phases via condensation and evaporation, heterogeneous
reactions of gaseous molecules on the particle surfaces,
reactions that occur within the condensed particle phase,
and homogeneous nuclcation of particles from low-
vapor-pressure gaseous precursors. In addition to those
processes, our ability to predict the ambient concentration,
composition, and si/c distribution of atniosplieric PM is
dependent on the accuracy of emission inventories, the
numerical representation of aqucous-plia.se cloud and fog
processes, and an accurate treatment of particle deposition.
A correct simulation of atmospheric aerosol properties
is vital for the prediction of PM, 5 concentration changes
resulting from emission reductions and for calculations
of regional haze. Two emerging priorities in the FY 10-
FY11 time period will be (1) to accurately simulate the
mass concentration and composition of coarse particles
(PMCoaise), for which a new NAAQS is anticipated, and (2)
to simulate cxcccdances of the new 24-h PM, 5 NAAQS
(35 fig m 3), which occur predominantly in urban areas.
Recent evaluation studies have revealed that the
largest biases in CMAQ PM, 5 results arc driven by
ovcrprcdictions of the unspeciatcd PM, 5 (Appel ct
al.. 2008; Matliur ct al.. 2008), referred to hereafter as
PM . Observations of PM , arc obtained from the
Other. Other
gravimetric PM, 5 mass measurement minus the sum of
several chemically-speciatcd measurements (sulfatc|SO,|.
NO,, ammonium, OC, and elemental carbonfEC]). Model
predictions of PMother consist of unspeciatcd primary PM,,
plus any noncarbonaccous material (e.g., oxygen and
hydrogen atoms) associated with the organic aerosol. In
FY09-FY11, three efforts will be undertaken to mitigate
the incommensurability in PMofter evaluations and improve
model performance for this important quantity.
Research Activities and Accomplishments
As new configurations have been added to the CM AQ
modeling system (e.g., CMAQ-MP, CMAQ-TX,
CM AQ-Hg. CMAQ-DDM. sulfate tracking, carbon
apportionment), it has become increasingly difficult
to maintain the code archive because of some subtle
inefficiencies in the aerosol module architecture.
13
-------
In FY 10-FY11, problematic portions of the aerosol code
will be incrementally updated and tested with a goal of
removing the most egregious inefficiencies before the
FY 11 release of CMAQ.
In FY 10, the new secondary organic aerosol (SOA)
treatment released in CMAQ v4.7 was described in a
journal publication by Carlton et al. (in press). In FY 10-
FY 11, the precursor-specific SOA results will be evaluated
against organic-traccr-based measurements collected
by HEASD in Research Triangle Park, the Midwest
(LADCO sites), and Southeast (SEARCH sites). Based
on our findings from those evaluations and reports from
any ongoing National Center for Environmental Research
(NCER)-funded extramural research, we will refine the
model treatment of biogenic SOA. In FY 10. HE ASD
delivered an APM summarizing their new experimental
findings on anthropogenic SOA. Based on their report,
combined with the results of ongoing extramural research,
the anthropogenic SOA formation pathways in CMAQ
will be ovcrliauled in FY 11-FY 12. Specifically, we
expect to revise the treatment of alkane SOA and add new
SOA formation pathways from light-weight PAHs (e.g.,
naphthalene, alky 1 iiapthalcnes).
Coordinating with investigators from ENVIRON and
the Electric Power Research Institute (EPRI), a new
SOA module using the volatility basis set (VBS) will be
implemented in CMAQ and tested against the base SOA
module in CMAQ. If those tests arc successful, we will
strive to release the VBS code as a user-specified option
in FY 11. In the future, after sufficient testing, this may
become the default option.
A major focus of the FY 11 APM for CMAQ model
development is improved PM, 5 predictions in "residual"
nonattainnicnt areas, several do/en counties which arc
forecasted to remain in nonattainnicnt of the NAAQS even
after all of the national-scale emission control programs
(e.g., Clean Air Interstate Rule [CAIR], NOx SIP Call,
diesel retrofit, etc.) have taken effect. The vast majority
of these residual problems arc likely to be dominated
by primary PM,Thus, a detailed examination of the
emission sources contributing to primary PM, 5 will be
critical. In FY 10, the primary carbon source apportionment
module tliat has been released with the past three public
releases of CMAQ will be upgraded to CMAQ v4.7
and released to the public. In FY 10-FY 11. the source
apportionment capability in CMAQ will be extended
to treat other primary PM, 5 species, such as the trace
metals emitted from industrial facilities. The resulting
instrumented model will be used to assess and
improve PM, 5 emission inventories in residual
nonattainnicnt areas.
In CMAQ, gas/particle equilibrium for inorganic species
has been computed using the ISORROPIA module since
the release of CM AQ \4.2 in 2002. The current version
of that module, ISORROPI A v 1.7. treats the SOr/NO,-/
' y 4 3
C1 -/NH4 +/Na+/H,0 system in a computationally efficient
manner. but it suffers from numerical instabilities under
certain extreme conditions. Over the past several years,
three new thermodynamic modules have been developed as
promising alternatives to ISORROPI A. Through a National
Oceanic and Atmospheric Administration (NOAA)-
funded contract with the Georgia Institute of Technology.
ISORROPI A \ 2.0 was developed and released in FY09.
The main advancement in ISORROPI A \ 2.0 is that it treats
three additional species that arc abundant in sea salt and
soil dust: (1) Ca,+, (2) K+, and (3) Mg,+. In FY 11, we will
implement the latest ISORROPI A module in the FY 11
release of CM AQ.
Next Steps
A new CM AQ subroutine will be developed to produce
model estimates of total PM,At present, model users who
seek PM, 5 predictions from CMAQ must perform a linear
combination of several do/en species in the model output.
Inconsistencies in the postprocessing methodology can lead
to discrepancies in the model results reported by different
users, even when the exact same CM AQ configuration
is applied to identical inputs. In addition to bringing
consistency to the method of summing individual chemical
components, the new subroutine will simulate changes
in mass (i.e., artifacts) tliat occur when ambient PM, 5 is
sampled and analyzed by the Federal Reference Method
(FRM). The measurement artifacts to be simulated arc
volatilization losses of ammonium nitrate during sampling,
adsorption of semi-volatile organic gases during sampling,
and retention of particle-phase water during gravimetric
analysis of the filters (with consultation from Neil Frank
[OAQPS]). The main output of this new subroutine will
be gridded and time-resolved mass concentrations of PM,,
FRM which, for the first time, can be compared directly with
the large surface network of FRM monitors (i.e., State and
Local Air Monitoring Stations [SLAMS]). Once the key
measurement artifacts arc incorporated into the CM AQ
calculations, the final modeled values of PM, 5 FRM can
be used without hesitation in data fusion efforts (e.g., for
CDC-PHASE). The SLAMS data are the basis for NAAQS
attainment demonstrations so the modeled PM, 5 FRM output
values should prove very useful to OAQPS. EPA regions,
and the States.
PM, 5 emission inputs will be augmented to explicitly
speciate a number of chemical components that arc
currently part of PMofter. At present, the Sparse Matrix
Operator Kernel Emissions (SMOK.E) processor generates
gridded files that include PM, 5 emissions of sulfate (S042 ),
(NO} ), OC. and EC, where as all of the remaining mass
is lumped together as PMother. The PMother component
constitutes over half of the National Emission Inventory
(NEI) for PM,5. Using detailed speciation profiles derived
from our previous work (ReIT et al.. 2009), we will modify
the SMOKE processor to subdivide the emissions of
PM0ther into primary ammonium (NH4+), sodium (Na+),
chloride (C1-), selected trace elements (niagncsiuni|Mg|.
-------
aluminumfAl], silicon [Si], potassiumfK], caliuni |Ca|.
titanium [Ti|, manganese [Mn], and iron [Fe|h non-carbon
organic mass (NCOM), and a greatly-reduced quantity of
P^other- Tlic CMAQ code will be modified to read these
new species from the emission files and incorporate them
into the chemistry and transport operators.
The CMAQ aerosol module will be enhanced to treat
the oxidation of OC that occurs in the atmosphere. It
has been shown that OC oxidation increases NCOM
(Turpin and Lim, 2001), thereby enhancing ambient
PM2 5 mass concentrations. Our preliminary analyses of
ambient data indicate that NCOM constitutes 10%-25%
of PM2 5. However, the oxidation of secondary OC is
treated very crudely in CMAQ v4.7 (using a zcrotli order
oligomcrization rate constant) and the oxidation of primary
OC is neglected entirely (a constant OM/OC ratio of 1.2 is
assumed for all sources of OC). These shortcomings will
be addressed in FY 11, and the resulting model predictions
of NCOM will be evaluated against measurement-based
estimates of the same quantity.
The net result of the research described above will be 10
newly predicted quantities from the CMAQ model
(PM2 5 FRM and fine-particle Mg, Al, Si. K, Ca, Ti, Mn, Fe,
and NCOM) and more accurate predictions of NH4+, Na+,
and CI- (because their direct emissions from anthropogenic
sources will be accounted). When the total modeled
concentrations of individual PM25 species (i.e., SO,:, NO,,
NH4+, EC, and OC) arc subtracted from the new model
estimates of PM2 5 FRM, we will obtain model results for
PM„(, that arc commensurate with the difference-based
Other
observations of PMother collected across the Interagency
Monitoring of Protected Visual Environment Network
(IMPROVE) and Chemical Spcciation Networks. We then
will be equipped to assess and address the CMAQ model
biases for PMother in a rigorous manner.
Building on the findings published in FY08 (Davis
et al., 2008) plus recent experimental evidence from
researchers at NO A A, the University of Washington, and
the University of Illinois. CMAQ model treatment of the
heterogeneous reaction between N205 and particle surfaces
will be updated in FY 11 -FY 12 to account for the influence
of organic content and/or organic coatings.
EPA is considering a new ambient standard for PM ^
defined as PM10 minus PM,The largest sources of
PMCoarse are sea salt, fugitive dust from anthropogenic
activities (e.g., road traffic, agricultural tilling,
construction), and windblown dust from arid land.
All of these sources arc treated in CMAQ except for
natural windblown dust. In FY 11, efforts will be made
to incorporate new algorithms for estimating windblown
dust emissions (Tong et al.. in preparation) directly into
the CMAQ model and incorporate those new species
into the chemistry and transport operators. With these
advancements, the model infrastructure will be in place to
numerically predict PM,, concentrations at various times
J L Coarse
and locations.
CMAQ docs a reasonably good job of predicting fine
particle mass concentrations (e.g., PM, 5), but a much
poorer job representing the number concentration of
atmospheric particles. In FY 11. we will coordinate with
members of the external CMAQ user community to
improve the treatment of nuclcation and primary particle
emissions in the FY 11 release of CMAQ. These two
updates will be the first steps toward refining CMAQ
predictions particle number.
3.3 CMAQ Gas-Phase Chemistry
Introduction
The treatment of gas-phase chemistry, including the
production of gas-phase pollutants; production of semi-
volatile components, which can form aerosols; and the
associated gas-phase chemistry solvers, will always be at
the core of any advanced air quality model, particularly
those that are formulated with the onc-atmosphcrc.
niultipollutant concept. As such, periodic revisions to the
representation of atmospheric chemistry in air quality
models arc needed to incorporate new scientific findings
as they become available and to address emerging air
pollution issues that may arise. Also, it is often necessary
to revise existing chemical mechanisms and solvers to be
consistent with changes that arc made in other parts of the
modeling system to enhance computational performance,
case of use, and functionality . This is an ongoing process
that is necessary to keep advanced air quality models, such
as the CMAQ model, accurate, scientifically relevant, and
fully operational.
In the past, gas-phase chemical mechanism development
focused on single-pollutant issues, but since largely it has
become clear that it is more appropriate to treat chemistry
in an integrated, multiphase. niultipollutant manner. The
general goal of our research in this area is to develop,
refine, and implement the gas-phase chemical mechanisms
for use in the CMAQ model to accomplish the following
objectives:
• Ensure that CMAQ and other models tliat arc used for
regulatory and research purposes have scientifically
justifiable gas-phase chemical representations, arc
appropriate for the application being studied, and arc
consistent with our most up-to-date knowledge of
atmospheric chemistry
• Ensure that interactions between gas-phase chemistry
and the chemistry occurring in aqueous- and particle-
phases arc accounted for adequately, so that we can
truly predict multimedia chemical effects of
emissions changes
• Develop techniques, tools, and strategics, so that we arc
able to efficiently expand current mechanisms to predict
the chemistry of additional atmospheric pollutants that
we anticipate will become important in the future
15
-------
Research Activities and Accomplishments
In FY 10, we completed the incorporation of one new,
statc-of-thc-science atmospheric chemical mechanism
into CMAQ, the SAPRC-07TB mechanism, and initiated
incorporation of the Regional Atmospheric Chemistry
Mechanism, version 2 (RACM2).
The RACM2 mechanism is an updated and expanded
version of RACM and includes more detailed aromatic and
isoprcne chemistry. RACM2 was developed in 2007 and
contains 349 chemical reactions. One hallmark of R ACM2
is its detailed representation of peroxv radical chemistry in
rural and remote environments. By implementing RACM2
in CMAQ, we will provide a computational laboratory
for examining the influence of gas and aerosol production
in these types of environments and their influence on air
quality. This quality of RACM2 also may provide a basis
for future extensions to hemispheric simulations.
The SAPRC-07 mechanism is an updated version of the
SAPRC-99 mechanism, which has been used widely
in CMAQ and other air quality models for many years.
In FY 10, we have completed the incorporation of a
customized, niultipollutant version of this mechanism,
S APRC-07TB. which provides an explicit description of
the most important species and reactions for a number of
current EPA applications. SAPRC-07TB enables criteria.
HAPs. and niultipollutant applications to give consistent
results. In addition, the detailed organic chemistry in the
original version of SAPRC-07 provides the capacity for
changing the way that individual VOCs arc represented by
our version of the mechanism. In the coming fiscal year,
we will complete our analyses of the behavior of SAPRC-
07TB in CMAQ and incorporate any updates in inorganic
and organic reactions. In addition, as we improve our
ability to describe important SO A and cloud chemistry, we
can use this information to develop more explicit versions
of SAPRC-07TB that include these details.
The CB05 mechanism will continue to be a major
mechanism used in regulatory applications where many
simulations are performed and for research studies done
over long time periods. This nicclianisni uses structural
lumping techniques to provide a highly compact
representation, which makes it the most computationally
efficient mechanism currently available for use in CMAQ
niultipollutant studies. Recent published updates to the
aromatic chemistry (in specific the toluene chemistry)
were incorporated in the CB05 mechanisms, and extensive
testing of the impacts of the update on O, and PM
predictions was conducted. We will monitor research
in atmospheric chemistry and perform periodic, limited
updates to CB05 as appropriate, with a particular emphasis
on CB06.
Next Steps
In the next 3 years, we will finalize inclusion of RACM2.
We will make additional updates and improvements to
the standard versions of these mechanisms to make them
consistent with the latest research findings and provide
details needed by the aqueous and aerosol computations,
thus ensuring tliat the gas. aqueous, and aerosol chemistries
arc correctly integrated.
More importantly, we will start looking beyond the
standard, fixed-species, fixed-coefficients mechanism
suite to determine the best methodology for representing
atniosplieric chemistry in the future. Given the rapidly
changing research discoveries in atmospheric chemistry,
we will examine whether a more flexible approach to
mechanism creation is compatible with EPA's current
mode ling applications. We will explore collaborations with
groups in the United Kingdom, in particular the Master
Chemical Mechanism (MCM) developers. This mechanism
suite will allow CMAQ users to better match the type of
application with an appropriate mechanism. In addition to
these "base" mechanisms, we also arc developing extended
versions for both special research purposes, and versions
which will include the chemistry of additional H APs for
use in full, niultipollutant analyses, described in section
3.8 (Air Toxics). Research versions of the SAPRC-07TB
and CB05 mechanisms also will be constructed for special
studies. Two versions, in particular (see just below), will
be important components of our work during over the next
few years:
1. SAPRC-07TB with detailed primary pollutants for
reactivity studies—Current government-industry
joint efforts to study VOC reactivity have identified
the derivation of new reactivity scales as the most
important need to improve the scientific basis of
reactivity-based VOC regulations. To address this
need, we arc initiating efforts to rede rive reactivity
scales using CMAQ with the decoupled direct
method (DDM). This is a large effort, which requires
incrementally adding the more than 800 detailed
species in SAPRC-07, in groups of 20 to 50, to a
condensed version of SAPRC-07TB. The sensitivities
for each VOC then will be used to calculate a
wide variety of metrics that quantify potential
contributions of individual VOCs to O, formation
across the United States. The results of tliese
simulations will be used by States, local air pollution
agencies, and EPA to develop SIPs tliat better address
the role of VOCs in O, formation.
2. SAPRC-07TB with detailed, updated isoprcne
chemistry—Isoprcne is one of the most widespread
VOCs emitted in most of the world. A robust and
justifiable representation of isoprcne chemistry
and its effect on hydroxy 1 and hydro peroxv radical
formation, on cycling of nitrogen oxides and on
production of formaldehyde arc all essential to have
confidence that we adequately are representing O,
and organic aerosol formation and sensitivity. Groups
within AM AD arc collaborating arc reexamining
and potentially revising the treatment of isoprcne
-------
in SAPRC-07TB. Both a detailed and a condensed
update to the current isoprene chemistry will be
examined for potential inclusion in CMAQ-MP.
All new and modified mechanisms will continue to be
compatible with improved chemical representations for
SO A modeling being developed within HEASD and will
be linked closely with work performed under other AMAD
tasks.
The development of flexible, in-house chemical
mechanisms will be one area that we will begin explore
in FY 11. Under the U.S.-U.K.. collaborative agreement,
we will study the feasibility of incorporating condensed
representative versions of the MCM into a format
appropriate for CMAQ applications. With this approach,
we hope to more quickly incorporate new ground-breaking
science into the mechanisms without the long time lag (6
to 10 years) between official fixed mechanism versions.
We will monitor the scientific community for any new
methods for representing gas-phase chemistry with an
increased emphasis on interactions with aerosol formation
and greater accuracy of condensations. These new
methods arc at the frontiers of practice in the atmospheric
chemistry modeling community. We will keep up to date
with developments in this area and initiate efforts in the
later years of this task to implement advanced chemistry
representations in CMAQ.
3.4 Planetary Boundary Layer Modeling
Introduction
Air quality modeling systems arc essential tools for
air quality regulation and research. These systems are
based on Eulcrian grid models for both meteorology and
atmospheric chemistry and transport. They arc used for a
range of scales from continental to urban. A key process in
both meteorology and air quality models is the treatment
of subgrid-scale turbulent vertical transport and mixing of
meteorological and chemical species. The most turbulent
part of the atmosphere is the PBL. which extends from the
ground up to about 1 to 3 km during the daytime but is
only tens or hundreds of meters deep at night.
The modeling of the atmospheric boundary layer,
particularly during convcctivc conditions, long has been
a major source of uncertainty in numerical modeling of
meteorology and air quality. Much of the difficulty stems
from the large range of turbulent scales that arc effective
in the convcctivc boundary layer (CBL). Both small-
scale turbulence tliat is subgrid-scale in most nicsoscale
grid models and large-scale turbulence extending to the
depth of the CBL arc important for vertical transport
of atmosplieric properties and chemical species. Eddy
diffusion schemes assume that all of the turbulence is
subgrid-scale and, therefore, cannot simulate realistically
convcctivc conditions. Simple nonlocal-closure PBL
models, such as the Blackadar convcctivc model that has
been a mainstay PBL option in Fifth-Pennsylvania State
University (PSU) National Center Atmospheric Research
(NCAR) Mcsoscale Model (MM5) for many years, and the
original Asymmetric Convcctivc Model (ACM), also an
option in MM5, represent large-scale transport driven by
convcctivc plumes but neglect small-scale, subgrid-scale
turbulent mixing. A new version of ACM (ACM2) has been
developed that includes the nonlocal sclieme of the original
ACM combined with an eddy diffusion scheme. Thus.
ACM2 can represent both the supergrid-scale and subgrid-
scale components of turbulent transport in the CBL.
Testing ACM2 in onc-dinicnsional form and comparing
with large-eddy simulations (LES) and field data from
the second and third Global Energy and Water Cycle
Experiment (GEWEX) Atmospheric Boundary Layer
Study, known as the GABLS2 (CASES-99) and GABLS3
(Cabauvv. NL) experiments, demonstrate that the new
scheme accurately simulates PBL heights, profiles of fluxes
and mean quantities, and surface-level values. The ACM2
performs equally well for both meteorological parameters
(e.g., potential temperature, moisture variables, winds) and
trace cliemical concentrations, which is an advantage over
eddy diffusion models tliat include a nonlocal term in the
form of a gradient adjustment.
Research Activities and Accomplishments
The development and application of ACM2 in both WRF
and CMAQ now has achieved consistent PBL treatment
for meteorological and chemical species. Therefore, we
will pursue further development of the ACM2, particularly
for stable conditions. Very high concentrations of aerosols
have been observed in areas of winter cold pools. Many
western North American cities arc located in basins where
cold air can collect, resulting in extremely stable inversion
layers that trap locally emitted pollutants. Current
meteorology and air quality models have insufficient
vertical resolution and incomplete physical representations
to accurately simulate these conditions. In FY09 and FY 10,
AMAD was involved in a Regional Applied Research
Effect (RARE) project to model winter air pollution
episodes in Fairbanks. AK, where extreme surface
inversion layers can result in very high concentrations
of PM25. This modeling study uses a scries of nests with
horizontal grid resolutions of 12, 4, and 1.33 km and very
high vertical resolution, with the lowest layers of 4-m
thickness. We will continue to use the Fairbanks modeling
as a testbed for SBL model development.
We have continuing involvement with the GEWEX
GABLS. The ACM2 model was a participant in the
GABLS2 experiment that was a comparison of singlc-
colunin PBL models for a 2-day period of the CASES99
field experiment (Svcnsson. and Holtslag. 2006).
The ACM2 was also a participant in the GABLS3
experiment which is a single diurnal cycle at the Cabauw
meteorological tower in the Netherlands. The results of the
G ABLS3 experiment will be published soon in Boundary
Layer Meteorology. We have used tlie G ABLS3 experiment
and the LES data created for G ABLS 1. which was a SBL
case over sea ice in the arctic, to develop a new SBL model
17
-------
that was presented to the 19th Symposium on Boundary
Layers and Turbulence in August 2010. The new SBL
scheme also has been tested in WRF and evaluated for its
simulation of low-level nocturnal jets.
Next Steps
AM AD will have continued involvement in the Fairbanks
project through FY 11, in particular, an analysis of the
results of the air quality simulations. Also, an evaluation of
the new SBL scheme in WRF will continue in FY 11 and
be extended to testing in CMAQ and the coupled WRF-
CMAQ.
3.5 Meteorology Modeling for Air Quality
Introduction
Meteorology models arc very important components
of air quality modeling systems because they describe
the physical, dynamic, and thermodynamic state of the
atmosphere. The Division, therefore, has a vested interest
in developing components and capabilities that arc needed
for air quality applications based on existing meteorology
models. For example, the Division has been instrumental in
the development of nudging-based four dimensional data
assimilation (FDDA) in MM5 and more recently, in WRF
Model. FDD A has been shown to significantly decrease the
errors and biases of modeled meteorology that is a direct
input to air quality models. We also have developed a
land-surface model (LSM) and PBL schemes for MM5 and
transferred them to WRF for better simulations of the PBL
depth and surface fluxes and temperature, humidity, and
wind within the PBL. which arc also critical for air quality
modeling fidelity. MM5 has served as the Division's
primary source of meteorologically modeled input to the
air quality model for more than a decade, but. although
MM5-based meteorology is still being used in some limited
research, the Division now has transitioned fully to WRF
for all of our new model developments and applications.
Research Activities and Accomplishments
From FY 10 and into FY 11, a total of five annual WRF
simulations of the CON US at 12 km will be completed for
various CM AQ applications, which represent the official
transition away from MM5, as well as a transition from 36-
kni CON US to a more refined 12-kin scale.
Land-use and vegetation data arc important for (1) land
surface modeling within the meteorology model (WRF),
(2) dry deposition and bidirectional surface flux modeling,
and (3) modeling biogenic emissions of photochcniically
active chemical species. Thus, we have been working to
upgrade and update the land-usc/vcgctation data used in
our modeling systems. We already have accomplished the
first stage of this effort by integrating higher quality and
higher resolution land-use databases, the National Land
Cover Database (NLCD) for the U nitcd States and the
Moderate Resolution Imaging Spectroradiometer (MODIS)
land-use data for outside the United States, into WRF to
improve modeling of air-surface exchange processes. The
NLCD is based on 30-ni resolution L-enhanced thematic
mapping classified into 30 land use categories. The MODIS
data has 20 land-use categories at 1 km resolution. The
NLCD and MODIS datascts have been combined such that
the higher resolution NLCD is used preferentially where
available, with MODIS being used elsewhere. The Plcini-
Xiu LSM in WRF and the M3dry dry deposition model
in CM AQ have been modified to conform to this new
hybrid land-use datasct. We have tested the new NLCD and
MODIS data in a variety of applications in FY 10 and will
continue to analyze the result in FY 11. Although significant
differences in model performance between simulations
with NLCD and the old U.S. Geological Survey (USGS)
were not apparent in terms of surface meteorological
variables and precipitation at the 12-km scale, some
modest improvements were noted as the model grid scale is
decreased to 4 km and 1 km. We now have moved almost
exclusively to using NLCD (MODIS outside the United
States) for almost all modeling applications.
Another related area of research is the development of a
new vegetation datasct for biogenic emission modeling
also based on the NLCD/MODIS. The land-use data
will be combined with forest inventory data (FIA) and
agricultural data from the National Agricultural Statistical
Service (NASS) to create a high resolution datasct that
contains accurate information on tree species groups and
crop types that could be used by the Model of Emissions
of Gases and Aerosols from Nature (MEGAN) and
Biogenic Emission Inventor System (BEIS) biogenic
emission models. After testing and refinement of the
paranictcri/ations in WRF and CM AQ that depend on
land-use. the capability to use the hybrid NLCD/MODIS
land-use database will be released to the WRF and CM AQ
communities.
Another research area concerns the improvement
and application of data assimilation methods in the
meteorological model. Data assimilation is used in the
meteorology model to constrain the natural error growth
in the meteorology and air-quality simulations. Typically
for air-quality modeling, a nudging-based FDD A is
applied throughout the simulation period to create
dynamic analyses. We arc testing new FDDA techniques,
including the utilization of both nonstandard surface
and upper air observations that, in theory, will reduce
the uncertainty of the analyses used to drive the FDDA
and soil nudging. Tliese observation platforms include
radar wind profilers, aircraft, satellite wind, satellite sea
surface temperature, and a more dense set of surface
observations from the numerous incsonct sites around the
country. Furthermore, we have begun to explore using
the 3D Variational Analysis (3DVar) in the WRF system,
which has the potential to assimilate remotely sensed data
(e.g., satellite radiance. Dopplcr precipitation estimates).
We arc aggressively pursuing the testing of techniques
begining in mid-FY 10. and arc analyzing the results, which
-------
will be published in late 2011. The results, thus, far arc
encouraging in that systematic biases in transport wind arc
reduced through the use of some of these observational
datasets.
Next Steps
The meteorological model fields arc processed into
the CMAQ modeling system using the Meteorology-
Chemistry Interface Processor (MCIP). MCIP prepares the
meteorological fields in space and time for the emissions
processing and the chemistry transport model (CTM).
MCIP performs horizontal and vertical transformations,
diagnoses additional fields, and produces the output in a
format that is compatible with the CMAQ system. The
community release of MCIP supports model output from
MM5 and WRF-ARVV. In FY 11, MCIP will be updated on
an as-needed basis to support internal Division modeling
needs, as well as the CMAQ community. In particular,
scientific changes to MCIP have been completed to support
the NLCD/MODIS land-u.se. but further upgrades will
be required especially for the more detailed vegetation
work vet to be done. Furthermore. MCIP may need to be
modified to support the urban canopy paranictcri/ations in
WRF, to remove the dry deposition velocity calculations,
and to support upgrades to the Geostationary Operational
Environmental Satellites (GOES) processing. Releases
of MCIP will occur in conjunction with major CMAQ
releases and intermittently, as needed.
A few projects in which analy sis will continue through
FY 11 arc Northern Hemisphere annual simulations at a 108
km grid scale. We also arc using the high-resolution NLCD
at 4 km and 1 km over Houston. Cleveland and Atlanta.
In FY 11. we also will complete annual 2009 and 2010
CON US 12-km CMAQ simulations.
New guidelines for FDDA-based meteorological
simulations used for air quality models arc expected to be
released as a result of this 2010-2011 research. We will
rerun sonic of the longer term simulations to ensure the
results of the limited 4-day test case hold for seasonal and
annual simulations. Journal articles will document how the
simulations of meteorology and regional transport were
improved, which will build confidence in the inputs
for CMAQ.
Our meteorological modeling research program also
focuses on improving finer scale meteorological mode ling
simulations. The Division has embarked on a methodical
study of the efficacy of nicso-ganinia scale meteorological
mode ling simulations to address the suitability of
the resultant fields for air quality modeling. Several
simulations at fine scales have been or will be generated
in FY 11. We have executed a base simulation that uses a
very similar configuration as our 12-km scale simulations
in terms of the way surface processes arc handled (PX
LSM and ACM2 PBL schemes). We also have begun and
will complete several simulations that use a new and much
more advanced treatment of the urban canopy using the
same modeling domain as the base 1 km. The urban canopy
parameterization uses highly resolved urban morphology
data, like building height, building to surface fraction,
uiban fraction, and building area fraction, to represent
a number of influences cities have on atmospheric flow.
Not only will the meteorology of the simulations be
compared with the base simulation, but equivalent CMAQ
simulations will be executed to determine whether the
air quality predictions arc improved. A seasonal Atlanta
area 1 -km simulation is also another urban-related task
for 2011. The air quality simulations driven by this urban
meteorology will be used in a new technique to link
air quality with human exposure. The outcome of this
research will contribute toward the FY 11 APM "Improved
CMAQ mode ling system for use in urban-scale residual
non-attainment." as well as toward improving linkages
with human exposure models within the Laboratory. This
research also will help define a practical (horizontal scale)
limit of predictability for deterministic meteorological
models to be used for air quality modeling.
3.6 Integrated Meteorology-Chemistry
Modeling
Introduction
Traditionally . 3-D CTMs that arc used for air quality
research and regulation arc driven by 3-D meteorology
fields provided by a priori runs of a meteorology model.
Historically, the CTMs and meteorology models developed
over several decades along independent tracks with little
regard for computational, numerical, or even scientific
consistency between the two systems. In recent years,
however, there have been several efforts to combine
meteorology and chemical transport models into single
interactive systems. A primary driver for this trend has
been the need to include the direct and indirect feedback
effects of gases and aerosols on radiative forcing. Although
effects arc mainly important for climate applications, it
is becoming evident that they have substantial affects on
local meteorology and air quality in regions of extreme
air pollution. Zhang (2008) has provided an overview of
several coupled meteorology-clieniistry models including
the WRF/chem (Grcll, 2005) model in which chemistry
has been added into the WRF model (Skamarock, 2008)
at the science process level. Another approach is to couple
historically independent meteorology and chemical
transport models into a single executable. Advantages
of this approach include maintaining consistency with
existing separate sequential meteorology-chemistry
systems that are being continuously and extensively
applied and evaluated. Furthermore, the numerical and
computational techniques employed in meteorology
models and CTMs differ considerably because of the
greater need for strict mass conservation and positivc-
dcfinitcncss of transported scalars in the CTM. Also, CTMs
generally use fractional integration of various processes,
wheras meteorology models use generally integrate all
prognostic variables and parameterizations each main
time step. Because of these and other reasons we have
19
-------
coupled the WRF meteorology model and the CMAQ
model to create a coupled modeling system with two-way
interactions between chemistry and meteorology.
Research Activities and Accomplishments
A two-way meteorological and air quality coupled
modeling system, WRF-CMAQ, has been developed in
FY 10 to enhance model accuracy by eliminating excessive
interpolation of hourly meteorological data and by
incorporating aerosol feedback information from the air
quality model to the meteorological model. The two-way
coupled system consists of three components: (1) WRF, (2)
CMAQ, and (3) the coupler. The coupler is encapsulated in
Fortran 90 modules, so the details of the two-way coupling
system arc transparent to the users. This design enables
WRF or CMAQ to be detached from the system easily and
executed in standalone mode. The single-source coding
approach makes code maintenance less complicated.
Next Steps
Extensive testing of the two-way coupled WRF-CMAQ
is underway and will continue into FY 11. The testing
includes a wild fire episode in California in June 2008,
where emitted smoke was concentrated enough in the
central valley that SW radiation levels were much lower
than normal. Initial tests show the model with SW
feedback had lower radiation and daytime temperatures
that agreed much better with observations than did the
model with no feedbacks. Other tests arc being done on
a hemispheric scale. A manuscript that documents these
efforts is being prepared and will be published in FY 11-
2012.
3.7 Mercury Modeling
Introduction
As methods to measure mercury in the laboratory and
in the atmosphere continue to be developed, and as
further experimentation is performed, the basic scientific
understanding of atmospheric mercury behavior continues
to improve. As new information becomes available
from the peer-reviewed scientific literature, improved
formulations for simulating atmospheric mercury behavior
will be added to the CMAQ model. This likely will
involve the addition of new phvsicochcmical species
to the CMAQ model that arc found to have important
reaction with atmospheric mercury. Mercury is emitted
to air from a variety of environmental and industrial
sources. Industrial sources will continue to be treated using
model input data from emission inventories. However,
treating environmental sources of mercury will require
a multimedia modeling approach, where entirely new
modeling structures arc added to simulate media other than
the atmosphere, such as water bodies, soils, and vegetation.
It is assumed that atmospheric mercury model development
will be aided by the deployment of new ambient mercury
concentration monitors, and that dry deposition will be
added as a measured parameter of the Mercury Deposition
Network.
Research Activities and Accomplishments
Parameterizations were implemented that estimate the
surface water mercury oxidation and reduction reactions,
vegetation uptake of elemental mercury (HgO), and
reduction of soil-bound mercury to estimate the emission
potential of the natural surfaces. This allows CMAQ to
parameterize the air-surface water concentration gradient
dependence on the elemental mercury flux (Foley ct
al., 2010), the recycling of recently deposited divalent
mercury, and the enrichment of elemental mercury
concentrations in vegetation, soil, and surface waters that
have been documented recently in the literature (Bash
and Miller, 2009). These improvements capture observed
pulses of emissions following precipitation in arid
environments during measurement campaigns and are an
inline estimation of natural mercury emissions that remains
consistent with changes in modeled chemical mechanisms
and boundary conditions, changes that would require
updates to offline natural emission estimates (Bash. 2010).
Next Steps
The follow ing arc improvements and issues being
considered for the FY 11 CMAQ release.
• Chemical kinetics information—Kinetic rate constants
for gaseous and aqueous mercury reactions will be
updated as new information becomes available for
reactions of mercury with any other chemical species
included in the model provided there is sufficient time
to test the updated model before release.
• Multimedia treatment for air-surface exchanges of
elemental mercury—The CMAQ model will be
enhanced to include dry deposition and emission of
mercury to soils and vegetation (in addition to water
bodies) based on bi-directional flux treatments where
dry deposition and emission arc treated as a combined
process with cither an upward or downward net
flux. This modeling advance will be based on new
scientific information on the behavior of mercury
within those new media.
3.8 Air Toxics
Introduction
In the past, chemical mechanism and air quality
development have focused on O, and primary inorganic
PM; we arc expanding the scope of the atmospheric
photochemistry in CMAQ to include predictions for a
large number of H APs.
Research Activities and Accomplishments
We have expanded the base mechanisms of CB05 and
SAPRC-07, as described in the previous section, to
characterize the production and decay of approximately 40
-------
additional HAPs, including VOCs, inorganics, and metals.
These mechanisms arc used by the Agency to develop full,
niultipollutant analyses that examine the consequences
of emission control strategics on multiple types of
pollutants: O,. PM, H APs. visibility, acid deposition,
etc. Developing the chemistry for these H APs includes
appropriately accounting for compounds with a wide range
of atmospheric reactivity and source types, in all phases
(gas, aerosol, and aqueous).
Next Steps
In the next 3 years, further efforts in this area will overlap
with the developments in the base mechanism but with an
emphasis on H APs. We anticipate that, over the next few
years, the following efforts will be made.
• Characterization of explicit PAHs—This class of
compounds includes a large number of organics with
multiple benzene rings, possessing a wide range of
vapor pressures, toxicity, and reactivity. Because
some PAHs and their by-products can be highly
carcinogenic and/or produce SOA. we arc initiating
efforts to include their chemistry in upcoming
releases of CMAQ-MP.
• Description of arsenic compounds—Arsenic is a major
contributor to national air toxics assessments but has
potentially complex chemistry involving aqueous
transformations. We arc initiating efforts to better
predict the complex transport and transformations of
arsenic in CMAQ.
• Mechanism improvements for the most toxic air
pollutants—We will continue to monitor the scientific
literature and refine our representation of H APs
chemistry in CMAQ-MP. We work closely with
EPA, the Office of Air and Radiation (OAR), and
OAQPS to identify those compounds that pose the
largest risks for acute and chronic health effects
to the Unitcd States population, and we will focus
on refining our chemical representation of these
compounds. This will help us to ensure that the
chemistry and transport of these important pollutants
is characterized appropriately. In SAPRC-07TB, we
have incorporated additional pathways for highly
toxic acrolein formation that were not available in
previous chemical mechanisms. In the next planning
cycle, we will incorporate information from current
research studies on aromatic chemistry to help us to
refine our treatment of aromatic H APs and their toxic
by products.
• Description of new, emerging toxic air pollutants—
There is increasing evidence that a large number of
species, beyond those H APs listed in the Clean Air
Act 112(b), also can cause serious, adverse health
effects. For example, multifunctional carbonyl
compounds, formed as secondary pollutants from a
wide variety of nontoxic VOCs. may provide sources
of additional toxicity in atmospheric mixtures. As
more details on the exact structures of compounds
that can cause inflammatory responses to uiban
smog arc discovered, we will initiate efforts to better
represent these classes of compounds in CMAQ-MP.
• Incorporation of persistent organic pollutants.
pesticides and hydrolluorocarbons—Tlierc arc many
other pollutants that arc cither toxic per sc or have the
potential to cause toxic effects in aquatic systems, and
understanding the lifetime and distributions of these
compounds is of worldwide concern. CMAQ has
been used in past studies of two of tliese compounds
(atrazinc and tetrafiuoropropene), and we anticipate
that, as new information on emerging pollutants
becomes available. CMAQ-MP will provide an ideal
basis for studying these types of toxic pollutants.
Our efforts in air toxics mode ling will continue to
be closely coordinated with other gas. aqueous, and
aerosol chemistry being performed by AMAD, as
described in other sections of this report, as well as
other efforts in NERL/HEASD and in OAQPS.
3.9 Nanoparticles Modeling
As industrial production of nanoniatcrials increases, a
capability to accurately simulate the transport and fate of
these materials has risen to a high priority. As an initial
case study of nanoniatcrials, we arc investigating air
quality effects of doping dicscl fuel with nanoparticulatc
cerium oxide (nCcOj. nCcO, is used as a dicscl fuel
combustion cataly st to reduce fuel consumption and
reduce emissions of greenhouse gases and PM. In FY09,
we assembled a database of dicscl-cngine dynamometer
studies tliat have been conducted with and without nCcO,
in the fuel.
Research Activities and Accomplishments
In FY 10, we used the just-mentioned database to develop
a national emission scenario in which dicscl fuel across the
Unitcd States has been doped with a nCcO, additive. Those
emission estimates were used to assess the additive's effect
on regional-scale air quality using CMAQ.
Next Steps
In FY 11, we will simulate the dynamics of the particle
population emitted in dicscl cxliaust (with and without
nCcO,) near a major roadway to gain a detailed
understanding about the atmospheric transport and fate
of nanoniatcrials. Results from this 3-year effort will be
presented to clients in the Office of Transportation and Air
Quality (OTAQ) during FY 11.
3.10 Emissions Modeling Research
Introduction
Emissions arc among the most important drivers of the
CMAQ mode ling system. However, their estimates arc
subject to a large degree of uncertainty related to limited
knowledge on sources, processes, chemistry, locations, and
21
-------
temporal variability. The NARSTO Emission Invcntory
Assessment (www.narsto.org/section.src?SID=8) affirms
the large degree of uncertainty in emission inventories,
particularly for precursors of airborne fine PM and
for sources of OC. EC, and NH,. Most anthropogenic
emissions used in the CMAQ system arc derived from
EPA's National Emission Inventory (NEI) (www.epa.gov/
ttn/chicf/ciinforniation.htnil).
AMAD's emission modeling research focuses on the
evaluation and improvement of emission categories that
respond to meteorology and/or that arc natural or quasi-
natural in character, and that arc not readily available from
the NEI. As such, our research includes the development,
evaluation, and implementation of emission models for
biomass burning, fugitive dust, lightning, and biogenic
sources. These sources emit O, precursors (VOCs and
nitrogen oxides), particulate matter, and some air toxins.
Work performed in this area also supports the Division's
efforts in the Air Quality Model Evaluation International
Initiative (AQMEII), nanoparticlcs. and the Fairbanks, AK,
EPA R ARE project.
Research Activities and Accomplishments
During 2010, the Division focused on biomass burning,
biogenics, fugitive dust, and improved spcciation.
• Biomass burning—After working with EPA's
O AQPS in helping the release of an operational
satellite-based biomass burning emission estimation
system for the NEI, the Division turned its attention
to evaluating the emissions from this system in the
context of air quality mode ling and in working with
other researchers on improving areas of greatest
uncertainty. Collaborations with the National
Aeronautics and Space Administration (NASA) as
well as with researchers at Michigan Technological
University and the University of Kentucky, continued
under a N AS A-funded grant. A 2-day workshop
was held to continue the study of plume heights and
agricultural burning emission estimates. Annual
emission estimates being in 2003 have been provided
by our collaborator, collaborators at the University
of Louisville. These emission estimates arc being
evaluated for inclusion in future versions of the NEI.
• Biogenic emissions—For the 2011 release of CMAQ,
we plan to offer two alternatives for biogenic
emissions: NCAR's MEGAN and the BEIS version
3.14. Both models arc now in the modeling system,
and emissions from each have been compared. In
concert with the Division's ccosystcins-rclatcd
research, we worked with UNC's Center for
Environmental Programs (CEP) to incorporate
updated agricultural data and information from
the 30-in resolved National Land Cover Database
(NLCD). We plan to create an update to the Biogenic
Emission Landcovcr Database (BELD) that is based
on the NLCD. Undcr a N AS A-ROSES-fundcd grant
with the University of Maryland, we collaborated
on the use of satellite imagery to evaluate soil NO
emissions (sec presentation by H. Plata. "Soil NOx
model/satellite measurement intcrconiparisons." at
http://www.cmascenter.org/conference/2010/agenda.
cfm).
• Lightning NOx—Via a collaboration with NASA and
the University of Maryland, the Division continued
to explore the development and evaluation of an
algorithm to estimate nitric oxide production from
lightning using meteorological parameters available
from the MM5 and WRF meteorological models.
Results indicate that the NO profile simulated by
CMAQ in the middle troposphere (which had been
underestimated by CMAQ) compares much better
with observations when lightning-generated NOx is
included in the model.
• Wind-blown and fugitive dust—The Division
continued to interact with scientists in NOAA's air
quality forecast model research program to develop
and evaluate a wind-blown dust algorithm based on
land cover data and meteorological variables (notably
wind speed and precipitation). A wind-blown dust
module is being incorporated into the 2011 CMAQ
release. For fugitive dust from anthropogenic sources,
we have revised the methodology and improved the
spcciation of these emission sources. Specifically.
we have included the spcciation of trace metals from
all anthropogenic sources to better understand the
emission inventory and available measurements. We
have revised the transportable fraction applied to
the fugitive dust emission inventory to be consistent
across the entire inventory and have incorporated
temporal allocation adjustments to better account
for variations in the emission inventory. Finally,
we have included rain and snow events into the
emission processing to account for meteorological
effects on dust in the emissions processing. Two
of presentations on this research were given at the
2010 Community Modeling and Analysis System
(CMAS) conference: H. Simon ct al.. "Modeling the
trace-elemental composition of PM, 5 in CMAQ," and
(2) G. Pouliot ct al.. "Assessing the anthropogenic
fugitive dust emission inventory and temporal
allocation using an updated spcciation of particulate
matter." Both presentations may be viewed at http://
www.cmascenter.org/conference/2010/agenda.cfm.
• Spcciatcd emissions—The Division continued
to work with partners within EPA to improve
the SPEC I ATE database, which is central to the
spcciation of VOC and PM gas and aerosols for
emissions used in the CMAQ mode ling system.
Portions of this work arc summarized in Simon
ct al. (2010) and were presented by Simon and
colleagues at the 2010 CMAS conference.
-------
Next Steps
The Division plans to continue improving and evaluating
those components of the emission modeling system used
in CMAQ and areas in which other organizations, such as
OAQPS. arc able to provide support.
The Division's research is organized around several model
evaluation studies addressing O, and PM predictions of
CMAQ and characterization of CMAQ performance for
client groups, particularly OAQPS. Work is planned to
improve process-based emission algorithms and the use of
geographical data. Many of these improvements likely will
depend on outside funding and continued collaboration
with OAQPS and NRMRL. The NARSTO Emission
Inventory Assessment recommends that inventory builders.
"Develop and/or improve source profiles and emission
factors plus the related activity data to estimate emissions
for particulate matter, volatile organic compounds,
ammonia, and air toxics." Outputs from this research
will create tools for directly modeling hourly values from
dust and wild fires, VOCs from biogenic sources, and
from lightning NOx. The Division plans to further develop
and test emission modeling tools for episodic modeling
(hourly) of the emissions of biogenic emissions, wildland
fires, lightning NOx, and fugitive dust. In collaboration
with OAQPS, these advances will be incorporated into
the SMOKE modeling system. SMOKE was developed to
provide emission data to CMAQ. The emission modeling
advances arc in direct support of the major update of
CMAQ planned for FY 11.
Bio mass burning—We plan to continue our work
with OAQPS and the U.S. Forest Service to evaluate
information on fire activity, fuel loadings, and
climatological patterns associated with biomass burning
emission estimates. Sensitivity tests and model evaluation
of CMAQ arc planned to examine whether improvements
in the fire emission est i mat ion methods will improve air
quality model simulations. We plan to prepare one or
more publications for submittal to a pccr-rcvicwcd journal
related to this effort.
We also plan to continue our collaboration with scientists
at NASA Langlcy, as well as with NERL's Environmental
Sciences Division, to evaluate and possibly improve
plume-rise estimates for biomass burning events and to
improve temporal/spatial estimates of rangcland/cropland
burn emissions.
Biogenic emissions—We plan to continue work with
NRMRL and scientists at NCAR to integrate and evaluate
MEGAN in the CMAQ modeling system. Building off
previous progress, we plan to evaluate model performance
with MEGAN and submit a publication for consideration
to a pccr-rcvicwcd journal to report our findings and
recommendations. We intend to release the next version of
CMAQ with MEGAN and BEISv3.14.
Working with scientists at the University of North
Carolina, we will continue to explore updates of the
vegetation landcovcr with the 30-m resolved land cover
classes in the EPA/USGS NLCD. During 2011, we plan
to focus on collaborating with NCAR via UNC contract
to harnioni/c the vegetation cover datascts in MEGAN
and BEIS. Time and resources permitting, we will include
an updated vegetation cover datasct in BEIS after the fall
2011 release of CMAQ.
Lightning NOx—In collaboration with NASA, an algorithm
for estimating NO production from lightning in the CMAQ
modeling system will continue to be refined and tested.
NASA has indicated that a draft journal article on this
work is in preparation. We plan to incorporate the lightning
algorithm into the fall 2011 release of CMAQ.
Gcogcnic dust—As time permits, we will continue to
interact with NOAAs air quality mode ling forecast
research team to prepare a publication and to include an
algorithm for improved estimates of fugitive dust to be
integrated into the CMAQ modeling system. The Division
will continue to assess alternative temporal profiles and
to provide appropriate recommendations to OAQPS to
improve the NEI. We arc also testing an inline wind-
sblown fugitive dust emission algorithm in the CMAQ
code.
23
-------
24
-------
4.0
Air Quality Model Evaluation
4.1 introduction
To ensure the integrity of regional air quality models for
environmental decision-making, the Division conducts
extensive evaluation studies to rigorously assess air quality
model performance in simulating the spatial and temporal
features embedded in air quality observations. We analyze
the performance of meteorology, emissions, and chemical
transport models to characterize model performance
and also to identify what model improvements (inputs
or processes) are needed. Thus, the Division's model
evaluation efforts are tied tightly to its model development
research activities.
Our evaluations seek to answer four fundamental
questions: (1) Is the model getting the right answer? (2) Is
the model getting the right answer for the right reasons? (3)
How well does the model respond to changes in emissions
and/or meteorology? and (4) What is the uncertainty
associated with the performance of the model? We have
used these questions to construct a model evaluation
framework (Dennis et al., 2010) consisting of operational,
diagnostic, dynamic, and probabilistic techniques.
Operational evaluation is a comparison of model predicted
and observed concentrations of the end-point pollutant(s)
of interest and is a fundamental first phase of any model
evaluation study (Foley et al., 2010) Diagnostic evaluation
(b)
¦
Other
~
TC
~
NH/
¦
NO-,"
~
S042
20-1
15-
Northeast
c
O
(J
10-
5-
(H
21HI
12%
37ft
I52S
•4%
'16=,'=
Ł
±
.1
c
_
-
n
=
O
STN CM AO
20-
15-
10-
5-
0-
Atlantic
32%
1
. ———
¦31%
49Vi
STN CWAQ
— TO
Figure 4-1. Examples of two types of model evaluation techniques: (a) probabilistic—time series of daily maximum
8-h 03 concentrations from a 200-member CMAQ model ensemble at a monitoring site in an urban location; and
(b) diagnostic—percent contribution of individual aerosol species comprising the total average regional PM25 mass
concentrations predicted by CMAQ and measured by the Speciated Trends Network (STN) sites.
25
-------
investigates the processes and input drivers that a fleet
model performance (Bhavc et al.. 2007, see Figure4-lb
below). Dynamic evaluation focuses on assessing the
model's air quality response to changes in emissions and
meteorology, which is central to applications in air quality
management (Godovvitch et al., 2010). Probabilistic
evaluation characterizes the uncertainty of air quality
model predictions and is used to provide a credible range
of predicted values rather than a single "best-estimate"
(Pinder et al., 2009, see Figure 4-1 a). Because these four
types of model evaluation are not necessarily mutually
exclusive, the Division's model evaluation studies often
incorporate aspects from more than one technique
of evaluation.
Information on selected aspects of the Division's model
evaluation research program follows just below.
4.2 Atmosphepheric Model Evaluation Tool
Introduction
To conduct and communicate model evaluation requires
the availability of statistical and graphical software. Over
the past several years, the Division has been working
on the development and application of the Atmospheric
Model Evaluation Tool (AMET). A MET is an open-
source software package designed to aid in the evaluation
of meteorological and air quality models. AMET is a
combination of several publically available software
packages: specifically, the MySQL database software,
PERL, and R statistical software. The database is used to
store observed and modeled data, whereas the R software
is used to perform various statistical analyses and create
a number of different statistical plots. An example of
AMET's capability is shown above in Figure 4-lb.
Research Activities
Recent activities have focused on communicating
and releasing the software to the air quality modeling
community.
Accomplishments
Appel et al. (2011) reported on the design and public
release of AMET. AMETvl.l was made available through
the CM AS center Website at www.cmascenter.org.
Next Steps
Work is ongoing to create AMETv2.0, which should
include enhancements to the database and analysis codes to
provide greater support of new data sources (e.g., satellite
data). It is likely that the next release of AMET will occur
in late 2011 or early 2012.
4.3 Air Quality Model Evaluation Initiative
Introduction
The Air Quality Model Evaluation International Initiative
(AQMEII) is a joint research project between air quality
modeling groups in North America (U.S. and Canada) and
Europe (numerous European modeling groups from various
countries). AQMEII aims to promote research on regional
air quality model evaluation across the European and North
American atmospheric modeling communities through
the exchange of information on practices, the realization
of intercommunity activities and the identification of
research priorities, keeping policy needs in focus. The
activity is organized around the concept of Operational,
diagnostic, dynamic, and probabilistic evaluation defined
and discussed at the joint EPA/American Meteorological
Society workshop in Research Triangle Park in 2007
and the first AQMEII workshop in Stresa, Italy, in 2009.
AQMEII is coordinated by two co-chairs. Dr. Rao from
AMAD and Dr. Galmarini from the European Union's
Joint Research Center in Italy. In addition. Environment
Canada serves as a regional focal point with the U.S. and
Europe. The results of AQMEII will be available to the
scientific community in general at academic, institutional,
and private sector levels.
Research Activities
The Division is responsible for performing and evaluating
the CMAQ model simulation for North America. As one of
the coordinating members of AQMEII, AMAD has been
responsible for coordinating the availability of the input
data (e.g., meteorology, emissions) that were used for the
modeling applications for North America.
Recent activities have focused on communicating
and releasing the software to the air quality modeling
community.
Accomplishments
Over the course of the year, Division scientists have been
working closely with groups in Canada and Europe to
compare the results of their modeling efforts and identify
the strengths and weakness of each air quality model, with
the ultimate goal of the research project being to learn how
to improve those models and the methods of evaluating
results from air quality models. By the end of 2010, the
Division had completed the CM AQ modeling for North
America and submitted all output data to the European
Union's Joint Research Center for further analysis.
Next Steps
A special issue of Atmospheric Environment will be
dedicated to results from the AQMEII project, and AMAD
has committed to submitting a number of manuscripts
to the special issue on topics ranging from meteorology,
emissions, and boundary conditions to analyses of the
operational performance of the AQMEII CMAQ modeling
26
-------
results to principal component analysis (PCA) of the
modeling results. All of these papers are joint research
projects with many international groups contributing to
each of the manuscripts. Submission of manuscripts to the
special issue of Atmospheric Environment is due in May
2011. In addition, results from the AQMEII project will be
presented at the 2011 CMAS conference and the AQMEII
workshop to be held in Chapel Hill, NC, in October 2011.
4.4 Diagnostic Evaluation
Introduction
Diagnostic evaluation research during the year examined
the interaction of meteorology on chemical transport
modeling.
Research Activities
Our focus has been on photochemical O, formation and
horizontal transport processes because they strongly
influence the temporal evolution and spatial variability
of the simulated 3-D O, distribution within the lower
troposphere. Results from the CMAQ with the Carbon
Bond CB05 chemical mechanism are being evaluated
against surface-based routine measurements, as well as
1800
1600
1400
3"
O 1200
<
E
""" 1000
o
lu 800
X
600
400
200
0
0 2 4 6 8 10 12 14 16
WIND SPEED (m/s)
Figure 4-2. Modeled (gray) and observed (rose) wind
speed profiles averaged over the nocturnal periods of
August 11-15,2002, at Ft. Meade, MD. Boxes span the
25th to 75th percentiles, and whiskers extend from the
10th to 90th percentiles.
upper air profile measurements obtained by research
aircraft during intensive field studies from three summer
months in 2002, when several high-O, episodes occurred
in the eastern United States. Net O, production efficiency
(OPE) results suggested photochemical O, formation
was comparable between the model and observations,
with OPE values of 6.7 and 7.6, respectively, within the
afternoon PBL. Evaluation of wind profiles revealed wind
speeds generated by the WRF model with the FDD A base
approach displayed greater underestimation of observed
speeds in the nocturnal residual layer aloft. During a
multiday O, episode, when a low-level jet occurred nightly
in the Mid-Atlantic region modeled results with the base
FDDA underestimated the maximum jet speed by up to 3
m/s (Figure 4-2), and wind directions exhibited a southerly
bias of 20 or more relative to observed directions.
Trajectory analysis demonstrated that these differences in
winds can introduce large spatial displacements in modeled
and observed O, patterns within the region. A follow-up
sensitivity study with the WRF meteorological model
has been performed with different FDDA options, which
also included the use of additional available wind profile
data (i.e., boundary layer wind profiler measurements and
Doppler radar wind profiles). Results demonstrated that
model simulation of the Mid-Atlantic nocturnal low-level
jet winds was improved when more observed wind profiles
were assimilated.
Accomplishments
More accurate transport patterns based on these findings
are expected to lead to model improvement in
estimating air quality concentrations because wind fields
generated by the WRF model are directly applied in
the CMAQ chemical transport model simulations for
retrospective periods.
Next Steps
Journal manuscripts presenting results of this diagnostic
evaluation study and WRF/FDDA sensitivity effort are
being prepared for journal submission in 2011.
4.5 Dynamic Evaluation
Introduction
Dynamic evaluation examines a model's concentration
response caused by changes in emissions and/
or meteorology through comparisons to observed
concentration change. Therefore, historical periods must
be identified that provide observable changes in pollutant
concentrations that can be related closely to known and
appreciable changes in emissions and/or meteorology.
The time period spanning EPA's NOx SIP call program
provides a valuable test of model response using the
dynamic evaluation approach since substantial reductions
27
-------
in NOx emissions occurred in the major point source
sector during 2003-2004. Gego et al. (2007) and U.S. EPA
(2007) show examples of how observed O, levels exhibited
notable decreases after implementation of the NOx SIP
Call. Gilliland et al. (2008) documented the results of an
initial dynamic evaluation study based on CMAQ model
simulations with three different chemical mechanisms to
assess maximum 8-h O, changes between summer 2002
(pre-NOx SIP Call) and summers 2004 and 2005 (post-NOx
SIP Call).
Recognizing the large reduction in anthropogenic activity
that naturally occurs on weekends relative to weekdays,
we performed a dynamic evaluation of CMAQ modeling
system for the so-called "weekend O, effect" to determine
if observed changes in O, because of weekday-to-
weekend (WDWE) reductions in precursor emissions
could be accurately simulated. The weekend O, effect
offered another opportunity for dynamic evaluation, as it
is a widely documented historical phenomenon that lias
persisted since the 1970s.
Research Activities
During the past year, the Division expanded on the NOx-
SIP Call analysis by assessing the changes in maximum
8-h O, concentrations (> 95th%) from CMAQ simulations
spanning five consecutive summer periods (2002 to 2006)
against concentration changes at the rural-based Clean Air
Status and Trends Network (CASTNET) sites. Figure 4-3
below depicts variations in the modeled and observed O,
changes for each summer period relative to 5-year mean
values. The modeled results track the observed changes
rather closely, and considerable overlap exists during
most periods. However, the largest differences between
the modeled and observed values occur for the summer
2002 period, which are attributable to the notable model
underestimates of maximum O, levels during several
high O, episodes. A diagnostic study has been underway
to investigate horizontal transport and O, chemical
formation to examine whether either of these processes
is contributing to the underestimation. In addition.
Modeled and Observed Relative Changes in 8-h 03
Departure from Average over Five Summer Periods
ifi
*
2002 2003 2004 2005
2006 2007
Figure 4-3. Box/whisker plot of the percentage change
in modeled (gray) and observed (green) maximum 8-h
O, concentrations (>95th percentile) for each summer
period relative to a 5-year mean at the CASTNET
monitoring sites in the eastern United States. The boxes
span the 25th to 75th percentiles and whiskers extend
from the 10th to 90th percentiles.
(U
O)
c
co
-C
O
sz
o
"¦M
2
-*—<
sz
CD
o
c
o
o
i r
2002 2003 2004 2005 2006 2007
YEAR
Figure 4-4. Change in weekday 3-h average morning
NOx concentrations. Model results (gray) and
observations (white) are based on 42 urban sites. Each
box/whisker plot shows the median values (line inside
boxes). Boxes span the 25th to 75th percentiles, and
whiskers extend from the 10th to the 90th percentiles of
the concentration distributions.
28
-------
analyses arc proceeding with modeled and observed O,
profiles over multiple summers to determine the extent of
concentration change aloft.
Another dynamic model evaluation effort examined
weekday morning (high-traffic period) concentrations
of NOx to assess changes in modeled and observed
NOx levels that could be attributable to mobile emission
changes over this multi-year period. Although modeled
mobile emissions dropped by 25% over this period. CMAQ
results were governed by the 15% decrease evident in total
NO emissions, which also include notable contributions
x 7
from area, nonroad. and the minor point-source sectors.
Using 18 years (1988 to 2005) of observed and modeled
O, and temperature data across the northeastern United
States, we reexamined the long-term trends in the weekend
effect and confounding factors that could complicate the
interpretation of this trend and explored whether CMAQ
could replicate the temporal features of the observed
weekend effect. This work was performed in collaboration
with researchers from the State University of New York
and the University of Idaho.
Accomplishments
Three journal articles on dynamic evaluation were
published during 2010. Godovvitch and Rao (2010)
revealed that spatial patterns of percentage decreases in O,
between 2002 and 2006 showed strong similarities between
the modeled and observed results, although the modeled
changes were somewhat less than observed changes.
Godowitch et al. (2010) reported that observed urban NOx
concentrations decreased by nearly 25%, whereas modeled
concentrations declined by about 15% from 2002 to 2006
(see Figure 4-4). Pierce et al. (2010) reported that CMAQ
could replicate the decrease in amplitudes of the weekly
O, cycle that occurred during an 18-year period. However,
similar to the two other dynamic evaluation studies, the
modeled response of O, to weekday-weekend differences
in emissions was somewhat less than that observed.
These studies suggest that more attention should be given
toprobing uncertainties in emission estimates, reducing the
grid cell si/c in the lowest layer of CMAQ, and using time-
dependent and more realistic boundary conditions for the
CMAQ simulations.
4.6 Probabilistic Model Evaluation
Introduction
Most model evaluation studies have focused on assessing
the performance of a single set of deterministic modeling
results. However, it is well known that air quality
predictions arc sensitive to uncertainties in meteorology,
emissions, and processes represented within the chemical
transport model itself.
Research Activities
To support probabilistic model evaluations and the
application of air quality models in a probabilistic
manner, research is underway to achieve the following
accomplishments.
• Estimate the sensitivity of O, and aerosol predictions to
different model configurations and inputs (an AQMEII
activity)
• Develop methods to assess the uncertainty in CMAQ
predictions of O, and aerosols using an ensemble of
model configurations and input sets
• Apply these methods to assess the uncertainty in
CM AQ-predicted changes in O, and aerosols resulting
from emission changes
• Use statistical methods to postprocess the ensemble of
model runs based on observed pollutant levels
® Use the uncertainty information to help prioritize key
model development needs
In previous years, this work was based on a 12-mcmber
ensemble of CMAQ simulations that included six
different MM5 configurations and two different chemical
mechanisms. We expanded the membership of the
ensemble to include uncertainty from emissions by using
the CMAQ-DDM-3D to generate large member ensembles,
while avoiding the major computational cost of running
a regional air quality model multiple times. The resulting
ensemble of CMAQ-cstimated O, concentrations was used
to fit a probability distribution function, and the quality of
the ensemble-derived pdf was assessed using observations
and parametric (confidence interval) and nonparanictric
(ranked-histogram) methods. A journal article of this work
was published in Environmental Science and Technology
(Pinder et al.. 2009)
Accomplishments
During 2010, we began exploring statistical postprocessing
methods to weight the ensemble members and improve
estimates of model uncertainty. We applied a Baycsian
model averaging (BMA) technique to calibrate the
ensemble predictions by weighting each individual
ensemble member based on how closely it matches
observed O, values. We evaluated these methods based
29
-------
(A)
Maximum Daily Average 8-h 03
Concentration atTerre Haute, IN
(AQS#181670024)
8-
8-
O _
CN ^
2CC2 base mode
2005 base mode
2002 node dUMDJQDfi
2CCS mojei dtstribJBon
~ ~ 2002 ooier.attcns
* * 200S odise^vaflnni
-| 1 1 r
40 50 60 70
ppb
80
1
90
(B) Maximum Daily Average 8-h 03
Concentration at Detroit, Ml
(AQS#261630019)
CDIL base r-cdei
2Tji base r-ccei|_
Q0Q2 rcoel clrtrfcullor
rcacl clsfr fcultor
^^^2302 observation^
233z obsservatcrs
ppb
Figure 4-5. Base model, observational, and model ensemble empirical cumulative distributions of maximum daily
average 8-h O, concentrations for 2002 and 2005 at two AQS sites: (a) Terre Haute, IN; and (b) Detroit, MI. All
ensembles were constructed based on ±50% uncertainty in emissions of area and mobile NO and ±3% uncertainty
in emissions of point NOx. The wide spread of the ensemble at the Terre Haute site indicates greater sensitivity to NOx
emissions in comparison with the site in Detroit.
on observed O, in the Southeast in the summer of 2005.
A manuscript describing these results currently is being
drafted and is expected to be submitted for journal review
in 2011. As we continue to develop and improve these
methods for creating probabilistic model outputs, we are
also investigating how these tools can be used in policy-
relevant applications. Specifically, we are combining
probabilistic evaluation techniques with dynamic
evaluation of model-predicted O, concentrations. Using
DDM-produced sensitivity coefficients and building on
our previous model evaluation efforts, we estimated the
dynamic response of the model because of NOx reductions
between 2002 and 2006, while accounting for uncertainties
in NOx emission inputs (Figure 4-5). This work has been
described in a journal article accepted for publication by
Napelenok et al. (2011).
Next Steps
We currently are collaborating with Rice University and
N.C. State University to explore other Bayesian estimation
techniques for quantifying the uncertainty in model inputs
and outputs. In addition, we are collaborating with OAQPS
on a pilot project to compare a DDM-based approach
for examining emission control options with response
surface modeling methods that have been developed and
used by the regulatory community for the past several
years. This project will enable us to explore the strengths
and weaknesses of different probabilistic approaches for
supporting EPA's air quality management and policy needs.
30
-------
5.0
Climate and Air Quality Interactions
5.1 Introduction
Energy generation from fossil fuels contributes to
global warming and degrades air quality, both of which
profoundly impact human and ecosystem health.
Atmospheric levels of carbon dioxide (C02) and other
GHGs have increased dramatically since the Industrial
Revolution, and emissions from fossil fuel combustion
have been linked to human disease since the London
smog event in 1952. The Intcrgovcrninental Panel on
Climate Change (IPCC) Fourth Assessment Report (AR4)
(Solomon et al.. 2007) concluded that the continued
rise in GHGs from human activity is the primary cause
of the temperature increases observed over the 20th
century, and that global warming is likely to continue
over the next century even with significant mitigation of
greenhouse gas emissions. These temperature increases
also lead to changes in other climatic conditions, such as
changes in precipitation intensity and duration. Extreme
weather conditions could become more frequent, which,
in turn, can affect adversely human and ecosystem health.
Emissions from anthropogenic combustion sources arc also
the largest sources of PM, O,. CO, and NOx and sulfur.
These pollutants arc known to contribute to respiratory
and cardiovascular effects in human populations, as
well as ecological effects to aquatic (acidification and
eutrophication) and terrestrial ecosystems (damage to
agricultural and other vegetation).
The scientific issues of global climate change have been
studied for decades. However, in response to the 2007 U.S.
Supreme Court decision in Massachusetts vs. EPA and the
EPA Administrator's recent Endangcrnicnt Finding under
the CAA Act (U.S. EPA, 2009), the EPA must rapidly
prepare for rulemaking regarding control of emissions.
Additionally, the EPA Office of Air and Radiation (O AR)
recognizes that some "traditional" pollutants regulated
under the Clean Air Act also have radiative forcing
properties (e.g., the black carbon component of PM,,, O,)
tliat impact climate. The Agency has a need for integrated
policy approaches to both mitigate climate change and
manage air quality.
The urgency also has increased greatly to understand the
implications of climate change on local and regional scales,
given the numerous anticipated impacts on environmental
protection and regulatory responsibilities of the Agency.
For example, research already has demonstrated that future
climate conditions very likely will increase air quality risks
and decrease the effectiveness of emission control efforts
(e.g., Nolte et al.. 2008; Weaver et al.. 2009). In addition to
enhanced photochemical production in a generally warmer
environment, greater frequency of stagnation events could
exacerbate air quality problems. It also is anticipated that
increases in extreme precipitation events could lead to
additional water-borne disease outbreaks and degradation
of water quality conditions. Anticipated damages to aquatic
and terrestrial ecosystems could include impacts such as
encroachment by invasive species and bionie migration.
The Agency's needs fall into two broad questions from
the air. climate, and energy integrated transdisciplinary
research (ITR) plan that will be used to organize and
focus AMAD research activities described in this
chapter. The first regards risk assessment and adaptation:
How will climate change impact air and water quality,
water availability, and ecosystems'.' Then, there is risk
mitigation: How can the Agency best contribute to climate
change mitigation through U.S. controls of both long-lived
GHGs and short-lived pollutants'.'
Accomplishments
Because this research area represents a significant
broadening of scope beyond AMAD's traditional domain
of air quality, a white paper. "Integration of air quality
and climate change—mode ling connections from global
to regional scales." was undertaken to help plan and
guide the research. This white paper, together with white
papers on interactions between air quality and human
health and between air quality and ecosystems, was
reviewed by an external panel in March 2010. The white
papers contributed to the development of AMAD's 5-year
Strategic Plan and were published on the Division's Web
site in October 2010.
5.2 Regional Climate Modeling:
Dynamical Downscaling
Introduction
Global climate models (GCMs) arc used to simulate
gridded climate conditions with worldwide coverage.
Because GCMs often arc run for niulticcntury simulations.
GCMs typically generate output at fairly coarse temporal
(e.g., monthly or daily) and spatial (e.g., 2° latitude x
2.5° longitude or 1° latitude x 1° longitude) intervals.
To investigate and understand regional effects of climate
change (and particularly to capture extreme events),
higher temporal and spatial frequency of the gridded
fields is required. One broad category of methods to
create regional climate fields that arc influenced by GCM
simulations is "downscaling." Dow nscaling involves using
the fields from a GCM, in conibination with additional
information about the regional scale (e.g., topography,
land use, historical regional climate data) and physical,
mathematical, and/or statistical models, to extend the
GCM simulation to finer spatial and temporal granularity.
With increased horizontal texture, the downscaled
fields arc more sensitive than the GCMs to local spatial
31
-------
heterogeneities in weather and climate that result from
topographical changes, land/water interfaces, vegetation,
and population. Downscaled climate fields can then be
used to simulate the regional impacts of climate change
on water quality and availability, agriculture, ecosystems,
human health, air quality resulting from emissions control
strategies, and energy demand.
Dynamical downscaling involves using a regional climate
model (RCM), where the initial and lateral boundary
conditions arc obtained from a GCM. Dynamical
dow nscaling is fundamentally similar to regional weather
forecasting, except that a GCM is used for lateral boundary
conditions; there arc no observations to assimilate, and
simulation periods arc years or decades rather than days
or weeks. Furthermore, rather than evaluating the model
based on its skill in representing day-to-day weather
observations, the model's performance is evaluated in
terms of how well it represents climatic features, such
as seasonal means, the magnitude of extremes, or the
frequency of events exceeding some threshold.
Research Activities
For AMAD's dynamical downscaling, the WRF model
is being used as an RCM. Initial work is using coarse-
scale reanalysis fields (i.e., 2.5° x 2.5°) as input, which is
comparable in resolution to the GCM that will be used for
downscaling future climate. Focusing first on reanalysis
data enables comparison of the RCM output to a wide
variety of observational data, including both gridded data
at smaller spatial scale and point observations, to examine
the extent to which the RCM added value over the coarse-
scale input data.
Simulations over the continental United States have been
conducted using grid cells that arc 36-km x 36-km in
si/c. A key goal of this research is to develop optimal
techniques for downscaling the large-scale forcing resolved
by coarse grid global data while allow ing effective
small-scale forcing and dynamical response to be well
represented by the RCM. Comparative analysis of scale-
selective techniques for data assimilation (nudging) and
boundary condition specification arc important aspects of
this research. Thus, modeling experiments will compare
the effects of using no nudging (other than through the
lateral boundaries), analysis nudging, and spectral nudging.
Evaluation is planned against means and extremes of near-
surface fields, cloud and radiation fields, and spectra.
Accomplishments
During 2010, AM AD completed an initial study on the
use of WRF as an RCM. Three simulations of the year
1988 were conducted, and the results obtained using
analysis nudging, spectral nudging, and without any
nudging were compared to each other and to data from the
North American Regional Reanalysis (NARR) product. A
manuscript describing this study (Bow den ct al.. 2011) has
been completed and will be submitted for publication
in 2011.
32
Also during 2010, AMAD created its first continuous,
inultidccadal WRF simulations. Simulations were a.s
described above but covered the period from 1988 to 2007.
These simulations will be used for several evaluation
studies that arc planned for 2011.
Concurrent with the RCM simulations conducted
using reanalysis data, we have developed a dynamical
downscaling linkage between WRF and the NASA GISS
Model E GCM. A 6-vcar test datasct was acquired from
collaborators at GISS to establish the logistical connections
between Model E and WRF (i.e., vertical coordinates,
state variables, linkage of surface fields) and to test the
downscaling techniques.
Next Steps
Evaluation of the 20-ycar WRF simulations is in progress.
In addition to analyzing how well WRF represents
mean climatic variables, such as surface temperature,
precipitation, and radiation fields, we will focus on the
effect of analysis and spectral nudging on those mean
variables and on the simulated atmospheric circulation. In
addition, we plan to examine the ability of WRF to capture
the frequency of extreme events, such as droughts
and floods.
As GISS develops Model E fields for the IPCC AR5, those
data will be made available at high temporal frequency
(i.e., three-hourly or six-hourly) for "time slice intervals"
(i.e., approximately decadal subsets) of the GCM
simulation. Using knowledge gained from the reanalysis
testing, evaluated dynamical downscaling techniques will
be employed to create regional downscaled climate fields
from current and future decadal time slices that will be
used to analyze high-resolution regional climate change
effects.
Several aspects of dynamical downscaling need to be
refined to adapt WRF, which is a numerical weather
prediction model, to regional climate applications. For
example, the model must be able to realistically simulate
the surface energy budget over inultidccadal simulations.
Initial testing with different radiation sclieincs in WRF
has show n a persistent undcrprcdiction of cloud cover
that results in overheating of the land surface. Clearly,
improvements in the cloud and radiation physics
components of WRF arc needed for regional climate
applications. In 2011. we will be evaluating the modeled
radiation fields from the 20-ycar retrospective WRF
simulations against satellite retrievals from the Clouds
and Earth's Radiant Energy System (CERES) and MOD IS
instruments, which arc both aboard the Aqua and Terra
satellites, and against observations taken at the Department
of Energy's Atmospheric Radiation Measurement (ARM)
Climate Research Facility. These efforts to evaluate and
improve the cloud and radiation budgets in WRF arc being
coordinated with research to implement indirect aerosol
-------
radiative forcing in the WRF-CMAQ system (described
below and in Chapter 3), which also involves evaluation
using CERES and MODIS data. These 20-vcar WRF
simulations also will help with understanding how well
extremes and variability arc simulated by WRF, helping
build confidence in future-year estimates of extreme events
and variability under climate change conditions.
5.3 Regional Climate Modeling:
Statistical Downscaling
Introduction
Statistical downscaling methods use correlations among
historically observed and modeled meteorological variables
under current climate conditions to predict future regional
and/or local patterns and events from the broader-scale
GCM simulations. Typically, these approaches do not use
the same detailed information that is used in dynamical
downscaling, such as physical equations, orographic data,
or extensive land-use information. The advantages of
statistical downscaling methods lie in their efficiency and
speed, and these methods could be particularly attractive
if numerous climate scenarios need to be investigated.
Statistical methods arc not limited by the resolution
achievable by the nested regional dynamical model. Thus,
statistical methods possibly could be used to gain a better
understanding of fine-scale variability, even down to point
locations, given high-resolution training data.
It has been reported in the literature that the performances
of dynamical and statistical downscaling arc comparable
for current climatic conditions. However, it is questionable
whether statistical models can perform as well under
future conditions (Wilbv ct al., 2002) because statistical
downscaling methods rely on associations among
meteorological variables. These relationships do not
explain all of the inherent variability in atmospheric
phenomena; in fact, the choice of variables to be used as
the "predictors" in such approaches can be a difficult part
of the statistical downscaling process. Once a statistical
model has been developed for a particular time period
(e.g., using current climate), it is unclear whether the
relationships it incorporates will remain the same under
different climatic conditions (e.g., in future decades). (Sec
Schmith. 2008, for more discussion.) However, statistical
downscaling docs make this "stationarity" assumption as it
extrapolates to future conditions.
Research Activities
During the past year, AMAD has been reviewing,
improving, and testing statistical downscaling methods,
with an emphasis on evaluation and calibration. Most
of this research is being done using a regression-based
downscaling technique proposed by Hoar and Nychka
(2008), which models the response variable (i.e., surface
temperature) as a function of fine-scale temperature
estimates obtained by applying a thin-plate spline
interpolation technique to the output of a GCM (or other
coarse-resolution model). This enables us to bypass the
difficulties associated with predictor selection (a difficult
issue with many regression-based methods), so that we can
focus on issues associated with calibration, training period
selection, and quality and quantity of training data. If the
efficiency and accuracy of this method proves suitable,
this method could be adapted to use different response
variables or different predictors.
Accomplishments
In the past year, AMAD has implemented a regression-
based statistical downscaling method proposed by Hoar
and Nychka (2008). We currently arc finishing a period of
extensive testing, with a particular focus on understanding
the impact of such assumptions as stationarity and temporal
correlation. The downscaling met hod has been applied for
temperature and has been trained and evaluated using finc-
scalc output from the PRISM model, encompassing the
continental United States during the period. 1907 to 1996.
In tests with simulated data, we found that temporal
correlation docs have an impact on the downscaling
estimates and should be accounted for in the downscaling
model. Two statistical techniques to aid in detecting and
estimating temporal correlation, the Durbin-Watson test
and restricted maximum likelihood (REML), performed
well in our simulation studies. With tliesc adjustments,
the Hoar and Nychka method produced confidence
intervals for future years' temperature that were well
calibrated, with only a small loss in efficiency because
of the autocorrelation. These simulation studies provided
benchmarks of how well we could expect the method to
perforin in "ideal" circumstances, in which stationarity and
type of correlation structure can be assumed.
AMAD also applied the downcaling technique to the
PRISM temperature data. With a 30-year training period
(1907 to 1936), the met hod was used to obtain temperature
estimates for individual years (1937, 1996) and interval
averages (1937 to 1966 and 1967 to 1996). In contrast to
the simulated cases, the confidence intervals generated by
the technique were calibrated poorly. The magnitude of
this discrepancy (between simulated cases and cases using
PRISM) suggests that there arc fundamental problems with
the assumptions of the method when applied to "real-
world" cases and. particularly, with the assumption that the
statistical associations remain constant over time. Work
is currently underway to determine whether this problem
can be ameliorated (e.g., additional covariatcs. other data
sources, etc.).
33
-------
Next Steps
AMAD is planning future work to better understand
the relative strengths and weaknesses of statistical and
dynamical downscaling for various applications. Research
questions of particular interest include those that follow.
• Developing at least a basic understanding of how the
uncertainty may change when applied to future-year
GCM simulations
• Identifying the relative strengths/weaknesses of the
dynamical and statistical approaches to downscaling
• Determining whether hybrid downscaling approaches
may be able to capitalize on the strengths of both
dynamical and statistical methods
5.4 Development of Coupled Regional
Climate and Chemistry Modeling System
Introduction
Climate change is influenced by long-lived GHG and also
by short-lived radiativcly active gases and aerosols that
typically arc modeled by regional-scale air quality models
such as the CMAQ model. The direct and indirect feedback
effects of gases and aerosols on radiative forcing and cloud
microphysics have important climate impacts, particularly
at regional scales. Thus, both GCM and RCM components
of the climate modeling system need to include these air
quality feedback effects. Therefore, AMAD is developing a
coupled regional climate and chemistry modeling (RCCM)
system that can downscale climate and chemistry from the
GCM through a series of nested grids from hemispheric to
continental to local scales.
Research Activities
A two-way coupled WRF-CMAQ modeling system is
being developed and tested for air quality and climate
applications at scales ranging from hemispheric to local.
There are several advantages to an integrated meteorology
and chemistry modeling approach, including more frequent
data exchange for high-resolution modeling; better
integration of physical processes, such as cloud-chemistry
interactions; and feedback effects of gases and aerosols on
radiation and cloud microphysics. Thus, the WRF-CMAQ
system is an important development for air quality research
and assessment over multiple scales and regional climate
modeling, where direct and indirect feedbacks on radiation,
clouds, aerosols, and precipitation arc important.
Further research and development of more comprehensive
treatments of direct and indirect feedback of short-lived
gases and aerosols on radiation and cloud properties
arc needed. WRF-CMAQ currently includes direct
feedback of aerosols, accounting for si/c distribution and
chemical composition, on SW radiation in the Community
Atmospheric Model (CAM) radiation scheme and the
Rapid Radiative Transfer Model for GCMs (RRTMG)
SW scheme (lacono et al., 2008). The direct effects of
troposplieric O, on LW radiation in the CAM and RRTMG
LW schemes also have been implemented. Some initial
evaluation studies have shown an expected response to
these direct feedback effects for both typical summer
conditions in the eastern Unitcd States and for an outbreak
of multiple wildfires in California in June 2008. Aerosol
direct radiative effects associated with scattering and
absorption of incoming radiation result in a reduction of
SW radiation reaching the surface, which then translates
to reduction in temperature at the surface, as well as a
reduction in PBL height. For moderate pollution levels,
such as during typical summer conditions in the eastern
United States, impacts on SW reduction and subsequent
suppression in PBL heights arc relatively modest.
Significantly, larger changes in simulated surface radiation
were seen for the California wildfire case where smoke
from the wildfires resulted in much higher aerosol loading.
The reduced SW radiation caused much cooler surface air
temperatures (up to 4 to 5 K in the smoke plumes), lower
PBL heights (up to 400 m), and significantly higher ground
level concentrations of O, and PM25 (Mathur et al.. 2009).
Evaluation of direct aerosol forcing is continuing in 2011
by extending the initial 10-day test case for the eastern
United States to the full month of August 2006 using the
RRTMG radiation scheme and the current versions of both
WRF (v3.2) and CMAQ (v5.0). The California wildfire
case also is being updated with the latest model versions
and improved estimates of wild fire emissions.
New research aimed at improved modeling of aerosol
mixing state and optical properties has been initiated
through collaboration with UNC. The work will lead to
more accurate direct radiative effects of mixed composition
aerosols, particularly involving black carbon aerosol, in the
RRTMG SW scheme. Through this collaborative research,
we also arc investigating the feedback effects of aerosols
on LW radiation. It is hypothesized that some aerosols,
especially black carbon, may absorb LW radiation as well
as docs SW radiation.
Aerosols also affect radiation through their role as cloud
condensation nuclei (CCN). Greater concentrations of
CCN lead to a greater number of smaller cloud droplets,
causing greater SW cloud albedo. Activation of aerosols
as CCN alters cloud microphysics such that warm rain
processes arc less efficient, thereby extending cloud
lifetime. These effects of increased aerosol concentrations
on clouds arc known as the first and second indirect
effects, respectively.
Development and evaluation of the direct and indirect
feedback processes in the two-way coupled WRF-
CMAQ modeling system, and eventually the RCCM,
will advance through comparison with the latest release
version of the online coupled WRF/Chem model. WRF/
Clicni currently simulates aerosol, cloud, and radiative
interactions when employing its CBM-Z photochemical
mechanism and MOSAIC aerosol scheme (Chapman et
al., 2009). Sensitivity tests using present-day and future
-------
emission scenarios oriented toward climate change studies
will be conducted with both the WRF/Chem and the two-
way WRF-CMAQ modeling systems, and subsequent
simulation results will be analyzed to evaluate their
feedback processes.
Recent forecasting and retrospective CMAQ applications
over continental scales and for annual cycles have
pointed to the need for more robust methods to specify
chemical lateral boundary conditions (LBC). Although
linking with global-scale atmospheric chemistry models
has been pursued, it also is recognized that biases in
the global model can propagate and influence regional
CMAQ calculations and often confound interpretation
of model results. The successful extension of CMAQ to
hemispheric scales would provide several advantages: (1)
consistent LBCs for regional calculations focusing on air
quality issues in the United States (i.e., the biases arising
from process formulations are consistent at all scales); (2)
opportunities to explore emerging challenges related to
the processes regulating "background" levels (The recent
revisions to the O, NAAQS and possible further reductions
in the O, standard make the accurate representation of
background O, in the model even more imperative than
before); (3) lay the foundation for future applications of
the integrated WRF-CMAQ model to explore radiative
impacts of aerosols and radiatively active gases, such as
03, on the earth's energy budget and the indirect feedbacks
of aerosols on cloud microphysics; and (4) a framework for
assessing uncertainties in modeling regional distributions
of long-lived HAPs and toxics (e.g., mercury). When the
WRF-CMAQ model is configured for climate studies, the
hemispheric grid domain will be an important link between
the global climate model and the regional nested grids over
North America and local areas
Accomplishments
During 2010, an initial implementation of aerosol indirect
effects using the CCN activation scheme from the two-
way, or online, coupled WRF/Chem model (Abdul-Razzak
and Ghan, 2000) was adapted to WRF-CMAQ.
Initial testing of the WRF-CMAQ over the Northern
Hemisphere lias shown realistic ground-level concentration
distributions over the major emissions regions and realistic
aerosol optical depth (AOD) compared with MODIS
satellite retrievals. Preliminary analysis of the hemispheric
simulation for March 2006 showed remarkably good
agreement of modeled AOD with AERONET observations
in Beijing and Greenbelt. MD.
The proposal, "Development of a Next-Generation
Comprehensive Global Environmental Modeling System
for Assessing Air and Water Quality in Current and Future
Climates," was selected for funding as one of the initial
EPA ORD Pathfinder Innovation Projects. The proposal is
to extend CMAQ to the global scale by coupling it to the
Ocean-Land-Atmosphere Model (OLAM) GCM. OLAM
uses a telescopic finite-volume discretization scheme that
divides the globe into regular polygons of arbitrary size.
It lias the capability to provide greater spatial resolution
over areas of interest, while simulating the entire globe
Los Angeles
Focus Region
Houston
Focus Region
Figure 5-1. Example global telescoping grid structure.
35
-------
(see Figure 5-1). Simulating the entire globe at once would
avoid problems that can arise in downscaling a GCM
with an RCM. such as poor temporal resolution of lateral
boundary conditions, inconsistent physics treatments
between different modeling platforms, and the inability of
the RCM to feed back on the GCM solution.
3-year collaborative project with the Department of Energy
(DOE) entitled: Evaluation of the Interactions Among
Tropospheric Aerosol Loading. Radiative Balance, Clouds,
and Precipitation. This observational-modeling study
aims to address this issue by systematically investigating
changes in anthropogenic emissions of S02 and NO over
the past two decades in the United States, their impacts on
anthropogenic aerosol loading over the North American
troposphere, and subsequent impacts on regional radiation
budgets. The DOE funding of this interagency agreement
will be used to support three new postdoctoral fellows to
work on this project.
5.5 Decision Support Tools for Managing Air
Quality and Mitigating Climate Change
Introduction
Emissions of long-lived and short-lived GHGs and aerosols
have a substantial impact on both climate and air quality.
The goal of this effort is to develop an integrated suite of
models to examine the impacts and understand the key
uncertainties of emission mitigation strategies on both
climate and air quality. Ultimately, the goal is to answer
the question: What is the impact of U.S. emission control
strategies on air quality and climate? Figure 5-2 describes
the two classifications of decision-support tools. On the
left side of the diagram are screening tools to enable
decisionmakers to rapidly identify emission scenarios
and policy options of interest. On the right is the series
of models, spanning global and regional scales, needed
to fully assess the impacts of a given policy option. The
screening tools can be used to identify a narrow but critical
subset of emission scenarios or policy options that then
can serve as input to the more complete models. The
work in this section describes AMAD efforts relevant to
the screening tools and the Atmosphere-Ocean General
Circulation Modeling.
This schematic describes the models and linkages from
changes in policy decisions to changes in SLCF, GHG,
and CAPs, and their local impacts on human health and
ecosystems.
Research Activities
Decision support tools for scenario analysis. A key
finding of the Climate Impacts on Regional Air Quality
(CIRAQ) project is that future air quality largely is driven
by changes in future emissions. Forecasting emission
changes over decades is impossible, so decision-makers
rely on scenarios to help guide their choices. However,
scenarios are comprised of many assumptions. Discovering
policy options that deliver robust results across a range
of assumptions requires investigating many scenarios.
Unfortunately, state-of-the-science global and regional
climate models require tremendous computational
resources. Exploring more than a handful of scenarios is
not tenable. Decisionmakers need approximation tools
that can be used to rapidly screen the costs and benefits of
climate mitigation strategies.
Emission & Land Use
Scenarios
Reflecting U.S. Policy
Options for GHGs,
SLCF, and CAPs
Atmosphere Ocean
General Circulation Model
GISS ModelE & other models
with online chemistry
and aerosols
It
Screening Tools
MAGIC C
GHG
A Global T
Adjoint
GEOS-Chem
SLCF ->
ARad. Forcing
Adjoint
CMAQ
CAPs
AAir Quality
1
GHG & SLCF A Climate
(Temperature, precipitation, etc.)
2° * 2.5° horizontal scale
Weather Research and
Forecasting Model (WRF)
~
WRF-CMAQ
with online
chemistry
and aerosols
Downscaling to
regional climate
at finer horizontal
scale
Assessments: Human Health, Air Quality,
Water Resources, Ecosystem Impacts
Figure 5-2. Schematic describing model linkages.
36
-------
For long-lived GHGs. the emissions arc proportional
to the climate impacts. However, short-lived cliniatc-
forcing gases and aerosols undergo chemical and physical
transformations that substantially change their climate
impacts. Radiative forcing can be used as an approximate
surrogate for the impact of these species on climate. We arc
developing an integrated assessment model composed of
four parts: (1) the GEOS-Chem global chemical transport
model, (2) the LIDORT radiative transfer model, (3) the
adjoint of these two models coupled together, and (4)
the market allocation (MARKAL) model. This project is
named GLIMPSE.
GEOS-Chem and LIDORT can estimate the impact of
emissions, chemical fate, and transport on direct radiative
forcing. The adjoint of these models can be used to build
a reduced form model that can rapidly assess the impacts
of a change in emissions on radiative forcing. Finally.
MARKAL connects technologies and energy demands
to emissions. With GLIMPSE it is possible to examine
combined constraints of GHG emissions, short-lived
species radiative forcing, and relative cost to examine the
trade-oils between different policy options.
This tool is being developed in collaboration with our
colleagues in NRMRL who arc extending MARKAL to
include more sources of short-lived, radiativcly active
gases and aerosols. This is in collaboration with University
of Colorado professor Davcn Henze, who is the original
developer of GEOS-Chem Adjoint.. Finally, successful
completion will require frequent interaction with our
colleagues in the EPA program offices to match the
decision-support tool to their needs.
The first application of the tool will be to examine impacts
of emission reductions from black carbon and sulfate.
Next, it will be extended to include radiative forcing and
chemical precursor emissions for Ov Finally, the adjoint
version of CMAQ will be included as part of GLIMPSE to
simultaneously assess air quality and radiative forcing.
Global modeling of U.S. Emission control strategies;
impacts of short-lived, radiativety active gases and
aerosols. Although greenhouse gas emissions and short-
lived species radiative forcing arc useful approximations,
they do not fully prescribe the change in climate. Scenarios
and policy options that merit further consideration will be
used as input to the global climate model for a complete
assessment of the global climate impacts of U.S. climate
mitigation strategics.
The first phase of this work will use the GISS Mode IE
simulations prepared for the regional climate downscaling
ta.sk. These climate simulations simulate present day and
future climate under multiple Representative Concentration
Pathways (RCPs). the scenarios under development for
the IPCC AR5. GISS Model E simulations will provide
detailed diagnostic information for 10-ycar timcsliccs,
beginning in 1980, 2000, 2030, and 2100.
In the second phase, a few scenarios informed by
policy options developed using GLIMPSE will be used
as input to additional climate simulations using GISS
Model E coupled with online chemistry to resolve the
climate impacts of U.S. emission control strategics on
global climate. A key purpose of these simulations is
to contrast the relative impact of GHG mitigation with
emission reductions of short-lived climate-forcing species.
The results from these simulations will be used with
the regional downscaling methods described in Task
AMAD11-108 to determine the regional climate impacts
of U.S. emission changes.
Accomplishments
We have developed a prototype of the GLIMPSE system,
presented the results at scientific meetings and conferences,
met with stakeholders in the program offices, and
incorporated their feedback.
Next Steps
In 2011, we will use GLIMPSE as a tool to analyze
the impacts of U.S. emission control strategics on
concentrations and global radiative forcing of short-lived,
radiativcly active aerosols, including S04, black carbon,
and OC and to assess the impact of possible climate change
mitigation policies on radiative forcing and air quality.
Also in 2011, we will conduct a review of proposed
feedback mechanism to assess the impacts of reactive
nitrogen emissions on climate cliangc. and we will
conduct a bounding study using models and observations
to examine the extent to which anthropogenic emissions
enhance O, in the upper troposphere.
5.6 Biosphere-Atmosphere Interactions:
Improving the Treatment of Isoprene
Oxidation
Introduction
The scenarios described above characterize emissions
trajectories for sources that can be controlled directly.
However, climate and land management changes arc
likely to impact future emissions of VOCs as well. With
warming temperatures and increased CO;. isoprene
emissions arc expected to increase. A review of the NCER
STAR-fundcd projects on climate and air quality identified
a key uncertainty in the future O, production efficiency as
the chemical treatment of isoprene and the formation of
isoprene nitrate compounds.
Isoprene is emitted naturally by many species of trees and
shrubs. With the exception of methane, which reacts much
more slowly in the atmosphere, isoprene is the most highly
emitted VOC, both worldwide and within the United
States. It is an important precursor for O, formation and
is thought to be a substantial contributor to SO A. These
species arc critical for understanding interactions between
37
-------
air quality and climate for two reasons. First, in the future,
we expect there to be much more isoprene relative to
NOx. As demonstrated by recent field measurements,
the isoprene, chemistn and formation of O , under these
conditions is highly uncertain. Second, unlike the rest of
the United States, the climate in the Southeast has not
wanned. To what extent have light-scattering aerosols from
isoprene contributed to this trend?
We are partnering with collaborators making field
measurements and conducting laboratory studies to
improve our understanding of isoprene oxidation and to
assess the relative impacts of uncertainties in isoprene
emissions, chemical mechanism and reaction rates, and dry
and wet deposition rates.
Research Activities
Recent laboratory and theoretical studies have provided
new insight into the oxidation of isoprene. including
more detailed understanding of the formation and fate of
isoprene nitrates (INs; Paulot et al., 2009), formation of
epoxides under low-NOx conditions, and reformation of
HOx via isomerization reactions. We have incorporated
these reactions into the SAPRC07 chemical mechanism
and simulated the time period of the INTEX-NA
field campaign to investigate the impact of INs on 03
production and the oxidized nitrogen budget.
Accomplishments
The new scheme has been evaluated against smog chamber
experiments from three different laboratories. The results
show improved predictions of O, for the experiments with
the lowest NOx concentrations. We also have implemented
the new scheme into the CMAQ model and have conducted
simulations for the INTEX-NA field campaign period
during summer 2004. When compared with ambient
measurements, we find this updated chemistry provides
a more accurate simulation of the isoprene oxidation
products methacrolein, methyl vinyl ketone, and alkyl
nitrates. Including limited HOx radical regeneration in
the oxidation of isoprene improves simulated hydroxy
radical (OI-I) concentrations, especially when isoprene
concentrations are high. Over the southeastern United
States, a region of high isoprene emissions, simulations
including this chemistry show a 2 to 4-ppb increase in
O, (see Figure 5-3) and 30% increase in SOA.
Next Steps
We now are testing updates to the nitrate radical oxidation
pathway for isoprene, as well as experiments to examine
the sensitivity of simulated O, concentrations to key
reaction rates in the chemical mechanism. A follow-on
study is planned for 2011 that will focus on assessing
the relative impact of chemical mechanism uncertainties
with uncertainties in emissions and deposition rates on
simulated O, concentrations.
n
4.00 112
\
0.00 1
ppbV
148
July 5, 2004 0:00:00
Min*=-0.00 at (147,9). Max=4.32 at (103,40)
Figure 5-3: Change in simulated mean surface-level 03 concentration because of updated chemical
mechanism.
38
-------
6.0
Linking Air Quality to Human Health
6.1 Introduction
The need for improved dispersion models to support
human exposure assessments in urban areas has led to
a significant effort to design and build a new modeling
const met to efficiently provide near-field concentration
estimates to support exposure and health related studies.
Included in this effort is the ongoing near-roadway
research program that has been working toward improving
the urban-scale line-source (mobile source) dispersion
algorithms. Additionally, this task involves development
and application of model evaluation techniques and
acquiring, assessing, and processing databases for model
development and evaluation purposes. The new modeling
system will be applied in ongoing exposure and health
studies in collaboration with clients within and outside
of EPA.
Along with the need for better models is the need to
understand the optimal accuracy and precision in exposure
estimates needed to conduct human exposure and health
studies. In the absence of personal exposure measurements,
epidemiology studies traditionally have relied on imperfect
surrogates of personal exposures, such as area-wide
ambient air pollution levels based on readily available
outdoor concentrations from central monitoring sites. The
direct use of monitoring data inherently assumes that they
are representative of the air quality over a broad area.
However, there is increasing evidence tliat the current
monitoring network is not capturing the sharp gradients in
exposure caused by high concentrations near, for example,
major roadways. Yet, providing more detailed exposure
information can be resource intensive and may result in
greater (or undefined) uncertainty. Research in the Division
has focused on improving the understanding of the optimal
level of accuracy and precision for use in exposure
estimates.
6.2. Research Description
EPA's Clean Air Research niultivcar plan increases the
emphasis on research to better understand the linkages
between sources and health outcomes. This research area
is focused on the critical modeling link between sources
of air pollutants in urban areas and health outcomes.
Modeling the wide variety of pollutant sources in the
meteorologically and topographically complex urban
environment will require the development of an improved
mode ling system tliat is both complete in its handling of
the urban setting yet computationally efficient because
of the hundreds and possibly thousands of sources
tliat arc considered in exposure analyses. Additionally,
model development and evaluation databases and model
evaluation techniques will be developed to accomplish
this task. The EPA Fluid Modeling Facility is uniquely
designed for conducting laboratory studies characterizing
the often steep concentration gradients expected in
urban areas and is a major source of development and
evaluation data.
The new dispersion modeling system will be based
in a significant way on the previous development of
the American Meteorological Sociey/EPA Regulatory
Model (AERMOD) urban-scale dispersion model.
Although AERMOD is designed for a wide variety of
regulatory applications, the new model will benefit from
the science and algorithm basis of AERMOD without
the computational burden often required for regulatory
analyses. The new model will be an integral part of a
hybrid mode ling system tliat links observations with both
regional- and urban-scale modeled concentrations to
provide the necessary concentration distributions required
for human exposure modeling and associated liealtli risk
assessments. Although the new dispersion model is not
expected to account for complex chemistry, it is anticipated
that, for selected pollutants (e.g., NO), near-field chemistry
may be accounted for with simple paranictcri/ations.
In urban areas, mobile sources arc a significant contributor
to air pollution exposure. This research area includes an
effort to further understand the atmosplieric transport
and dispersion of mobile source pollutants within the
first few hundred meters of the roadway, a region often
characterized by complex flow (e.g., vehicle-induced
turbulence, noise barriers, road cuts, buildings, vegetation).
Examination of traffic emissions and potential population
exposures near roadways includes field and laboratory
measure incuts of concentration distributions near simulated
roadways. This task involves the investigation of near-
road dispersion through numerical, wind tunnel, and tracer
studies, with the goal of characterizing the near-road
spatial distribution of mobile source pollutants and the
development of improved modeling algorithms within the
new urban-scale dispersion model.
Research in the Atmospheric Exposure Integration Branch
(AEIB) also includes a project to compile the results of
a suite of tiered exposure studies to capture the critical
elements relevant to characterizing exposure (e.g.,
variability of pollutant. Ileal tli study design), common
findings resulting from the studies, and remaining
challenges. This work requires close coordination
across NERL's air quality mode ling and measurement
groups (AMAD and HEASD) to tap their collective
knowledge of the chemistry and physical processes
involved with the spatial and temporal properties of
multiple pollutants. In addition, close collaboration with
the epidemiology community (HEASD, the National
Center for Environmental Asscssniant [NCEA], NHEERL
39
-------
and academia) is critical in understanding the different
demands of the various health study designs and to
demonstrate the impact of the tiered exposure definitions
on the health effect measure. This research provides critical
information on the level of spatial and temporal data
needed to support human exposure and health studies. It
will result in greater efficiencies as more refined exposure
information can require more resources, yet result in more
(or undefined) uncertainty.
6.3. Accomplishments
Wind Tunnel Modeling of Flow and Dispersion
near Roadways
A series of physical modeling experiments was
performed in the Division's Meteorological Wind Tunnel,
investigating 14 common roadway configurations,
including a flat roadway with no surrounding obstacles
(base case), noise barriers of varied height and distance
from the roadway, two different porous barriers intended
to simulate the dispersive effects of roadside vegetation,
and depressed and elevated roadways. Results show that
the configuration of the roadway can have a substantial
effect on the concentrations both on and downwind of the
roadway (Heist et al., 2009a). The results are being used to
develop algorithms for inclusion in dispersion models to
account for these effects.
As part of the EPA and FHA field study of mobile source
pollutant dispersion near Interstate 15 in Las Vegas, NV,
a laboratory study was conducted in the Meteorological
Wind Tunnel on a 1:200 scale model of the study area
to examine the effect of the road depression associated
with a railroad overpass at the field site and to explicitly
define the flow and concentration fields surrounding
this major roadway (Figure 6-1). Wind tunnel data show
that the roadway depression (of about 5.5 m maximum
depth) results in a decrease in maximum ground-level
concentration of about 25% over that observed with a flat
roadway (Heist et al.. 2009b).
Dispersion Model Algorithm Development
A new line-source algorithm (AERL1NE) has been
developed that efficiently computes dispersion from
sources, such as roadways, considering multiple lanes of
traffic, low wind, meander conditions and varying wind
directions (including those parallel to the traffic flow)
(Venkatram et al., 2009). Comparison of AERLINE against
the Raleigh 2006 highway field study for both downwind
and upwind concentrations is in progress. A draft algorithm
formulation was developed that accounts for the influence
of roadways that are depressed below grade. This algorithm
is currently undergoing evaluation against wind tunnel data
and will be included in AERLINE in FY 11.
Analysis of the Raleigh 2006 Pilot Field Study
The 2006 Raleigh field stud}', conducted in the summer
2006, was designed to provide data to characterize the
influence of traffic-generated emissions in the near-road
environment, especially to help assess the impact on
air quality and particle toxicity near a heavily traveled
highway. Pollutant measurements were synchronized with
real-time traffic and meteorological monitoring devices
Figure 6-1. Wind tunnel model of Las Vegas field study area built at 1:200 scale.
40
-------
Sonie-100
Smiic-Zl
Traffic ram
Figure 6-2. Aerial photograph of study location in Raleigh, NC.
to provide continuous and integrated assessments of
the variation of near-road pollutant concentrations and
particle toxicity with changing traffic and environmental
conditions, as well as distance from the road (Figure 6-2).
This research helped to demonstrate the temporal and
spatial impact of traffic emissions on near-road air quality
(Isakov et al., 2007; Venkatrain et al., 2007).
Tracer Field Study of Flow and Dispersion near
Roadway Noise Barriers
An interagency agreement was developed between NERL
and NOAA's Field Research Laboratory to conduct a field
study in Idaho Falls of flow fields and tracer dispersion
from a line source (simulating roadway emissions)
immediately upwind of a 6-m-high barrier (typical of
near-road noise barriers found in urban and suburban
areas) (Figure 6-3). Simultaneous line-source tracer release
and measurements were conducted without the barrier.
The results currently are being analyzed to be used for
model algorithm development to account for the effects
of noise barriers in AERLINE. Unlike the associated wind
tunnel studies, which are limited to neutral conditions, the
tracer studies were designed to examine dispersion for all
atmospheric stabilities (Finn et al.. 2010).
Computational Fluid Dynamics (CFD) Modeling
of Near-Road Configurations
To complement the wind tunnel and tracer studies of
near-road dispersion, a series of CFD simulations has
been initiated in collaboration between NERL (AMAD
and HEASD) and NRMRL. In FY09, the FLUENT CFD
code was used to simulate the six-lane flat roadway
scenario, with conditions identical to the wind tunnel
study. The wind tunnel results were used to suggest the
most appropriate on-road initial dispersion conditions and
the most appropriate turbulence parameterization, as well
as other flow parameters to be used for other roadway
scenarios.
Gulf of Mexico Oil Fire Plume Modeling
During remediation efforts following the British Petroleum
oil spill, more than 400 separate controlled burns were
conducted over a 2-m period, resulting in the removal
of approximately 11 million gallons of crude oil from
the Gulf (see Figure 6-4). AEIB scientists co-authored a
journal article describing a risk assessment of the airborne
dioxin that was a by-product of these controlled burns
(Schaum et al., 2010). The paper describes a screening-
level assessment of the exposure and potential cancer
risk to nearby workers and residents potentially exposed
41
-------
Figure 6-3. Photograph of bag samplers and anemomete
Falls field site.
to dioxin via inhalation. The authors also examined the
potential for this population to be exposed to dioxins in fish
harvested for consumption. Using available information
about the size and rate of each burn, meteorology from
nearby buoys, and the NOAA North American meso-
meteorological model, scientists determined that the
worst case concentrations occurred within 50 m of burns
when wind speeds were at their highest (10 m per s).
These dispersion modeling results and the limited on-
site monitoring provided critical input to the dioxin risk
assessment. The estimated cancer risks from the predicted
air concentrations and deposition were found to be lower
than the EPA's typical range when consideration is given
to require additional actions, such as cleanup or the
establishment of regulatory policy.
Accountability Indicators Project
In its mission to protect human health and the environment,
the EPA implemented the NO Budget Trading Program
(NBP) to reduce the emissions of NOx and secondarily-
formed 0,. These pollutants and their precursors can be
transported downwind, contributing to pollutant levels
at locations much farther from the emission sources,
potentially impacting human health in downwind areas.
In this study, we investigated the health impacts in
New York State from exposure to polluted air parcels
transported from the Midwest. We developed and applied
a methodology to identify and target the transport of
polluted air parcels and demonstrated that the risk for
hospital admission because of respiratory-related illness
was increased in New York State on those days that the
air parcel originated over the Ohio River Valley, where
several high-emitting power plants are located (Figure
downwind of a simulated noise barrier at the Idaho
6-5). More specifically, we examined air parcels moving
through a boundary drawn around high-emitting power
plants in the Midwest using back-trajectories and found
significant associations with these transported air parcels
(risk estimates ranged from 1.06 to 1.16 for the entire study
time period) for six of the eight New York State regions).
The approach developed by this research can be used as an
indicator of exposure for transported air pollution.
6.4. Next Steps
Model Development
AEIB will focus on the design and development of a new
urban-scale modeling construct that is similar in nature to
AERMOD but without the regulatory structure. Instead,
the new model will address the specific needs of human
exposure and health analyses. The overall model design
will emphasize computational efficiency, user friendliness,
use of readily available inputs, and the characterization
of uncertainty, and it will make use of existing model
algorithms in existing dispersion models. Wind tunnel and
field studies will feed into the development and testing of
near-road algorithms in relation to atmospheric stability,
wind direction, and vehicle-induced turbulence.
Model Evaluation
The new dispersion model will be the air quality core
of the human exposure modeling system (including
observations and regional modeling) in the approach to
link urban sources and human exposure assessments and
human health outcomes. Evaluation of both individual
model algorithms and the overall modeling approach in
human exposure and health-related scenarios is needed.
-------
Figure 6-4. Example of a controlled bum of spilt oil in
the Gulf of Mexico.
Databases for use in model evaluation will include the
MESA AIR (Baltimore) study, the NEXUS (Detroit) study.
Urban 2003 (Oklahoma City), and the California Energy-
Commission urban tracer study (Southern California). In
addition, wind tunnel study data and the Idaho Falls tracer
study also can serve in an evaluation capacity, particularly
for individual model algorithms. Through collaboration
with U.K. scientists, the DAPPLE urban study data-base in
London also will be considered for the evaluations.
Applications of New Modeling System and
Client Collaboration
Once the new modeling system lias been developed and
tested, inclusion of it or its components will be considered
in the HEASD's Community and Environmental Justice
modeling effort (i.e., C-FERST/STREETS software).
Additionally the new dispersion model will be used to
support the design and interpretation of the air quality and
health study planned for the Research Triangle Park, area
in 2013 or 2014. Collaboration with the Federal Aviation
Administration on the development of specific applications
at airports will be initiated because of the significance of
these urban source areas. Finally, collaborations with the
Office of Transportation and Air Quality and the FHA
relative to characterization of mobile source emissions and
the FHA/EPA field studies will continue.
A large part of AEIB's local-scale modeling application
efforts will be focused on characterizing the concentration
distributions related to the NEXUS research project
currently underway to characterize the impact of mobile
sources on near-road air quality and exposures for children
with persistent asthma who live near major roadways
in Detroit. Exposure metrics developed in this project
will be coupled with the Childhood Health Effects from
Roadway and Urban Pollutant Burden Study (CHERUBS).
Modeled and monitored air quality and exposure data
will be coupled with assessments of respiratory effects
to investigate the relationships between traffic-related
exposures and observed health effects. Air quality
modeling is being conducted with the current AERMOD
dispersion model for a screening level analysis. AERLINE
and the new dispersion model will be used in a refined
analysis once actual measurements of traffic and air quality-
are available from the study.
Synthesis Analysis
Various levels of air pollution characterization (e.g.,
statistical combination of observed and modeled values,
hybrid modeling approaches) will be provided for several
tiered exposure studies, including the Cooperative
Agreements with the University of Washington. Rutgers
University, and Emory University, and the collaborative
effort with the New York State Department of Health. A
synthesis analysis will be conducted to summarize the
results from the tiered studies described above, as well
as other studies, to dociunent the findings in a systematic
way. The analysis will identify the spatial and temporal
resolution needed for the various human exposure and
health study designs commonly used, the characteristics
of the various air pollutants of concern (e.g., variability,
reactivity), and our current knowledge regarding the state-
of-science estimates that currently can be provided and the
uncertainty surrounding these estimates.
Demonstration Project
The concepts summarized and examined in the synthesis
analysis will be incorporated into a dispersion model
being developed specifically for use in human exposure
and health studies. For example, pollutants that must
be characterized at finer spatial resolutions also might
need simple chemistry (e.g., NO,), as well as line-source
capability. Thus, such approaches will be integrated into
the new model and applied in studies such as NEXUS
through C-FERST. C-FERST is a framework developed
for use by communities and individuals to estimate human
exposure to toxic substances and to identify and prioritize
risks from these substances.
. 1,440
— Til in t~r
Figure 6-5. Graphic on left shows high NO. emissions
in Midwest (July 1997) and graphic on right shows
transport of pollution from Midwest into New York
State. Red squares are locations from which back-
trajectories were initiated.
43
-------
-------
7.0
Linking Air Quality and Ecosystem
7.1 Introduction
EPA is charged with protecting the health and welfare of
the Nation. The CAA definition of welfare effects includes,
but is not limited to, effects on soils, water, wildlife,
vegetation, visibility, weather, and climate. Terrestrial and
aquatic ecosystems represent an integration of all these
factors. Atmospheric deposition is an important nonpoint
source of pollution impacting ecosystems. Deposition of
both nitrogen and sulfur is the main source of acidification
of fresh water and terrestrial ecosystems. Atmospheric
deposition of nitrogen is a contributor to nitrogen-
nutrient enrichment affecting Western ecosystems and
eutrophication affecting estuaries. From 15% to 40% of the
nitrogen load to coastal estuaries is estimated to come from
atmospheric deposition. Atmospheric deposition is a major
source of mercury that is methylated and subsequently bio-
accumulated in food chains, affecting wildlife and humans.
AMAD's goal for air-ecosystem linkage is the consistent
interfacing of weather, climate and air quality models with
aquatic and terrestrial ecosystem models to provide the
local atmosphere-biogeochemical drivers of ecosystem
exposure and resultant effects. A goal is also to harmoni/c
the connection of the local ecosystem scale (tens of square
kilometers) with the regional airshed scale (thousands to
millions of square kilometers). The physically consistent
linkage of atmospheric deposition and exposure with
aquatic/watershed and terrestrial models is central, has
not received adequate attention to date, and needs further
development.
Achieving these goals requires a coordinated research
strategy covering a variety of activities to best use
AMAD's limited resources. The strategy involves:
(1) improvements in CMAQ's bi-directional and uni-
directional air-surface exchange characterizations,
(2) improvements in the WRF/CMAQ land surface
characterizations and air-surface exchange in complex
terrain. (3) improvements in precipitation fields. (4) linkage
of hydrologic models with WRF/CMAQ to provide a
consistent connection between precipitation and hydrology
for a system-consistent linkage to ecosystem models, and
(5) application of expanded model capability to support
assessments of management scenarios The research is
organized into three interrelated tasks.
1. Air Deposition and Ecosystem Services Assessments
2. Air-Ecosystem Linkage Studies
3. Model and Tool Development
Relevance. The National Research Council (NRC) and the
Clean Air Act Advisory Committee (CAA AC) have urged
EPA to pay increased attention to ecosystem protection
and develop its capacity in this direction. In response,
the Agency has undertaken a joint review of the existing
secondary (welfare-based) NAAQS for O,. NOx and sulfur
oxides (SOX). This is the first time secondary standards
have been addressed separately from the primary standards.
The secondary NAAQS review is focused on ecological
effects from exposure to O, concentrations and from
atmospheric deposition associated with acidification sulfur
and nitrogen and nutrient enrichment nitrogen. Because
the CAA stipulates that the NAAQS are for ambient air
concentrations, not deposition, the air quality models serve
a critical function to connect air quality and deposition
to support the secondary NAAQS. In addition, the NRC
and CAAAC recommended that EPA explore the use of
critical loads in the development of the secondary NAAQS.
The Office of Air Programs (OAP) is exploring the use of
critical loads (associated with atmospheric deposition) as
a management tool. In addition, a niultiagcncy group on
critical loads as an approach to ecosystem management
for land management agencies has been established. The
agencies arc the National Park Service (NPS), the United
States Forest Service (USFS), USGS. and EPA. Modeling
of critical loads by these agencies is focusing in the
eastern United States on acidification and is focusing on
acidification in the eastern United States and on nitrogen
nutrient enrichment nationally, but with a focus on the
western United States on nitrogen nutrient enrichment.
The climate change research program at EPA has begun to
be equally concerned about ecosystem response, as well
as air quality and human health. The goal is to identify
and evaluate mitigation and adaptation strategies that will
sustain or improve current ecosystem functioning and
provision of services. This requires that changes in climate
and land use arc expressed in ways that can be translated
by models to changes in such ecosystem functions and
services. EPAs climate change program also puts a priority
on assessing and predicting water quality, not just water
quantity. Using models to investigate alternative future
scenarios facilitates comparing and contrasting the effects
of land use cliange with the effects of climate cliange
on air-surface exchange of nutrients and pollutants and.
ultimately, ecosystem health.
In parallel. EPAs ESRP has a focus on ecosystem services
and their relationship to human health and well-being.
The overarching ESRP research questions arc; What arc
the effects of multiple stressors on ecosystem services,
at multiple scales, over time; and What is the impact of
different plausible changes in these services on human
well-being and on the value of the services. Land use
change and climate change arc considered to be two
45
-------
important drivers to be addressed in the context of
changing ecosystem sendees. Tansdisciplinary research
programs that explore interactions between air, water,
and land are central to addressing these questions and
developing the science and modeling tools necessary
to simulate alternative future scenarios and mitigation
strategies.
The objective of the ecology theme research is to advance
the development and application of state-of-the-scienee
atmospheric tools to determine the exposure of ecosystem
receptors to atmospheric pollution causing and contributing
to ecological and ecosystem sendee degradation. The intent
is to provide an improved ability to connect forecasts of
the effect of air management actions on future atmospheric
deposition to ecosystem models. The air-ecosystem
research within AMAD aims to provide better atmospheric
connections for three major ecosystem assessment areas
described below.
1. Atmospheric deposition that better supports SOx-NOx
welfare standards and critical load assessments and
determinations
2. Air-ecosystem exposure linkage of atmospheric
stressors to ecosystem models to support ecosystem
assessments coupled with air management drivers,
including the assessment of ecosystem services
3. Atmospheric/meteorological-ecosystem exposure
linkage to support ecosystem impact assessment of
climate change
7.2 Air Deposition and Ecosystem
Services Assessments
7.2.1 Introduction
The thrust of this research is to define, develop and
investigate targeted CMAQ linkage applications for
study and learning and to develop guidance on how to
best establish chemical and meteorological linkages with
ecosystem models, including improvement of linkages
to address climate impacts on ecosystems. Three major
areas of research are (1) connection of WRF precipitation
to consistent surface hydrology. (2) linkage to water
quality/watershed models, and (3) linkage to ecosystem
welfare and critical load models. Target ecological end
points include fresh water aquatic, estuarine aquatic, and
terrestrial systems. A key focal point is to develop a model
representation of the hydrologic overland flow with other
agency partners that is consistent with the precipitation
from WRF/CMAQ. This focal point addresses a critical
issue identified by the most recent linkage research.
The thrust of linkage to water quality model research
is to address air-water model cross-media (linkage/
coupling) connection issues and develop approaches to
reconcile space-time mismatches for management support.
Development of guidance for critical load estimation is the
third part of the research thrust. Accuracy requirements for
effective model-to-model communication are explored and
defined and approaches are developed to more effectively
provide atmospheric deposition to ecosystem models for
stressor relationships for critical load management support.
7.2.2 Research Direction
The objectives of this research are: (1) to develop a linkage
between meteorological-model-produced precipitation and
associated modeled surface hydrology to address air-water
paradigm mismatches and support ecosystem exposure
and loading estimates under climate change. (2) to actively
use the CMAQ model to research and explore climate
and air-ecosystem linkage issues to develop guidance
on how to improve the linkage, (3) to illuminate model
functioning and examine areas of model uncertainty to
identify areas for further model development, and (4)
to improve CMAQ and tool development and inform
application approaches in support of multimedia modeling
and ecosystem management. Primary attention is focused
on the elements of the hydrological cycle that impact
ecosystem loading and exposure to atmospheric pollutants.
Two focal areas are the NOx-SOx air quality standard
review and the study of critical loads (deposition-based
tipping points). Aspects of weather and climate beyond
that can combine with the hydrological cycle to impact
ecosystem loading and exposure to atmospheric pollutants
also will be explored. The lessons are to be synthesized
to inform research planning for linkage research, CMAQ
and ecosystem model development (including ecosystem
services assessment), and integrated multimedia model
development and application.
Base {solid kre) S0?v3 (dash line) Bi-directional (circles)
1
Advection
0.8
0.6
Total
0 4 Dm,
0.2
SO.Vrf
0
100
200
300
400
500
Distance (km)
Figure 7-1. Range of influence of NH3 emissions from
a single, isolated Sampson County, NC cell in 2002
CMAQ simulations using unidirectional, the deposition
velocity as a surrogate, and bidirectional NH3 surface
exchange parameterizations, Dennis et al. (2010)
-------
N = 146 N = 159 N = 164 N = 161 N=159 N=158 N = 163 N = 1SS N=170 N=I64 N=167 N = 160
~ Base Case
~ Bidi Case
~t
00
0
i
I
m
00
nl
QQ
uu
0
I
01
Jan
Feb
Mar Apr May
Jun
Jul
Aug Sep Oct Nov Dec
Figure 7-2. Monthly NH4 wet deposition bias when compared to NADP deposition for an annual 2002 simulation for
a base case unidirectional exchange and a bidi case with bidirectional NH3 exchange. The red diamond in the box plot
represents the mean monthly deposition bias. Monthly adjustments to precipitation biases have been applied where
precipitation errors significantly correlate with deposition error; a star above the box plot indicates that this correction
was not applied.
7.2.3 Accomplishments
CMAQ sensitivity runs for Chesapeake Bay with the new
DDM-3D capability for NOy deposition were completed
for an emissions sector analysis and a six-Bay-State
relative-contribution analysis for the 2020 CAIR scenario.
The Chesapeake Bay Modeling Subcommittee was briefed
on the results. Using the DDM-3D corrected earlier
problems with the brute-force approach.
A new NH, budget analysis at 12 km, using a prototype
CMAQ with NH, bi-directional air-surface exchange
incorporated, showed that incorporating bi-directional
exchange of NH, will have an important impact of
reducing the local dry deposition and an impact on the
estimates of the range of influence of NH, emissions,
almost doubling the range (Figure 7-1). These results
underline the importance of incorporating NH,
bi-directional exchange in CMAQ (Dennis et al., 2010).
WRF meteorology for 12-km CONUS domain for 2002
was completed in FY 10 for ESRP and general use.
Preliminary annual CMAQ runs with inline lightning NOx
algorithm simulations were completed in FY 10 for 2002
using WRF 12-km CONUS 2002 meteorology.
A pilot simulation of the bi-directional CMAQ was
completed to show proof of capability to meet the FY11
deliverable to the Nitrogen Team of bidirectional CMAQ
total nitrogen deposition fields. Preliminary evaluation
against limited ambient NH3 observations and NADP wet
deposition estimates suggests that the changes from the
unidirectional CMAQ are reasonable and result in reduced
CMAQ simulation bias (Figure 7-2).
Preliminary estimates of county-level nitrogen-fertilizer
input uncertainty, expressed as a coefficient of variation
and preliminary FEST-C denitrification estimates were
shared with Nitrogen Team members assembling an
N-input meta-database.
7.2.4 Next Steps
The objectives of future research in multimedia
modeling for air-ecosystem linkages and ecosystem
service assessments are: (1) to study and improve the
connection between meteorological and atmospheric
models and ecosystem models to address air-water
paradigm mismatches; (2) to actively use the CMAQ
model to research and explore climate and air-ecosystem
linkage issues to develop guidance on how to improve
transdisciplinary modeling; (3) to illuminate model
functioning and examine areas of model uncertainty
to identify areas for further model development to
inform future research focusing on the development and
evaluation of future modeling techniques and tools for
CMAQ; and (4) to improve the capacity of CMAQ to
address drivers of change of particular interest to the
ESRP assessments of ecosystem services, such as land use
and land cover change in response to population growth
or bio-energy policy and climate change in support of
transdisciplinary modeling and ecosystem management.
Two focal areas are the NOx-SOx secondary air quality
standard review and the study of critical loads (deposition-
based tipping points). The lessons are to be synthesized
to inform research planning for CMAQ and ecosystem
model development and for integrated transdisciplinary
47
-------
model development and application. The final objective is
to provide atmospheric inputs (e.g., wet and dry deposition
of nitrogen and ambient particulate concentrations) to the
ESRP national and place-based research teams to support
overall ecosystem service assessment activities.
7.3 Air-Ecosystem Linkage Studies
7.3.1 introduction
The thrust of this research is to develop CMAQ modeling
research and applications in multimedia modeling
assessments associated with ESRP. using CMAQ as
a laboratory. The ESRP teams with which AMAD is
interacting most arc: the Future Midwest Landscapes
(FML) place-based team, the Nitrogen Team involving
national and regional assessments, the Mapping Team
involving provision of deposition-related data layers,
and the Coastal Carolinas place-based team. Provision of
CMAQ physical and chemical outputs to link to ecosystem
models for scenarios representing current and future
conditions arc major objectives.
7.3.2 Research Direction
An underlying objective is to identify and to elucidate
atmosphere-ecosystem services of relevance to the ESRP.
in collaboration with O AQPS. The role of these services is
then communicated to the broader ecosystem community
through participation in national and place-based ESRP
teams, research planning committees, and peer reviewed
publications. A fundamental objective is to assess the
relevance and capacity of CMAQ to address and respond
to drivers of change of particular interest to ESRP. such as
land use and land cover change in response to population
growth or bio-energy policy and climate change. Coupled
with this is the objective to provide atmospheric inputs
(e.g., wet and dry deposition of nitrogen and ambient
particulate) concentrations to the ESRP national and place-
based research teams to support overall ecosystem service
assessment activities. A final objective is to develop the
capacity of CMAQ to address secondary welfare NAAQS
for SO, and NOx.
7.3.3 Accomplishments
Activities in FY 10 focused on continued development
of the APVVS implementation plan and coordination
with watershed model development for the Cape Fear
basin. Planning focused on extending the coordination
of air-water linkage testing with air-uatcr/tcrrestrial
research being performed for the Cape Fear basin with the
Ecosystem Research Division (ERD).
A beta version of FEST-C was completed and was
transferred to the ESRP Nitrogen, mapping, and FML
Team scientists (Cooter et al., 2010). We continued to
interact with FML SPARROW, SWAT, and BMP modelers
regarding their use of CMAQ deposition and FEST-C
Nitrogen budget estimates. We participated in regular
research planning conference calls and coordinated
FEST-C output plans with FML modelers.
The FML was identified by AMAD as a pilot for improved
cross-divisional cooperation. A NERL pool postdoctoral
position was developed with ESD to explore the more
effective use of their planned and on-going field Nitrogen
and landu.sc/landcovcr research in CMAQ model
development and evaluation for the ESRP.
Chesapeake Bay place-based research activities were
aimed at contributing to an overall understanding of
Chesapeake Bay nitrogen issues for development of a Bay
implementation plan.
7.3.4 Next Steps
Several research topics arc planned for the future research
in multimedia modeling for air-ecosystem linkages and
ecosystem service assessments organized under the three
headings below.
1. Multimedia Linkage and Transdisciplinary Studies
- Linking to water quality and terrestrial models
- Linking to ecosystem welfare and tipping points
- Linking hydrology model to WRF/CMAQ system
- Linking climate, air quality and ecosystem services
2. Using CMAQ as a Laboratory for Deposition
Research
- Gulf hypoxia and regional watershed modeling
with SPARROW
- Chesapeake Bay TMDL and presidential directive
- Secondary N A AQS-drivcn research
3. Ecosystem Services Research Program
- National pollutant-driven research (Nitrogen Team)
- Place-based driven research
- FML
- Albemarle-Pamlico Watershed and
Estuary Study (APWES)
- Chesapeake Bay ESRP research
7.4 Model and Tool Development
7.4.1 Introduction
The thrust of this research area is to further develop and
advance CMAQ and supporting tools for ecosystem
applications. The research is subdivided into five key areas:
(1) to improve the parameterizations of the air-surface
exchange of atniosplieric pollutants, both bi-directional
and unidirectional for aquatic and terrestrial landscapes;
(2) to develop subgrid scale land-usc-specific deposition
estimates and to develop parameterizations of subgrid
variation in dry deposition because of complex terrain; (3)
to advance the WRF/CMAQ land surface parameterizations
and leaf area index (LAI) estimates to improve connections
to ecosystem models; (4) in coordination with the model
-------
development team, to improve the modeling of the amount,
location, and duration of precipitation by WRF to reduce
the error in wet deposition estimates from CMAQ as well
as to extend WRF precipitation predictions to finer grid
scales in complex terrain; and. finally. (5) to develop and
advance tools to facilitate the linkage of CMAQ
with ecosystem.
7.4.2 Research Direction
A major objective is to improve unidirectional and
bi-directional exchange process descriptions in CMAQ,
with an emphasis on incorporation of bi-directional
exchange algorithms in CMAQ for NH, and mercury. This
objective also includes evaluation of CMAQ deposition
predictions to test improvements; to suggest further model
improvements, including the search for missing pathways
in CMAQ; and to reduce uncertainty in deposition
calculations. A second major objective is also to develop
subgrid-scale approaches to better represent dry deposition
variability within a grid. The third major objective is
to advance the land surface paranictcri/ations used by
WRF/CMAQ. A final major objective is identification,
development, and enhancement of air-surface interfaces
and tools needed to facilitate the use of CMAQ deposition
outputs by analysts and researchers not familiar with the
atmospheric deposition model and enable CMAQ to better
support ecosystem studies and support CMAQ process
enhancements.
7.4.3 Accomplishments
To improve the understanding of bidirectional exchange
processes. AMAD collaborated with field scientists
from EPA/NRMRL and the NO A A ARL in planning and
conducting a field campaign where NH, surface exchange
was measured over corn, unfertilized grassland, and forest
canopies. Data from the field studies have been used in
several model development efforts, during including
first-order and siniplificd-first-order closure models for
flux from a vegetative canopy (Bash et al.. 2010) and a
process-based soil NH, emissions model (Cooter et al..
2010). A bi-directional flux pilot study for the eastern
United States was conducted by Robin Dennis. Ellen
Cooter. Megan Gore (NCSU graduate student). Viney
Aiieja (graduate students supervising professor), and Jon
Plcini. The purpose of the pilot is to facilitate the step-wise
implementation of the new bi-directional flux algorithms
developed and constrained by observations collected in
collaboration with NRMRL and NO A A into CMAQ. As
a first step, the new canopy model and emission potential
from the soil emission model were incorporated into
CMAQ 4.7.1. Testing and evaluation began in 2010, and
several manuscripts arc being prepared for submission in
2011.
The CMAQ deposition model was changed to enable
mercury-specific deposition processes. Consideration of
the nicsophyll resistance was added, which is important
for modeling elemental mercury. For reactive gaseous
mercury (RGM), the chemical parameters were changed
to be consistent with HgC,,. Previously. RGM had been
modeled using HNO, as a surrogate. This enables CMAQ
to parameterize the air-surface water concentration
gradient dependence on the elemental mercury flux
(Foley et al.. 2010), the recycling of recently deposited
divalent mercury, and the enrichment of elemental mercury
concentrations in vegetation, soil, and surface waters tliat
have been documented recently in the literature (Bash
and Miller 2009). Tliese improvements capture pulses
of emissions observed following precipitation in arid
environments observed during measurement campaigns
and arc an inline estimation of natural mercury emissions
that remains consistent with changes in modeled chemical
mechanisms and boundary conditions, clianges that would
require updates to offline natural emission estimates
(Bash 2010).
AMAD scientists arc collaborating with USD A to build
CMAQ modeling capability to simulate pesticide transport
in a project that was initiated in FY 10. CMAQ with
pesticide fate and transport capability will be evaluated
with special observations collected by the U.S. Department
of Agriculture (USDA). A coordination meeting was held
in March 9, 2010, in Research Triangle Park among the
USDA Principal Investigators (Cathlccn Hapenian and
Laura McConncll). the USDA postdoctoral fellow assigned
to the project (Cody Howard), and interested AMAD
scientists. The first phase of the project was to design a
field study to be conducted during spring/summer 2011.
Dr. Howard was provided Pesticide Emissions Model
(PEM) code by AMAD which he is now using on USDA
machines. He completed both the SMOKE and CMAQ
training offered by CM AS. Subsequent to training, he
requested and was provided AAA connection to the HPCC
to facilitate his continued exploration of CMAQ.
Unidirectional Exchange
In FY09, as part of a collaborative effort with NRMRL,
deposition velocity estimates from CMAQ for ultrafine
particles were compared against field measurements taken
over a loblolly pine forest stand in FY06 and FY07. The
results were presented at the 2009 NADP meeting.
In FY09-10, staff collaborated with scientists from the
U nitcd States and Canada on a study of sulfur budgets of
watersheds in Southeastern Canada and the Northeastern
Unitcd States Wet deposition is measured at these sites,
but dry deposition estimates arc not available. A number
of approaches for estimating the contribution of dry
deposition were investigated.
The wet deposition results and ambient NH, concentrations
from CMAQ version 4.7.1, with the changes made during
implementation of the NH, bi-directional cxcliange pilot
project were evaluated against network observations.
49
-------
Grain Corn Managment Schedule
3.5
3
2.5
2
1.5
0.5
0
^ >$> & ^ K
-------
During FY09, 2002 USD A county-level crop data were
assembled and provided to the contractor (UNC) for
update of BEIS agricultural crop fractions. This work
was a collaborative effort with the emissions modeling
team. Results of the updated crop information have been
QA'd and accepted. During FY 10, the updated BELD
agricultural data was used to distribute crop area in the
beta version of the FEST-C tool, and issues were identified
that will be resolved during FY 11. Work plan delays have
pushed completion of the BELD update (BELD 4.0)
into FY11.
Watershed Deposition Tool
The Watershed Deposition Tool (WDT) was developed
to provide an easy to use tool for mapping the deposition
estimates from CMAQ to watersheds to provide the linkage
between air and water needed for TMDL (total maximum
daily load) and related nonpoint-source watershed
analyses. This software tool takes gridded atmospheric
deposition estimates from NOAA/EPA's regional,
multipollutant air quality model, CMAQ, and allocates
them to 8-digit HUCs (hydrologic cataloging units of rivers
and streams) within a watershed. State, or region. WDT
was delivered to AMAD in FY07 and released to the public
via the AMAD Web site at http://www.epa.gov/asmdnerl/
EcoExposure/depositionMapping.html. The software and
sample CMAQ data files can be downloaded from the Web
site. In FY08. the WDT was updated and continues to be
used widely by the community to map CMAQ deposition
estimates to watershed delineations, (Figure 7-4). The
WDT was developed by Argonne National Laboratory
and contains proprietary code. In keeping with EPAs
efforts to provide open source products to the extent
possible, the capabilities of the WDT were migrated into
the Visualization Environment for Rich Data Interpretation
Figure 7-4. Watershed deposition tool output showing
the average (per unit area) annual total (wet+dry)
oxidized nitrogen deposition (kg-H/ha) estimated for
each 12-digit HUC in the Albemarle-Pamlico Basin for
2002.
(VERDI) during FY08 and FY09. A description of the
WDT and example application of the tool was published in
Schwede et al. (2009). In FY10, files for 2002-2006 from
the CDC PHASE study were made available for download
and use with the WDT
Visualization Environment for Rich Data
interpretation
In FY07, the initial version of VERDI was delivered to
AMAD. VERDI is an open source Java tool for visualizing
CMAQ and other environmental data. It is designed as
an update/replacement for the Package for Analysis and
Visualization of Environmental Data (PAVE), which
currently is used for visualizing CMAQ data. An initial
version of the VERDI was released formally to the public
in FY08 for use in visualizing CMAQ output. The software
and associated documentation are available from littp://
verdi-tool.org. The distribution of the tool is managed by
CMAS. In FY08-09, additional capabilities were added
to VERDI, including an alternate tile plot routine, an
areal interpolation plot that provides the capability of
the WDT, and the ability to display CMAQ data in polar
stereograpliic and lat-long projections. A demonstration of
this enhanced version of VERDI was given at the 2009
CMAS meeting. Numerous enhancements were made
to VERDI in FY10, including improving the scripting
capability and overlay capabilities, adding a remote file
browser, and adding the capability to export shapefiles. A
poster describing the enhancements was presented at the
2010 CMAS meeting.
Spatial Allocator
The Spatial Allocator tool was revised to include raster
tools for converting NLCD data to the gridded format
required for CMAQ modeling and to process data produced
by the FEST-C tool.
7.4.4 Next Steps
The objectives in air-surface exchange model and tool
development is to further improve and advance CMAQ
and supporting tools for ecosystem applications. The
research is subdivided into six key areas: (1) to improve
the parameterizations of the air-surface exchange
of atmospheric pollutants, both bi-directional and
unidirectional for aquatic and terrestrial landscapes;
(2) to develop subgrid-scale land-use-specific deposition
estimates, to develop parameterizations of subgrid
variation in dry deposition because of complex terrain,
and to develop techniques to account for deposition of
pollutants via cloud water deposition; (3) to advance and
modernize the WRF/CMAQ land surface parameterizations
to improve connections to ecosystem models; (4) to
develop a model representation of the hydrologic overland
How, with other Agency partners that is consistent with the
precipitation from WRF/CMAQ; (5) in coordination with
the model development team, to improve the modeling of
amount, location, and duration of precipitation by WRF to
51
-------
reduce the error in wet deposition estimates from CMAQ
as well as to extend WRF precipitation predictions to finer
grid scales in complex terrain; and. finally, (6) to develop
and advance tools to facilitate the linkage of CMAQ, with
ecosystem models based on lessons learned from the model
applications.
Several research topics are planned for the future research
in the air-surface exchange and model tool development
organized under the two headings below.
1. Air-Surface Exchange
- Bi-directional exchange
oNH3
o Mercury
o Pesticides
- Unidirectional exchange
oNO
0S0J
- Subgrid variability
o Land use variability
o Terrain variability and occult deposition
2. Data and Tool Development
- Fertilizer Emissions Scenario Tool for
CMAQ (FEST-C)
- Watershed Deposition Tool (VVDT)
- Visualization Environment for Rich Data
Interpretation (VERDI)
52
-------
References
Abdul-Razzak, H.. and S. Ghan. 2000. A parameterization
of aerosol activation. 2. Multiple aerosol
types. J. Geophys. Res., 105(D6), 6837-6844,
doi: 10.1029/1999JD901161.
Appcl. K.W., Gilliam. R.C.. Davis. N., andZubrow, A.:
Overview of the Atmospheric Model Evaluation Tool
(AMET) vl.l for evaluating meteorological and air
quality models. Environ. Modell. Softw., 26, 4, 434-443,
2011.
Appcl. K.W., P.V. Bhavc, A.B. Gilliland, G. Sarvvar, and
S.J.Roscllc. Evaluation of the community multiscalc
air quality (CMAQ) model version 4.5: Sensitivities
impacing model performance; Part 11 - particulate matter.
Atmospheric ENvironment, 42(24): 6057-066, (2008).
Araujo. R.. R.S. Dyer. R.C. Fortniann. F.A. Fulk, F.S.
Hauchnian, D.T. Hcggcni. S.T. Rao. M R. Rodgcrs. L.S.
Sheldon and E.J. Weber. A Conceptual Framework for
U.S. EPAs National Exposure Research Laboratory.
U.S. Environmental Protection Agency. Washington.
DC. EPA/600/R-09/003 (NTIS PB2010-103951) (2009).
Bash. J O. 2010. Description and initial simulation of a
dynamic bi-directional air-surface exchange model for
mercury in CMAQ. J. Geophys. Res.. Vol 115. Issue
D06305, doi: 10.1029/2009/JD012834.
Bash, J.O, J.T. Walker, G.G. Katul, M.R. Jones,
E. Nemitz, and W. Robarge. 2010. Estimation of
in-canopy ammonia sources and sinks in a ZeaMays
Canopy. Environmental Science and Technology,
44:1683-1689.
Bash. J.O. and D R. Miller; 2009. Growing season total
gaseous mercury (TGM) flux measurements over an
Acer nib rum L. stand. Atmospheric Environment.
43: 5953-5961.
Bhavc. P. V., Pouliot. G. A., and Zheng. M. 2007.
Diagnostic model evaluation for carbonaceous PM, 5
using organic markers measured in the southeastern US.
Environ. Sci. Technol, 41(5): 1577-1583.
Bowden, J.H., T.L. Otte. and C.G. Nolte. An examination
of mean error and added variability for interior grid
nudging techniques over an annual cycle using the WRF
model for regional climate modeling (to be submitted to
J. Clim., 2011).
Chapman, E.G., W.I. Gustafson. Jr.. R.C. Easter, J.C.
Barnard. S.J. Ghan, M.S. Pckour, and J.D. Fast. 2009.
Coupling acrosol-cloud-radiativc processes in the
WRF-Chem model: Investigating the radiative impact of
elevated point sources. Atmos. Chem. Phys., 9, 945-964.
Cooter, E.J., J.O. Bash, J.T. Walker. M R. Jones, and
W. Robarge. 2010. Estimation of NH3 bi-directional
flux from managed agricultural soils. Atmospheric
Environment, 44:2 107-2 115, doi: 10.1016/j.
atmosenv.2010.02.044.
Dennis. R.L., R. Matliur. J. Plcini. and J. Walker. 2010.
Fate and transport of ammonia at the local to regional
scale as simulated by the Community Multiscalc Air
Quality model. Atmospheric Pollution Research,
1. 207-214.
Finn. D.. K. L. Clawson, R. G. Carter. J. D. Rich. R. M.
Ecknian. S. G. Perry, V. Isakov, and D. Heist. 2010.
Tracer Studies to Characterize the Effects of Roadside
Noise Barriers on Near-Road Pollutant Dispersion
under Varying Atmospheric Stability Conditions. Atmos.
Environ., 44(2):204-214.
Foley, K. M„ S. J. Roscllc. K. W. Appcl. P. V. Bhavc, J. E.
Plcini. T. L. Otte. R. Matliur, G. Sarwar, J. O. Young. R.
C. Gilliam. C. G. Nolte, J. T. Kelly, A. B. Gilliland. and
J. O. Bash. 2010. Incremental testing of the Community
Multiscalc Air Quality (CMAQ) mode ling system
version 4.7. Gcosci. Model Dev., 3, 205-226.
Gcgo, E., PS Porter, A. Gilliland. S.T. Rao. Observation-
based assessment of the impact of nitrogen oxides
emissions reductions on O, air quality over the eastern
United States. Journal of Applied Meteorology and
Climatology. 46, 994-1008, 2007.
Gilliland. A. B.. Hogrcfc, C, Pindcr. R. W., Godowitch, J.
M., Foley, K. L., Rao. S. T. (2008). Dynamic evaluation
of regional air quality models: Assessing changes in 0-3
stemming from changes in emissions and meteorology.
Atmos. Environ. 42(20): 5110-5123.
Godowitch, J.M., and S.T. Rao. Assessing multi-year
changes in modeled and observed maximum 8-hour
o/onc with a dynamic evaluation approach. 12th
Conference on Atniosplieric Chemistry. Paper #163, on-
line at http://anis.confcx.coni/anis/htscarch.cgi. 2010.
Godowitch, J. M„ Pouliot. G. A., Rao, S. T. (2010).
Assessing multi-year changes in modeled and observed
urban NOX concentrations from a dynamic model
evaluation perspective. Atmos. Environ. 44(24): 2894-
2901. doi:1016/j.atmosenv.2010.04.040
Golden. H E., C D. Knightcs. E.J. Cooter. R.L. Dennis,
R.C. Gilliam. K.M. Foley, 2010. Linking air quality
and watershed models for environmental assessments:
Analysis of the effects of model-specific precipitation
estimates on calculated water flux, Environmental
Modelling & Software, 25, 1722-1737.
53
-------
Grell ct al.. 2011: WRF/Chem Version 3.3 User's Guide.
http://ruc.noaa.gov/wrl7WG 11/Users_guide.pdf
Heist. D.. S. Perry, L. Brixev. 2009a. A Wind Tunnel
Study of the Effect of Roadway Configurations on the
Dispersion of Traffic-Related Pollution. Atmos. Environ.,
43: 5101-5111.
Heist. D.K., Perry, S C.. Brixev, L.A., 2009b. Dispersion
of Traffic-Related Pollutants in a Wind-Tunncl Model
of a Las Vegas Field Site. International Workshop on
Physical Modeling of Flow and Dispersion Phenomena.
Brussels. Belgium. Aug. 24-26.
Hoar. T., Nychka, D., 2008. Statistical downscaling of the
community climate system model (CCSM) monthly
temperature and precipitation projections. htto://www.
gisclimatechange.org/Downscaling.pdf.
lacono. M.J., IS. Delaine re. E.J. Mlawer, MVV. Shephard,
S.A. Clough. and W.D. Collins. Radiative forcing by
long-lived greenhouse gases: Calculations with the
AER radiative transfer models. J. Geophys. Res.. 113,
D13 103. doi: 10.1029/2008JD009944, 2008.
Isakov, V., J. Touma. and A. Khlystov. 2007. A method
of assessing air toxics concentrations in urban areas
using mobile platform measurements. J. Air & Waste
Manage. Assoc. 57:1286-1295, DOI: 10.3155/1047-
3289.57.11.1286.
Mathur. R., J. Plcim. D. Wong, T. Otte. R. Gilliam, S.
Rosclle. J. Young, F. Binkowski, and A. Xiu, 2009. The
WRF-CMAQ Integrated On-line Modeling System:
Development, Testing, and Initial Applications, Air
Pollution Modeling and Its Application XX, Springer,
The Netherlands, in press.
Mathur, R.. S. Yu, D. Kang. and K.L. Schere. Assessment
of the winter-time performance of developmental
particulate matter forecasts witht the Eta-CMAQ
modeling system. Journal of Geophysical Research-
Stmospheres (JGR-Atmospheres). 113(D02303): 1-15,
(2008)
Napelenok, S.L, Foley, K.M., Kang. D„ Mathur, R.. Pierce,
T.„ Rao, S.T., 2011. Dynamic evaluation of regional air
quality model's response to emission reductions in the
presence of uncertain emission inventories (in press).
Nolte, C.G., A.B. Gi Hi land, C. Hogrefe. L.J. Mickley,
2008, Linking global to regional models to assess
future climate impacts on surface o/one levels in
the United States. J. Geophys. Res.. 113, D14307.
doi: 10.1029/2007JD008497.
Paulot. F., J.D. Croun.se. H.G. Kjaergaard. J.H. Kroll. J.H.
Seinfeld. P.O. Wcnnberg, 2009, Isoprene photooxidation:
New insights into the production of acids and organic
nitrates, Atmos. Chem. Phys. 9, 4, 1479-1501.
Pierce, T„ C. Hogrefe, S.T. Rao. PS. Porter. J. Ku. 2010.
Dynamic evaluation of a regional air quality model:
assessing the emissions-induced weekly o/one cycle,
Atmos. Environ. 44:3583-3596.
Pinder. R. W„ Gilliam. R. C., Appel. K. W., Napelenok,
S. L„ Foley, K. M„ Gilliland. A. B. (2009). Efficient
Probabilistic Estimates of Surface O/one Concentration
Using an Ensemble of Model Configurations and Direct
Sensitivity Calculations. Environ. Sci. Technol. 43(7):
2388-2393.
Schaum. J., M. Cohen. S. Perry. R. Art/.. R. Draxler, J. B.
Frithsen. D. Heist. M. Lorber. and L. Phillips . 2010.
Screening level assessment of risks due to dioxin
emissions from burning oil from the BP Deepwater
Horizon Gulf of Mexico spill. Environ. Sci. Technol. 44:
9383-9389.
Schwede, D.. R. Dennis. M. Bit/. 2009. The Watershed
Deposition Tool: A tool for incorporating atmospheric
deposition in watershed analyses. Journal of the
American Water Resources Association, 45(4), 973-985.
Schmith. T„ 2008. Stationarity of regression relationships:
Application to empirical downscaling. J. Clim., 21,
4529-4537.
Skamarock. W. C, and Coauthors, 2008: A description of
the advanced research WRF version 3. NCAR Tech Note
NCAR/TN-4751STR, 125 pp. [ Available from UCAR
Communications, P.O. Box 3000, Boulder. CO 80307.]
U.S. EPA, 2009, Endangerment and Cause or Contribute
Findings for Greenhouse Gases under Section 202(a) of
the Clean Air Act. http://www.epa.gov/climatechange/
endangernient.html.
US EPA, 2007. NOx Budget Trading Program. EPA-
430-R-07-009. http://www.epa.gov/airmarkets
Venkatram, A., V. Isakov. E. Thoma, and R. Baldauf.
2007. Analysis of air quality data near roadways using a
dispersion model. Atmos. Environ., 41(40): 9481-9497.
Venkatram. A., V. Isakov, R. Seila. and R. Baldauf. 2009.
Modeling the impacts of traffic emissions on air toxics
concentrations near roadways. Atmos. Environ.. 43(20):
3191-3199.
Weaver, C., X. Liang. J. Zhu. P. Adams, P. A mar. J.C.
Avise, M. Caughey. J. Chen, R.C. Cohen, E. Cooter. J.
Dawson. R.C. Gilliam. A. Gilliland. A.H. Goldstein.
A.E. Grambsch. A. Guenther. W.I. Gustafson. R.A.
Harley. S. He. B.L. Hemming, C. Hogrefe, H. Huang. S.
Hunt. D.J. Jacob. PL. Kenny, K. Kunkel. J. Lamarque.
B. Lamb. N.K. Larkin. L.R. Leung. K. Liao. J. Lin. B.H.
Lynn, K. Manomaiphiboon, C.F. Mass. D. McKinzie,
L.J. Mickley. S. O'Neill. C.G. Nolte. S.N. Pandis. P.N.
Racherla. C. Roscnzwcig. A. Russell. E. Salathe. A.L.
Steiner. E. Tagaris. Z. Tao. S. Ton.se. C. Wiedinmver, A.
Williams, D. Winner, J. Woo. S. Wu, and D.J. Wuebbles.
A Preliminary Synthesis of Modeled Climate Change
-------
Impacts on U.S. Regional O/onc Concentrations.
Bulletin of the American Meteorological Society.
American Meteorological Society, Boston, MA,
90(12):1843-1863 (2009)
Wilby, R. L„ Dawson, C. W., and Barrow, E. M„ 2002.
SDSM - A decision support tool for the assessment of
regional climate change impacts. Environ. Modelling
Software, 17. 147-159.
Zhang, Y, 2008: Online-coupled meteorology and
chemistry models: history, current status, and outlook.
Atmos Chein and Phys.. 11. 1680-7316.
-------
-------
APPENDIX A
Atmospheric Modeling and Analysis Division Staff Roster
(as of December 31, 2010)
Office of the Director
S.T. Rao. Director
David Moblcv, Deputy Director
Patricia McGhee, Assistant to the Director
Sherry Brown
Wanda Payne (SEEP1)
Ken Schere, Science Advisor
Gary Walter, IT Manager
Emissions and Model Evaluation Branch
Tom Pierce. Chief
Wyat Appcl
Brian Eder
Kristen Foley
Jim Godowitch
Steve Howard
Roger Kwok (NRC2 Postdoctoral Fellow)
Sergey Napclenok
George Pouliot
Havala Pye (ORISE3 Postdoctoral Fellow)
Alfrcida Torian
Atmospheric Exposure Integration Branch
Val Garcia. Chief
Jesse Bash
Jason Ching
Jim Crooks (Postdoctoral Fellow)
Robin Dennis
Megan Gore (Contractor)
Vlad Isakov
Gill-Ran Jeong (Visiting Scientist)
Donna Schwede
Myrto Valari (NRC2 Postdoctoral Fellow)
David Heist. Fluid Modeling Facility
Asliok Patel (SEEP1), Fluid Modeling Facility
Steve Perry, Fluid Modeling Facility
Bill Petersen (Contractor). Fluid Modeling Facility
John Rose (SEEP1), Fluid Modeling Facility
Atmospheric Model Development Branch
Roliit Matliur. Chief
Shirley Long (SEEP1), Secretary
Prakash Bliavc
Garnet Erdakos (NRC2 Postdoctoral Fellow)
Rob Gilliam
Bill Hut/ell
Deborah Lucckcn
Martin Otte (Postdoctoral Fellow)
Shawn Roselle
Golam Sarwar
John Streicher
David Wong
Jeff Young
Shaocai Yu (Postdoctoral Fellow)
Applied Modeling Branch
Jon Pleini, Acting Chief
Mclanie Ratteray (SEEP,), Secretary
Farhan Aklitar (ORISE3 Postdoctoral Fellow)
Bill Bcnjcy
Jared Bowden (NRC2 Postdoctoral Fellow)
Russ Bullock
Dan Cohan (Visiting Scientist)
Barron Henderson (ORISE3)
Jerry Hcrwchc
Chris Nolte
Tanya Otte
Rob Pinder
Jenise Swall
Ben Wells (Contractor)
Ying Xie (NRC2 Postdoctoral Fellow)
'SEEP - Senior Environmental Employee Program
-NRC - National Research Council
'ORISE - Oak Ridge Science and Education Program
57
-------
APPENDIX B
Division and Branch Descriptions
Atmospheric Modeling Analysis Division
The Division leads the development and evaluation of
atmospheric models on all spatial and temporal scales for
assessing changes in air quality and air pollutant exposures,
as affected by changes in ecosystem management and
regulatory decisions, and for forecasting the Nation's
air quality . AMAD is responsible for providing a sound
scientific and technical basis for regulatory policies to
improve ambient air quality . The models developed
by AMAD arc being used by EPA, NOAA, and the air
pollution community in understanding and forecasting
the magnitude of the air pollution problem and also in
developing emission control policies and regulations for air
quality improvements. AMAD applies air quality models to
support key integrated, niultidisciplinary science research.
This includes linking air quality models to other models
in the source-to-outcome continuum to effectively address
issues involving human health and ecosystem exposure
science.
Atmospheric Model Development Branch
AMDB develops, tests, and refines analy tical, statistical,
and numerical models used to describe and assess
relationships between air pollutant source emissions and
resultant air quality , deposition, and pollutant exposures
to humans and ecosystems. The models arc applicable
to spatial scales ranging from local/urban and nicsoscale
through continental, including linkage with global models.
AMDB adapts and extends meteorological models to
couple effectively with chemical-transport models to create
comprehensive air quality modeling systems, including
the capability for two-way communication and feedback
between the models. The Branch conducts studies to
describe the atmospheric processes affecting the transport,
diffusion, transformation, and removal of pollutants in
and from the atmosphere using theoretical approaches,
as well as from analyses of monitoring and field study
data. AMDB converts these and other study results into
models for simulating the relevant phy sical and chemical
processes and for characterizing pollutant transport and
fate in the atmosphere. AMDB conducts model exercises
to assess the sensitivity and uncertainty associated with
model input databases and applications results. AMDB's
modeling research is designed to produce tools to serve
the Nation's need for science-based air quality decision-
support systems.
Emissions and Model Evaluation Branch
EMEB develops and applies advanced methods for
evaluating the performance of air quality simulation
models to establish their scientific credibility. Model
evaluation includes diagnostic assessments of modeled
atmospheric processes to guide the Division's research
in arcas such as land-use and land cover characterization,
emissions, meteorology, atmospheric chemistry, and
atmospheric deposition. The Branch also advances the use
of dynamic and probabilistic model evaluation techniques
to examine whether the predicted changes in air quality
arc consistent with the observations. By collaborating
with other EPA offices that provide data and algorithms
on emissions cliaractcri/ation and source apportionment
and the scientific community, the Branch evaluates the
quality of emissions used for air quality modeling and.
if warranted, develops emission algorithms that properly
reflect the effects of changing meteorological conditions.
Atmospheric Exposure Integration Branch
AEIB develops methods and tools to integrate air quality
process-based models with human liealtli and ecosystems
exposure models and studies. The three major focus arcas
of this Branch arc (1) linkage of air quality with human
exposure. (2) deposition of ambient pollutants onto
sensitive ecosystems, and (3) assessment of the impact of
air quality regulations (accountability ). AEIB's research
to link air quality to human exposure includes urban-scale
modeling, atmospheric dispersion studies, and support of
exposure field studies and epidemiological studies. The
urban-scale modeling program (which includes collection
and integration of experimental data from its Fluid
Modeling Facility ) is focused on building "hot-spot" air
toxic analy sis algorithms and linkages to human exposure
models. The deposition research program develops tools
for assessing nutrient loadings and ecosystem vulnerability,
and the accountability program develops techniques to
evaluate the impact of the regulatory strategics that have
been implemented on air quality and conducts research to
link emissions and ambient pollutant concentrations with
exposure and human and ecological health end points.
Applied Modeling Branch
AMB uses atmosplieric mode ling tools to address emerging
issues related to air quality and atmospheric influences
on ecosystems. Climate change, growing demand for
biofucls. emission control programs, and growth all affect
air quality and ecosystems in various ways that require
integrated assessment. Fundamental to these studies is the
development of credible scenarios of current and future
conditions on a regional scale and careful consideration
of global-scale influences to air pollution and climate.
Scenarios of climate, growth and development, and
regulations will be used with regional atmospheric models
to investigate potential changes in exposure risks related to
air quality and meteorological conditions.
58
-------
APPENDIX C
2010 Awards and Recognition
EPA Gold Medal
William Bcnjcy, Ellen Cootcr. Alice Gilland. and
Robert Gilliam—Climate Change and Air Quality
Assessment Team
ORD Bronze Medal
VIad Isakov. Joe Touma and others—Cummulative Air
Accountability Team
ORD Bronze Medal
Prakash Bhavc. Ann Carlton, Rohit Mathur. Sergey
Napelenok. Robert Pinder, George Pouliot. Golam Sarwar,
and others-Organic Aerosol Science Team
ORD Technical Organic Assistance to the Regions or
Program Offices Award
Kenneth Schere, Jonathan Plcim. Rohit Mathur. George
Pouliot. Rogcrt Gilliam, and Thomas Pierce—Fairbanks
Modeling Team
ORD Diversity Leadership Award
Prakash Bhavc—For recognition of his various approaches
to model diversity ideals and for his active support of
diversity programs
NERL Special Achievement Award
Wyat Appel and Robert Gilliam—Goal 1, Support the
Agency "s Mission: Developing the Atmospheric Model
Evaluation Tool (AMET) and overseeing its successful
transfer to the meteorological and air quality modeling
community
S.T. Rao and David Mob ley—Goal 2, Be a High
Performing Organization: Efforts to improve integration
and coordination of research across HEASD
Jonathan Plcim—Goal 3, Be a Leader in the Environmental
Research Community: Demonstrating excellence in
advancing the state of the science in atmospheric modeling
Prakash Bhavc, Ann Carlton. Deborah Lucckcn,
Rohit Mathur and Golam Sarwar-Goal 4, Integrate
Environmental Science and Technology to Solve
Environmental Problems: Recognizing NERL's
Atmospheric Chemistry and Modeling Team, who
effectively integrated research across research disciplines
and organizations
Kristen Foley—Meritorious Research Support Award:
Providing much needed statistical and programming
support toward the advancement of credible air
quality models
Ken Schere, Alice Gilliland and others—Teamwork Award:
Modeling Linkage Team
David Heist and Steven Perry—Blue Ribbon Paper
Award: The Effect of a Tall Tower on Flow and Dispersion
Through a Model Urban Neighborhood. Part 1: Flow
Characteristics and Part 2: Pollutant Dispersion
Robert Pinder. Robert Gilliam. Wyat Appel. Sergey
Napelenok, Kristen Foley, and Alice Gilliland - Blue
Ribbon Paper Award: Efficient Probabilistic Estimates of
Surface Ozone Concentration Using an Ensemble of Model
Configurations and Direct Sensitivity Calculation
Prakash Bhavc. Heather Simon. George Pouliot. and David
Mobley - Blue Ribbon Paper Award: Emissions Inventory
of PM25 Trace Elements Across the U.S.
Shawn Rosclle—Leadership Award: Demonstrating
leadership and assembly of various CMAQ versions and
for providing enthusiastic support of its use by the program
offices, contract stall, and internal division stall
Bill Hut/ell—Collaboration Award: Essential role in the
improvement and maintenance of the computer code for
the Community Multiscale Air Quality Model
AMAD Awards
Blue Ribbon Paper: David K. Heist and Steven Perry—
Effect of a Tall Tower On Flow and Dispersion Through a
Model Urban Neighborhood. Part 1: Flow Characteristics
and Part 2: Pollutant Dispersion
Blue Ribbon Paper: Robert Pinder. Robert Gilliam.
Wyat Appel. Sergey Napelenok. Kristen Foley, and Alice
Gilliland—Efficient Probalisitic Estimates of Surface
Ozone Concentration Using an Ensemble of Model
Configurations and Direct Sensitivity Calculation
Blue Ribbon Paper: Prakash V. Bhavc. Heather Simon.
George A. Pouliot. and David Mobley—Emissions
Inventory PM, 5 Trace Elements Across the Unitcd States.
Leadership Award: Shawn Rosclle—Demonstrating
leadership in coordinating the development and assembly
of various CM AQ versions and for providing enthusiastic
support of its use by the program offices, contract stall, and
internal division staff.
Collaboration Award: Bill Hutzcll. For his essential role
in the improvement and maintenance of the computer code
for the Community Multiscale Air Quality Model
59
-------
APPENDIX D
2010 Publications(Division authors are in bold.)
Journal Articles
Appel, W., S.J. Rosette, R.C. Gilliam, J.E. Pleim and
R.C. Gilliam. Sensitivity of the Community Multiscale
Air Quality (CMAQ) Model v4.7 Results for the
Eastern Unitcd States to MM5 and WRF Meteorological
Drivers. Geoscientific Model Development, Copernicus
Publications. Katlenburg-Lindau. Germany. 3(1): 169-188,
(2009).
Bao. H„ S. Yu and D.Q. Tong. Massive Volcanic SO,
Oxidation and Sulphate Aerosol Deposition in Ceno/oic
North America. Nature, Macmillan Publishers Ltd..
London. UK, 465(7300):845-974, (2010).
Bash, J.O., J.T. Walker, G.G. Katul. M.R. Jones, E.
Nemitz and W. P. Robarge. Estimation of In-Canopy
Ammonia Sources and Sinks in a Fertilised Zea mays
Field. Environmental Science & Technology, American
Chemical Society. Washington. DC. 44(5): 1683-1689,
(2010).
Bash, J.O. Description and Initial Simulation of a
Dynamic Bidirectional Air-Surface Exchange Model for
Mercury in Community Multiscale Air Quality Model.
Journal of Geophysical Research-Atmospheres, American
Geophysical Union, Washington. DC. 115(D06305):
1-15, (2010).
Carlton, A.G., R.W. Finder, P. Bhave and G. Pouliot.
To What Extent Can Biogenic SO A Be Controlled?
Environmental Science & Technology, American Chemical
Society. Washington. DC. 44(9):3201-3646, (2010).
Carlton, A.G., P. Bhave, S. Napelenok, E.O. Edney,
G. Sarwar, R.W. Finder, G. Pouliot and M. Houyoux.
Model Representation of Secondary Organic Aerosol
in CMAQ v4.7. Environmental Science & Technology,
American Chemical Society, Washington, DC.
44(221:8553-8560, (2010).
Cooter, E., J. O. Bash. J.T. Walker. Jr., M R. Jones and
W. Robarge. Estimation of NH3 Bi-Directional Flux from
Managed Agricultural Soils. Atmospheric Environment,
Elsevier Science Ltd.. New York. NY, 44( 17):2107-2115.
(2010).
Dennis, R.L., R. Mathur, J.E. Pleim and J.T. Walker. Jr.
Fate of Ammonia Emissions at the Local to Regional Scale
as Simulated by the Community Multiscale Air Quality
Model. Atmospheric Pollution Research, Turkish National
Committee for Air Pollution Research and Control. Izmir,
Turkey. 1(4):207-214, (2010).
Dennie, R.L., T. Fox, M. Fucntes. A. Gilliland, S. Hanna.
C. Hogrefe, J. Irwin, S.T. Rao, R. Scheffe, K.L. Sehere,
D. Steyn and A. Venkatram. A Framework for Evaluating
Regional-Scale Numerical Photochemical Modeling
Systems . Environmental Fluid Mechanics, Springer, New
York. NY, 10(4):471-489, (2010).
Djalalova. I., J.M. Wilczak. S. McK.een, G. Grell.
S. Pcckhani. M. Pagowski. L. DellcMonachc. J. McQueen,
P. Lee, Y. Tang. J. McHenry, W. Gong. V. Bouchet and
R. Mathur. Ensemble and Bias-Correction Techniques
for Air-Quality Model Forecasts of Surface O, and PM, 5
during the TEXAQS-II Experiment of 2006. Atmospheric
Environment, Elsevier Science Ltd.. New York. NY,
44(41:455-467, (2010).
Driscoll. C.T., E.B. Cowling, P. Grcnnfelt, J.M. Galloway
and R.L. Dennis. Integrated Assessment of Ecosystem
Effects of Atmospheric Deposition. EM: Air and Waste
Management Associations Magazine for Environmental
Managers, Air & Waste Management Association,
Pittsburgh, PA, (11):6-13, (2010).
Finn. D.. K.L. Clawson. R.G. Carter, J.D. Rich. R.M.
Ecknian. S.G. Perry, V. Isakov and D. Heist. Tracer
Studies to Characteri/e the EITects of Roadside Noise
Barriers on Near-Road Pollutant Dispersion under
Varying Atmospheric Stability Conditions. Atmospheric
Environment, Elsevier Science Ltd.. New York. NY,
44(2):204-214, (2010).
Foley, K., S.J. Roselle, W. Appel, P. Bhave, J.E. Pleim,
T.L. Otte, R. Mathur, G. Sarwar, J.O. Young, R.C.
Gilliam, C.G. Nolte, J. T. Kelly, A. Gilliland and J.O.
Bash. Incremental Testing of the Community Multiscale
Air Quality (CMAQ) Modeling System Version
4.7. Geoscientific Model Development, Copernicus
Publications. Katlenburg-Lindau. Germany. 3(l):205-226,
(2010).
Gantt, B.. N. Mcskhidzc and A.G. Carlton. The
Contribution of Marine Organics to the Air Quality of
the Western Unitcd States. Atmospheric Chemistry and
Physics, Copernicus Publications, Katlenburg-Lindau.
Germany, 10( 15):7415-7423, (2010).
Ghosh. S.K., H. Lee, J. Davis and P. Bhave. Spatio-
Teniporal Analysis of Total Nitrate Concentrations Using
Dynamic Statistical Models. Journal of the American
Statistical Association, American Statistical Association,
Alexandria. VA, 105(490):538-551, (2010).
Gilliam, R.C. and J.E. Pleim. Performance Assessment of
New Land-Surface and Planetary Boundary Layer Physics
in the WRF-ARW. Journal of Applied Meteorology and
Climatology, American Meteorological Society, Boston,
MA, 49(4):760-774, (2010).
60
-------
Godowitch, J.M., G. Pouliot and S.T. Rao. Assessing
Multi-year Changes in Modeled and Observed Urban
NOx Concentrations from a Dynamic Model Evaluation
Perspective. Atmospheric Environment, Elsevier Science
Ltd.. New York. NY. 44(24) :2894-2901, (2010).
Golden. H E., C D. Knights. E.J. Cooter, R.L. Dennis,
R.C. Gilliam, K.M. Foley. Linking Air Quality and
Watershed Models for Environmental Assessments:
Analysis of the Effects of Model-Specific Precipitation
Estimates on Calculated Water Flux. Environmental
Modelling & Software, 25:1722-1737 (2010).
Horton, R.. C. Roscn/weig. R. Ramaswamv, PL. Kinney,
R. Mathur and J.E. Pleini. Integrated Climate Change
Information for Resilient Adaptation Planning. EM:
Air and Waste Management Associations Magazine for
Environmental Managers, Air & Waste Management
Association. Pittsburgh. PA, (11): 14-25, (2010).
Johnson, M.M., V. Isakov, J. Tounia. S. Mukcrjec and
H.A. O/kavnak. Evaluation of Land Use Regression
Models Used to Predict Air Quality Concentrations in an
Urban Area. Atmospheric Environment, Elsevier Science
Ltd. New York, NY, 44(30):3660-3668, (2010).
Kang. D., R. Mathur and S.T. Rao. Real-Time Bias-
Adjusted O, and PM2 5 Air Quality Index Forecasts and
their Performance Evaluations over the Continental United
States. Atmospheric Environment, Elsevier Science Ltd.
New York. NY. 44(18):2203-2212, (2010).
Kang G. D.. R. Mathur and S.T. Rao. Assessment
of Bias-Adjusted PM Air Quality Forecasts over the
Continental United States During 2007. Geoscientific
Model Development, Copernicus Publications. Katlenburg-
Lindau. Germany. 3(l):309-320, (2010).
Kelly, J., P. Bhave. C.G. Nolte, U. Shankar and K. Foley.
Simulating Emission and Chemical Evolution of Coarse
Sea-Salt Particles in the Community Multiscale Air Quality
(CMAQ) Model. Geoscientific Model Development,
Copernicus Publications. Katlenburg-Lindau. Germany,
3(l):257-273, (2010).
Lim. C.Y, M. Stein, J.K. Ching and R. Tang. Statistical
Properties of Differences between Low and High
Resolution CMAQ Runs with Matched Initial and
Boundary Conditions. Environmental Modelling &
Software, Elsevier Science. New York, NY, 25( 1): 158-169.
(2010).
Luecken, D.J., R.L. Waterland. S. Papasav va, K.
Taddonio, W.T. Hutzell, J.P. Rugh and S. Andersen. O/one
and TFA Impacts in North America for Degradation of
2,3,3,3-Tetrafluoropropene (HFO-1234yf), a Potential
Greenhouse Ga.s Replacement. Environmental Science &
Technology, American Chemical Society. Washington. DC,
44(11:343-348, (2010).
Nielsen-Gammon, J.W., X. Hu. F. Zhang and J.E.
Pleini. Evaluation of Planetary Boundary Layer Scheme
Sensitivities for the Purpose of Parameter Estimation.
American Meteorological Society Monthly Weather
Review, American Meteorological Society. Boston. MA,
138(9):3400-3417, (2010).
Otte, T.L. and J.E. Pleini. The Meteorology-Chemistry
Interface Processor (MCIP) for the CMAQ Modeling
System: Updates through MCIP\ 3.4.1. Geoscientific
Model Development, Copernicus Publications, Katlenburg-
Lindau. Germany, 3(l):243-256, (2010).
Pierce, T.E., C. Hogrefe, S.T. Rao. P. Porter and J. Ku.
Dynamic Evaluation of a Regional Air Quality Model:
Assessing the Emissions-Induced Weekly O/one Cycle.
Atmospheric Environment, Elsevier Science Ltd, New
York, NY. 44(29):3583-3596, (2010).
Rao, S.T. and D. Mobley. Moving Toward an Integrated
Transdisciplinary Approach to Solving Environmental
Problems. EM: Air and Waste Management Associations
Magazine for Environmental Managers, Air & Waste
Management Association. Pittsburgh. PA, (11):5, (2010).
Simon, EL, L. Beck. P. Bhave, F. Divita. Y. Hsu. D.J.
Luecken, D. Mobley, G. Pouliot. A.H. Re IT. G. Sarwar
and M. Strum. The Development and Uses of EPA's
SPEC I ATE Database. Atmospheric Pollution Research,
Turkish National Committee for Air Pollution Research
and Control. Izmir, Turkey, 1(4): 196-206, (2010).
Simon, EL, Y. Kimura, G. McGaughey, D. Allen. S. S.
Brown, D. Coffman, J. Dibb. H. D. OstholT, P. Quinn, J.M.
Roberts. G. Yarwood. S. Kemball-Cook, D.W. Byun and D.
Lee. Modeling Heterogeneous CINO Formation. Chloride
Availability, and Chlorine Cycling in Southeast Texas.
Atmospheric Environment, Elsevier Science Ltd. New
York. NY, 44(40):5476-5488, (2010).
Tian, D.. D. Cohan, Y. Hu. S. Napelenok, M.E. Chang
and A. Russell. Uncertainty Analysis of O/one Formation
and Response to Emission Controls Using Higher-Order
Sensitivities. Journal of the Air & Waste Management
Association, Air & Waste Management Association,
Pittsburgh. PA, 60(7):797-804, (2010).
Yu, S., R. Mathur, G. Sarwar, D. Kang, D. Tong, G.
Pouliot and J.E. Pleini. Eta-CMAQ Air Quality Forecasts
for O, and Related Species Using Three Different
Photochemical Mechanisms (CB4, CB05, SAPRC-99):
Comparisons with Measurements During the 2004
IC ARTT Study. Atmospheric Chemistry and Physics,
Copernicus Publications, Katlenburg-Lindau. Germany,
10(6):3001-3025, (2010).
Zhang. Y, P. Liu. X. Liu. M.Z. Jacobson. PH. McMurry,
F. Yu, S. Yu and K.L. Schere. A Comparative Study of
NucleationParameterizations: 2. Thrce-Dimensional
Model Application and Evaluation. Journal of Geophysical
Research, American Geophysical Union, Washington, DC.
115(020213): 1-26, (2010).
61
-------
Book Chapters
Clawson. K.L.. D. Finn, R.G. Carter. J.D. Rich. R.M.
Eckman. S.G. Perry, V. Isakov, D. Heist and T.E. Pierce.
NOAA EPA Near-Roadway Sound Barrier Atmospheric
Tracer Study 2008. Chapter 1, Douw G. Steyn, S.T. Rao
(ed.), Air Pollution Modeling and its Applications XX .
Springer Netherlands, Netherlands. C(1.15):27-31, (2010).
Garcia, V.. E. Gego, R. Jones, S. Lin. C. Pantea. S.T. Rao
and A. Woottcn. Examining the Impact of Regional-Scale
Air Quality Regulations on Human Health Outcomes.
Chapter 7, Douw G. Steyn and S.T. Rao (ed.), Air Pollution
Modeling and its Applications XX. Springer Netherlands.
Netherlands. 7.5(C):545-548, (2010).
Godowitch, J.M., G. Pouliot and S.T. Rao. On the Use
of a Dynamic Evaluation Approach to Assess Multi-
year Change in Modeled and Observed Urban NOx
Concentrations. Chapter 4, Douw G. Steyn and S.T. Rao
(ed.), Air Pollution Modeling and Its Application XX.
Springer Netherlands, Netherlands. C(4.3):337-341,
(2009).
Kang. D„ R. Mathur and S.T. Rao. Implementation
of Real-Time Bias-Adjusted O, and PM, 5 Air Quality
Forecasts and their Performance Evaluations during 2008
over the Continental United States. Chapter 3, Douw G.
Steyn and S. T. Rao (ed.), Air Pollution Modeling and
its Applications XX. Springer Netherlands. Netherlands,
C(3.1):283-288, (2010).
Lin. S., R. Jones, C. Pantea. V. Garcia, S.T. Rao.
S. Hwang and N. Kim. Impact of the NOx SIP Call on
Respiratory Hospitalizations in New York State. Chapter 7,
Douw G. Steyn and S.T. Rao (ed.), Air Pollution Modeling
and its Application XX. Springer Netherlands. Netherlands.
7.6(0:549-552, (2010).
Mathur, R., J.E. Pleim, B.C. Wong, T.L. Otte, R.C.
Gilliam, S.J. Roselle, J.O. Young, F.S. Binkowski and
A. Xiu. The WRF-CMAQ Integrated On-Line Modeling
System: Development. Testing, and Initial Applications.
Chapter 2, Douw G. Steyn and S. T. Rao (ed.), Air
Pollution Modeling and its Applications XX. Springer
Netherlands. Netherlands. C(2.9): 155-159. (2010).
Napelenok, S., J. Arnold. K. Foley and D. K. Henze.
Regional Background Fine Particulate Matter. Chapter
2, Douw G. Steyn and S.T. Rao (ed.), Air Pollution
Modeling and its Applications XX. Springer Netherlands,
Netherlands. 2.32(C):277-280, (2010).
Pierce, T.E., D. Heist, V. Isakov, S.G. Perry, K. Clawson
and R. Eckman. Towards an Improved Characterization
of Dispersion Near Major Roadways. Chapter 1, Douw
G. Steyn and S.T. Rao (ed.), Air Pollution Modeling and
its Applications XX. Springer Netherlands. Netherlands.
C(1.16):95-98, (2010).
Pleim, J.E., R.C. Gilliam and S. Yu. Atmospheric
Boundary Layer Modeling for Combined Meteorology and
Air Quality Systems. Chapter 1, Douw G Steyn, S.T. Rao
(ed.), Air Pollution Modeling and its Applications XX.
Springer Netherlands. Netherlands. C(1.8):45-49, (2010).
Porter. S., C. Hogrefe. E. Gego. K. Foley, J.M.
Godowitch and S.T. Rao. Application of Wavelet Filters
in an Evaluation of Photochemical Model Performance.
Chapter 4, Air Pollution Modeling and its Applications
XX. Springer Netherlands. Netlierlands. 4.17(0:415-420,
(2010).
Ran. L„ J.E. Pleim and R.C. Gilliam. Impact of High
Resolution Land-Use Data in Meteorology and Air
Quality Modeling Systems. Chapter 1, Douw G. Steyn
and S. Trivikrania Rao (ed.), Air Pollution Modeling and
its Applications XX. Springer Netherlands. Netherlands.
C(l.l):3-7, (2010).
Rao, S.T., K.L. Schere, S. Galniarini and D. Steyn.
AQMEII: A New International Initiativc on Air Quality
Model Evaluation. Chapter 4, Douw G. Steyn and S.T.
Rao (ed.), Air Pollution Modeling and its Applications
XX. Springer Netherlands. Netlierlands, 4.11(0:385-389,
(2010).
Sarwar, G., R. Joseph and R. Mathur. Influence of
Chlorine Emissions on Ozone Levels in the Troposphere.
Chapter 2, Douw G. Steyn and S.T. Rao (ed.), Air Pollution
Modeling and its Applications XX. Springer Netherlands.
Netherlands. 2.23(C):237-240, (2010).
Published Reports
Rao, S.T., J.O. Bash, S. Brown, R.C. Gilliam, D.
Mob ley, S. Napelenok, C.G. Nolte, T.E. Pierce and R.W.
Finder. Summary Report of the Atmospheric Modeling
and Analysis Division's Research Activities for 2009.
U.S. Environmental Protection Agency. Washington. DC.
EPA/600/R-10/058 (NTIS PB2011-100862). 2010.
Rao S.T., J.O. Bash, R.C. Gilliam, D. Mobley, S.
Napelenok, C.G. Nolte, T.E. Pierce, R.W. Finder and
C.G. Nolte. Summary Report of Atmospheric Modeling
and Analysis Divisions Research Activities for 2008.
U.S. Environmental Protection Agency. Washington, DC.
EPA/600/R-10/014, 2010.
Other
Hamilton, W.J., D.W. Byun, W. Chan, J.K. Ching, Y. Han,
R. A. Lope/. V.F. Coarfa and D. Lee. A Pilot Study using
EPA's CM AQ Model and Hospital Admission Data to
Identify Multipollutant "Hot Spots" of Concerns in Harris
County. Texas. Mickey Leland National Urban Air Toxics
Research Center (NUATRC), Houston, TX, 2010.
62
-------
APPENDIX E
Acronyms and Abbreviations
3D 3-Dimensional
4D 4-Dimensional
AAA Authentication Authorization Audit
ACM Asymmetric Convective Model
AEIB Atmospheric Exposure Integration Branch
AERLINE
AERMOD American Meteorological Society/EPA
Regulatory Model
AERDNET AErosol RObotic NETwork
AFMS Agricultural fertilizer modeling system
AMAD Atmospheric Modeling and Analysis
Division
AMB Applied Modeling Branch
AMDB Atmospheric Model Development Branch
AMET Atmospheric Model Evaluation Tool
AMNet Ambient Mercury Network
AOD Aerosol Optical Depth
APM Annual Performance Measure
APWS Albemarle-Pamlico Water Shed
AQMEII Air Quality Model Evaluation International
Initiative
AQS Air Quality System
AR4 Fourth Assessment Report
AR5 Fifth Assessment Report
ARL Air Resources Laboratory
ARM Atmospheric Radiation Measurement
BEIS Biogenic Emission Inventory System
BELD Biogenic Emissions Landuse Data
Bidi Bidirectional
BMA Bayesian Model Averaging
BMP Best Management Practice
BRACE Bay Regional Atmospheric Chemistry
Experiment
C-FERST/ Community-Focused Exposure and Risk
STREETS Screening Tool / Screening Tool for
Roadway Emissions and Exposure to Toxics
CAA Clean Air Act
CAAAC Clean Air Act Advisory Committee
CAIR Clean Air Interstate Rule
CAM Community Atmospheric Model
CMAQ-MP
CAP Criteria Air Pollutants
CAPMON Canadian Air and Precipitation Monitoring
Network
CAST NET EPA's Clean Air Status and Trends Network
CB05 Carbon Bond 2005
CBL Convective Boundary Layer
CBM-Z Carbon Bond Mechanism-Zaveri
CCN Clouds and the Earth's Radiant Energy
System
CDC Centers for Disease Control and Prevention
CEP UNC's Center for Environmental Programs
CERES Clouds and Earth's Radiant Energy
System
CFD Computational Fluid Dynamics
CHERUBS Childhood Health Effects from Roadway
and Urban Pollutant Burden Study
CIRAQ Climate Impacts on Regional Air Quality
CMAQ Community Multiscale Air Quality Model
CMAQ-TX Community Multiscale Air Quality Model-
Texas
CMAS Community Modeling and Analysis System
CO Carbon Monoxide
C02 Carbon Dioxide
CONUS Continental US
CTM Chemical Transport Model
DAPPLE Dispersion of Air Pollution and Penetration
into the Local Environment
DDM Decoupled Direct Method
DOE Department of Energy
DDM-3D Decoupled Direct Method-3D
EC elemental carbon
EMEB Emissions and Model Evaluation Branch
EPA U.S. Environmental Protection Agency
EPIC Environmental Policy Integrated Climate
Model
EPRI Electric Power Research Institute
ERD Ecosystem Research Division
ESD Environmental Sciences Division
ESRP Ecological Services Research Program
FDDA 4D Data Assimilation
FEST-C Fertilizer Emissions Scenario Tool for
CMAQ
FIA Forest Inventory Data
FLUENT Computational Fluid Dynamics Sofware
produced by ANSYS, Inc. (its not an
acronym, just the name of the software)
FML Future Midwestern Landscapes
FRM Federal Reference Method
FY Fiscal Year
GCM Global Climate Model
63
-------
GABLS GEWEX Atmospheric Boundary Layer Study
GEOS-Chem Global 3-D chemical transport model
(CTM) for atmospheric composition driven
by meteorological input from the Goddard
Earth Observing System (GEOS) of the
NASA Global Modeling and Assimilation
Office
GEWEX Global Energy and Water Cycle Experiment
GHG Greenhouse Gas
GISS Goddard Institute for Space Studies
GLIMPSE Geos-CHEM LIDORT Integrated with
MARKAL for the Purpose of Scenario
Exploration
GEOS Goddard Earth Observing System
HAP Hazardous Air Pollutant
HEASD Human Exposure & Atmospheric Sciences
Division
Hg[ll] Oxidized Mercury
HgC12 Mercuric Chloride
HgtO] Elemental Mercury
HN03 Nitric Acid
HONO Nitrous Acid
H02 Hydroperoxyl Radical
H202 Hydrogen Peroxide
HUC Hydrologic Unit Code
HPCC High Performance Computers and
Communication
I AG Interagency Agreement
ICARTT International Consortium for Atmospheric
Research on Transport and Transformation
IE Institute for the Environment (UNC-CH)
IMPROVE Interagency Monitoring of Protected Visual
Environment Network
INs Isoprene Nitrates
INTEX Intercontinental Chemical Transport
Experiment
INTEX-NA Intercontinental Chemical Transport
Experiment-North America
IPCC International Panel On Climate Change
ISORROPIA Thermodynamics Partitioning Module
ITM International Technical Meeting
ITR Integrated Transdisciplinary Research
LADCO Lake Michigan Air Directors Consortium
LAI Leaf Area Index
LBC Lateral Boundary Condition
LES Large-Eddy Simulations
LIDAR Light Detection And Ranging
LIDORT Linearized Discreet Ordinate Radiative
Transfer
LSM Land Surface Model
Lur Land-Use Regression
LW Longwave
M3dry
MACT
MAE
MCIP
MARKAL
MCM
MDA
MDN
MEGAN
MESA AIR
MLBC
MM5
MODIS
MOSAIC
MP
MPI
MYSQL
NAAQS
NADP
NAM
NAMMIS
NARR
NARSTO
NAS
NASA
NASS
NATO
NBP
NCAR
NCEA
nCe02
NCER
NCOM
NCSU
NEI
NERL
NEXUS
NGA
NH,
Models 3 Dry Deposition Model
Maximum Achievable Control Technology
Mean Absolute Error
Meteorology-Chemistry Interface Processor
Maximum Achievable Control Technology
Master Chemical Mechanism
Maximum Daily Average
Mercury Deposition Network
Model Of Emissions Of Gases And Aerosols
From Nature
Multi-Ethnic Study of Atherosclerosis and
Air Pollution
Multilayer Biochemical Model
Fifth Generation Of The Penn State/Ucar
Mesoscale Model
Moderate Resolution Imaging
Spectroradiometer
Model for Simulating Aerosol Interactions
and Chemistry
Multipollutant
message passing interface
open source database software
National Ambient Air Quality Standard
National Acid Deposition Program
North American Mesoscale
North American Mercury Model
Intercomparison Study
North American Regional Reanalysis
Formerly the North American Research
Strategy for Tropospheric Ozone
National Academy of Sciences
National Aeronautics and Space
Administration
National Agricultural Statistical Service
North Atlantic Treaty Organization
NO Budget Trading Program
National Center for Atmospheric Research
National Center for Environmental
Assessment
Nanoparticulate Cerium Oxide
National Center for Environmental
Research
Graphically Reduced Carbon Organic Mass
North Carolina State University
National Emission Inventory
National Exposure Research Laboratory
Near-road Exposure to Urban air pollutants
Study
National Geospatial Agency
Ammonia
64
-------
NH4
Ammonium
OTHER
Model Species Representing Pm25 That Is
NHEERL
National Health & Environmental Effects
Not Chemically Speciated
Research Laboratory
PMML
Predictive Model Markup Language
NLCD
National Land Cover Data
PRISM
Parameter-Elevation Regressions on
NMM
Nonhydrostatic Mesoscale Model
Independent Slopes Model
NO
Nitrogen Oxide
PX LSM
Pleim-Xiu Land Surface Model
no2
Nitrogen Dioxide
QUIC
Quick Urban Industrial Complex
N°3
Nitrate
Qv
Water vapor mixing ratio
N 0
Dinitrogen Pentoxide
RACM2
Regional Atmospheric Chemistry
2 5
Mechanism, version 2
NOx
Oxides Of Nitrogen
RARE
Regional Applied Research Effort
NO
Oxidized Nitrogen
y
RCCM
Regional Climate and Chemistry Modeling
NOAA
National Oceanic And Atmospheric
System
Administration
RCM
Regional Climate Model
NOAH
Noaa's Land Surface Model
RCP
Representative Concentration Pathways
NPS
National Park Service
RELMAP
Regional Lagranian of Air Pollution
Nr
Reactive Nitrogen
REML
Durbin-WatsDon test and restricted
NRC
National Research Council
maximum likelihood
NRMRL
National Risk Management Research
Laboratory
REMSAD
Regional Modeling System for Aerosols and
Deposition
NSF
National Science Foundation
RGM
reactive gaseous mercury
NUDAPT
National Urban Database And Access
RMSE
root mean squared error
Portal Tool
ROSES
Research Opportunities in Space and Earth
03
Ozone
Sciences
OAP
Office Of Air Programs
RRTMG
Rapid Radiative Transfer Model for GCMs
OAQPS
Office Of Air Quality Planning And
Standards
SCIAMACHY
Scanning Imaging Absorption Spectrometer
for Atmospheric Cartography
OAR
Office Of Air & Radiation
SEARCH
South Eastern Aerosol Research and
OC
Organic Carbon
Characterization Study
OH
Hydroxy Radical
SGV
subgrid variability
OLAM
Ocean-Land-Atmosphere Model
SHEDS
Stochastic Human Exposure and Dose
Simulation
OM
Organic Mass
SIP
State Implementation Plan
OPE
ozone production efficiency
SLAMS
State or Local Air Monitoring Station
ORD
Office of Research and Development
SLCF
short-lived climate forcers
OTAQ
Office of Transportation and Air Quality
SMOKE
Sparse Matrix Operator Kernel Emissions
PAH
Polycyclic Aromatic Hydrocarbon
SO,
sulfur dioxide
PAN
Peroxyacyl Nitrate
S04
sulfate
PAVE
Package For Analysis And Visualization Of
4
sulfur oxides
Environmental Data
sox
PBL
Planetary Boundary Layer
SOA
secondary organic aerosol
PCA
Principal Component Analysis
SOAcld
secondary organic aerosol formed in clouds
PEM
Pesticides Emissions Model
SPARROW
Spatially Referenced Regressions on
Watershed Attributes
PM
Particulate Matter
SPS
Science for Peace and Security
PM,,
Particulate Matter Smaller Than 2.5
2.5
Microns In Diameter
SST
Sea Surface Temperature
PM2.5FRM
PM25 is defined above and FRM is:
Federal Reference Method
STAR
Science To Achieve Results
STENEX
Stencil Exchange
PM10
Particulate Matter Smaller Than 10
STN
Speciated Trends Network
Microns In Diameter
SW
Shortwave
^ COARSE
Particulate Matter Between 2.5 And 10
Microns In Diameter
SWAT
TBEP
Soil & Water Assessment Tool
Tampa Bay Estuary Program
-------
TEAM Trace Element Analysis Model
TES Tropospheric Emission Spectrometer
TexAQS Texas Air Quality Study
TM Thematic Mapper
TMDL Total Maximum Daily Load
UCP Urban Canopy Parameter
UNC-CH University of North Carolina at Chapel Hill
USDA United States Department of Agriculture
USFS United States Forest Service
USGS U.S. Geological Survey
VBS Volatility Basis Set
VERDI Visualization Environment For Rich Data
Interpretation
VOC Volatile Organic Compound
WDT Watershed Deposition Tool
WDWE Weekday-To-Weekend
WRF Weather Research And Forecasting
WSOC Water Soluble Organic Compound
YSU Yonsei University
66
-------
SCIENCE
SEPA
United States
Environmental Protection
Agency
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Recycled/Recyclable Printed on paper that contains a minimum of
50% posteonsumer fiber content processed chlorine free
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO.G-35
------- |