EPA/600/R-10/058 June 2010 www.epa.gov/ord
Summary Report of the
Atmospheric Modeling and Analysis Division's
Research Activities for 2009
ST. Rao, Jesse Bash, Sherry Brown, Robert Gilliam, David Heist, David Mobley,
Sergey Napelenok, Chris Nolte, Tom Pierce, and Rob Pinder
Atmospheric Modeling and Analysis Division
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 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 does not constitute endorsement or recommendation for
use.
-------
Abstract
The research presented here was performed by the Atmospheric Modeling and Analysis Division (AMAD) 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 reduce exposures to sensitive populations and ecosystems. AMAD is responsible for
providing a sound scientific and technical basis for regulatory policies to improve ambient air quality. The models
developed by AMAD are being used by EPA and the air pollution community in understanding and forecasting not
only the magnitude of the air pollution problem but 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 source-to-outcome continuum 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 the AMAD for
calendar year 2009.
-------
-------
Table of Contents
List of Tables vii
List of Figures viii
1. Introduction 1
2. Summary of Accomplishments for the Division 3
2.1 Division-Wide Accomplishments 3
2.2 Model Development and Diagnostic Testing 3
2.3 Air Quality Model Evaluation 4
2.4 Climate and Air Quality Interactions 5
2.5 Linking Air Quality to Human Exposure 6
2.6 Linking Air Quality and Ecosystems 7
3. Model Development and Diagnostic Testing 9
3.1 Introduction 9
3.2 CMAQ Aerosol Module 10
3.3 CMAQ Gas and Aqueous Chemical Mechanisms 10
3.4 Planetary Boundary Layer Modeling for Meteorology and Air Quality 13
3.5 Multiscale Meteorological Modeling for Air Quality 14
3.6 Coupled WRF-CMAQ Modeling System 15
3.7 Mercury Modeling 18
3.8 CMAQ for Air Toxics and Multipollutant Modeling 20
3.9 Emissions Modeling Research 21
4. Air Quality Model Evaluation 25
4.1 Introduction 25
4.2 Operational Performance Evaluation of Air Quality Model Simulations 25
4.3 Diagnostic Evaluation of the Oxidized Nitrogen Budget Using Space-Based, Aircraft, and Ground
Observations 26
4.4 Diagnostic Evaluation of the Carbonaceous Fine Particle System 26
4.5 Inverse Modeling To Evaluate and Improve Emission Estimates 28
4.6 Probabilistic Model Evaluation 28
4.7 Statistical Methodology for Model Evaluation 29
4.8 Dynamic Evaluation of a Regional Air Quality Model 29
5. Climate and Air Quality Interactions 32
5.1 Introduction 32
5.2 Climate Impact on Regional Air Quality 32
5.3 Emission Scenario Development 32
5.4 Regional Climate Downscaling 33
5.5 Statistical Climate Downscaling 33
5.6 Integrated Tools for Scenario Discovery 34
6. Linking Air Quality to Human Health 38
6.1 Introduction 38
6.2 Near-Roadway Environment 38
6.3 Evaluating Regional-Scale Air Quality Regulations 39
6.4 Linking Local-Scale and Regional-Scale Models for Exposure Assessments 40
6.5 National Urban Database and Access Portal Tool 42
7. Linking Air Quality and Ecosystems 43
7.1 Introduction 43
7.2 Linking Air Quality to Aquatic and Terrestrial Ecosystems 43
7.3 Linking Ecosystem Services 46
7.4 Air-Surface Exchange 49
7.5 CMAQ Ecosystem Exposure Studies 54
7.6 Software Tool Development 58
-------
Table of Contents (cont'd.)
References 61
Appendix A: Atmospheric Modeling and Analysis Division Staff Roster 63
Appendix B: Division and Branch Descriptions 64
Appendix C: 2009 Awards and Recognition 65
Appendix D: 2009 Publications 66
Appendix E: Acronyms and Abbreviations 69
VI
-------
List of Tables
3-1 Base Photochemical Mechanisms in CMAQ and the Species Commonly Predicted by Each
Mechanism
3-2 Summary of Surface-Based Model Performance Statistics for Each Simulation 15
3-3 Evaluation Statistics from the North American Mercury Model Intercomparison Study 20
3-4 Hazardous Air Pollutants Represented in the Current CMAQ Multipollutant Model 21
VII
-------
List of Figures
1-1 The Division's role in the source-exposure-dose-effects continuum from the atmospheric science
perspective 1
1-2 The Division's structure and organization 2
3-1 A flowchart that outlines the various components of the CMAQ modeling system 9
3-2 Representation of PM size and composition in CMAQ v4.7 10
3-3 Current Interactions between gas, aqueous, and aerosol chemistry in CMAQ base and multipollutant
models 11
3-4 Comparison of average modeled vertical profiles of sulfate with NOAA WP-3D aircraft measurements. ... 12
3-5 Layer-averaged vertical profiles of OC and WSOC on August 14, 2004 13
3-6 The most direct measure of success for a PEL model for both meteorology and air quality is its ability
to accurately simulate the vertical structure of both meteorological and chemical species 15
3-7 Spatially distributed root mean square error difference between the WRF and MM5 for August 2006 16
3-8 Mean absolute error profiles of model-simulated temperature, wind speed, and wind direction for
August 2006 16
3-9 Diurnal mean wind speed profiles for January and August 2006 17
3-10 Two sets of initial simulations have been conducted to test the evolving coupled WRF-CMAQ
modeling system and to systematically assess the impacts of coupling and feedbacks 18
3-11 CMAQ multipollutant model predictions for ozone, as maximum 8-h value, and formaldehyde, as
monthly average for July 2002 21
3-12 AMAD's research contributed to the NEI's Wildfire Emissions Inventory 22
3-13 Comparison of isoprene emissions estimated by BEIS and MEGAN 23
4-1 Model outputs are compared to observations using various techniques 25
4-2 Scatter plot of observed versus CMAQ-predicted sulfate for August 2006 created byAMET 26
4-3 Vertical profile of the ratio of nitric acid to total oxidized nitrogen, as sampled during the
Augusts, 2004, ICARTT flight over the northeastern United States 27
4-4 Source contributions to the modeled concentrations of fine-particulate carbon in six U.S. cities 27
4-5 Comparison of modeled and observed NO2 column concentrations 28
4-6 Spatial plots of ozone and probability of exceeding the threshold concentration for July 8, 2002,
at 5 p.m. EOT 29
4-7 Assessment of CMAQ's performance in estimating maximum 8-h ozone in the northeastern
United States on June 14, 2001 30
4-8 Example of dynamic evaluation showing observed and air quality model-predicted changes from
differences between summer 2005 and summer 2002 ozone concentrations from Gilliland et al.
(2008) 31
5-1 Differences in mean and 95th percentile maximum daily 8-h average ozone concentrations 33
5-2 Seasonally averaged wind fields at 300 hPa as simulated by North American Regional Reanalysis,
WRF without nudging, WRF with analysis nudging, and WRF with spectral nudging 34
5-3 Mean July 500-hPa geopotential height for GISS ModelE; base WRF run, without any interior nudging;
WRF with analysis nudging; and WRF with spectral nudging 35
5-4 Mean July 2-m temperature for GISS ModelE, base WRF run without any interior nudging, WRF with
analysis nudging, and WRF with spectral nudging 36
5-5 GLIMPSE data flow: GEOS-Chem LIDORT Adjoint model is used to attribute radiative forcing changes
to U.S. emission sectors 37
6-1 Linking local-scale and regional-scale models for exposure assessment characterizing special
variation of air quality near roadways assessing the effectiveness of regional-scale air quality
regulations 38
6-2 The Fluid Modeling Facility houses the Division's meteorological wind tunnel used to study the effect
of roadway configuration and wind direction on near-roadway dispersion 39
VIM
-------
List of Figures (cont'd.)
6-3 Assessing the impact of regulations on ecosystems and human health end points showing the
indicators and process linkages associated with the NOX Budget Trading Program 40
6-4 Schematics of the hybrid modeling approach showing local impact from stationary sources, near-road
impact from mobile sources, and regional background from CMAQ 41
6-5 Urban canopy effects 42
7-1 A Venn diagram representing ecosystem exposure as the intersection of the atmosphere and
biosphere 43
7-2 Fractional deciduous forest coverage as represented in the 30-m resolution 2001 NLCD based on
Landsat 7 satellite imagery and in the 1-km resolution 1992 NLCD based on Landsat Thematic Mapper
satellite imagery 44
7-3 Receptor-specific ozone deposition velocities to croplands 45
7-4 Receptor-specific ozone deposition velocities to forested ecosystems 45
7-5 Left panel is a map of the Deep River and Haw River watersheds within the Cape Fear River Basin 46
7-6 Future Midwestern landscapes study area superimposed on the Midwest ecoregions 47
7-7 Flow chart of AMAD's role in FML model development 48
7-8 2002 Annual total nitrogen deposition 49
7-9 2002 Annual acidifying dry deposition of sulfur and oxidized and reduced nitrogen 50
7-10 Air-surface exchange resistance diagrams of unidirectional exchange, bidirectional exchange of
ammonia, and bidirectional exchange of mercury and ammonia using the FEST-C tool 51
7-11 Mean air-surface exchange of NH3 for the month of July estimated by CMAQ v4.7 using MM5 with the
PX land surface scheme for unidirectional exchange of NH3 and bidirectional exchange of NH3 52
7-12 Daily Harnett County, NC, NEI soil emission estimates and simplified process model estimates plotted
with Lillington, NC, observations 53
7-13 Ammonia exchange budget estimated from the analytical closure model 54
7-14 TES transect locations and surface observations overlaid on a map of the estimated NH3 emission
density in Eastern North Carolina 54
7-15 CMAQ is a source of data for ecosystem managers that is not available in routine monitoring data,
such as complete dry and wet deposition estimates, and the "one atmosphere" concept of CMAQ is
needed to understand the balance between uncertainties in atmospheric reaction rates and deposition
pathways 55
7-16 Airsheds and watershed for Narragansett Bay, Chesapeake Bay, Pamlico Sound, Mobile Bay, Lake
Pontchartrain, and Tampa Bay 56
7-17 Model-predicted contributions of six Bay States account for 50% of the 2020 oxidized nitrogen
deposition to the Chesapeake Bay Watershed 57
7-18 Fraction of total oxidized nitrogen deposition to Tampa Bay explained by local emission in the
watershed 58
7-19 Examples of VERDI used to visualize and evaluate CMAQ output 59
7-20 Screen shot of the 2002 annual CMAQ total reduced nitrogen deposition mapped to watersheds
draining into the Albemarle-Pamlico Sound displayed in CIS mapping software 60
7-21 Spatial Allocator output from raster tools on North Carolina 1-km grids for factional tree canopy
coverage and impervious surfaces from NLCD data 60
IX
-------
-------
CHAPTER 1
Introduction
The research presented here was performed by the
Atmospheric Modeling and Analysis Division (AMAD) 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. This report summarizes the research and
operational activities of the Division for calendar year
2009.
The Division structure includes four research
branches:
(1) the Atmospheric Model Development Branch
(AMDB),
(2) the Emissions and Model Evaluation Branch
(EMEB),
(3) the Atmospheric Exposure Integration Branch
(AEIB), and
(4) the 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 in this report (Appendix E).
The Division's role within EPA's National Exposure
Research Laboratory's (NERL's) "Exposure Framework"
and the EPA Office of Research and Development's
(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 research branches, the Division provides
atmospheric sciences expertise, air quality forecasting
support, and technical guidance on the meteorological
and air quality modeling 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 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 fwww.epa.qov/amad/l.) 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.
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?
Source-to-Outcome Continuum
Ambient
Concentrations
Exposure
Figure 1-1. The Division'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
-------
• 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 calendar year
2009. It includes descriptions of research and
operational efforts in air pollution meteorology, in
meteorology and air quality model development, and in
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).
Sound Science for 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
AMAD Structure and Organization
Figure 1-2. The Division's structure and organization.
-------
CHAPTER 2
Summary of Accomplishments for the Division
As a summary of and introduction to the annual
report for 2009, the following Division accomplishments
are highlighted.
2.1 Division-Wide Accomplishments
1. Issue: Strategic thinking regarding air quality
monitoring and modeling in the next decades
Accomplishment: Coordination of papers for the
October 2009 issue of the Air and Waste
Management Association's Environmental Manager
(EM) on Monitoring and Modeling Needs in the 21st
Century
Findings: Four-dimensional air quality, emissions,
and meteorological data are needed with increased
spatial and temporal resolution for improving air
quality models and future policy decisions.
Impact: This EM special issue provides a thought-
provoking set of articles for managers to consider in
improving monitoring, modeling, meteorological,
emission characterization, and data analysis
programs to meet future challenges of the air quality
management program. This work led to the
preparation of an inter-Divisional collaborative
research proposal to AMI, involving program and
regional offices, for obtaining 3-D air quality data
over the United States using commercial aircrafts.
2. Issue: Milestone anniversary meeting of three
decades of international cooperation on air pollution
modeling (It is the United States' turn to host this
North Atlantic Treaty Organization (NATO)
meeting.)
Accomplishment: Development and host of the
program for the May 18-22, 2009 Meeting of the
30th NATO/Science for Peace and Security (SPS)
International Technical Meeting (ITM) on Air
Pollution Modeling and its Applications, in San
Francisco, CA
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: Over 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. A book under
the NATO banner was published by the conference
organizers.
3. Issue: January 2009 peer review of NERL's AMAD
Accomplishment: Preparation of extensive
handbook and poster book documentation and
posters for the 2009 AMAD Peer Review
Findings: The draft report of the Peer Review
Committee was complimentary of the Division's
research and included constructive suggestions.
Impact: The 2009 AMAD Peer Review confirmed
the wisdom of AMAD's strategic research directions
and excellence of past accomplishments. Based on
the peer review, we prepared three white papers on
the Division's new research initiatives.
4. Issue: June 2009 Board of Scientific Counselors
(BOSC) Review of ORD's Clean Air Research
Program
Accomplishment: Preparation of posters and
abstracts for the Air Quality and Multipollutant
Sessions; AMAD co-chair of Air Quality Session
Findings: Multipollutant air quality management is
needed.
Impact: The 2009 BOSC Peer Review of the ORD
Air Research Program highlighted AMAD's
modeling and analysis contributions to the air quality
and multipollutant themes of the program. The
Division's contributions to air quality modeling were
viewed very favorably by BOSC.
5. Issue: Systematic intercomparisons and
evaluations are needed for regional air quality
models over different continental regions.
Accomplishment: Initiation of collaborations for the
Air Quality Model Evaluation International Initiative
(AQMEII) with Canadian and European partners;
development of program for first AQMEII Workshop,
April 27-29, 2009, Stresa, Italy
Finding: The AQMEII modeling initiative was begun
with a workshop in April 2009, during which North
American and European perspectives on model
evaluation were discussed.
Impact: A model intercomparison 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 using the model evaluation framework
developed by AMAD.
2.2 Model Development and Diagnostic Testing
1. 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: Extension of the Community
Multiscale Air Quality (CMAQ) model to hemispheric
-------
scales: initial demonstration of the concept for Hg
and aerosol radiative effects
Findings: Air quality modeling results for ozone,
particulate matter (PM), Hg, and other pollutants
over the United States are sensitive to the
specification of boundary concentrations.
Impact: CMAQ modeling capability now has been
extended to the full Northern Hemisphere, enabling
consistent specification of North American boundary
concentrations and helping understand how the
intercontinental transport of pollution affects air
quality over the United States.
2. Issue: Chemical kinetic mechanisms are at the
heart of air quality models used for National
Ambient Air Quality Standard (NAAQS)
implementation.
Accomplishment: Testing and initial incorporation
of new chemical kinetic mechanisms in CMAQ:
SAPRC07 and RACM2
Findings: The latest generation lumped species
chemical mechanisms have been tested against
smog chamber data and evaluated for incorporation
into the CMAQ model. 2.
Impact: The CMAQ model will contain versions of
three state-of-the-science chemical mechanisms for
use in air quality modeling (CB05, SAPRC07, and
RACM2) for testing the robustness of emission
control strategies by the program office and States.
3. Issue: Engineered nanomaterials can lead to
ambient exposures of nanoparticles and health
effects.
Accomplishment: Development of a joint AMAD-
Human Exposure and Atmospheric Sciences
Division (HEASD) research plan for predictive
models for the transport, transformation, and fate of
engineered nanomaterials
Findings: The initial focus of study will be on
cerium oxide, a possible diesel fuel stabilization
additive, and titanium dioxide, which is used in paint
and other surface coatings.
Impact: The joint research plan will lead to studies
on the chemical and physical attributes of these
nanomaterials, as well as initial ambient modeling
studies.
2.3 Air Quality Model Evaluation
1. Issue: EPA-National Oceanic and Atmospheric
Administration (NOAA) collaboration in air quality
model forecasting has developed an initial capability
for PM2 5 forecast guidance across the United
States. This capability needs comprehensive 3.
evaluation before operational deployment.
Accomplishment: Completed Annual Performance
Measure (APM) 154: Analysis and evaluation of
developmental PM forecast simulations over the
continental United States. (This APM reflects the
development, deployment, and detailed evaluation
of a "developmental" PM forecast modeling system
for the continental United States and approaches to
produce reliable forecast of air quality index [AQI].)
Findings: Developmental forecast simulations
during 2004-2008 continuously were analyzed and
evaluated against near real-time measurements
from the AIRNOW network. In addition, forecasts of
fine-PM speciation were compared against
measurements from a variety of other surface PM
networks. The systematic errors found in model
predictions of both total PM2 5 and its constituents
have provided guidance for future research and
further model development.
Impact: To improve the accuracy and utility of PM2 5
forecast guidance obtained from comprehensive
atmospheric models in the short-term,
postprocessing bias-adjustment techniques that
combine the model forecast with near real-time
observations from the AIRNOW network were
developed to provide reliable operational AQI
forecasts. If the proposed method is operationalized
by NOAA and EPA, it would enable the
development of credible air quality, AQI, and
exposure surfaces for the continental United States
on a daily basis.
Issue: CMAQ v4.7 was released to the public in
October 2008. Extensive incremental testing was
conducted on the model prior to release. Results of
the testing and evaluation need to be documented.
Accomplishment: Documentation of extensive
process testing and evaluation of CMAQ v4.7 to
support its release in October 2008 and multiyear
(2002-2006) model evaluations of CMAQ v4.7 in
support of the Centers for Disease Control and
Prevention (CDC) collaboration on the PHASE
project
Findings: The continued evaluation of CMAQ v4.7
has led to the correction of several performance
issues with the new model. In addition, as part of
the CDC PHASE project, annual CMAQ v4.7
simulations were performed for 2002-2006. This
multiyear simulation provided an opportunity to
evaluate the CMAQ model under numerous
meteorological conditions. The model evaluation
revealed several systematic model performance
issues that occur each year, while other
performance issues appear to occur under specific
meteorological conditions.
Impact: Model deficiencies identified from process
testing and annual 2002-2006 simulations were
corrected and implemented in an interim release of
CMAQ v4.7.1 in late 2009, enhancing the scientific
credibility for CMAQ.
Issue: The quantification of uncertainty in air quality
modeling results has been an important goal, but
there has been little progress in this area.
Accomplishment: Demonstration of probabilistic
model evaluation of the CMAQ model using an
ensemble of model configurations and direct
sensitivity analysis
Findings: Advances in probabilistic modeling
approaches include improved methods for
characterizing and understanding the sources of
-------
uncertainty. Using Bayesian Parameter Estimation,
advanced methods have been developed for
translating an ensemble of CMAQ model
simulations into a probability distribution.
Impact: The Direct Decoupled Method (DDM) has
been incorporated in CMAQ v4.7, which is used to
calculate the sensitivity of ozone to specific
emission sources and model parameters. Further,
these techniques have been applied to identify
emission source sectors that have significant
contributions to ozone sensitivity. This work is
helping us in testing the robustness of the response
of CMAQ to emission reductions.
2.4 Climate and Air Quality Interactions
1. Issue: Future air quality is expected to be affected
both by climate change and by emissions changes.
Phase 1 of the Climate Impacts on Regional Air
Quality (CIRAQ) project focused on the potential
impacts of climate change on air quality. Phase 2 3.
has added regional emissions projections for the
future on top of climate change.
Accomplishment: Completed APM 258: The
impact of climate change on U.S. PM
concentrations: Model sensitivity tests and PM
concentration changes in the United States under a
future climate scenario with and without future
emission scenarios
Findings: In Phase 1 (climate change only), CMAQ
modeling under the future (2050) scenario resulted
in average ozone increases of approximately 2 to
5 ppb and 95th percentile (i.e., fourth highest) ozone
increases greater than 10 ppb in some regions. In
Phase 2, with emissions projections added, it was
found that the magnitude of the decrease in ozone
resulting from changing emissions is much larger
than the increase resulting from climate change.
The effect of climate change on PM concentrations
appears to be driven primarily by changes in
precipitation patterns, which are highly uncertain.
For these simulations, increased future precipitation
leads to decreased PM concentrations, so that the
effect of changing emissions and climate are in the
same direction.
Impact: This initial CIRAQ study has laid the
foundation for future air quality-climate change
assessments. Large uncertainties exist in future 4.
projections from any single global climate model
(GCM), so research planning has taken into account
the use of up to three GCMs from which to simulate
regional climate. Various downscaling techniques
will be tested, including dynamical and statistical.
Screening tools and comprehensive modeling tools
will be developed to assess the potential impacts of
air quality on global and regional climate.
2. Issue: Regional downscaling of GCM results must
begin with global model data. AMAD must establish
strong working relationships with global modeling
groups to acquire the appropriate data for
downscaling.
Accomplishment: Establishment of collaboration
with the National Aeronautics and Space
Administration (NASA) Goddard Institute for Space
Studies (GISS) on global to regional downscaling of
upcoming GCM simulations covering the 21st
century
Findings: NASA/GISS, under the leadership of
Dr. James Hansen, is one of the premier global
climate modeling centers in the world. Their latest
global model, Model E, will be used for simulations
to inform the next International Panel on Climate
Change (IPCC; the fifth) Assessment Report.
Impact: An interagency agreement with NASA is
being established, with a postdoctoral fellow to work
between both NERL/AMAD and NASA/GISS, to
obtain high temporal resolution Model E results for
AMAD's regional model downscaling with weather
research and forecasting (WRF). This demonstrates
the value of cross-agency collaboration.
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: Regional Climate Downscaling
with WRF has been tested using both global
reanalysis data and output from the GISS Model-E
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 has shown much better correspondence
between global and regional meteorological
patterns. Results are sensitive to the nudging
parameters; thus, more testing is needed to
determine best configuration.
Impact: AMAD's experiments with data assimilation
in the process of downscaling from global to
regional climate models (RCMs) 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.
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 pollution 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 ozone feedback on longwave
(LW) radiation. Indirect feedback is under
development.
Findings: The WRF-CMAQ coupled meteorology-
chemistry model has been tested, including direct
-------
aerosol feedback on SW radiation and ozone
feedback on LW radiation. Indirect feedback is
under development.
Impact: The 2-way coupled WRF-CMAQ system 2.
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. This work led to
the preparation of a research proposal to build
EPA-U.S. Department of Energy collaboration in the
climate change arena.
2.5 Linking Air Quality to Human Exposure
1. Issue: Methods are needed for verifying the impact
of emissions control programs on air quality ambient
concentrations, human exposures, and health
outcomes
Accomplishment: Completed ARM 155: Develop a
Mesoscale Pilot of Approaches for Identifying and
Tracking Regulatory Impacts (This ARM reflects the
culmination of several research projects that have
resulted in approaches for identifying and tracking
air quality impacts of regional-scale regulatory
emissions control programs. These approaches 3.
were applied to examine the impact of the NOX
Budget Trading Program [NBP].)
Findings: The CMAQ model was used to
characterize air quality before and after the
implementation of the NBP and to evaluate
correlations between changes in emissions and
pollutant concentrations. Model simulations were
used to estimate the anthropogenic contribution to
total ambient concentrations and the impact of not
implementing the regulation. Methods were
developed to differentiate changes attributable to
emission reductions from those resulting from other
factors, such as weather and seasonal variations.
Trajectory models were used to investigate the
transport of primary and secondary pollutants from
their emission sources to downwind regions. In
addition, research has focused on relating NOX
emissions and ambient ozone concentrations to
human exposure and health end points.
Impact: Combined modeled/measured high-
resolution air quality surfaces were used in human
exposure models, epidemiological health studies,
and health risk assessments. The preliminary
results indicate that the NBP might have contributed
to reduced respiratory-related hospital admissions in
some regions of New York State. This effort led to
the development of an innovative method to
understand air quality and human health linkages.
Issue: Ambient air pollutant concentrations are
needed to assess exposures but are not equivalent
to them. Methods are needed to develop exposure
estimates informed by modeled ambient
concentrations.
Accomplishment: Development and demonstration
of a methodology to link regional- and local-scale air
quality models with human exposure models for
improving community level environmental health
studies, involving near-source exposures to multiple
pollutants
Findings: A 2009 Journal of the Air & Waste
Management Association paper by Isakov and
co-investigators presents an innovative
methodology to link regional- and local-scale air
quality models with human exposure models. It
shows the existence 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: The complexity in the spatial variation of
exposures among different population cohorts,
especially in the context of cross-sectional or
intra-urban analysis of air pollution health effects,
could be quite challenging. The information derived
from this study will be used by EPA as a resource
for future air accountability research planning.
Through this effort, the Division has helped to
advance exposure science.
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 of such exposure
is needed.
Accomplishment: For the near-road research
program, developed wind tunnel and field study
databases and improved algorithms for urban
roadways in the American Meteorological Society
(AMS)/EPA Regulatory Model (AERMOD) in
support of human exposure and health
assessments.
Findings: The new line source algorithm
significantly advances the assessment tools for
near-road application. To be approved for inclusion
in AERMOD, this algorithm and the work described
underwent extensive internal and external peer
review. This review and approval process included
input from the AMS/EPA Regulatory Model
Improvement Committee that provided scientific
advice and support to EPA in the area of near-
source/short-range dispersion modeling. This work
supports EPA offices and programs needing to
simulate short-range dispersion in relation to permit
applications, exposure research, and human health
risk assessments by ensuring that short-range
-------
dispersion programs incorporate peer-reviewed
science. This research assists with the transfer of
state-of-the-art science and modeling techniques
into practical, workable tools applicable to key
programs, such as regulatory modeling. The
improved model provides EPA and other
stakeholders with the information needed to identify
potential health risks for near-road populations and
to develop air pollution control programs to address
these risks. In addition, it enables the modeling of
air quality impacts for regulatory programs under the
Federal Highway Administration's (FHA's)
Transportation Conformity Rule and the National
Environmental Policy Act.
Impact: The importance of this work was
recognized by EPA and external stakeholders by its
inclusion in the proposed nitrogen dioxide (NO2)
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 litigation. The FHA 3.
requested EPA's guidance and expertise in
implementing their near-road research requirements
as part of this litigation, and an inter-agency
agreement 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.
2.6 Linking Air Quality and Ecosystems
1. Issue: Existing treatment of ammonia (NH3) flux in
the CMAQ model consists of a specified emissions
term and a computed deposition. More realistic
treatment is needed considering the compensation
points for NH3 in soil and the plant canopy allowing
for two-way flux.
Accomplishment: Development of new CMAQ NH3
bidirectional exchange algorithms through joint
AMAD-National Risk Management Research
Laboratory (NRMRL) collaboration on field data
analyses and model development
Findings: Critical data needed to parameterize a
two-layer deposition model was collected, and it
was shown that it was feasible to parameterize a
model that accounts for bidirectional exchange and
include it in CMAQ. The need for a fertilization 4.
model to provide an estimate of the soil
compensation point was identified.
Impact: The foundation is laid for a more
sophisticated approach to air-surface exchange
within CMAQ, and a strong rationale is provided to
bring air-surface exchange calculations fully into
CMAQ. Incorporating bidirectional exchange of NH3
is expected to significantly impact the range of
influence of NH3 emissions.
2. Issue: Biases and/or errors in the Fifth-Generation
Pennsylvania State University/National Center for
Atmospheric Research (NCAR) Mesoscale Model
(MM5) or WRF modeled precipitation can cause
problems for calibrated watershed models that
typically use observed precipitation data for
calibration.
Accomplishment: Identification of the need for
WRF-consistent hydrology to address linkage
disparities through collaboration between AMAD
and the Ecological Research Division (ERD) on
analysis of the effect of MM5 precipitation errors on
watershed hydrology
Findings: Errors in MM5 or WRF modeled
precipitation timing, location, and amount are too
large to be handled by calibrated watershed models,
making the direct use of CMAQ wet deposition for
air-water linkage very problematic.
Impact: A key new AMAD research area is
identified, linking a hydrology model to WRF/CMAQ,
which is needed for CMAQ to successfully support
atmosphere-ecosystem linkage. This effort would
help advance ecological exposure assessments.
Issue: Future deposition is expected to be
significantly reduced by Clean Air Act (CAA)
regulations that address ozone and PM2 5
attainment. Finer resolution grids (12-km) match
better to watershed segments and better resolve
coastal estuaries for linking atmospheric deposition
to coastal systems.
Accomplishment: Delivery of nitrogen deposition
futures scenarios for 2009, 2020, and 2030 to
Chesapeake Bay Program
Findings: The CAA Amendments are anticipated to
make major reductions (>50%) in oxidized nitrogen
deposition to coastal estuaries across the eastern
United States. Such reductions are very important to
restoration efforts. However, these gains are offset
to a significant degree by the expected future
increases in ammonia emissions.
Impact: Linked the latest CMAQ with the latest
Chesapeake watershed model—both at higher
spatial resolution. The Division provided
Chesapeake Bay Program a complete set of
deposition scenarios that provides a best estimate
of the benefits of CAA regulations on deposition for
Chesapeake Bay Program Office total maximum
daily load (TMDL) analyses and other management
analyses.
Issue: MM5 or WRF modeled precipitation errors
cause a problem for providing deposition inputs to
critical loads models that require the most accurate
deposition inputs possible for their biogeochemical
mass balance calculations.
Accomplishment: Development of approach to
postprocess CMAQ wet deposition to reduce errors
and delivery of postprocessed CMAQ deposition
fields to EPA and the National Park Service (NPS)
for national critical loads analysis
Findings: Use of Parameter-Elevation Regressions
on Independent Slopes Model (PRISM) data to
correct for modeled precipitation error, plus simple
-------
bias corrections, enables reduction and smoothing applied successfully to 2002 annual deposition data
out wet deposition error and inclusion of orographic to create acceptable national deposition fields for
effects on wet deposition. This approach appears to EPA and NPS critical loads analyses. The critical
be preferable to data fusion for providing modeled loads models for the first time used CMAQ wet and
fields to better fill in for sparse monitoring of wet dry deposition fields for input, successfully
deposition. demonstrating the capability to use CMAQ for these
Impact: The postprocessing approach was critical loads analyses.
-------
CHAPTER 3
Model Development and Diagnostic Testing
3.1 Introduction
EPA and the States are responsible for
implementing the NAAQS for ozone and PM. New
standards for 8-h average ozone and daily average PM2 5
concentrations recently have been promulgated. Air
quality simulation models, such as the CMAQ modeling
system, are central components of the air quality
management process at the national, State, and local
levels. CMAQ, which is used for research and regulatory
applications by the EPA, States, and others, must have
up-to-date science to ensure the highest level of
credibility for the regulatory decisionmaking process. The
research goals under the CMAQ model development and
evaluation program are as follows.
• Develop, evaluate, and refine scientifically credible
and computationally efficient process simulation and
numerical methods for the CMAQ air quality modeling
system
• Develop the CMAQ model for a variety of spatial
(urban through continental) and temporal (days to
years) scales and for a multipollutant regime (ozone,
PM, air toxics, visibility, and acid deposition)
• Adapt and apply the CMAQ modeling system to
particular air quality/deposition/climate-related
problems of interest to EPA and use the modeling
system as a numerical laboratory to study the major
science processes or data sensitivities and
uncertainties related to the problem
• Evaluate the CMAQ modeling system using
operational and diagnostic methods and to identify
needed model improvements
• Use CMAQ to study the interrelationships between
different chemical species, as well as the influence of
uncertainties in meteorological predictions and
emission estimates on air quality predictions
• Collaborate with research partners to include up-to-
date science process modules within the CMAQ model
system
• Pursue computational science advancements (e.g.,
parallel processing techniques) to maintain the
efficiency of the CMAQ modeling system
The CMAQ modeling system outlined in Figure 3-1
initially was released to the public by EPA in 1998.
Annual updated releases to the user community and the
creation of a Community Modeling and Analysis System
(CMAS).center, which provides user support for the
CMAQ system and holds an annual CMAQ users
conference, have helped to create a dynamic and
diverse CMAQ community of over 2000 users in
90 countries. CMAQ has been and continues to be used
extensively by EPA and the States for air quality
management analyses, by the research community for
studying relevant atmospheric processes, and by the
international community in a diverse set of model
applications. Future research directions include
development of an integrated WRF (meteorological
Meteorological Model
(WRForMMS)
Met-Chem Interface
Processor (MCIP)
Met. Data Processing
CMAQ AQ Model
Chemical-Transport
Computations
Weather Data
EPA Missions
Inventory
SMOKE
Anthropogenic and Biogenic
Emissions Processing
Hourly 3-D Gridded Chemical
Concentrations
Figure 3-1. A flowchart that outlines the various components of the CMAQ modeling system.
9
-------
model)-CMAQ model for two-way feedbacks between
meteorological and chemical processes and models and
extension of the CMAQ system to hemispheric scales for
global climate-air quality linkage applications and to the
neighborhood scale for human exposure applications.
3.2 CMAQ Aerosol Module
Atmospheric PM is linked with acute and chronic
health effects, visibility degradation, acid and nutrient
deposition, and climate change. Accurate predictions of
the PM mass concentration, composition, and size
distribution are necessary for assessing the potential
impacts of future air quality regulations and future
climate on these health and environmental outcomes.
The objective of this research is to improve predictions of
PM mass concentrations and chemical composition
(Figure 3-2) by advancing the scientific algorithms,
computational efficiency, and numerical stability of the
CMAQ aerosol module.
To achieve this objective, we have focused efforts in
five areas to improve previous versions of the CMAQ
aerosol module were deficient. First, we doubled the
computational efficiency of the aerosol module by
improving the computations of coagulation coefficients
and secondary organic aerosol (SOA) partitioning.
Second, we worked with the developer of ISORROPIA,
CMAQ's thermodynamic partitioning module for inorganic
species, to smooth out discontinuities. Third, we
developed a new parameterization of the heterogeneous
hydrolysis of dinitrogen pentoxide (N2O5) as part of a
larger effort to mitigate model overpredictions of
wintertime nitrate aerosol concentrations. Fourth, we
vastly improved the treatment of SOA by incorporating
several new SOA precursors and formation pathways.
Fifth, we implemented an efficient scheme to treat the
dynamic interactions between inorganic gases and the
coarse PM mode.
•Trimodal size distribution
• Fine modes in equilibrium w.gas
•Coarse modes: dynamic transfer
• Fine modes coagulate
SVGCs
2 FINE MODES
Figure 3-2. Representation of PM size and composition in CMAQ v4.7.
As a result of this research, the CMAQ aerosol
module has been enhanced and greatly improved over
the past 5 years. During that time, the aerosol module
has been used for regulatory and forecasting
applications (e.g., EPA's Clean Air Interstate Rule
[CAIR], NOAA's National Centers for Environmental
Prediction) because it is scientifically credible,
computationally efficient, and numerically stable. With
the recent scientific enhancements, our clients have
increased confidence in the utility of CMAQ predictions
of PM for future regulatory applications (e.g., Renewable
Fuel Standard rulemaking). Meanwhile, the community of
CMAQ users outside EPA continues to grow rapidly.
3.3 CMAQ Gas and Aqueous Chemical
Mechanisms
An accurate characterization of atmospheric
chemistry is essential for developing reliable predictions
of the response of air pollutants to emissions changes, to
predict spatial and temporal concentrations, and to
quantify pollutant deposition. In the past, air quality
modelers have focused largely on single-pollutant
issues, but it since has become clear that it is more
appropriate to treat chemistry in an integrated,
multiphase, multipollutant manner (National Research
Council, 2004). For example, both inorganic and organic
aqueous-phase chemistry can influence formation of
10
-------
SOA through cloud processing (Carlton et al., 2006,
2007). High-NOx versus low-NOx conditions influence
ozone, SOA, and secondary toxics formation (Ng et al.,
2007; Luecken et al., 2008). Our research and
implementation program for chemical mechanisms
accounts for production of pollutants in the gas and
aqueous phase, as well as for precursors to aerosol
production, as shown in Figure 3-3.
Glyoxal
Methylglyoxal
Nitric and sulfuric acid
OH, 03, H202
Gas-phase photochemistry:
Ozone, HAPs (including mercury),
POPs, carbonyls, peroxides, etc.
Organic peroxides
Hydroxycarbonyls
Hydroxynitrates
Aromatic nitrates, etc.
In addition, the requirements for air quality modeling
also have changed: The new NAAQS for ozone and
PM25 have shifted our focus from urban-scale ozone
episodes (~7 days) to regional/continental-scale
simulations over longer time periods (1 mo to 1 year). In
addition, our chemical mechanisms must adapt quickly to
address emerging issues of high importance, such as
changing climatic conditions and the impacts of biofuels.
Aqueous chemistry:
sulfate, nitrate plus
organics
Organic material
Mn, Fe
Aerosol chemistry:
Sulfate, nitrate, SOA
Figure 3-3. Current interactions between gas, aqueous, and aerosol chemistry in CMAQ base (criteria pollutants) and
multipollutant (including HAPs) models.
The goal of our research in this area is to develop,
refine, and implement chemical mechanisms for use in
the CMAQ model to
• ensure that CMAQ and other regional models that are
used for regulatory and research purposes have
scientifically justifiable chemical representations, are
appropriate for the application being studied, and are
consistent with our most up-to-date knowledge of
atmospheric chemistry;
• ensure that interactions between gas-, aqueous-, and
particle-phase chemistries are accounted for
adequately, so that we can better predict multimedia
chemical effects of emissions changes;
• develop techniques, tools, and strategies, so that we
are able to efficiently expand current mechanisms to
predict the chemistry of additional atmospheric
pollutants that we anticipate will become important in
the future.
Our efforts to improve the chemical mechanisms in
CMAQ have resulted in more complete and up-to-date
descriptions of the important chemical pathways that
influence concentrations of the criteria pollutants ozone
and PM. These efforts are linked closely to the research
that we perform in developing the secondary organic
aerosol module. We continue to improve the base
photochemical mechanisms that drive the oxidant and
radical chemistry. Because our models are used for both
research and regulations, we constantly strive for a
balance between stability and response to new scientific
information. We partner with other EPA researchers and
outside experts to develop state-of-the-science chemistry
descriptions that we implement in CMAQ to provide more
accurate descriptions of important chemical pathways.
Table 3-1 shows the base photochemical mechanisms
currently maintained and released in CMAQ and some of
the most important species predicted by these CMAQ
mechanisms. In addition, variants of other mechanisms,
including portions of the Master Chemical Mechanism,
are being used in CMAQ by outside groups. The different
mechanisms predict slightly different values of ozone
and other gas phase species and also can affect PM
formation, as shown in Figure 3-4, where the sulfate
production pathways differ widely depending on the
particular chemical mechanism used.
Clouds cover roughly 60% of the Earth's surface,
yet aqueous phase cloud chemistry is poorly understood
and not well characterized in atmospheric models.
Recently, CMAQ's aqueous chemistry was expanded to
11
-------
Mechanism
CB05
SAPRC-99
SAPRC07
RACM2
CB4
Notes
Standard with chlorine
chemistry, used for regulatory
application
Used for research applications
Customized version with
chlorine chemistry, in testing
phase
Customized version, currently
in testing phase
To be phased out in 201 1
•elease of CMAQ
Table 3-1. Base Photochemical Mechanisms in CMAQ and the
Species Commonly Predicted by Each Mechanism
Major Species Predicted in CMAQ
Ozone (O3)
Nitrogen oxide and nitrogen dioxide (NO and NO2)
Other oxidized nitrogen (PAN, MONO, N2O5, and organic
nitrates)
Fine and coarse participate matter (PM2s and PM10)
Sulfur dioxide and sulfate (SO2 and SO4)
Nitric acid and nitrate (HNO3and NO3.)
Hydrogen peroxide (H2C>2)
Carbon monoxide (CO)
Biogenic VOCs (isoprene, pinene, and sesquiterpenes)
Aromatic compounds (benzene, xlenes, and toluence)
Radicals (such as OH, HO2, and NO3)
Number of gas phase species: 86 (CB05), 94 (SAPRC-99)
Number of aerosol species: 75
SCr^g/m3)
6 8 10
SO/ftigAn3)
Figure 3-4. Comparison of average modeled (bars) vertical profiles of sulfate with NOAA WP-3D aircraft measurements
(black line; July-August 2004).
include cloud production of SOA via two in-cloud organic
reactions: (1) glyoxal with hydroxyl radical (OH) and
(2) methylglyoxal with OH.
The cloud processing hypothesis for SOA formation
is that water-soluble oxidation products of reactive
organic compounds partition into cloud droplets, oxidize
further, and create low-volatility compounds that remain,
in part, in the particle phase on droplet evaporation
(>90% of cloud droplets evaporate).
When SOA formation from these organic aqueous
phase reactions was added, CMAQ model performance
for particulate organic carbon (OC) improved. This is
most noticeable when comparing the vertical profile of
CMAQ-predicted OC with WSOC measurements from a
NOAA P3 "cloud experiment" flight during the
International Consortium for Atmospheric Research on
Transport and Transformation (ICARTT) in 2004, as
shown in Figure 3-5.
The inclusion of chlorine reactions and the explicit
chemistry for 43 Hazardous Air Pollutants (HAPs) has
helped to expand the applications for which CMAQ can
be used. More detail on the HAPs portion of the CMAQ
mechanism can be found in section 2.8 of this chapter.
The inclusion of additional chemical detail in the aqueous
and aerosol modules is providing pathways for more
complete descriptions of secondary organic aerosol
formation and decay.
Future Directions
Because atmospheric chemistry is central to air
quality models, our future efforts in atmospheric
chemistry mechanisms will continue to evolve and fully
employ our expertise in gas, aqueous, and aerosol
chemistry. Future efforts will involve reducing known
uncertainties in current chemical mechanisms and
improving gas-aerosol-aqueous chemistry linkages.
We will continue to monitor in-house and external
research in atmospheric chemistry, toxic air pollutants,
aerosol formation, and aqueous chemistry. We will
assess the robustness and importance of new
discoveries, and partner with leading researchers to
direct research in areas that will provide the greatest
improvements in air quality model predictions. We will
modify the mechanisms to include new information (such
as new reactions) to keep our mechanisms at state of
the science.
12
-------
o
o
o
in
o
o
o
o
o
0
CO
J3 o
<] o
o
C\J
o
o
o
o -
X--
0.1
0.2 0.5
OC
1.0
2.0
Figure 3-5. Layer-averaged vertical profiles of OC and WSOC on August 14, 2004. Normalized mean bias for layer-average
values for this flight was reduced from -65% to -15% when SOAcid was included. Note: Dashed line and "x" indicate layer-
averaged base CMAQ OC prediction. Solid line and "o" indicate CMAQ OC prediction with cloud-produced SOA included.
WSOC observations from the NOAA P3 flight are indicated with "A". The x-axis is log scale. (Adapted from Carlton et al.
[2008])
We also anticipate that our future efforts will involve
extending the chemistry beyond "traditional" pollutants to
address newly emerging issues such as biofuels,
pesticides, and chemicals that contribute to global
warming.
3.4 Planetary Boundary Layer Modeling for
Meteorology and Air Quality
Air quality modeling systems are essential tools for
air quality regulation and research. These systems are
based on Eulerian grid models for both meteorology and
atmospheric chemistry and transport. They are 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 planetary
boundary layer (PEL), which extends from the ground up
to ~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 convective 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 are effective
in the convective boundary layer (CBL). Both small-scale
turbulence that is subgrid-scale in most mesoscale grid
models and large-scale turbulence extending to the
depth of the CBL are important for vertical transport of
atmospheric properties and chemical species. Eddy
13
-------
diffusion schemes assume that all of the turbulence is
subgrid-scale and, therefore, cannot simulate convective
conditions realistically. Simple nonlocal-closure PEL
models, such as the Blackadar convective model, which
has been a mainstay PEL option in NCAR's mesoscale
model (MM5) for many years, and the original
Asymmetric Convective Model (ACM), also an option in
MM5, represent large-scale transport driven by
convective plumes but neglect small-scale, subgrid-scale
turbulent mixing. A new version of the ACM (ACM2) has
been developed that includes the nonlocal scheme of the
original ACM combined with an eddy diffusion scheme.
Thus, ACM2 can represent both the super-grid-scale and
subgrid-scale components of turbulent transport in the
CBL. Testing ACM2 in one-dimensional form and
comparing to large-eddy simulations (LES) and field data
from the second and third Global Energy and Water
Cycle Experiment Atmospheric Boundary Layer Study,
known as the GABLS2 (CASES-99) and GABLS3
(Cabauw, The Netherlands) experiments demonstrate
that the new scheme accurately simulates PEL heights,
profiles effluxes and mean quantities, and surface-level
values. ACM2 performs equally well for both
meteorological parameters (e.g., potential temperature,
moisture variables, winds) and trace chemical
concentrations, which is an advantage over eddy
diffusion models that include a nonlocal term in the form
of a gradient adjustment.
ACM2 is in the latest releases of the WRF model
and the CMAQ model and is being used extensively by
the air quality and research communities. Comparisons
to data from the TexAQS II field experiment show good
agreement with PEL heights derived from radar wind
profilers and vertical profiles of both meteorological and
chemical quantities measured by aircraft spirals.
3.5 Multiscale Meteorological Modeling for Air
Quality
Air quality models require accurate representations
of airflow and dispersion, cloud properties, radiative
fluxes, temperature and humidity fields, boundary layer
evolution and mixing, and surface fluxes of both
meteorological quantities (heat, moisture, and
momentum) and chemical species (dry deposition and
evasion). Thus, meteorological models are critical
components of the air quality modeling systems that
evolve with the state of science. Because of this
evolution, there is a need to frequently challenge our
established models and configurations; this includes
examining not only new physics schemes but also data
assimilation strategies, which serve to lower uncertainty
in air quality model output. It is also necessary to
develop and refine physical process components in the
models to address new and emerging research issues.
Each of these research objectives has the overarching
goal to improve meteorological model simulations to
ultimately reduce uncertainty in air quality simulations.
Our meteorology modeling research program involves
several key projects that have led to improved
meteorological fields. The first is the transition from the
MM5 mesoscale model system to the WRF model that
represents the current state of science. Part of this effort
was to implement in WRF the land-surface (Pleim-Xiu
[PX]), surface-layer (Pleim), and PEL (ACM2) schemes
that have been used in MM5 and are designed for
retrospective air quality simulations. Part of this effort
included improving the PX land-surface physics that
included a deep-soil-nudging algorithm and snow cover
physics that dramatically improved temperature
estimations in the winter simulations and in areas with
less vegetation coverage. An additional effort was to
work toward implementing, in WRF, the nudging-based
4D data assimilation (FDDA) capability that had been
available in MM5. Another effort has been a
reexamination of FDDA techniques, including the use of
an objective reanalysis package for WRF (OBS-GRID) to
lower the error of analyses that are used to nudge the
model toward the observed state. RAWINS was the
equivalent package used by MM5.
Current results of the implementation of new
physics in WRF show that our configuration is
comparable to or exceeds the level of MM5 in terms of
the uncertainty or error in near-surface variables like 2-m
temperature, 2-m moisture, and 10-m wind as indicated
in Table 3-2. This is true only when the new analysis
package is used to improve analyses used for FDDA and
soil moisture and temperature nudging in WRF.
Figure 3-7 shows error differences between WRF and
MM5, where both models were configured as similar as
possible (i.e., PX land surface model [LSM], ACM2 PEL,
etc.). The large number of dark blue and purple areas
indicate WRF has a much lower temperature error than
MM5. In Table 3-2 and Figure 3-9, PXACM is the
simulation that used the PX LSM and ACM2 PEL
scheme, whereas the terminology NOAHYSU indicates
the simulation that used NOAA's land surface model
(NOAH) LSM and Yonsei University (YSU) PEL scheme.
A new evaluation method that utilizes both wind
profiler and aircraft profile measurements provides a
routine method to examine not only the uncertainty of
simulated wind in the PEL but also the less examined
temperature structure. The WRF model has low error in
temperature (median absolute error of 1.0 to 1.5 K or
less), wind speed (<2.0 m/s), and wind direction
(<30 deg) in the PEL, which is generally less than the
error near the surface (Figure 3-8). The model also
simulates the evolution of the wind structure, including
features like nocturnal jets and the convective mixed
layer (see Figure 3-9), with low error (<2.0 m s"1). Our
current configuration of WRF has met the requirements
for the transition from MM5.
14
-------
Table 3-2. Summary of Surface-Based Model Performance Statistics for Each Simulation
(Also provided is the RMSE [2-m temperature only] of analysis dataset that was used for the
indirect soil moisture and temperature nudging of the PX LSM.)
RMSE
2-m Temperature (K)
January
August
2-m Mixing ration (g kg"1)
January
August
10-m Wind speed (m s"n)
January
August
10-m Wind direction (deg)
January (MAE)
August (MAE)
WRF
PXACM
2.48
1.94
0.92
1.86
1.64
1.47
21
30
MM5
PXACM
2.52
2.00
0.84
1.92
1.79
1.49
25
33
WRF
NOAHYSU
2.33
2.31
0.78
2.11
1.78
1.60
23
32
Obsgrid
Analysis
1.29
1.22
RAWINS
Analysis
1.47
1.31
P3-Tliet?
WRF-ARW-Thsta
t 2 3 4 5 6 7 29ft 300 302 M4 3W 308 310 0
Oft (g/kg) Thela
23
NGf-new
Figure 3-6. The most direct measure of success for a PBL model for both meteorology and air quality is its ability to
accurately simulate the vertical structure of both meteorological and chemical species. The figure above shows an
example of WRF and CMAQ profiles (both use the ACM2 scheme) compared with aircraft measurements. The top of the
PBL mixed layer is well defined and modeled for both meteorology variables (Qv and Theta) and chemical variables (NOy).
Although such simultaneous measurements of vertical profiles of meteorology and chemistry are very rare, these limited
results are encouraging.
3.6 Coupled WRF-CMAQ Modeling System
Although the role of long-lived greenhouse gases in
modulating the Earth's radiative budget long has been
recognized, it now is acknowledged widely that the
increased tropospheric loading of aerosols also can
affect climate in multiple ways. Aerosols can provide a
cooling effect by enhancing reflection of solar radiation,
both directly (by scattering light in clear air) and indirectly
(by increasing the reflectivity of clouds). On the other
hand, organic aerosols and soot absorb radiation, thus
warming the atmosphere. Current estimates of aerosol
radiative forcing are quite uncertain. The major sources
of this uncertainty are related to the characterization of
atmospheric loading of aerosols, the chemical
composition and source attribution of which are highly
variable both spatially and temporally. Unlike
greenhouse gases, the aerosol radiative forcing is
spatially heterogeneous and estimated to play a
15
-------
Figure 3-7. Spatially distributed root mean square error (RMSE) difference (2-m temperature) between the WRF and MM5
for August 2006. Negative values indicate WRF has a lower error, and positive values indicate MM5 has a lower error.
0
0_
o
0
o"
o
10
0
0-
KD-
h rnrf
I--CTI •--<
• T~
1
I
M^^M
|
• LD <
i ...{]
1 i •
l f. (
1
,rr-}....|
1 i l
Aug 2006^ '
o
r
Op
8-
*~
8-
«.
8-
0.6 1.3 1.4 1.8
>"
t-
,.
H
I--
H
H
!•••
D-
— i. , .
]•••••< •
>.....
3... ,
T— • i »
D" ' •
33 -'
1 1-1
' CD-1
i.,.jMn,...i
t-'OD1 •
h.rfy^i
,..fn -'
Aug 2006
1 3 5
c
1"
0
o.
c
c-
8-
".aa'-.
•••ID -
'•CD <*
1
>-fn •
l"l]-#
f[J].«l •
"Jj <
'o--^
>-rj~i • •
'-n — f-H
-azi} •
>-H 1 ~|-
-az •
i-az
Aug 2006
10 30 50 70
Absolute Error (K)
Absolute Error im s")
Absolute Error (Deg.)
Figure 3-8. Mean absolute error (MAE) profiles of model-simulated temperature, wind speed, and wind direction for
August 2006. The observations used to compute MAE include 19 NOAA wind profilers located in the central United
States.
16
-------
Wind Profilers (August 2006)
" WRF PXACM (August 2006)
Time of Day (UTC]
tvrs 'Jvic
Time of Day (UTCJ
Figure 3-9. Diurnal mean wind speed profiles (height above ground level) for January and August 2006. The left column
represents the mean observed wind speed computed using 19 NOAA wind profilers located in the central United States.
The right column is the corresponding model-simulated mean wind speed using the grid points closest to the wind
profiler sites. Small dots indicate the mean PBL height of the WRF.
significant role in regional climate trends. The accurate
regional characterization of the aerosol composition and
size distribution is critical for estimating their optical and
radiative properties and, thus, for quantifying their
impacts on radiation budgets of the Earth-atmosphere
system.
Traditionally, atmospheric chemistry-transport and
meteorology models have been applied in an off-line
paradigm, in which archived output describing the
atmosphere's dynamical state, as simulated by the
meteorology model, is used to drive the transport and
chemistry calculations of the atmospheric chemistry-
transport model. A modeling framework that facilitates
coupled online calculations is desirable because it
(1) provides consistent treatment of dynamical processes
and reduces redundant calculations; (2) provides the
ability to couple dynamical and chemical calculations at
finer time steps and, thus, facilitates consistent use of
data; (3) reduces the disk-storage requirements typically
associated with off-line applications; and (4) provides
opportunities to represent and assess the potentially
important radiative effects of pollutant loading on
simulated dynamical features. To address the needs of
emerging assessments for air quality-climate interactions
and for finer scale air quality applications, AMAD recently
began developing a coupled atmospheric dynamics-
chemistry model: the two-way coupled WRF-CMAQ
modeling system. In the prototype of this system, careful
consideration has been given to its structural attributes to
ensure that it can evolve to address the increasingly
complex problems facing the Agency. The system design
is flexible regarding the frequency of data communication
between the two models and can accommodate both
coupled and uncoupled modeling paradigms. This
approach also mitigates the need to maintain separate
versions of the models for online and off-line modeling.
In the prototype coupled WRF-CMAQ system, the
simulated aerosol composition and size distribution are
used to estimate the optical properties of aerosols, which
then are used in the WRF radiation calculations. Thus,
the direct radiative effects of absorbing and scattering
tropospheric aerosols estimated from the spatially and
temporally varying simulated aerosol distribution can be
fed back to the WRF radiation calculations as
demonstrated in Figure 3-10. This results in a "two-way"
coupling between the atmospheric dynamical and
chemical modeling components. This extended capability
provides unique opportunities to systematically
investigate how atmospheric loading of radiatively
important trace species affects the Earth's radiation
budget. Consequently, this modeling system is expected
to play a critical role in the Agency's evolving research
and regulatory applications exploring air quality-climate
interactions.
17
-------
Case 1: Eastern U.S., August 2-11, 2QG6, 12 km resolution
Surface PWb
Aerosol Optical Depth
Optical
properties of
aerasofe
Reduction In PEL
Reduction in shortwave radisficn
reaching the surface in regions of
aerosol owling
10DO
• Observed
• vtfo Feedback
' wilh Feedback
10
Bondville • August 6, 200S
12 14 15 18 20 22
Time (UTC)
24
Figure 3-10. Two sets of initial simulations have been conducted to test the evolving coupled WRF-CMAQ modeling
system and to systematically assess the impacts of coupling and feedbacks. The upper panels in the figure above
demostrate the impact that aerosols estimated by CMAQ have on the meteorological models' estimates of planetary
boundary layer (PBL) height and downward shortwave radiation. The lower panel of the figure above is verification that
the simulation, which includes these feedbacks, agrees better with the observed shortwave radiation.
3.7 Mercury Modeling
AMAD has been working on the development of
atmospheric mercury models since the early 1990s,
when the Regional Lagrangian Model of Air Pollution
(RELMAP) was adapted to simulate mercury in support
of EPA's Mercury Study Report to Congress. As the
scientific understanding of atmospheric mercury
continued to develop in the late 1990s, it became
apparent that Lagrangian-type models, also known as
"puff models, would have difficulties simulating the
complex chemical and physical interactions of mercury
with other pollutants that were being discovered. Thus,
18
-------
AMAD's focus for atmospheric mercury model
development was moved to the CMAQ model. That
model simulates atmospheric processes within a
3-D array of predefined finite volume elements and can
model complex interactions between all of the pollutants
that might exist within each volume element. The CMAQ
model was developed to simulate photochemical
oxidants, acidic and nutrient pollutants, and aerosol PM,
all of which have been shown to interact with mercury in
air and in cloud water and influence its deposition to
sensitive aquatic ecosystems. The "multipollutant"
approach of CMAQ, where all pollutants are simulated
together just as they exist in the real atmosphere, is
applied in atmospheric mercury modeling at AMAD.
A number of modifications were made to the
standard CMAQ model to enable it to simulate
atmospheric mercury; these are described in detail in
Bullock and Brehme (2002). Because new information
about chemical and physical processes affecting
atmospheric mercury continually is being published,
refinement of the model code is an ongoing process.
Further modification of the CMAQ-Hg chemical
mechanisms for mercury in both the gaseous and
aqueous phases is expected as additional chemical
reactions are identified and studied. The latest public
release of CMAQ provides the ability to simulate
atmospheric mercury in the multipollutant version of the
model. We found this to be the most efficient way to
maintain and disseminate mercury simulation capabilities
in CMAQ because of the increasing number of pollutants
with which mercury is known to react.
AMAD has participated in two major model
intercomparison studies for atmospheric mercury. The
first was the Intercomparison of Numerical Models for
Long-Range Atmospheric Transport of Mercury,
sponsored by the European Monitoring and Evaluation
Programme (EMEP) and organized by EMEP's
Meteorological Synthesizing Center-East in Moscow,
Russia. The first phase of this EMEP study involved the
simulation of mercury chemistry in a closed cloud volume
given a variety of initial conditions, and results were
published in Ryaboshapko et al. (2002). The second
phase of the study involved full-scale model simulations
of the emission, transport, transformation, and deposition
of mercury over Europe for two short periods of 10 to
14 days each. Model simulations were compared to field
measurements of elemental mercury gas, reactive
gaseous mercury, and particulate mercury in air. Results
from this phase of the study were reported in
Ryaboshapko et al. (2007a). The third and final phase of
the EMEP intercomparison involved model simulations
for longer periods of time (up to 1 year) and comparisons
to observations of the wet deposition of mercury. Results
from this phase of the study were reported in
Ryaboshapko et al. (2007b).
As the EMEP study was nearing completion, AMAD
organized a second mercury model intercomparison
study, this time with a focus on North America. The North
American Mercury Model Intercomparison Study
(NAMMIS) took advantage of standardized weekly wet
deposition samples taken by the Mercury Deposition
Network (MDN) as described in Vermette et al. (1995)
and separate event-based precipitation samples taken at
Underhill, VT (Keeleret al., 2005). In addition to CMAQ,
two other regional models were tested in the NAMMIS;
the Regional Modeling System for Aerosols and
Deposition (REMSAD), and the Trace Element Analysis
Model (TEAM). All three models were each applied to
simulate the entire year of 2001 three times, each time
using initial condition and boundary conditions developed
from a different global model. The NAMMIS provided not
only a comparison between regional atmospheric
mercury models but also a measure of the sensitivity of
each regional model to uncertainties regarding
intercontinental transport. The NAMMIS evaluated each
regional model for its agreement to observations of wet
deposition of mercury from the MDN and Underhill
observations. Results from the NAMMIS statistical model
evaluation are shown in the Table 3-3. For most of the
evaluation metrics, CMAQ was found to have superior
agreement to the observations.
Results from all three of the regional models tested
(CMAQ, REMSAD, and TEAM) varied depending on the
global model used to define lateral boundary
concentrations for mercury. These global models
included Chemical Tansport Model for Mercury
(CTM-Hg), Goddard Earth Observing System's
Chemistry (GEOS-Chem) model, and the
Global/Regional Atmospheric Heavy Metals (GRAHM)
model. All of the regional models used meteorological
data provided by the MM5 model. Statistics for the
precipitation data obtained from MM5 also are shown in
the table. Obviously, the level of accuracy one can
expect from the regional air quality models is limited by
the accuracy of the input precipitation data. It does
appear that the best performing air quality simulations
have about the same level of accuracy as the
precipitation data provided to those simulations. Thus, it
can be reasoned that significant improvements in the
simulation of mercury wet deposition are contingent on
improvements in the modeling of physical meteorology.
Complete descriptions of the NAMMIS study design,
participating models, and modeling results are available
in two articles published in the Journal of Geophysical
Research (Bullock et al., 2008; Bullock et al., 2009).
CMAQ mercury modeling capabilities have been
applied to support various EPA regulatory actions for
mercury. They also have been used to provide
information regarding mercury deposition from global
background concentrations to tribal, State, and regional
environmental authorities in the development of their
water quality protection strategies. EPA currently is
working with the United Nations Environment Program
toward the development of international treaties to
reduce mercury emissions to the environment. AMAD is
working to expand the CMAQ modeling domain to cover
the Northern Hemisphere. This will allow CMAQ to
provide modeling assessments of the intercontinental
transport of mercury and its importance as a global
pollutant.
19
-------
Table 3-3. Evaluation Statistics from the North American Mercury Model Intercomparison Study
Model
Test Case
r"
Mean Bias
(ng m"2)
Mean
Normalized
Bias
Mean
Normalized
Error
Normalized
Mean Bias
Normalized
Mean Error
Mean
Fractional
Bias
Mean
Fractional
Error
CMAQ
CTM
0.15
-12.2
178.1
1.580
-0.049
0.708
0.142
0.725
GEOS-
Chem
0.12
46.9
213.0
2.031
0.187
0.847
0.279
0.771
GRAHM
0.14
40.2
207.0
2.020
0.160
0.823
0.247
0.771
REMSAD
CTM
0.16
67.8
226.0
2.292
0.270
.899
0.167
0.839
GEOS-
Chem
0.15
10.2
248.7
2.593
0.399
0.989
0.242
0.861
GRAHM
0.16
41.3
213.8
2.133
0.164
0.850
0.103
0.835
TEAM
CTM
0.14
164.2
278.8
3.298
0.653
1.109
0.602
0.885
GEOS-
Chem
0.12
220.3
326.3
3.804
0.876
1.298
0.670
0.928
GRAHM
0.18
155.2
264.8
3.091
0.617
1.053
0.593
0.867
MM%
Precip.
0.35
1.9
(mm)
15.3
(mm)
1.681
0.078
0.620
0.098
0.641
3.8 CMAQ for Air Toxics and Multipollutant
Modeling
In the past, chemical mechanism and air quality
development have focused on ozone and primary
inorganic PM, we are expanding the scope of the
atmospheric photochemistry in CMAQ to include
predictions for a large number of HAPs. More information
on air toxics and EPA's important role in identifying and
mitigating high concentrations of air toxics can be found
at EPA's air toxics Web site
(http://www.epa.gov/ttn/atw/index.html).
We build on the base photochemical mechanisms in
CMAQ by adding explicit chemical characterizations for
HAPs. The multipollutant version of CMAQ (CMAQ-MP)
currently predicts the 44 individual HAPs shown in
Table 3-4.
In addition to HAPs explicitly listed in the CAA
Section 112(b), research versions of CMAQ have been
modified to model additional, potentially toxic compounds
that are emerging pollutants, such as pesticides (dioxin),
herbicides (atrazine), and hydrofluorocarbons
(tetrafluoropropene).
In CMAQ-MP, the chemistry was harmonized with
the regulatory model for ozone and PM2 5, allowing the
Agency to analyze simultaneous effects of emission
control strategies on all high-priority pollutants. This
chemistry accounts for interactions and feedbacks
between multiple pollutants, which would not be possible
in separate simulations. CMAQ-MP provides a tool that
can be used to help answer the following questions.
• What tools can we provide to help the Agency to
evaluate the true overall effects of an emission control
strategy, and, therefore, develop strategies that
optimize human and ecological health?
• How do we ensure that the chemistry that is used in
regulatory and research models is rigorous and state
of the science?
• How do changes in one air pollutant affect other
pollutants?
• What is the best way to incorporate flexibility into the
chemistry, so that the Agency can quickly respond to
emerging issues and new atmospheric pollutants?
Two examples of output from one multipollutant
modeling simulation are shown in Figure 3-11.
Future Directions
Because chemistry impacts every component of air
quality models, our future efforts in atmospheric
chemistry mechanisms will continue to evolve and fully
employ our expertise in gas, aqueous, and aerosol
chemistry. Future efforts will involve reducing known
uncertainties in current chemical mechanisms and
improving gas-aerosol-aqueous chemistry linkages.
We will continue to monitor internal and external
research in atmospheric chemistry, toxic air pollutants,
aerosol formation, and aqueous chemistry. We will
assess the robustness and importance of new
discoveries and partner with leading researchers to
direct research in areas that will provide the greatest
improvements in air quality model predictions. We will
modify the mechanisms to include new information (such
as new reactions) to keep our mechanisms at state of
the science.
We also anticipate that our future efforts will involve
extending the chemistry beyond "traditional" pollutants to
20
-------
Table 3-4. Hazardous Air Pollutants Represented in the Current CMAQ Multipollutant Model
Gas-Phase HAPS
Formaldehyde
1,3-butadiene
Naphthalene
Acrolein
Acetaldehyde
1,3-Dichloropropene
Quinoline
Vinyl chloride
Acrylonitrile
Trichlorethylene
Benzene
1,2-Dichloropropane
Ethylene oxide
1 ,2-Dibromoethane
1,2-Dichloroethane
Tetrachlo^oethylene
Carbon tetrachloride
Dichloromethane
1 ,1 ,2,2-Tetrachloroethane
Chloroform
2,4-Toluene diisocyanaate
Hexamethylene 1-6-diisocyanate
Maleic anhydride
Triethylamine
Chlorine
Hydrazine
Hydrochloric acid
p-Dichloro benzene
Xylene (o,m, and p explicitly)
Toluene
Methanol
Multiphase and Aerosol HAPS
Diesel PM
Beryllium compounds
Cadmium compounds
Lead
Manganese compounds
Nickel compounds
Chromium 3
Chromium 6
Elemental mercury
Reactive gaseous mercury
Particulate mercury
eon
60,0
40 JO
20.0
PpbV
OJ)
4.0
1JJ
1«ppbV
0,0
Figure 3-11. CMAQ multipollutant model predictions for ozone (left), as maximum 8-h value, and formaldehyde (right), as
monthly average for July 2002.
address new, emerging issues, such as biofuels,
pesticides, and chemicals that contribute to global
warming.
3.9 Emissions Modeling Research
Emission data input is one of the principal drivers of
the CMAQ modeling system. However, estimates of
emissions data are subject to a large degree of
uncertainty, as noted in the NARSTO (formerly the North
American Research Strategy forTropospheric Ozone)
Emission Inventory Assessment
(http://www.narsto.org/section.src?SID=8), particularly
for precursors of airborne fine PM and for sources of
organic and elemental carbon (EC) and ammonia. Most
anthropogenic emissions used in the CMAQ system are
available from EPA's National Emissions Inventory (NEI)
(http://www.epa.gov/ttn/chief/eiinformation.html). AMAD
focuses on the evaluation and improvement of emission
categories that respond to meteorology and/or that are
natural or quasi-natural in character, and that are not
readily available in the NEI. Our work includes the
development, evaluation, and implementation of
emission models for biomass burning, fugitive dust,
lightning, and biogenic sources. These sources emit
ozone precursors (volatile organic compounds [VOCs]
and nitrogen oxides), PM, and some air toxins.
After working with EPA's OAQPS to release an
operational satellite-based biomass burning emission
estimation system for the NEI, the Division focused on
evaluating the emissions from this system in the context
of air quality modeling and in working with other
researchers in improving areas of greatest uncertainty.
We continued to compare emissions from alternative
methodologies and to evaluate CMAQ model
performance with these alternative emissions, and we
began to collaborate with the NPS to compare carbon
aerosol concentrations from two different air quality
modeling systems with IMPROVE measurement data.
Collaborations with NASA (as well as with researchers at
Michigan Tech and the University of Kentucky) also
21
-------
began under a NASA-funded grant. Objectives of the
NASA research include evaluation of plume rise and
refinement of rangeland/cropland biomass burning
emission estimates.
The 2011 release of CMAQ is scheduled to offer two
alternatives for biogenic emissions: NCAR's Model for
Emissions of Gases and Aerosols from Nature (MEGAN)
and Biogenic Emissions Inventory System (BEIS) v3.14.
Both models are now being tested in CMAQ. In concert
with the Division's ecosystems-related research, we
worked with UNC's Institute for the Environment (IE) to
incorporate updated agricultural data and information
from EPA and the U.S. Geological Survey's (USGS's)
30-m National Land Cover Database (NLCD). We then
worked further with UNC-IE to design a plan for
incorporating updated forest inventory data and for
possibly harmonizing the vegetation cover data in
MEGAN with that in BEIS v3.14. Under a NASA-funded
grant with the University of Maryland, we collaborated on
the use of satellite imagery to evaluate soil NOX
emissions.
Via a collaboration with NASA and the University of
Maryland, we 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. Early results indicate that the
NOX 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.
The Division continued to interact with 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). In addition, we began to assess
the use of alternative temporal profiles for computing
fugitive dust emissions to possibly correct for temporal
biases that have been observed with urban PM2 5
measurement data.
The Division is continuing to work with other
partners in EPA to improve the SPECIATE database,
which is central to speciating VOC and PM gas and
aerosols for emissions used in the CMAQ modeling
system.
In future years, the Division's priorities in emissions
research will be on improving and evaluating
components of the emission modeling system used in
CMAQ and where other organizations, such as OAQPS,
are unable to provide support. Where resources permit,
we will improve the scientific content, accuracy, and
efficiency of emission models that are required for the
development, testing, and evaluation of the CMAQ
modeling system.
Future Directions
The Division's research is organized around several
model evaluation studies addressing ozone 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 of PM (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
Sparse Matrix Operator Kernel Emission (SMOKE)
modeling system, which processes emissions data for
CMAQ. All of the planned emissions research directly
supports the major release of CMAQ in 2011.
Biomass burning emissions. 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 are planned to examine whether
improvements in the fire emission estimation methods
will improve air quality model simulations. Figure 3-12 is
an example of biomass burning emissions. We plan to
prepare one or more publications forsubmittal to a peer-
reviewed journal related to this effort.
Annual Average PM; $ Wildland Fire Emission Density
(2006 - 2008)
tons per square mile
Figure 3-12. AMAD's research contributed to the NEI's
Wildfire Emissions Inventory. (Plot courtesy of S. Raffuse,
STI, Inc.)
We also plan to continue our collaboration with
scientists at NASA in Langley, VA, as well as with
NERL's Environmental Sciences Division, to evaluate
and possibly improve plume rise estimates for biomass
22
-------
burning events and to improve temporal/spatial
estimates of rangeland/cropland burn emissions.
Biogenic emission modeling. Biogenic emission
estimates can strongly affect the assessment of the
anthropogenic activities on tropospheric chemistry. Yet,
large uncertainties persist in biogenic emission
estimates. Figure 3-13 shows the differences of isoprene
between MEGAN and BEIS, for example. We plan to
continue work with EPA's 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 peer-reviewed journal
to report our findings and recommendations. We intend
to include MEGAN and BEIS v3.14 in the 2011 release
of CMAQ.
2003 Annual BEIS Emissions
Isoprene
2003 Annual MEGAN Emissions
Isoprene
12500
10000
7500
5000
2500
0
Mg/yr
Min= Oat(1,1), Max=13G48 at(20,49)
Min= 0 at (1,1), Max=25524 at (85,45)
Figure 3-13. Comparison of isoprene emissions estimated by BEIS and MEGAN.
Working with scientists at the University of North
Carolina, we will continue to explore updates of the
vegetation landcoverwith the 30-m resolved land cover
classes in the EPA/USGS NLCD. During 2011, we plan
to focus on collaborating with NCAR via the UNC
contract to harmonize the vegetation cover datasets in
MEGAN and BEIS. Time and resources permitting, we
will include an updated vegetation cover dataset in BEIS
for the 2011 release of CMAQ—but, at the time of this
writing, achieving this goal appears to be a challenge.
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. As of the winter 2009/2010, NASA has
provided the Division with initial estimates, so that we
can perform testing with CMAQ. NASA has indicated that
a draft journal article on this work is in preparation. We
plan to incorporate an online version of the lightning NOX
algorithm in the 2011 release of CMAQ.
Geogenic dust. Depending on curability and the
time available to interact with NOAA's air quality
modeling forecast research team, a publication will be
prepared and an algorithm for improved estimates of
fugitive dust will 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 are
also testing an in-line windblown fugitive dust emission
algorithm in the CMAQ code. Accelerated progress in
this area, particularly to support hemispheric and/or
global climate research, may require allocation of
additional resources.
Speciation of emissions. The Division plans to
continue to champion improvements in the speciation of
VOCs and PM. This work will be accomplished largely
through collaborative work with NRMRL and OAQPS.
Meanwhile, scientists in the Division will attempt to use
the CMAQ modeling system to assess the contributions
from and the uncertainties of various aspects of the NEI.
An emissions inventory of fine-particulate trace elements
(e.g., calcium, iron, silver, tin, antimony, etc.) has been
developed using the 2001 NEI in combination with
emission profiles in the SPECIATE v4.0 database. This
inventory is now being evaluated against trace-elemental
measurements collected at urban sites in the Speciated
Trends Network (STN). The inventory will be refined as
necessary and then used as input to the CMAQ source-
apportionment model to compute atmospheric
concentrations of various trace elements in PM25. These
modeled concentrations will be compared against
corresponding measurements taken across the major
monitoring networks (e.g., IMPROVE, STN, SEARCH,
and NADP).
23
-------
Fairbanks, Alaska. Based on raw emissions stable, wintertime conditions in Alaska. This effort will
information supported under a contract by EPA require innovative approaches with different source
Region 10, we plan to assess and integrate emissions categories and at fine vertical resolution
for fine-scale CMAQ modeling of fine particulates during
24
-------
CHAPTER 4
Air Quality Model Evaluation
4.1 Introduction
To ensure that we provide quality products to
regulatory, academic, and other end users, we conduct
extensive evaluation studies to rigorously assess air
quality model performance in simulating the
spatio-temporal features embedded in the air quality
observations. We comprehensively analyze the
performance of meteorology, emissions, and chemical
transport models to not only characterize model
performance but also identify what model improvements
(inputs or processes) are needed. Thus, model
evaluation efforts are tied directly with model
development.
The Division has developed a framework (Dennis
et al., 2010) to classify the different aspects of model
evaluation under four general categories: (1) operational,
(2) diagnostic, (3) dynamic, and (4) probabilistic.
(a)
i .:•••
=•_•
=>:•
Jo
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. Diagnostic evaluation
investigates the processes and input drivers that affect
model performance. 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. 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." Because these four types of model
evaluations are not necessarily mutually exclusive,
research studies often incorporate aspects from more
than one category of evaluation.
K
(b)
• Other
D TC
U NHf
• NO/
D SO,,2
20-,
i -is-
JJ
% r
c 5-
o
(J
n_
Northeast
Atlantic
22%
c
45V, o
15-
10-
r.-
STN
"H ••••••
i— H—i™
sjs= H^i*
M% H9V,
STN CyAQ
Figure 4-1. Model outputs are compared to observations using various techniques, including (a) time series of daily
maximum 8-h ozone concentrations from a 200-member CMAQ model ensemble at a monitoring site in an urban location
and (b) percent contribution of individual aerosol species comprising the total average regional PM2.s mass
concentrations predicted by CMAQ and measured by the Speciated Trends Network (STN) sites.
4.2 Operational Performance Evaluation of Air
Quality Model Simulations
Two of the three main components of an air quality
model (e.g., CMAQ) simulation are the input meteorology
and the air quality model simulation itself, with the third
being the input emissions. Meteorological data are
provided by models, such as MM5 and WRF. The quality
of the meteorological data, specifically how well the
predicted values (e.g., temperature, wind speed, etc.)
compare with the observed state of the atmosphere, is
critical to the performance of the air quality model, which
is highly dependent on the meteorological data to
25
-------
accurately simulate pollutants in the atmosphere. As
such, an important aspect of any air quality simulation is
the evaluation of the quality of the predicted
meteorological data. This is accomplished by comparing
model-simulated values against observed data. This type
of evaluation is referred to as operational evaluation.
A similar evaluation of the air quality model simulation is
also performed using available observed air quality
measurements.
As the developer of the CMAQ model, AMAD is
frequently evaluating CMAQ simulations as part of the
testing process as the model evolves with state-of-the-art
science. Examples of changes to the modeling system
that may require testing include updates/corrections to
the model code, changes in the model inputs (e.g.,
meteorology, emissions), and any other changes that
may impact the model predictions. As computing power
has increased (and continues to increase) overtime, the
frequency of model simulations has increased, whereas
the time required to run a simulation has decreased.
Additionally, the duration of model simulations has
increased from a week or several weeks to multiple
months and multiple years. With this increase in the
number and duration of air quality simulations comes an
increase in the time required to thoroughly evaluate each
simulation. To evaluate a simulation within a reasonable
amount of time, AMAD developed the Atmospheric
Model Evaluation Tool (AMET), which aids researchers
in evaluating the operational performance of a
meteorological or air quality simulation. A brief
description of AMET is given below.
AMET is a combination of an open-source database
software (MYSQL), the R statistics software, and
Ml b_12km_34L SOU fix August MM
: IMPflOVF .;M1 b 17km Ml)
Ł BTN (M1b_12km_34L)
Uunlhly Avenage
SO4 (ug/mS)
1 *
Mlb_1iS-n_MI-
IWROVE 1.79 -67 M.1 -56 S7.B
•"•-. 2.0S -If- :•?.?- -QS 27.7
CA3T14ai 1.22 -117 17 fl -134 141
0 5 10
Observation
Figure 4-2. Scatter plot of observed versus CMAQ-
predicted sulfate for August 2006 created by AMET.
FORTRAN and PERL scripts that, together, provide an
organized and powerful system for processing
meteorological and air quality model output and, then,
evaluating the performance of model predictions. AMET
uses FORTRAN and PERL scripts to pair observed
meteorological and air quality data with model
predictions, then populates a MYSQL relational database
with the paired data, and, finally, uses R statistics scripts
to create statistics and plots to show the operational
model performance. Many R scripts are already available
with the release version of AMET, but users familiar with
R can modify existing scripts or create new scripts to suit
their evaluation needs.
4.3 Diagnostic Evaluation of the Oxidized
Nitrogen Budget Using Space-Based, Aircraft,
and Ground Observations
Recent studies have shown that, when compared
with field observations, chemical transport models make
significant errors in the simulated partitioning of NOy
between NO2, HNO3, and PAN. This impacts the long-
range transport of ozone precursors, misrepresents the
relative effectiveness of local versus regional emission
control strategies, and distorts the spatial and temporal
distribution of nitrogen deposition. In this research, we
use a combination of modeling tools equipped with
process analysis; satellite data; aircraft observations
from the ICARTT, INTEX-NA, and TexAQS 2006 field
campaigns; and surface observations to better
understand and improve the simulated fate and transport
of oxidized nitrogen species. We are applying this
analysis to better quantify the relative impact of local
versus regional NOX emission control strategies, the
contribution of lightning NOX to atmospheric chemistry,
and the long-range transport and deposition of NOy to
remote ecosystems.
4.4 Diagnostic Evaluation of the Carbonaceous
Fine Particle System
Routine measurements of speciated PM2§ (e.g.,
IMPROVE, STN) are often insufficient to diagnose the
causes of model errors in OC concentrations because
they cannot distinguish the origin of OC between primary
versus secondary, anthropogenic versus biogenic, or
mobile sources versus area sources. Through
identification of the sources and processes contributing
the OC, the necessary improvements in the modeled
processes or emission inputs can be identified. Current
diagnostic evaluation work is listed below that will
support better understanding of the carbonaceous
aerosol system.
Estimating how much OC observed is
secondary. Routine measurements of EC and OC can
be used in conjunction with model predictions of EC and
primary OC to estimate concentrations of secondary OC
(Yu et al., 2007). These estimates can be used as a
preliminary assessment of model performance for
secondary OC.
26
-------
HN03: NO,
^—^
Ł
CD
+;
15
O
o
i"
o
o
0 ~
CO
o
o
o ~
(M
o
o
0 ~
0
o O'
\ /y
\ A
\ s
oeŁ
o uu
1
OOO
J 1 \
0,00
fBi a
0.2 0.4
O-O
observations
CMAQ-SAPRC
CMAQ-CB05
0.6 0.8 1.0
ratio
Figure 4-3. Vertical profile of the ratio of nitric acid (HNOs) to total oxidized nitrogen (NOy) as sampled during the August
8, 2004, ICARTT flight over the Northeastern United States. When the observations are paired in time and space with the
CMAQ simulations, we find that the chemical mechanisms used in CMAQ over-estimate the contribution of nitric acid to
total NOy, especially in the free troposphere.
D Slog, Secondary
D Anthrcp Secondary
r. Diner Primary
• Vsc Industrial
• Paved Road
• Food Cooking
D Naluial Gas Comb
r Oil Gomb
nCoal Comb
• Waste Co TTS
• Wild fifes
• Anthrop
L'Aircraft Exhaus!
• Nonroaa Gasoline
• Gasoline Exhaust
• Nanroad Diesel
• Dies* i
Figure 4-4. Source contributions to the modeled concentrations of fine-particulate carbon in six U.S. cities.
Primary OC predictions from different sources.
Measurements of individual organic compounds that are
specific to certain primary emission sources may be used
to evaluate model predictions of primary OC on a
source-by-source basis. Measurements of this type at
the SEARCH monitoring sites have been used to
evaluate model results during the July to August 1999
period in the southeastern United States. (Bhave et al.,
2007).
Fossil fuel versus modern carbon predictions.
Measurements of radiocarbon (14C isotope) enable one
to distinguish fossil fuel carbon (e.g., motor vehicle
exhaust, coal and oil combustion) from modern carbon
(e.g., biomass combustion, biogenic SOA). summer of
1999 (Lewis et al., 2004) are being used to evaluate
model predictions of these two types of carbon.
Tracers of anthropogenic and biogenic
secondary organic aerosol (SOA). Novel analytical
techniques for quantifying individual organic compounds
that are unique tracers of anthropogenic and biogenic
SOA have been developed by EPA scientists. These
compounds were measured at an RTP site throughout
the 2003 calendar year (Kleindienst et al., 2007) and
27
-------
have been used to evaluate recent improvements to the
CMAQ SOA module (Bhave et al., 2007).
Many of these exploratory projects are in
collaboration with scientists in NERL HEASD.
4.5 Inverse Modeling To Evaluate and Improve
Emission Estimates
Although continuously updated and improved,
emission inventories are still considered to be one of the
largest sources of uncertainty in air quality modeling. It is
often difficult to measure the emission factors, activity
information, or both for various emitting processes, such
as forest fires, animal husbandry practices, and motor
vehicles. Therefore, bottom-up inventories for such
Measurements of this type at Nashville, TN, in the
processes often are based on estimates and averages.
To complement, evaluate, and better inform
bottom-up emission inventories, we develop and apply
inverse modeling methods. These types of "top-down"
approaches employ observational data from continuously
operating pollutant measurement networks, intensive
Continental US
field campaigns, and remote sensing technologies to
infer emission inventories based on current state-of-the-
science understanding of physical and chemical
processes in the atmosphere.
In one specific application, we use the satellite-
observed NO2 column density to attempt to identify any
possible bias in the NOX emission inventories over
several regions in the southeastern United States.
Figure 4-5 shows a model comparison of satellite
observations (from SCIAMACHY retrieval) and CMAQ
prediction. This application relies on the adaptive-
iterative Kalman filter as an inverse method and
decoupled direct method in 3D (Decoupled Direct
Method [DDM]-3D) as a way to quantify the relationship
between emission rates of NOX and atmospheric
concentrations of NO2. We find that urban emissions in
Atlanta, GA, and Birmingham, AL, are likely to be
overestimated, whereas more rural concentrations of
NO2 are likely to be low because of missing emissions
and chemical processes aloft in the CMAQ model.
Southeast US
<2
15 -Z
NO2(10 molecules cm )
2-4 4-6 6-8 8-10
Figure 4-5. Comparison of modeled and observed NO2 column concentrations.
4.6 Probabilistic Model Evaluation
When weighing the societal benefits of different air
quality management strategies, policymakers need
quantitative information about the relative risks and
likelihood of success of different options to guide their
decisions. A key component in such a decision support
system is an air quality model that can estimate not only
a single "best-estimate" but also a credible range of
values to reflect uncertainty in the model predictions.
Probabilistic evaluation of CMAQ seeks to answer these
questions.
• How do we quantify our uncertainty in model inputs
and parameterizations?
• How do we propagate this uncertainty to the predicted
model outputs?
• How do we communicate our level of confidence in the
model-predicted values in a way that is valuable and
useful to decisionmakers?
To address these questions, we have deployed a
combination of deterministic air quality models and
statistical methods to derive probabilistic estimates of air
quality. For example, an ensemble of deterministic
simulations is frequently used to account for different
sources of uncertainty in the modeling system (e.g.,
emissions or meteorological inputs, boundary conditions,
parameterization of chemical or physical processes).
A challenge with ensemble approaches is that chemical
28
-------
transport models require significant input data and
computational resources to complete a single simulation.
We have applied the CMAQ-DDM-3D to generate large
member ensembles while avoiding the major
computational cost of running the regional air quality
model multiple times. We also have used statistical
methods to postprocess the ensemble of model runs
based on observed pollutant levels. Maximum likelihood
estimation is used to fit a finite mixture statistical model
to simulated and observed pollutant concentrations. The
final predictive distribution is a weighted average of
probability densities, and the estimated weights can be
used to judge the performance of individual ensemble
members, relative to the observations.
These approaches provide an estimated probability
distribution of pollutant concentration at any given
location and time. The full probability distribution can be
used in several ways, such as estimating a range of
likely or "highly probable" concentration values or
estimating the probability of exceeding a given threshold
value of a particular pollutant. For example,
Figure 4-6 shows the estimated probability of exceeding
an ozone threshold concentration of 60 ppb over the
Southeastern United States for current conditions (top)
and with a 50% reduction in NOx emissions (bottom).
Compared with the single base CMAQ simulation (far
left), the spatial gradients provided by the ensemble-
based estimates (middle and right) more accurately
reflect the observed exceedances under current
conditions.
Single Simulation
Ensemble Mean
Ensemble Prob.
0.0 0.2 0.4 0.5 0.8 1,0
Figure 4-6. Spatial plots of ozone and probability of exceeding the threshold concentration for July 8, 2002, at 5 p.m. EOT.
Observations are shown in white circles.
4.7 Statistical Methodology for Model Evaluation
Model evaluation efforts often include graphical
comparisons of monitoring data paired with the output for
the model grid cells in which the monitors lie and
statistical summaries of the differences that exist. If
certain differences or regions are of particular interest,
the investigator may narrow the evaluation's focus to a
limited area and time period. Advanced statistical
methods can aid the evaluator by making the best use of
the limited monitoring data available, accounting for the
differences between point-based measurements
(monitors) and grid cell averages (model output) and
assessing the model output for grid cells in which no
monitors are located.
Although a variety of approaches reasonably could
be utilized, we have focused on methods that allow us to
better understand and utilize the spatial correlation of
pollutant fields, such as kriging-based methods. For
example, we have used Bayesian kriging to investigate
the relationship between ammonium wet deposition and
precipitation and kriging with adjustments for anisotropy
to better understand ozone and PM2 5 concentrations in
the northeastern United States. In addition, recent work
(Figure 4-7) has explored the impact on model
evaluation of incommensurability (i.e., the mismatch
between point-based measurements and areal averages
(model output).
4.8 Dynamic Evaluation of a Regional Air
Quality Model
The dynamic evaluation approach explicitly focuses
on assessing the model-predicted pollutant responses
stemming from changes in emissions or meteorology.
However, the emergence of the dynamic evaluation
approach introduces new challenges. In particular,
retrospective case studies are needed that provide
observable changes in air quality that can be related
closely to known changes in emissions or meteorology.
The NOX State Implementation Plan (SIP) Call has
29
-------
I
I "•
• :•;.
•
•_ .*„•"
= '•oS'l
s:
8"
^'
1400 icoo laoo 2000
isoa \sxi
Distncs(krn)
(a) Observed concentrations
(b) Modeled concentrations
p
1SOO 1930
Distnce (km I
UCD 1COO 1330 2030
(c) Block kriging estimates based on observations
(d) Grid cells of interest for further investigation
Figure 4-7. Assessment of CMAQ's performance in estimating maximum 8-h ozone in the northeastern United States on
June 14, 2001, by Swall and Foley.
offered a very strong initial case study to test model
responses via dynamic evaluation.
EPA's NOX SIP call required substantial reductions in
NOX emissions from power plants in the eastern United
States during summer ozone seasons, with the emission
controls being implemented during 2003 through May 31,
2004. Gego et al. (2007) and USEPA (2007) show
examples of how observed ozone levels have decreased
noticeably after the NOX SIP call was implemented.
Because air quality models are applied to estimate how
ambient concentrations will change because of possible
emission control strategies, the NOX SIP call was
identified as an excellent opportunity to evaluate a
model's ability to simulate ozone response to known and
quantifiable observed ozone changes. An example of a
dynamic evaluation study is described in Gilliland et al.
(2008), where air quality model simulation results with
the CMAQ model were evaluated before and after major
reductions in NOX emissions. Figure 4-8 provides an
example from this prototype modeling study, where
changes in maximum 8-h ozone are compared from the
summer 2005 period (after the NOX controls) with those
from the summer 2002 period (before the NOX emission
reductions occurred). The spatial patterns of percentage
decreases in ozone derived from observations and the
model exhibit strong similarities. However, these results
also revealed model underestimation of ozone
decreases as compared to observations, especially in
the northeastern States at extended downwind distances
from the Ohio River Valley source region. This may be
attributed to an underestimation of NOX emission
reductions or a dampened chemical response in the
model to those emission changes or other factors.
Analysis methods, such as the e-folding distances
(Gilliland et al., 2008; Godowitch et al., 2008), have been
used to show that NOX emissions in these simulations
are not impacting ozone levels as far downwind as
observations suggest, which could be a factor here. Next
steps must involve further diagnostic evaluation to
identify what chemical, physical, or emission estimation
30
-------
5
-5
-10
-15
-20
-25
-30
{a)Obs
(b) CMAQ v4.6 CB05
Figure 4-8. Example of dynamic evaluation showing (a) observed and (b) air quality model-predicted changes (%) from
differences between summer 2005 and summer 2002 ozone concentrations from Gilliland et al. (2008). The results
illustrate the relative change in ozone when comparing the 95th percentile daily 8-h maximum levels between the two
summers.
uncertainties are contributing to these initial results from
the model. Findings from additional analysis of this case
study ultimately can lead to model improvements that are
directly relevant to the way air quality models are used
for regulatory decisions.
31
-------
CHAPTER 5
Climate and Air Quality Interactions
5.1 Introduction
AMAD has been working on improving our
understanding of the interactions between air pollution
and climate change. Below are some of the science
questions we are addressing.
• How will future climate change affect air quality?
• How do short-lived air pollutants impact atmospheric
dynamics on regional and global scales?
• What will be the regional-scale impact of climate
change on precipitation patterns?
• How will emission controls implemented for air quality
management impact climate change?
• What are the most cost-effective ways to mitigate
climate change by reducing concentrations of
pollutants that contribute to radiative forcing while
meeting air quality goals?
The first phase of the CIRAQ pilot study has been
completed. Other projects that are in progress include
the the ones noted below.
• Developing alternative scenarios for future U.S.
emissions of ozone precursors and species that form
atmospheric PM
• Developing methods to generate a range of future
regional-scale climate scenarios via dynamic
downscaling and statistical downscaling
• Developing integrated decision support tools for rapid
assessment of emission scenarios designed for
improving air quality and mitigating climate change
• Using the coupled WRF-CMAQ meteorology and
chemistry model to investigate feedbacks of future
emission scenarios on radiative budget
• Developing improved atmospheric chemistry models
for understanding the impact of biogenic isoprene and
anthropogenic NOX on short-lived, radiatively active
species.
5.2 Climate Impact on Regional Air Quality
Air quality is determined both by emissions of air
pollutants, including ozone and PM precursors, and by
meteorological conditions, including temperature, wind
flow patterns, and the frequency of precipitation and
stagnation events. For air quality management
applications, regional-scale models are used to assess
whether given emission control strategies will result in
compliance with the relevant NAAQS. These modeling
applications typically assume present meteorological
conditions, which means that potential changes in
climate are not included in air quality assessment. With
emission controls that are implemented over several
decades, however, future climate trends could impact the
effectiveness of these controls.
AMAD initiated the CIRAQ project in 2002 to
develop a pilot modeling study to incorporate regional-
scale climate effects into air quality modeling. It involved
collaboration across multiple Federal agencies and with
academic groups with global-scale modeling expertise,
who were supported through the EPA Science To
Achieve Results (STAR) grant program.
The GISS GCM v2' was used to simulate the period
from 1950 to 2055 at 4° latitude x 5° longitude resolution.
Historical values for greenhouse gases (as CO2
equivalents) were used for 1950 to 2000, with future
greenhouse gas forcing following the IPCC's A1B
scenario. Colleagues at the Pacific Northwest National
Laboratory downscaled the GCM outputs using the Penn
State/NCAR MM5 model to simulate meteorology over
the continental United States at 36-km resolution for two
10-year periods centered on 2000 and 2050.
For the first phase of this project, the effect of
climate change alone was considered, without
attempting to account for changes in emissions of ozone
and PM precursors. Hourly emissions were simulated
using the SMOKE model. Anthropogenic emissions were
based on the EPA 2001 modeling inventory, projected
from the 1999 National Emission Inventory (NEI)
version 3. Biogenic emissions were calculated using the
BEIS model and the simulated future meteorology. Air
quality was simulated for two 5-year periods (1999 to
2003 and 2048 to 2052) using CMAQ v4.5. Figure 5.1
shows changes in simulated average and 95th percentile
values of the maximum daily 8-h average (MDA8) ozone
concentrations for both summer and fall.
5.3 Emission Scenario Development
For the first phase of the CIRAQ study, AMAD
examined air quality under a future climate scenario with
anthropogenic emissions of ozone and aerosol
precursors fixed at 2001 levels and biogenic emissions
from vegetation and soils allowed to vary with the
simulated meteorology (Nolte et al., 2008). For the
second phase of CIRAQ, future air quality is simulated
using the same meteorology from phase 1 and
alternative projections of future anthropogenic emissions.
Emission projections for different scenarios of
economic growth and technological utilization have been
developed by colleagues at NRMRL using the
EPA 9-region MARKAL energy system model. MARKAL
outputs were converted to source classification code-
specific growth factors, which then were used with the
SMOKE model to generate emissions inputs for use by
the CMAQ chemical transport model.
Air quality simulations using these emissions
projections and the climatological meteorology described
above have been conducted using CMAQ v4.7. Analysis
of these simulations is in progress.
32
-------
1 Jun - 31 Aug
1 Sep - 31 Oct
Figure 5-1. Differences (5-year future - 5-year current) in mean (top) and 95th percentile (bottom) maximum daily 8-h
average (MDA8) ozone concentrations. Results show summertime increases of 2 to 5 ppb in mean MDA8 concentrations
in Texas and parts of the eastern United States and even larger increases in 95th percentile concentrations, suggesting
increased severity of ozone episodes. Still larger increases are predicted for the September-October time period,
suggesting a lengthening of the ozone season (Nolte et al., 2008).
5.4 Regional Climate Downscaling
To meet EPA's growing need for regional climate
projections to support impact assessments, AMAD is
developing climate downscaling capabilities using both
dynamic downscaling and statistical downscaling
techniques. AMAD is developing a methodology for
using the WRF model to downscale GCM simulations
provided by colleagues at NASA's Goddard Institute for
Space Studies.
When using coarse-scale data (either from a
reanalysis or a GCM) as lateral boundary conditions
(LBCs) for a regional model without any further
constraint, the interior meteorological fields simulated by
the regional model can deviate significantly from those of
the driving fields. Four-dimensional data assimilation
(FDDA) techniques provide one way to constrain the
RCM and keep it from diverging too far from the coarse-
scale fields. If the regional model is constrained too
strongly to the GCM fields, however, there is the
possibility that the benefit of using the higher resolution
RCM will not be realized. What is needed is a delicate
balance between the amount of constraint given to the
RCM and the freedom of the RCM to simulate its own
mesoscale features.
Analysis nudging and spectral nudging are two
forms of interior nudging available within the WRF model.
These methods have been applied in the literature (e.g.,
Miguez-Macho et al., 2004; Lo et al., 2008), but they
rarely have been compared to each other for climate
simulations. Our research will apply each nudging
method to reanalysis- and GCM-driven WRF model
simulations, with physics options chosen for air quality
applications.
Preliminary simulations (Figures 5-2, 5-3, and 5-4)
indicate that nudging is likely needed for both reanalysis-
and GCM-driven simulations to maintain large-scale
consistency between the driving fields and those
simulated within the WRF model.
5.5 Statistical Climate Downscaling
Statistical downscaling methods use correlations
among observed and modeled meteorological variables
to predict regional and local patterns and events that are
likely to occur based on 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 are not limited by the
resolution achievable by the nested regional dynamical
model. Thus, statistical methods possibly could be used
33
-------
NARR
n!w iSSv lewawiwS* iw TW
No Nudging
i5« iiS tiw
«S *1« iffv Sv SS Tir S»
10
12
13
14
15
17
18
19
20
22
23
Analysis Nudging
Spectrgl Nudging
Figure 5-2. Seasonally-averaged (April-June) wind fields at 300 hPa as simulated by (a) North American Regional
Reanalysis, (b) WRF without nudging, (c) WRF with analysis nudging, and (d) WRF with spectral nudging. Analysis
nudging improves WRF's ability to simulate the location and intensity of the jet stream.
to gain a better understanding of fine-scale variability,
even down to point locations.
It has been reported in the literature that the
performances of dynamical and statistical downscaling
are comparable for current climatic conditions. However,
it is questionable whether statistical models can perform
as well under future conditions (Wilby et 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 is a difficult part of the statistical
downscaling process. Once a statistical model has been
developed fora 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). However, statistical
downscaling makes this assumption as it extrapolates to
future conditions.
Current research interests in statistical downscaling
include the following.
• Evaluating the performance of statistical downscaling
methods in estimating the frequency, duration, and
intensity of extreme meteorological events
• Developing at least a rough understanding of how the
uncertainty affects estimates, and, particularly, how
the uncertainty may change when applied to future-
year GCM simulations
• Identifying the relative strengths and weaknesses of
the dynamical and statistical approaches to
downscaling
• Determining whether hybrid downscaling approaches
may be able to capitalize on the strengths of both
methods
5.6 Integrated Tools for Scenario Discovery
Because climate change occurs over decades,
scenarios are used to understand the impacts of policy
decisions on a range of future outcomes. However, fully
assessing the air quality and climate change impacts of a
given emission scenario requires extensive
computational modeling and analysis. Tools that can
rapidly inform decisionmakers and stakeholders are a
first-order need.
34
-------
GISS Model
No Nudi
5600 5650 5700 5750 5800 5850 5900 5950 6000
Analysis Nudging
Spectrgl Nudging
Figure 5-3. Mean July 500-hPa geopotential height (m) for (a) GISS ModelE, (b) base WRF run without any interior
nudging, (c) WRF with analysis nudging, and (d) WRF with spectral nudging. Although both nudging techniques are
applied only above the planetary boundary layer, both serve to keep the 500-hPa geopotential height simulated by WRF
closer to that simulated by ModelE.
To meet this need, we are developing GLIMPSE
(GEOS-CHEM LIDORT Integrated with MARKAL for the
Purpose of Scenario Exploration), a framework for
connecting atmospheric chemistry, radiative forcing, and
energy-economy models to rapidly understand the
integrated air quality and climate change impacts of U.S.
emission scenarios. Its four components, as depicted in
Figure 5-5, are as follows.
(1) GEOS-Chem, global chemical transport model to
simulate the global impacts of U.S. emissions
(2) LIDORT, a radiative transfer model to calculate the
radiative forcing impacts from short-lived species,
such as black carbon
(3) Adjoint calculations of GEOS-Chem LIDORT to
explicitly attribute the contribution from U.S. emission
sources to global changes in radiative forcing
(4) EPA 9-Region MARKAL energy system model to
discover the technologies, activities, and policy
options that jointly achieve our air quality and climate
change goals
In the first version of GLIMPSE, we will use the
adjoint version of GEOS-Chem LIDORT adjoint model
developed by Daven Henze at the University of
Colorado. This model will calculate the change in sulfate
and black carbon direct radiative forcing resulting from
emissions from U.S. sources. These data will be used by
MARKAL to find emission scenarios that achieve a given
reduction in radiative forcing for minimal cost. The key
assumptions driving these emission scenarios will be
further analyzed to find emission scenarios that robustly
achieve reductions in radiative forcing despite
uncertainties in future projections. Once such a subset of
robust emission scenarios is determined, it will be used
as input to more complete global and regional climate
models to fully quantify the impacts.
35
-------
GISS Model E
No Nudi
273 276 275 282 285 2B8 2S1 244 297 300 302
Analysis Nudging
Spectrgl Nudging
Figure 5-4. Mean July 2-m temperature (K) for (a) GISS ModelE, (b) base WRF run without any interior nudging, (c) WRF
with analysis nudging, and (d) WRF with spectral nudging. Without nudging, average near-surface temperatures
simulated by WRF for the Pacific Northwest are more than 6 K warmer than in the GCM.
36
-------
The GLIMPSE
Integrated Framework
GEOS-Chem
LIDORT Adjoint
Model
Concentrations,
Senstivities,
Radiative Forcing
Emission Changes
Environmental Impact
Constraints
Technologies,
Emissions,
Costs
MARKAL
Energy System
Model
Future Emissions
Scenario
1
Atmosphere -
Ocean Coupled
General
Circulation
Model
Policy Assessments
on Human and
Ecosystem Health
Figure 5-5. GLIMPSE data flow: GEOS-Chem LIDORT Adjoint model is used to attribute radiative forcing changes to U.S.
emission sectors. These data are used in conjunction with greenhouse gas emissions as constraints for the MARKAL
model, which, in turn, is used to generate scenarios that meet these constraints.
37
-------
CHAPTER 6
Linking Air Quality to Human Health
6.1 Introduction
This research theme applies existing models and
tools and develops new tools and approaches to link air
quality to human exposure and human health. Typically,
epidemiological studies rely on ambient observations
from sparse monitoring networks to provide metrics of
exposure. Yet, for many pollutants in urban areas, large
spatial variations exist, particularly near roads and major
industrial sources. Further complicating the issue,
ambient concentrations do not necessarily represent
actual exposures, which can be influenced by the
infiltration of ambient concentrations into indoor facilities
(such as automobiles, homes, schools, and workplaces)
and the activity of individuals (such as outdoor exercise,
walking, commuting, etc.). Finally, populations also are
impacted by the transport of pollutants. These multiple
factors affecting exposure require approaches that scale
from regional to local environments and to the individuals
experiencing the exposure (Figure 6-1). Thus, this
research provides analytical and physical modeling
approaches that provide the spatial and temporal detail
of concentration surfaces needed to understand the
relationships among pollutants emitted, the resulting air
quality, and exposure of humans to these pollutants.
Regional scale
Understanding the
relationships among air quality,
human exposure and health
endpoints requires the
incorporation of modeling tools
and methods at different scales
Human scale
Figure 6-1. Linking local-scale and regional-scale models for exposure assessment characterizing spatial variation of air
quality near roadways assessing the effectiveness of regional-scale air quality regulations. (Source: Stein et al., 2007)
Research conducted under this theme focuses on
developing analytical tools and methods based on
models and observations to improve the characterization
of human exposure, evaluate the effectiveness of control
strategies with respect to health outcomes, and address
exposure issues, such as exposure to multiple pollutants
and for multiple scales.
6.2 Near-Roadway Environment
Recent studies have identified increased adverse
health effects in the population that lives, works, and
attends school near major roadways. EPA's Clean Air
Research multiyear plan, therefore, emphasizes air
research to better understand the linkages between
traffic pollutant sources and health outcomes. The
purpose of the effort described here is to better
understand the atmospheric transport and dispersion of
emissions within the first few hundred meters of the
roadway, a region often characterized by complex flow
(e.g., noise barriers, depressed roads, buildings,
vegetation) and where steep gradients of concentration
have been observed. Work within AMAD has focused on
developing and improving various numerical modeling
tools necessary for assessing potential human exposure
near roadways.
The AERMOD dispersion model is one of the
modeling approach that is being used to link between
urban sources (particularly mobile emissions) and human
exposure assessments and human health outcomes. As
part of ORD's Near-Road Research Program, laboratory,
field, and numerical modeling studies are underway to
better characterize the concentration distributions
surrounding the wide variety of complex roadway
configurations found in urban areas. These studies
include an examination of wind direction and roadway
configuration effects in the Division's meteorological wind
tunnel located at the Fluid Modeling Facility (Figure 6-2).
38
-------
Figure 6-2. The Fluid Modeling Facility houses the Division's meteorological wind tunnel used to study the effect of
roadway configuration and wind direction on near-road dispersion.
A research project has been initiated 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, Ml. Exposure metrics
developed in this project will be coupled with health
outcomes determined in the Childhood Health Effects
from Roadway and Urban Pollutant Burden Study
(CHERUBS). Modeled and monitored air quality and
exposure data will be used with assessments of
respiratory effects to investigate the relationships
between traffic-related exposures and observed health
effects. Air quality modeling will be conducted with the
AERMOD dispersion model. Additionally, wind tunnel
simulations of flow and dispersion near roadway
configurations characteristic of area in the health study
will be conducted at the Division's Fluid Modeling
Facility. Wind tunnel studies will support the
development and evaluation of the AERMOD model for
urban, near-road applications and assist in the
interpretation of site-specific monitoring. The air quality
modeling and wind tunnel simulations of the Detroit area
are critical links between traffic-related emissions and
human exposures and health outcomes.
6.3 Evaluating Regional-Scale Air Quality
Regulations
A core objective of the CAA is to "protect and
enhance the quality of the Nation's air resources so as to
promote the public health and welfare and the productive
capacity of its population." To achieve this goal, billions
of dollars are spent annually by the regulated community
and Federal and State agencies on promulgating and
implementing regulations intended to reduce air pollution
and improve human and ecological health. Historically,
the impact of air pollution regulations has been
measured by tracking trends in emissions and ambient
air concentrations. Now, however, EPA is exploring the
potential of extending the concept of measuring impact
to a more complete understanding of the relationships
along the entire source-to-outcome continuum.
Assessing whether air quality management activities are
achieving the originally anticipated results from sources
through outcomes requires (1) the development of
indicators that capture changes in source emissions,
ambient air concentrations, exposures, and health
outcomes; and (2) the ability to characterize the
processes that impact the relationships among these
indicators. This research moves beyond characterizing
emission and ambient concentration changes because of
regulatory control actions to linking these changes to
human exposure and health end points.
The NOX SIP Call recently was implemented by EPA
to reduce the emissions of NOX and the secondarily
formed ozone and to decrease the formation and
transport of ozone across State boundaries. Over the
past 3 years, AMAD's research has demonstrated
reductions in observed and modeled ozone
concentrations resulting from the NOX SIP Call
39
-------
(Figure 6-3). The CMAQ model was used to characterize
air quality before and after the implementation of the NOX
SIP Call and to evaluate correlations between changes
in emissions and pollutant concentrations. Model
simulations were used to estimate the anthropogenic
contribution to total ambient concentrations and the
impact of not implementing the regulation. Methods were
developed to differentiate changes attributable to
emission reductions from those resulting from other
factors, such as weather and annual and seasonal
variations. Trajectory models were used to investigate
the transport of primary and secondary pollutants from
their sources to downwind regions.
Power Industry HOx Reductions
Ozone Season (2002 vs. 2004)
Linking ambient
concentrations to exposure
Exposure Estimates
ror Ozone
Linking exposure
to human health
Linking directly
between indicators
Monthly Rates of Respiratory
NYS
\ A
•T=|:H=l.H; 14; 'WiM-MR |.|
I ••' I ™ I *~ I j« I ••' I "• I '• I
Figure 6-3. Assessing the impact of regulations on ecosystems and human health end points showing the indicators
(boxes) and process linkages (arrows) associated with the Nox Budget Trading Program. (Source: Garcia et a!., 2008)
We will continue to develop ways to systematically
track and periodically assess progress in attaining
national, State, and local air quality goals, particularly
those related to criteria pollutants regulated under the
NAAQS and related rules. Current research is focused
on relating NOX emissions and ambient ozone
concentrations to human exposure and health end
points. Improved air quality surfaces that combine
observed and modeled data are being generated for use
in exposure models, epidemiological health studies, and
risk assessments. These studies will examine the
benefits of using improved air quality surfaces versus
central monitoring approaches and of using exposure
probability factors versus ambient ozone concentrations
in health studies. In addition, these studies will evaluate
changes in predicted exposure and risk assessments
and actual changes in health end points (e.g., respiratory
diseases) between the pre- and post-NOx SIP Call time
periods. Finally, research is moving beyond the NOX SIP
Call to assess upcoming regulations. An approach for
evaluating the CAIR is being investigated to establish
and integrate "metrics" (predictions of changes
associated with the promulgation of CAIR) and
"indicators" (actual levels of the same or closely related
parameters observed during the implementation of
CAIR).
6.4 Linking Local-Scale and Regional-Scale
Models for Exposure Assessments
EPA and State and local governments increasingly
need urban-scale air quality assessments that capture
spatial heterogeneity, identify highly exposed
subpopulations, and support public health studies. Air
quality modeling estimates should account for local-scale
40
-------
features, long-range transport, and photochemical
transformations. Therefore, a hybrid air quality modeling
approach is under development to integrate results from
a grid-based chemical-transport model with a local plume
dispersion model to provide these spatially and
temporally resolved air quality concentration estimates
(Figure 6-4). Such capabilities are also critical to support
human exposure and environmental health studies and
to help identify air pollutant sources of greatest risk to
humans. The coupling and appropriate application of
these models will improve estimates, demonstrate utility
in environmental health accountability programs, assist
in the development of risk mitigation strategies, and
improve epidemiology and community health studies.
Local impact from stationary sources
Near-road impact from mobile sources
Regional background from CMAQ
Combined model results
for multiple pollutants for
all receptors (census
block group centroids}
in the study area
Figure 6-4. Schematics of the hybrid modeling approach showing (a) local impact from stationary sources, (b) near-road
impact from mobile sources, and (c) regional background from CMAQ. (Source: Isakov et al., 2009)
AMAD scientists currently are involved in several
activities to develop and evaluate techniques in support
of exposure and health studies. This research is focused
on integrating air quality modeling into exposure and
health studies. Critical to that is the improvement of fine-
scale air quality models. A new method to enhance air
quality and exposure modeling tools has been advanced
to provide finer scale air toxics concentrations to
exposure models. This hybrid modeling approach
combines the results from regional- and local-scale air
quality models (the CMAQ chemistry-transport model
and the AERMOD dispersion model). An important
component of this research is an EPA feasibility study
conducted in New Haven, CT, that examines the
cumulative impact of various air pollution reduction
activities (at local, State, and national levels) on changes
in air quality concentrations, human exposures, and
potential health outcomes in the community. In
conjunction with local data on emission sources,
demographic and socioeconomic characteristics and
indicators of exposure and health, the methodology can
serve as a prototype for providing high-resolution
exposure data in future community air pollution health
studies. For example, the methodology can be used to
provide the baseline air quality assessments of impacts
resulting from regional- or local-scale air pollution control
measures. It also can be applied to estimate the likely
impact of future projected air pollution control measures
or urban or industrial growth on human exposures and
health in the community.
AMAD's scientists also are participating in several
cooperative research projects to test the newly
developed techniques in support of exposure and health
studies involving three major academic institutions:
(1) Emory University, (2) Rutgers University, and (3) the
University of Washington. NERL also has initiated
another cooperative research project (CHERUBS) with
the University of Michigan. This project is focused on
health effects of near-roadway exposures to air pollution.
The overall goal of these activities is to enhance the
41
-------
results from epidemiologic studies of ambient PM and
gaseous air pollution through the use of more reliable
approaches for characterizing personal and population
exposures.
6.5 National Urban Database and Access Portal
Tool
Based on the need for advanced treatments of high-
resolution urban morphological features (e.g., buildings,
trees) in meteorological, dispersion, air quality, and
human exposure modeling systems, a new project was
launched called the National Urban Database and
Access Portal Tool (NUDAPT). The prototype NUDAPT
was sponsored by EPA and involved collaborations and
contributions from many groups, including Federal and
State agencies and private and academic institutions
here and in other countries. It is designed to produce
gridded fields of urban canopy parameters (UCPs) to
improve urban simulations, given the availability of new
high-resolution data of buildings, vegetation, and land
use (Figure 6-5). Urbanization schemes have been
introduced into MM5, WRF, and other models and are
being tested and evaluated for grid sizes on the order of
1 km or so. Additional information includes gridded
anthropogenic heating and population data, incorporated
to further improve urban simulations and to encourage
and facilitate decision support and application linkages to
human exposure models. An important core-design
feature is the utilization of Web portal technology to
enable NUDAPT to be a "community-based system.
This Web-based portal technology will facilitate
customizing of data handling and retrievals.
O
Urban canopy effects
anthn
heating"
Houston
Sky View Factor
0451-075 Hj
0 751 - 0 85 r~M
0 851 - 09 | I
0.901. 0 95 ^H
0951 • 1 )•
Figure 6-5. Urban canopy effects. (Source: Ching et al., 2009)
High-resolution building information is being
acquired by the National Geospatial Agency (NGA;
formerly the National Imagery and Mapping Agency).
When completed, NGA will have obtained data from as
many as 133 urban areas. Building data can be acquired
by extractions from paired stereographic aerial images
by photogrammetric analysis techniques or from digital
terrain models (DTMs) acquired by airborne Light
Detection and Ranging (LIDAR) data collection. LIDAR
data are acquired by flying an airborne laser scanner
over an urban area and collecting return signals from
pairs of rapidly emitted laser pulses and processed to
Modeling Requirement
To capture the grid
average effect of detailed
urban features in nicso-
scale atmospheric models
Solution
Defined and implemented
Urban canopy parameterizations
such as height-to-width ratios and
sky view factors
into their model formulations
produce terrain elevation data products, including full
feature digital elevation models (OEMs) and bare-earth
DTMs. Subtracting the DTM from the DEM produces
data of building and vegetation heights above ground
level. Currently, NUDAPT has acquired datasets and
hosts 33 cities in the United States with different degrees
of coverage and completeness. Data are presented in
their original format, such as building heights, day and
night population, vegetation data, and land-surface
temperature and radiation, or in a "derived" format, such
as the UCPs for urban meteorology and air quality
modeling applications.
42
-------
CHAPTER 7
Linking Air Quality and Ecosystems
7.1 Introduction
Ecological resources are exposed to atmospheric
pollutants through wet and dry deposition processes.
A long-term goal of multimedia environmental
management is to achieve sustainable ecological
resources. Progress toward this goal rests on a
foundation of science-based methods and data
integrated into predictive multimedia, integrated
multidisciplinary, multistressor, open architecture
modeling systems. The strategic pathway aims at
progressing from addressing one stressor at a time to a
comprehensive multimedia-multistressor assessment
capability for current and projected ecosystem health.
Over the next several years, the 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
harmonize the connection of the local ecosystem scale
(tens of kilometers) with the regional airshed scale
(thousands to millions of 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.
7.2 Linking Air Quality to Aquatic and Terrestrial
Ecosystems
Ecosystem exposure occurs when stressors and
receptors occur at the same time and place (Figure 7-1).
To model the exposure, models for different media (e.g.,
air, water, land) must be linked together. Linkages
among models for air, water, and land can occur through
the use of consistent input data, such as land use and
meteorology, and through the appropriate exchange of
data at relevant spatial and temporal scales.
Atmospheric
Transformation
Transport
and Fate
of Stressor
In Space and Time
Aquatic and
Terrestrial
Receptor
Biogeochemical
Functioning
In Space and Time
Figure 7-1. A Venn diagram representing ecosystem exposure as the intersection of the atmosphere and biosphere
(http://www.epa.gov/amad/EcoExposure/index.html).
Improved Spatial Distribution of Terrestrial
Receptors
Dry deposition velocity varies with underlying
vegetation type because of differences in leaf area index,
canopy height, and plant characteristics, such as
minimum stomatal resistance. CMAQ v4.7 relies on the
1992 National Land Cover Dataset to identify the location
of land cover types. USGS 2001 National Land Cover
Database (NLCD) and 2001 to 2006 NOAA coastal lands
(C-CAP) databases provide higher resolution
information. The Spatial Allocator Raster Tool can be
used to compute CMAQ modeling-domain-gridded land
use information based on these input image data. As an
example, Figure 7-2 illustrates our improved ability to
identify the extent of deciduous forest cover areas in
North Carolina over earlier lower resolution estimates.
Colors indicate the percentage of each 1-km rectangular
grid containing deciduous trees.
The second stage of this spatial improvement is to
update the 1-km resolution Biogenic Emissions
Landcover Database v3 (BELD3) dataset agricultural
species distributions. The current distribution is based on
1995 National Agricultural Statistics Service surveys.
These estimates are being updated to reflect 2001 crop
distributions in combination with the 2001 NLCD
imagery. At present, the BELD3 data are used to
determine bioemission input for CMAQ. We anticipate its
more extensive use in the estimation of species-specific
exposure to atmospheric nitrogen and mercury
deposition.
Improved Estimates of Receptor-Specific
Atmospheric Deposition
The deposition velocity calculation for CMAQ v4.7 is
a combination of processes modeled in the
meteorological model and the chemical transport model.
Because CMAQ is a grid-based model, the influence of
the different land covers that comprise a grid cell are
averaged in the meteorological model for use in the
deposition velocity calculations. These grid-average
values are carried forth from the meteorological model to
the chemical transport model where chemical specific
43
-------
NLCD 2001 30rn Deciduous Forest
NLCD 1992 1km Deciduous Forest
Figure 7-2. Fractional deciduous forest coverage as represented in the 30-m resolution 2001 NLCD based on Landsat 7
satellite imagery (right panel) and in the 1-km resolution 1992 NLCD based on Landsat TM satellite imagery (left panel).
deposition velocity calculations are done. Ecological
applications need information regarding the amount of
deposition to the individual land cover categories. To be
able to provide this information without requiring
modification of the meteorological model, an approach
has been implemented in CMAQ that disaggregates
these grid-average values within CMAQ to allow output
of deposition estimates for each land cover type within a
grid in a manner consistent with meteorological model
flux calculations. Figures 7-3 and 7-4 illustrate deposition
velocity dependence on vegetation type for ozone.
7.2.1 Linking Air and Water Quality Models
Linking Air Quality and Watershed Models—
Collaborative Research with the ERD
AMAD and ERD have been collaborating to explore
air-water model linkage. The present focus is on how
best this might be accomplished now that multiple years
of CMAQ deposition are feasible, and grid sizes are
shrinking because of increasing computational capability.
Watershed models are calibrated to multiple years of
observed hydrology and precipitation. Chemical
simulations are generated using the same inputs, as well
as drawing on current monitored deposition fields from
the National Acid Deposition Program (NADP).
Scenarios of changes in deposition, however, are drawn
from CMAQ simulations (e.g., Sullivan et al., 2008).
Unfortunately, temporal and spatial agreement between
the modeled meteorological data used to drive CMAQ
deposition estimates and observed precipitation used to
drive the water quantity and quality simulations can be
poor, so that the base case and the future cases are not
consistent. This raises several questions: How sensitive
are watershed models to this error? Can the watershed
models tolerate these errors in scenario mode? Can we
create greater consistency by using the CMAQ
meteorological inputs for all watershed simulations?
As a first step, we have explored with 2001-2003
data (1) the use of daily cooperative station data to
perform a monthly calibration of the Grid Based Mercury
Model (GBMM; Tetra Tech, 2006), (2) calibrated model
runoff volume response to 36-km simulated daily
precipitation and mean daily temperature fields,
(3) response to 12-km simulated daily precipitation and
mean daily temperature fields, and (4) response to 4-km
Parameter-Elevation Regressions on Independent
Slopes Model (PRISM)-generated precipitation data.
Figure 7-5 shows preliminary results for the hydrologic
response to these various sources of precipitation data.
Errors in the simulated meteorology related to
timing, spatial coverage, magnitude, and suppressed
interannual variability can be observed. The benefit of
higher scale meteorological simulations can be noted in
cases where model runoff volume driven by the 12-km
precipitation is much closer to the USGS observed runoff
than that driven by the 36-km simulation. Exceptions
occur where there is little or no runoff response
difference between the two meteorological datasets. This
happens most often during the fall months and has been
traced to a failure of the analysis model used to nudge
the meteorological simulation to capture the
development of tropical storms off the coast of North
Carolina. MM5 precipitation errors were found to be a
serious problem when linking MM5 to calibrated
watershed models, indicating the need to develop
hydrology that is consistent with MM5/WRF precipitation.
The PRISM database (Daly et al., 2002) contains 4-km
gridded monthly precipitation generated via a set of
regression expressions and cooperative station data and
represents a more spatially complete dataset. PRISM
44
-------
Ozone Deposition Velocity - Crop
231
221
211
201
191
181
171
151
151
141
131
> 121
111
1D1
91
61
71
81
51
41
31
21
t1
1
1 21 41
B1 1(11 121 U1 161 IE: 2D1 221 241 261
X
1.53
1.14S
0.765
ossa
0.191
Nil 1241. E2X . - I. HlK I2UX IW- - 1 42S
Figure 7-3. Receptor-specific ozone deposition velocities to croplands.
Ozone Deposition Velocity - Forest
231
22 I
211
sai
191
181
171
181
151
141
131
>- 121
111
101
M
81
71
41
51
41
31
21
11
1
21 41 81 101 141 1C1 161 2D1 221 241 261
JUV2!. Bat ISMftIO lire
. I.MlKIJU 22S- 1.1
Figure 7-4. Receptor-specific ozone deposition velocities to forested ecosystems.
data were found to be useful in adjusting modeled facilitate better hydrologic linkage with watershed
precipitation errors. models. The use of higher resolution simulations (4-km)
Ongoing research within AMAD is focusing on ways nudged using analyses that include more extensive data
to improve meteorological precipitation simulations to assimilation (OBS-GRID) or that employ more advanced
45
-------
A Stream gage
Stream network
| | Watershed Boundaries
Elevation (m)
High: 326
vy"V-v-.~..---,-V
Deep: 36 km MM5
AA
~
6/01 12/01 6/02 12/02 6/03 12/03
Simulated Runoff
6/01 12/01 6/02 12/02 6/03 12/03
-Observed Runoff • Precipitation
Figure 7-5. Left panel is a map of the Deep River and Haw River watersheds within the Cape Fear River Basin. Right panel
shows time series of the simulated monthly runoff for the Deep and Haw watersheds during the 2001-2003 period for the
different precipitation datasets. Runoff for each precipitation dataset is compared to the USGS gage value for each
watershed.
data assimilation techniques, such as 3D variational
analysis, are being explored. Outcomes of these
experiments will be evaluated and, if significant
improvement is noted, will be tested within the GBMM.
CMAQ deposition datasets are being developed for
terrestrial and aquatic critical loads assessments and for
linking with USGS's SPARROW model. CMAQ
deposition datasets are planned to transition to those
with the land use mosaic approach and bidirectional
ammonia deposition. The initial emphasis for a core
capability would be off-line approaches to atmospheric
deposition that address bidirectional exchange and land
use. To further support trend analysis, sensitivity testing
to illustrate the response of atmospheric deposition to
various land use changes is planned.
7.3 Linking to Ecosystem Services
Humankind benefits from a multitude of resources
and processes that are supplied by natural ecosystems.
Collectively, these benefits are known as ecosystem
services and include products like clean air and clean
water. Ecosystem services are distinct from other
ecosystem functions because there is human demand
for and benefit from these natural assets.
Measurement of ecosystem services is the new
strategic focus for EPA's Ecological Services Research
Program (ESRP). It is believed that making the
evaluation of these services a routine part of
decisionmaking will transform the way we understand
and respond to environmental issues. The ESRP's
mission is to conduct innovative ecological research that
provides the information and methods needed by
decisionmakers to assess the benefits of ecosystem
services to human well-being and, in turn, to shape
policy and management actions at multiple spatial and
temporal scales. The overarching ESRP research
questions are as follows.
• What are the effects of multiple stressors on
ecosystem services, at multiple scales, overtime?
• What is the impact of various plausible changes in
these services on human well-being and on the value
of the services?
7.3.7 Future Midwestern Landscapes
The Future Midwestern Landscapes (FML) Study is
being undertaken as part of ESRP. The study examines
the variety of ways in which the landscapes of the
Midwest, including working lands, conservation areas,
wetlands, lakes, and streams, contribute to human well-
being. The FML goal is to quantify the current magnitude
of those contributions, and to examine how ecosystem
services in the Midwest could change over the next 10 to
15 years, given the growing demand for biofuels, as well
as the growing recognition that many different ecosystem
services are valuable to society and need to be
encouraged. The FML study will examine how the overall
46
-------
complement of ecosystem services provided by the
Midwest may be affected. The study will characterize a
variety of ecosystem services for a 12-State area of the
Midwest (see Figure 7-6).
ATLANTIC
Hf.HI AMCF.
WEST-CENTRA*.
SEM.ARJD
I
WESTERN
CGKDILLIHA
Ml* ED
IVOOO
PLAINS
CENTRAL USA
H ANS
SO Nnflhwn Forwl
•$ i sorrvrooo SHIELD
!~5 2 MIXED WOOD SHIELD
s s ATLANTIC HKJHLANDS
6.0 Northweslern Fotesled Mountain*
• 6 2 WESTERN CORDaLCRA
ft.D Eastern Temperate Forests
•« 1 uweo viooo
• 82 CENTRAL USA PLAINS
* 3 SOUTHEASTERN USA PLAINS
84OZARK OUACWTA-APPALACMIAN
FORESTS
85 MISSISSIPPI ALLUVIAL AND
SOUTHEAST USA COASTAL FLAWS
9.0 Grut PlaNn
92 TIMPERATI PRAJRieS
d 9 S i« ST-C ENTRAL SEMPRA PRAIRIE 5
_ » 4 SCUTH CEHTRAi SEMi-ARID PRAIRIES
SOOTH CEMTRAL
SEMI.
ARID PRAIRIES
OUACHITA
AFVAUSCHWN
FORESTS
SISSIPPI ALL
AN050UTHEA5TU
COASTAL PLAINS
Mi)9fTr*uUr»«
h»» Pomte Oc«n
I _ ISace Prov
Figure 7-6. Future Midwestern landscapes study area (thick black line) superimposed on the Midwest ecoregions.
Alternative future scenarios will be used to contrast
the current path (i.e., the policy-driven ramp-up of biofuel
production) with an alternative path, in which
hypothetical incentives are directed toward land uses
that produce a wider range of services. Conceptual
models of these scenarios will be used to explore the
nature and magnitude of changes to ecosystems and
human well-being expected for each scenario and to set
priorities for research. Detailed land use/land cover maps
will be constructed for the baseline and alternative future
scenarios, and computational models will be employed to
simulate the effects of land use changes in terrestrial,
atmospheric, and aquatic environments. In addition, a
socioeconomic framework and set of indicators will be
developed for evaluating the ecological changes in each
scenario, in terms of societal well-being.
The FML approach defines a linked-modeling
system to address the issues posed by the alternative
scenarios. Figure 7-7 illustrates the specific role of
AMAD research and model development. In particular,
the FML study will examine projected landscape
changes and subsequent changes in ecosystem
services. This task will make use of advances in CMAQ
modeling of land use change (mosaic) and bidirectional
ammonia flux to explore the combined impact of land use
changes on the deposition of nitrogen to underlying
watersheds in the Midwest. Ongoing research will help
elucidate the communication of these results in terms
that are relevant to ecosystem exposure assessment
(e.g., mosaic output and WDT utilization). Planned
analyses for the FML include changes in regional
ambient concentrations of ozone, oxidized and reduced
nitrogen species, sulfur dioxide, sulfate, and fine PM. We
also will provide changes in the magnitude and spatial
and temporal distributions of ozone and nitrogen flux
(emission and deposition) to FML ecosystems defined by
NLCD vegetation class.
47
-------
Meteorology from
a weather model
Spatial and
Temporal
Allocation
(SMOKE)
Emissions from
the EPA National
Inventory and
MARKAL
• BenMap
-Biogenic Emission
-Re-emission
-Transport
-Transformation
Gas Chemistry
Aqueous Chemistry
-Loss Process
e.g., deposition
NO
NO2
N205
HN03(gas)
IMO3(aerosol)
Organic N03
PAIMs
NH3(gas)
NH4 (aerosol)
S02 (gas)
SO, (aerosol)
SO4 (wet)
Others
•SWAT
Figure 7-7. Flow chart of AMAD's role in FML model development.
7.3.2 ESRP Nitrogen Pollutant Specific Study
ESRP Pollutant Specific Studies: Nitrogen
Regulating Services
The significance of Nr, which includes oxidized,
reduced, and organic forms, to the environment stems
from the duality of its environmental impacts. On the one
hand, Nr is one of life's essential nutrient elements. It is
required for the growth and maintenance of all of Earth's
biological systems. For humans, there are several sets of
services provided by natural and anthropogenic sources
of Nr, including the production of plant and animal
products (food and fiber) for human consumption and
use and the combustion of fuels that supports our energy
and transportation needs. Increasing demands for
energy, transportation, and food lead to greater demand
for Nr. Although releases of nitrogen are associated with
societal benefits, Nr is a powerful environmental
pollutant. Over the past century, human intervention in
the nitrogen cycle and use of fossil fuels has led to
substantial increases in production of Nr and in human
and ecosystem exposure to Nr. The amount of Nr
applied to the Nation's landscape and released to the
Nation's air and water has reached unprecedented
levels, and projections show that Nr pollution will
continue to increase for the foreseeable future. These
increases in Nr pollution are accompanied by increased
environmental and human health problems. The ESRP
Nitrogen Team will address its broad goal of connecting
Nrto ecosystem services through a two-pronged effort
with national work, where possible, and with smaller
scale, regional studies tackling specific problems and
ecosystem types.
National Scale Nitrogen Studies
Mapping at the national scale is being developed
with an initial focus on selected studies of nitrogen inputs
to the landscape. This work is being conducted in a
collaborative manner with the ESRP Mapping Team. The
ESRP Mapping Team is taking the lead on creating the
layers, whereas the Nitrogen Team will provide data and
model outputs and will contribute to designing the
mapping approach. Three major Nr inputs and transfers
have been selected as initial cases for the national
mapping: fertilizer input, atmospheric deposition, and
nitrogen transfer from land to water.
Nutrient Loading and Atmospheric Deposition
Atmospheric deposition is an important source of
nitrogen to terrestrial and aquatic landscapes. There is
direct deposition to the landscape and transfer of the
deposition from the terrestrial landscape to water bodies.
Atmospheric deposition of sulfur, oxidized nitrogen,
reduced nitrogen, and ozone will be simulated by CMAQ
for a 12-km grid size for the eastern United States and
the continental United States. Typical compilations of
deposition are monthly and annual accumulated
deposition amounts. A base year of 2002 is available to
represent current conditions (Figures 7-8 and 7-9).
CMAQ simulations for 2006 also may be available.
CMAQ projections of deposition for 2020 and 2030 that
represent the implementation of nitrogen oxide controls
to meet health standards for ozone and PM2§ under the
48
-------
2002 Annual 36km CMAQ
kgflia |
Figure 7-8. 2002 Annual total nitrogen deposition (wet and dry oxidized and reduced species).
1990 CAA Amendments (CAAAs) will be available for
mapping as well. Such projections show a significant
reduction in oxidized nitrogen deposition across the
eastern United States. The 12-km CMAQ grid can be
mapped to 12-digit hydrologic unit codes (HUCs) or any
other desired set of polygons. The CMAQ data will be
augmented by National Acid Deposition Program
(NADP) wet deposition data in the mapping exercise.
The use of CMAQ dry deposition combined with
precipitation-corrected and NADP-augmented CMAQ wet
deposition will be examined for the national mapping of
nitrogen deposition.
Regional Scale Nitrogen Studies
A regional approach will be pursued for several
questions in the ESRP Nr research program that
currently cannot be approached nationally. Case studies
for the regional approach have been selected that have
national significance and for which we desire to develop
a national approach. The objective is to extend the
regional case studies through a synthesis of methods to
be able to encompass a national perspective. CMAQ
deposition results will be used in several studies, in
particular, the study of tipping points.
Tipping Points in Ecosystem Condition and Services
The critical loads or tipping points approach can
provide a useful lens through which to assess the results
of current policies and programs and to evaluate the
potential ecosystem protection and ecosystem services
values of proposed policy options. A major stressor of
concern with serious consequences for freshwater
aquatic and terrestrial systems is acidification from
atmospheric deposition of Nr and sulfur. Several Federal
agencies are working together on regional pilot projects
across the United States to explore the possible role a
critical loads (or tipping points) approach can have in air
pollution control policy in the United States. The ESRP
Nr research program has selected three of the regional
pilot projects that provide an excellent opportunity for the
ESRP program to work within and build onto their efforts.
They are the Blue Ridge Mountains aquatic systems, the
Adirondacks terrestrial systems, and the Rocky Mountain
aquatic systems. CMAQ deposition outputs and NADP
data will be used to provide deposition inputs to the
ecosystem models used in these projects. CMAQ
projections to 2020 and beyond of deposition also will be
used to assess vulnerable ecosystems. These studies
are expected to come to fruition in 2010, after which a
synthesis effort will be undertaken to determine how best
to create national critical load mapping capabilities for
the EPA Office of Air Programs (OAP). Major players in
these pilots are EPA, NPS, and the U.S. Forest Service.
This research will involve close coordination among ORD
(AMAD and the National Health and Environmental
Research Laboratory's Western and Atlantic Ecology
Divisions), the OAP Clean Air Markets Division, and the
Office of Air and Radiation's OAQPS.
7.4 Air-Surface Exchange
The interaction between the atmosphere and the
underlying surface increasingly is recognized as
important in ecosystem health and in air pollution
transport processes. Just as there has been a movement
away from assessing human exposure to air pollutants
one chemical species at a time toward an integrated
one-atmosphere approach, so too should there be an
integrated one-atmosphere approach to assessing
ecosystems exposure to air pollutants. With this in mind,
we propose that now is the time to advance from simply
a one-atmosphere to a one-biosphere approach that
includes integration across multiple media and
biogeochemical processes to more effectively address
49
-------
2002 Annual Total Sulfur Dry Deposition
2002 Annual Oxidized-Nitrogen Dry Deposition
2002 Annual Reduced-Nit rag en Dry Deposition
MO) Mir
Figure 7-9. 2002 Annual acidifying dry deposition of sulfur and oxidized and reduced nitrogen (eq ha'1 year'1)
ecological interactions with the atmosphere as well as
human systems.
A deposition-based assessment of the impact of air
pollution on ecosystem health is more appropriate than
the existing concentration-based standards used to
protect human health. However, there is an extreme
paucity of measured and monitored dry deposition
estimates for use with ecosystem management
modeling. The estimates from the atmospheric models fill
a critical gap.
Improved Dry Deposition Algorithms for CMAQ
A targeted focus on creating state-of-the-science
dry deposition algorithms for the air quality models has
significant importance to ecosystem exposure to air
pollution. A major objective is to reduce uncertainty in
deposition/air-surface exchange calculations by
discovering and including missing pathways and by
creating a more ecosystem-compatible surface-layer link
with water quality and terrestrial models (Figure 7-10).
Model air-surface exchange uncertainty has led to
collaborations with measurement groups and the design
of experiments at field campaigns to refine and develop
mechanistic air-surface exchange algorithms. This has
resulted in the refinement of coarse-mode particulate
nitrate aerosol deposition and bidirectional exchange
algorithms for NH3 from soils following fertilizer
application and the impact of vegetation canopies on the
atmosphere-biosphere exchange.
Improved Dry Deposition for Network Applications
One of the ways EPA assesses the results of air
pollution control is through the Clean Air Status and
Trends Network (CASTNET). Dry deposition estimates
from CASTNET are inferred from measured atmospheric
concentrations and a dry deposition velocity estimated
from the physical characteristics of the ecosystem and
wind velocity measurements. The Multilayer Model
(Clarke et al., 1997; Finkelstein et al., 2000; Meyers
et al., 1998) is used to predict deposition velocity, which
then is paired with the measured concentration to
calculate the pollutant flux. Air-surface exchange
research will continue to develop better models for
predicting deposition velocity for network operations.
Providing better estimates of deposition flux will improve
our ability to forecast ecosystem sustainability.
50
-------
Unidirectional exchange
Rj^
R*
Mf
R" ! —
Soil Deposition
a)
Canopy Deposition
R.
, Jrn-i
Bbt
\f3 !SS
o
^ r- _
Vfi|^7:*-i3-S
» R*
Bidirectional exchange
"« •
R*i
4
R !C-
Ra | *
1 X,:, _
* "rw — 1
R*
s».
Air-Soil Exchange
b)
Air-Canqoy Excnange |
P ^ ^
\^SSm:
^tne
I I p
"E
Bidirectional exchange with soil capacitance
H,
i
**:
«v.l
<
;
R t C"
• 1:
1 Jtn
H,
'Cft,
Is
l*s
r^ Air-Soil Exchange
Air-Cancpy EKchange
. __R,
Js,
XLnMin'n'nt
R v
n- •tri-
r"'-*He
up^cca=-L'-
• • R
• .•
"W^^
Hf
**»
R,
R«
^
R.A
R-,
Rrn
ca
x*
x,
XD
Cft,
^d
Rcsisiancc to
deposition
Capacitance for re-
emission
Atmospheric
rcsis Luiicc
Lajninar bnundan
layer resistance
ln-cflnop>
ainiosplicnc
rcsislwicc
Boundary layt-i
rttSLSLiUlK IKM tJIC
soil surface
Culicular rcsisinnec
Stain iiial resistance
Mcsoplis 11
rtsisuiiiK
Atmospheric
ciMivcniratiDii
t'anopv equilibrium
at /.-•*,, (soil +
mcsoptiyl])
Soil equilibrium
conccniration
MfisopliyJ]
equilibrium
ijuiK-oiiiiatn»i
Measured in-canopy
couccnlraticHi at r. =
aim
Soil resistance t= ft
fordepctsition)
Figure 7-10. Air-surface exchange resistance diagrams of unidirectional exchange (a), bidirectional exchange of ammonia
(b), and bidirectional exchange of mercury and ammonia using the FEST-C tool (c).
7.4.1 Nitrogen Surface Exchange
Excessive loading of nitrogen from atmospheric
nitrate and ammonia deposition to ecosystems can lead
to soil acidification, nutrient imbalances, and
eutrophication. Accurate nitrogen deposition estimates
are important for biogeochemical cycling calculations
performed by ecosystem models to simulate ecosystem
degradation and recovery. Because of the lack of
available monitoring data, creating these estimates is a
high priority for water and soil chemistry modeling of
nutrient loading, soil acidification, and eutrophication.
In collaboration with the atmospheric measurement
community, we have conducted work to advance
nitrogen air-surface exchange (dry deposition and
evasion from soil and vegetation surfaces), modeling of
ammonia, and the treatment of coarse-mode nitrate
51
-------
chemistry in the CMAQ model. This process has
included the following steps.
(1) Develop testable hypotheses from the literature in
the form of new modules or routines for CMAQ
(2) Assist in the design of the field campaign needed to
collect measurements of the parameters required to
2.000240
1.000
0.000
-1.000
-2.000
kg/ha
further develop these algorithms and to conduct
robust evaluations of them
(3) Use the resulting field measurements to refine and
evaluate the model algorithms for the development
of an operational model (Figure 7-11)
2.000240
1.000
o.ooo
-1.000
279
-2.000
kg/ha i
279
Figure 7-11. Mean air-surface exchange of NH3 for the month of July estimated by CMAQ v4.7 using MM5 with the PX land
surface scheme for (a) unidirectional exchange of NH3 and (b) bidirectional exchange of NH3 (positive values indicate net
evasion and negative values indicate net deposition).
The development of the bidirectional ammonia
exchange and coarse-nitrate model algorithms improved
the modeled oxidized and reduced nitrogen budgets and
the partitioning between gas and size-segregated
aerosol phases. Mechanistic model algorithms
developed in collaboration with measurement groups
enhance the credibility of the CMAQ nitrogen budget for
ecosystem assessments. Results from the bidirectional
ammonia exchange model helped prioritize current and
future measurement needs in field experiments.
7.4.2 Soil NH3 Emissions
CMAQ representation of the regional nitrogen
budget is limited by its treatment of NH3 soil emission
from and deposition to underlying surfaces as
independent rather than tightly coupled processes and
by its reliance on soil emission estimates that do not
respond to variable meteorology and ambient chemical
conditions. The present study identifies an approach that
addresses these limitations, lends itself to regional
application, and will better position CMAQ to meet future
assessment challenges. These goals were met through
the integration of the resistance-based flux model of
Nemitz et al. (2001) with elements of the U.S.
Department of Agriculture Environmental Policy
Integrated Climate (EPIC) model. Model integration
centers on the estimation of ammonium and hydrogen
ion concentrations in the soil required to estimate soil
NH3 flux. The EPIC model was calibrated using data
collected in collaboration with NRMRL and N.C. State
University during an intensive 2007 field study in
Lillington, NC. A simplified process model based on the
nitrification portion of EPIC was developed and
evaluated. It then was combined with the Nemitz et al.
(2001) model and measurements of near-surface NH3
concentrations to simulate soil NH3 flux at the field site.
Finally, the integrated flux (emission) results were scaled
upward and compared to recent national ammonia
emission inventory estimates. The integrated model
results are shown to be more temporally resolved (daily),
while maintaining good agreement with established soil
emission estimates at longer time scales (monthly)
(Figure 7-12). Although results are presented for a single
field study, the process-based nature of this approach
and NEI comparison suggest that inclusion of this flux
model in a regional application should produce useful
assessment results if nationally consistent sources of soil
and agricultural management information are identified.
Fertilizer Scenario Tool for CMAQ (FEST-C)
Enhancements to the CMAQ bidirectional flux model
require additional, nationally consistent information
regarding fertilizer application timing, amount, and mode
of application, as well as soil characteristics and surface
losses in runoff. Research (Cooter et al., 2010) has
demonstrated that a well-vetted agricultural management
model can provide this information. A work assignment
has been drafted for the development of a nationally
consistent version of this model, designed to run either in
stand-alone mode for independent analyses or in
52
-------
If) -
cs
•o
O>
C\J -
o -
Observed Soil Flux
NEI2002af
Flux Model
190
195
200
Day of Year
205
210
Figure 7-12. Daily Harnet County, NC, NEI soil emission estimates and simplified process model estimates plotted with
Lillington, NC, observations.
conjunction with SMOKE to produce CMAQ-ready input
information to address "what-if questions associated
with current and alternative land use, land cover, and air
quality changes in response to population growth,
bio-energy, and air quality regulatory policy and climate
change. The stand-alone version of the model will output
files that can be displayed and analyzed using the
VERDI visualization tool.
7.4.3 Canopy NH3 Exchange
Regional and global estimates of the impact of
ammonia emissions on climate change, air-quality, and
human and ecosystem health must be scaled up with air
quality models. The effect of soil emission processes and
in-canopy sources and sinks on the net ecosystem flux
need further quantification (Button et al., 2007).
Scientists from AMAD have collaborated with field
scientists from NRMRL and Duke University to estimate
in-canopy and soil ammonia exchange processes based
on field measurements and modeling theory and have
designed experiments to elucidate a better process level
understanding of the biological, chemical, and
mechanical processes influencing the soil-vegetation-
atmosphere exchange of nitrogen over managed and
natural ecosystems. An analytical in-canopy scalar
transport closure model based on the mixing length
theory developed by Prandtl (1925) that estimates
in-canopy sources and sinks by using measured
concentration and wind speed profiles was developed.
In-canopy sources and sinks were estimated, and above-
canopy micrometeorological fluxes, soil chemistry, and
leaf chemistry measurements were collected in a
fertilized corn, Zea mays, field in Lillington, NC, during
the 2007 growing season. Estimates of in-canopy
sources and sinks were inferred using measured
in-canopy concentration profiles and a simple closure
model. Ammonia concentrations were measured at four
heights in the canopy and at one height above the
canopy using manually collected denuders in addition to
three collocated above-canopy continuous Ammonia
Measurement by Annular Denuder with Online Analysis
(AMANDA) concentration. Vertical profiles of wind speed,
heat, and momentum fluxes were made from inside the
canopy to a height of 10 m using an array of 3D sonic
anemometers. Ancillary vertical profiles of temperature
were measured using copper/constantan thermocouples
for model evaluation.
Modeled ammonia and sensible heat fluxes agreed
well with above-canopy micrometeorological flux
measurements. The soil at this site was found to be a
consistent emission source, whereas the vegetation
canopy was typically a net ammonia sink with the lower
portion of the canopy being a constant sink
(Figure 7-13). The upper portion of the canopy was
dynamic, exhibiting periods of local deposition and
evasion. The use of simple Eulerian-based, in-canopy
exchange estimates allowed for a physically descriptive
partitioning of atmospheric-soil and atmospheric-
vegetation exchange of measured scalars. These
detailed source and sink estimates are being used to
constrain NH3 soil emission estimates and the influence
of the vegetation canopy on the net flux for managed
agricultural land types in CMAQ.
A goal is also to harmonize the local ecosystem
scale (tens of square kilometers) with the regional
airshed scale (thousands to millions of square
kilometers). Surface NH3 concentrations were measured
beginning early in the 2009 fiscal year along transects in
53
-------
Average model fluxes - daytime
Figure 7-13. Ammonia exchange budget estimated from
the annalytical closure model.
North Carolina and with NASA Tropospheric Emission
Spectrometer (TES) retrievals to collect data on a
regional scale to evaluate the regional application of
these local mechanistic models (Figure 7-14). These
observations and observations from monitoring networks
will develop a new continental-scale ammonia emission
inventory using model inversion techniques for CMAQ
with and without ammonia bidirectional surface
exchange. NH3 bidirectional model parameters, soil and
vegetation emission potentials (f), and point sources will
be optimized in the bidirectional model inversion.
7.5 CMAQ Ecosystem Exposure Studies
Guidance and Advice to the Ecosystem Management
Community Using CMAQ as a Laboratory
Atmospheric deposition of sulfur and nitrogen is a
key contributor to ecosystem exposure and degradation,
causing acidification of lakes and streams and
eutrophication of coastal systems. Reductions in
atmospheric deposition of sulfur and oxidized nitrogen
resulting from regulations in the 1990 CAAAs are
expected to significantly benefit efforts to improve water
quality. However, water quality managers are not taking
advantage of information on anticipated deposition
reductions in developing their management plans.
Managers need to understand what to expect from
atmospheric emissions and deposition. This
understanding must come from an air quality model
utilized as a laboratory; it cannot come just from
measurements. The goal is to bring air quality into
ecosystem management through regional air quality
modeling and to facilitate the air-ecosystem linkage.
Ammonia Emission Density • TES Transect
kg NHl'sq km Q CAMNet Wlonitaing Site
Figure 7-14. TES transect locations and surface observations overlaid on a map of the estimated NH3 emission density in
Eastern North Carolina.
54
-------
Through identification of basic management
questions, we define what research and tool
developments for the air quality modeling system are
needed to make the linkage functional and the air-
ecosystem modeling applicable and useful.
Our approach is to collaborate with select,
motivated air-water partners who are willing to work
together to provide a test laboratory with the atmospheric
model to explore, assess, and apply improved
techniques to advance water quality management goals
and test linkage approaches. We develop an
understanding of the needs of the water quality
managers through real-world experience and
participation with model applications. We then design
model analyses and sensitivity studies to identify and
direct what atmospheric science needs to deliver.
Results help provide answers to nearly universal
questions uncovered in the course of the application
studies: How much is depositing? Which anthropogenic
or natural source is responsible for and where is the
deposition from? How much will deposition change
because of air quality regulations and population and/or
economic growth? Guidance on several fronts has been
developed; for example,
• a combination of local emission sources and long-
range transport of pollutants requires both local and
regional approaches,
• the uncertainty in ammonia emissions and
concentrations is very important, and
• CAAA reductions have been significant.
Air deposition reductions are now a vital component
of the Chesapeake Bay Program's restoration efforts.
Critical air deposition information also has been provided
to the Tampa Bay Estuary Program to address its total
maximum daily load (TMDL) needs and assessment
goals. Our efforts have opened the door for water quality
managers to include air deposition and make their
management plans more efficient and effective. The
work has paved the way for using CMAQ in national
NOX-SOX regulatory assessments to protect ecosystems
and for using CMAQ in U.S. critical loads analyses.
An area where the one atmosphere approach of
CMAQ helped elucidate the connection between
modeled chemical mechanisms and ecosystem
exposure through dry deposition was heterogeneous
N2O5 conversion. The uncertainty in the heterogeneous
conversion of N2O5 to HNO3 was examined because it
impacts HNO3 concentrations and deposition. However,
this uncertainty has a minor impact on oxidized nitrogen
deposition because the deposition pathways among the
oxidized nitrogen species rebalance. Although zeroing
out this conversion reduces HNO3and NO3" deposition
by 18% and 26%, respectively, total oxidized nitrogen is
reduced by only 6% (Figure 7-15).
7.5.7 Airsheds
Long-Range Transport
Airsheds typically have a larger spatial extent than
estuaries, watersheds, and National Parks. For NOX
emissions, the range of influence is multi-State, leading
to airsheds that are multi-State in size. This is also true
for NH3 emissions, which is counter to conventional
wisdom in the ecological community. The airshed is
defined as the domain from which emissions would
account for a significant majority of the deposition to the
receptor watershed.
CMAQ Simulated Ratio of
Dry Deposition to Wet Deposition
of Total N: CMAQ 2002 Annual
Total Oxidized-N Deposition to
Chesapeake Bay Watershed
.-=, -
2.1° K.
Figure 7-15. CMAQ is a source of data for ecosystem managers that is not available in routine monitoring data, such as
(a) complete dry and wet deposition estimates, and (b) the "one atmosphere" concept of CMAQ is needed to understand
the balance between uncertainties in atmospheric reaction rates and deposition pathways.
55
-------
Airsheds: Oxidized-Nitrogen Deposition into Coastal
Estuaries
Using the procedure developed for the Chesapeake
Bay and outlined in Dennis (1997), airsheds for
20 coastal watersheds along the East and Gulf Coasts
were developed. Examples of oxidized nitrogen airsheds
are seen in Figure 7-16.
PRINCIPAL OXIDIZED NITROGEN AIRSHEDS FOR:
NARRAGANSETT BAY, CHESAPEAKE BAY,
R\MLICO SOUND, TAMPA BAY, MOBILE BAY,
LAKE POISTTCHARTRAIN
Like Colors Match Airshed to Watershed
Figure 7-16. Air sheds (solid lines) and watershed (solid areas) for Narragansett Bay (purple), Chesapeake Bay (green),
Pamlico Sound (blue), Mobile Bay (yellow), Lake Pontchartrain (brown), and Tampa Bay (red).
7.5.2 Chesapeake Bay Restoration
Chesapeake Bay is the largest estuary in the United
States and was the Nation's first estuary targeted by
Congress for restoration. Reversal of the rapid loss of
living resources resulting from excess nutrients (mainly
nitrogen), and restoration of the quality of the Bay has
been the goal of the Chesapeake Bay Program since its
inception in 1983. Atmospheric deposition of nitrogen to
the Chesapeake Bay watershed and Bay surface is a
major contributor to the Bay nitrogen load, affecting
current conditions and needing to be addressed in Bay
restoration efforts. The atmosphere is estimated to
contribute a quarter of the total nitrogen load delivered
from the watershed to the Bay. Direct atmospheric
deposition to the Bay's tidal waters increases the fraction
of the total load of nitrogen to the Bay from atmospheric
deposition by approximately a third. Chesapeake Bay
has been placed on EPA's list of impaired waters, with a
TMDL plan required in 2011. To provide the best
modeling science for the TMDL plan, a major upgrade of
the Chesapeake Bay Watershed model v5.1 is being
used, as well as CMAQ v4.7. This atmospheric modeling
will be a major update from the earlier use of the
extended Regional Acid Deposition Model (RADM). The
grid size is 12 km, better resolving the Bay, and the
effect of sea salt is included. The CMAQ modeling for the
Chesapeake Bay TMDL planning has the following three
major foci.
(1) Development of scenarios estimating the deposition
reductions expected by 2010 and 2020 because of
CAA regulations, such as the CAIR (as further
modified as the result of court actions [Figure 7-17])
56
-------
Figure 7-17. Model-predicted contributions of six Bay States account for 50% of the 2020 oxidized nitrogen deposition to
the Chesapeake Bay Watershed.
(2) A new NH3 budget analysis at 12 km, using a
prototype CMAQ with NH3 bidirectional air-surface
exchange incorporated, showed that incorporating
bidirectional exchange of ammonia will have an
important impact of reducing the local dry deposition
and an impact on the estimates of the range of
influence of ammonia emissions, almost doubling the
range.
(3) Estimation of the relative contribution the NOX
emissions from the six Bay States make to the
atmospheric deposition of oxidized nitrogen to the
Bay watershed and Bay surface after implementation
of (CAIR). The State allocation data form the basis
for a management decision rule for allocating State
emission reductions that are beyond the national
rules to watershed deposition reductions that can
count as State allocation reduction credits.
7.5.3 Tampa Bay
The Tampa Bay Estuary Program (TBEP) set
restoration of underwater seagrasses, an indicator of
overall Bay health, as a long-term natural resource goal.
Water quality targets and associated nitrogen loading
goals have been developed and adopted to support
attainment and maintenance of the seagrass restoration
goal. Atmospheric deposition of nitrogen is the largest
source type contributing to nitrogen loading to Tampa
Bay. Direct deposition to Tampa Bay is central and is
estimated to be second only to storm water runoff, but a
portion of storm water runoff is caused by atmospheric
deposition (wet and dry). Tampa is an excellent example
of a coastal bay where the existence of sea salt is a
significant factor in the rate of local nitrogen dry
deposition. Tampa Bay is unusual in that a large portion
of the watershed is urbanized and a major fraction of the
oxidized nitrogen deposition to Tampa Bay is estimated
to come from local sources (40% to 50%). Two of the
largest utility emitters of NOX emissions in the country in
2000 are located at the edge of Tampa Bay. They have,
through a consent decree, agreed to reduce their NOX
emissions by up to 95% by 2010. A research beta
version of CMAQ, CMAQ-UCD, incorporates sea salt.
The Florida Department of Environmental Protection
(FDEP) organized, with ORD help, the Bay Regional Air
Chemistry Experiment (BRACE) field study that took
place in Tampa during May 2002. One key objective of
BRACE is to provide field data to evaluate CMAQ-UCD.
The four major thrusts of the Tampa Bay Model
Evaluation and Application study are
(1) to evaluate CMAQ-UCD against the BRACE May
2002 data and make any model refinements that
may be required;
(2) to assess the relative contributions from the different
emissions sectors, particularly mobile sources and
utilities, to the annual oxidized nitrogen deposition to
Tampa Bay;
(3) to assess the change in annual deposition to Tampa
Bay that could be attributed solely to the NOX
emissions reductions by 2010 of the two power
plants on its shores (Figure 7-18); and
(4) to assess the change in annual deposition to Tampa
Bay that could be attributed to mobile source and
utility reductions under the CAIR in 2010.
The Tampa Bay assessment is being conducted in
concert with FDEP and TBEP, using grid sizes of 8 km
over Florida and 2 km over the Tampa region. The
57
-------
.n: i
Figure 7-18. Fraction of total oxidized nitrogen deposition to Tampa Bay explained by local emission in the watershed.
CMAQ-UCD also has been used as a benchmark model
for the development of the dynamic sea salt
parameterization in the 2008 CMAQ v4.7.
7.6 Software Tool Development
Linking air quality and ecosystems is inherently
transdisciplinary. Significant effort often is required to
analyze observations and model results and provide
them in a form required to support management
decisions. Most off-the-shelf tools do not address the
specialized needs or applications encountered in
analyzing data from a multimedia perspective, making it
more difficult than is necessary to link elements of the
multimedia components together. As such, it is
necessary to provide the larger ecosystem modeling and
management communities with tools designed to utilize
air quality modeling data. This primarily takes the form of
tools used to convert air quality model output to formats
used by ecologists and ecosystem managers and tools
to visualize and analyze model output. The need for
specialized tools is especially pertinent to bringing
atmospheric components together with watershed
components for multimedia management analyses.
7.6.7 Visualization Environment for Rich Data
Interpretaion (VERDI)
VERDI is a flexible, modular, Java-based program
for visualizing multivariate gridded meteorology,
emissions, and air quality modeling data created by
environmental modeling systems such as the CMAQ
model and the WRF model. VERDI offers a range of
options for viewing data, including 2D tile plots, vertical
cross-sections, scatter plots, line and bar time series
plots, contour plots, vector plots, and vector-tile plots.
Scripting capability in VERDI provides a powerful
interface for automating the production of graphics for
analyzing data (Figure 7-19).
VERDI was developed for EPA by Argonne National
Laboratory and currently is supported by the Community
Modeling and Analysis System (CMAS) Center, which is
hosted by the Institute for the Environment (IE) at the
University of North Carolina at Chapel Hill (UNC-CH) and
can be downloaded from the CMAS VERDI website
(http://www.verdi-tool.org/). VERDI is an open source
program, and community involvement in further
development is encouraged. VERDI is licensed under
the Gnu Public License (GPL) v3, and the source code is
available through SourceForge
(http://verdi.sourceforge.net/). In 2008 and 2009,
additional capabilities were added to VERDI, including
an alternate tile plot routine, an areal interpolation plot
that provides the capability of the Watershed Deposition
Tool, and the ability to display CMAQ data in polar
stereographic and lat-long projections.
7.6.2 Watershed Deposition Tool
Background. Atmospheric wet and dry deposition can
be important contributors to total pollutant loadings in
watersheds. Because deposition can be expensive to
monitor over an entire watershed, estimates of
deposition often are obtained from regional-scale air
quality models, such as the EPA's regional-scale,
multipollutant CMAQ. CMAQ can be used to estimate
deposition resulting from a number of scenarios,
including current conditions and future emissions
reductions that are expected because of rules, such as
CAIR and Clean Air Mercury Rule. CMAQ produces
gridded output with typical grid sizes of 36, 12, and 4 km.
Because watersheds do not conform to the grid layout of
CMAQ, additional tools must be used to map the results
from CMAQ to the watersheds to provide the linkage
between air and water needed for TMDL and related
nonpoint-source watershed analyses. This linkage then
enables water quality management plans to include the
reductions in atmospheric deposition produced by the air
regulatory community in their calculation of loadings to
the watershed.
58
-------
(19, 61| Layer: 0 O3[1]
ISIsf l!!!!l !!| ill!! Hill! 111!!! ii!l!!iliHi
i! s I iii 1III ill iff i1 i iiiiiii 11 it in i iiiiiii s
.. . :.-.
. .•. ,.«k'. ..»
• t *' -vv:-.V.
• ., •. A.-- >
Figure 7-19. Examples of VERDI used to visualize and evaluate CMAQ output: (a) VERDI tile plot of hourly surface ozone,
(b) VERDI scatter plot of annual oxidized nitrogen wet deposition versus oxidized nitrogen dry deposition, (c) VERDI time
series plot of hourly surface ozone for a selected cell, and (d) VERDI contour plot of hourly surface level ozone.
Overview of the Watershed Deposition Tool. The
Watershed Deposition Tool (WDT) was developed by the
AMAD 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 and related nonpoint-source watershed analyses.
This software tool takes gridded atmospheric deposition
estimates from EPA's regional, multipollutant air quality
model, CMAQ, and allocates them to 8-digit HUCs of
rivers and streams within a watershed, State, or Region
(Figure 7-20). The WDT can calculate the weighted
average CMAQ atmospheric deposition (wet, dry, and
wet + dry) across a selected HUC or a set of selected
HUCs for a given scenario. The WDT also can calculate
the average change in air deposition across an HUC
between two different air deposition simulations.
Calculations can be exported to comma-separated
values files. For experienced geographic information
system (CIS) users, the WDT also can export CIS
shapefiles of the CMAQ gridded outputs. The tool is
designed to work under the Microsoft Windows system.
Deposition Components Available from CMAQ
•Nitrogen
Dry Oxidized Nitrogen
Wet Oxidized Nitrogen
Total (Wet + Dry) Oxidized Nitrogen
Dry Reduced Nitrogen
Wet Reduced Nitrogen
Total (Wet + Dry) Reduced Nitrogen
Total Dry (Reduced + Oxidized) Nitrogen
Total Wet (Reduced + Oxidized) Nitrogen
Total (Reduced + Oxidized) Nitrogen
Sulfur
Total Wet Sulfur
Total Dry Sulfur
Total (Wet + Dry) Sulfur
•Mercury
Total Wet Mercury
Total Dry Mercury
Total (Wet + Dry)Mercury
59
-------
Met,
•
Watershed Segments (kg/ha)
27.0 to 30,0
24.0 to 27.0
21.0 to 21.0
18,0 to 21.0
15.0 to 1S.O
12.0 to 15.0
9.0 to 12.0
e.0to9.0
3.0 to 6.0
0.0 to 3.0
38 36ISM 77 18759W
Drag left mtxse button to zoom in, dick right mouse ixjttan to zoom out
Figure 7-20. Screen shot of the 2002 annual CMAQ total reduced nitrogen deposition mapped to watersheds draining into
the Albemarle-Pamlico Sound displayed in GIS mapping software.
7.6.3 Spatial Allocator
The Spatial Allocator was developed by the IE at
UNC-CH for EPA to provide tools that could be used by
the air quality modeling community to perform commonly
needed spatial tasks without requiring the use of a
commercial GIS (Figure 7-21). There are three
components to the Spatial Allocator.
(1) Vector tools: These tools process vector GIS data to
perform functions such as mapping data from
counties to grids and visa versa.
(2)
(3)
Raster tools: These tools process raster data to
perform functions such as converting NLCD land-use
data into gridded land use.
Surrogate tools: These tools use the Vector Tools
and additional Java tools to help manage the
creation and manipulation of spatial surrogates used
in emissions modeling.
The Spatial Allocator and associated documentation
is available for downloading from the CMAS Center
(http://www.ie.unc.edu/cempd/projects/mims/spatial/),
which is hosted by the IE at UNC-CH.
Tree Canopy Percent
Legena
CftNQm
Imperviousness Percent
Legend
Figure 7-21. Spatial Allocator output from raster tools on North Carolina 1-krn grids for fractional tree canopy coverage (a)
and impervious surfaces (b) from NLCD data.
60
-------
References
Bhave, P.V, G.A. Pouliot, and M. Zheng. Diagnostic Model
Evaluation for Carbonaceous PM2.5 Using Organic
Markers Measured in the Southeastern U.S.
Environmental Science & Technology, 41: 1577-1583
(2007).
Bullock, O.R., Jr. and K.A. Brehme. Atmospheric mercury
simulation using the CMAQ model: formulation description
and analysis of wet deposition results. Atmospheric
Environment, 36, 2135-2146 (2002).
Bullock, O.R., Jr., D. Atkinson, T. Braverman, K. Civerolo,
A. Dastoor, D. Davignon, J-.Y. Ku, K. Lohman, T.C.
Myers, R.J. Park, C. Seigneur, N.E. Selin, G. Sistla, and
K. Vijayaraghavan. The North American Mercury Model
Intercomparison Study (NAMMIS): Study description and
model-to-model comparisons. Journal of Geophysical
Research, 113, D17310, doi:10.1029/2008JD009803
(2008).
Bullock, O.R., Jr., D. Atkinson, T. Braverman, K. Civerolo,
A. Dastoor, D. Davignon, J-.Y. Ku, K. Lohman, T.C.
Myers, R.J. Park, C. Seigneur, N.E. Selin, G. Sistla, and
K. Vijayaraghavan. An Analysis of Simulated Wet
Deposition of Mercury from the North American Mercury
Model Intercomparison Study (NAMMIS). Journal of
Geophysical Research, 114, D08301,
doi: 10.1029/2008JD011224 (2009).
Carlton, A.G., H-J Lim, K. Altieri, S. Seitzinger, and B.J. Turpin.
Link between Isoprene and Secondary Organic Aerosol
(SOA): Pyruvic acid oxidation yields low volatility organic
acids in clouds. Geophysical Research Letters, 33,
L06822, doi:10.1029/2005GL025374 (2006).
Carlton, A.G., B.J. Turpin, K. Altieri, S. Seitzinger, A. Reff, H.-J.
Lim, and B.E. Ervens. Atmospheric Oxalic Acid and SOA
Production from Glyoxal: Results of Aqueous
Photooxidation Experiments. Atmospheric Environment,
41,7588-7602(2007).
Carlton, A. G., B. J. Turpin, K. Altieri, S. Seitzinger, R. Mathur,
S. Roselle, and R.J. Weber. CMAQ model performance
enhanced when in-cloud SOA is included: comparisons of
OC predictions with measurements, Environ. Sci.
Technol., 42(23): 8798-8802 (2008).
Ching, J., M. Brown, S. Burian, F. Chen, R. Cionco, A. Hanna,
T. Hultgren, T. McPherson, D. Sailor, H. Taha, and
D. Williams. National urban database and access portal
tool. Bulletin of the American Met. Soc., 90(8), 1157-1168
(2009).
Clarke, J.F., E.S. Edgerton, and B.E. Martin. Dry deposition
calculations for the clean air status and trends network.
Atmospheric Environment, 31(21): 3667-3678 (1997).
Cooter, E., Bash, J.O., Walker, J.T., Jones, M.R., and
Robarge, W. Estimation of NHs bi-directional flux over
managed agricultural soils. Atmos. Environ., 44, 2107-
2115(2010).
Daly, C., W.P. Gibson, G.H. Taylor, G.L Johnson, and
P. Pasteris. A knowledge-based approach to the
statistical mapping of climate. Climate Research,
22: 99-113(2002).
Dennis, R.L. Using the Regional Acid Deposition Model to
Determine the Nitrogen Deposition Airshed of the
Chesapeake Bay Watershed. In Joel E. Baker, editor,
Atmospheric Deposition to the Great Lakes and Coastal
Waters, Society of Environmental Toxicology and
Chemistry, Pennsacola, FL, pp. 393-413 (1997).
Dennis, R., T. Fox, M. Fuentes, A. Gilliland, S. Hanna, C.
Hogrefe, J. Irwin, ST. Rao, R. Scheffe, K. Schere, D.
Steyn, and A. Venkatram. A Framework for Evaluating
Regional-Scale Numerical Photochemical Modeling
Systems. Environmental Fluid Mechanics, Springer, New
York, NY, 10:471-489(2010).
Finkelstein, P.L., T.G. Ellestad, J.F. Clarke, T.P. Meyers, D.B.
Schwede, E.O. Hebert, and J.A. Neal. Ozone and sulfur
dioxide dry deposition to forests: observations and model
evaluation. Journal of Geophysical Research, 105(012):
15365-15377(2000).
Garcia, V,N. Fann, R. Haeuber, and P. Lorang. Assessing the
public health impact of regional-scale air quality
regulations. Air & Waste Management Association,
Environmental Manager, July 2008, 29-34 (2008).
Gego, E., P.S Porter, A. Gilliland, and ST. Rao. Observation-
based assessment of the impact of nitrogen oxides
emissions reductions on O3 air quality over the eastern
United States. Journal of Applied Meteorology and
Climatology, 46, 994-1008 (2007).
Gilliland, A.B., C. Hogrefe, R.W. Pinder, J.M. Godowitch, K.L.
Foley, ST. Rao. Dynamic evaluation of regional air quality
models: Assessing changes in O3 stemming from
changes in emissions and meteorology. Atmospheric
Environment, 42, 5110-5123 (2008).
Godowitch, J.M., C. Hogrefe, ST. Rao. Diagnostic analyses of
regional air quality model: Changes in modeled processes
affecting ozone and chemical-transport indicators from
NOx point source emission reductions. Journal of
Geophysical Research, 113, D19303,
doi:10.1029/2007JD009537 (2008).
Isakov, V., J.S. Touma, J. Burke, D. Lobdell, T. Palma,
A. Rosenbaum, and H.Ozkaynak. Combining Regional-
and Local-Scale Air Quality Models with Exposure Models
for Use in Environmental Health Studies. J. Air& Waste
Management Association, 59: 461-472 (2009).
Keeler, G. J, L. Gratz, and K. Al-Wali. Influences on the long-
term atmospheric mercury wet deposition at Underhill,
Vermont, Ecotoxicology, 14, 71-83 (2005).
Kleindienst, T.E., M. Jaoui, M. Lewandowski, J.H. Offenberg,
C.W. Lewis, P.V. Bhave, and E.O. Edney. Estimates of
the Contributions of Biogenic and Anthropogenic
Hydrocarbons to Secondary Organic Aerosol at a
Southeastern US Location. Atmospheric Environment,
41:8288-8300(2007).
Lewis, C.W., G.A. Klouda, and W.D. Ellenson. Radiocarbon
measurement of the biogenic contribution to summertime
PM-2.5 ambient aerosol in Nashville, TN. Atmospheric
Environment, 38: 6053-6061 (2004).
Lo, J.C., Z. Yang, and R.A. Pielke, Sr. Assessment of three
dynamical climate downscaling methods using the
Weather Research and Forecasting (WRF) model.
J. Geophys. Res., 113, D09112 (2008).
Luecken, D.J., S. Phillips, G. Sarwar, and C. Jang. Effects of
using the CB05 vs. SAPRC99 vs. CB4 chemical
mechanism on model predictions: ozone and gas-phase
photochemical precursor concentrations. Atmospheric
Environment, 42, 5805-5820 (2008).
Lo, J.C., Z. Yang, and R.A. Pielke Sr. Assessment of three
dynamical climate downscaling methods using the
Weather Research and Forecasting (WRF) model.
J. Geophys. Res., 113, D09112 (2008).
61
-------
Meyers, T.P., P. Finkelstein, J. Clarke, T.G. Ellestad, and P.P.
Sims. A multilayer model for inferring dry deposition using
standard meteorological measurements. Journal of
Geophysical Research, 103(017): 22645-22661 (1998).
Miguez-Macho, G., G.L. Stenchikov, and A. Robock. Spectral
nudging to eliminate the effects of domain position and
geometry in regional climate model simulations. Journal of
Geophysical Research, 109, 013104(2004).
National Research Council of the National Academies. Air
quality management in the United States.
http://www.nap.edu/catalog.php?record_id=10728#orgs
(2004).
Nemitz, E., C. Milford, and M.A Sutton. A two-layer canopy
compensation point model for describing bi-directional
biosphere-atmosphere exchange of ammonia
Q.J. R. Meteorol. Soc., 127: 815-833 (2001).
Ng, N. L, J.H. Kroll, A.W.H. Chan, A.W.H., P.S. Chhabra, R.C.
Flagan, and J.H. Seinfeld. Secondary organic aerosol
formation from m-xylene, toluene, and benzene. Atmos.
Chem. Phys. Discuss., 7, 4085-4126 (2007).
Nolte, C. G., A. B. Gilliland, C. Hogrefe, and L. J. Mickley.
Linking global to regional models to assess future climate
impacts on surface ozone levels in the United States.
Journal of Geophysical Research, 113, 014307 (2008).
Prandlt, 0. Uber die ausgebildete turbulenz. Z. Angew. Math.
Mech., 5, 136-139(1925).
Reff, A., P. Bhave, H. Simon, T. Pace, G. Pouliot, and
0. Mobley, M. Houyoux. Emissions inventory of PM2.5
trace elements across the United States. Environmental
Science and Technology, 43: 5790-5796 (2009).
Ryaboshapko, A., Bullock, R., Ebinghaus, R., llyin, I.,
Lohman, K., Munthe, J., Petersen, G., Seigneur, C., and
Wangberg, I. Comparison of mercury chemistry models.
Atmospheric Environment, 36, 3881-3898 (2002).
Ryaboshapko, A., Bullock, O.R., Christensen, J., Cohen, M.,
Dastoor, A., llyin, I., Petersen, G., Syrakov, 0., Artz, R.S.,
Davignon, 0., Draxler, R.R., and Munthe, J.
Intercomparison study of atmospheric mercury models:
1. Comparison of models with short-term measurements.
Science of the Total Environment, 376(1-3), 228-240
(2007a).
Ryaboshapko, A., Bullock, O.R., Christensen, J., Cohen, M.,
Dastoor, A., llyin, I., Petersen, G., Syrakov, D., Travnikov,
O., Artz, R.S., Davignon, D., Draxler, R.R., Munthe, J.,
and Pacyna, J. Intercomparison study of atmospheric
mercury models: 2. Modelling results vs. long-term
observations and comparison of country atmospheric
balances. Science of the Total Environment, 377(2-3),
319-333 (2007b).
Stein, A., Isakov, V., Godowitch, J., and Draxler, R. A hybrid
approach to resolve pollutant concentrations in an urban
area. Atmospheric Environment, 41: 9410-9426 (2007).
Sullivan, T.J., B.J. Cosby, J.R. Webb, R.L. Dennis, A.J. Bulger,
and F.A. Deviney, Jr. Streamwater acid-base chemistry
and critical loads of atmospheric sulfur deposition in
Shenandoah National Park, Virginia. Environ Monit
Assess., 137: 85-99 (2008).
Sutton, M.A., E. Nemitz, J.W. Erisman, et al. Challenges in
quantifying biosphere-atmosphere exchange of nitrogen
species. Environmental Pollution, 150: 125-139(2007).
Swall, J. and K. Foley. The Impact of Spatial Correlation and
Incommensurability on Model Evaluation. Atmospheric
Environment. Elsevier Science Ltd, New York, NY, 43(6):
1159-1376(2009).
Tetra Tech. Development of a second-generation of mercury
watershed simulation technology: Grid Based Mercury
Model, Users Manual, Version 2.0, Fairfax, Va. (2006).
USEPA. NOx Budget Trading Program. EPA-430-R-07-009.
http://www.epa.gov/airmarkets (2007).
Vermette, S., S. Lindberg, and N. Bloom. Field tests for a
regional mercury deposition network—sampling design
and preliminary test results. Atmospheric Environment,
29, 1247-1251 (1995).
Yu, S., P.V. Bhave, R.L. Dennis, and R. Mathur. Seasonal and
Regional Variations of Primary and Secondary Organic
Aerosols over the Continental United States:
Semi-empirical estimates and model evaluation.
Environmental Science & Technology, 41: 4690-4697
(2007).
Wilby, R.L., C.W. Dawson, and E.M. Barrow. SDSM - a
decision support tool for the assessment of regional
climate change impacts. Environmental Modelling &
Software, 17, 147-159(2002).
62
-------
APPENDIX A
Atmospheric Modeling and Analysis Division Staff Roster
(As of December 31, 2009)
Office of the Director
ST. Rao, Director
David Mobley, Deputy Director
Patricia McGhee, Assistant to the Director
Sherry Brown
Wanda Payne (SEEP1)
Ken Schere, Science Advisor
Gary Walter, IT Manager
Jeff West, QA Manager
Emissions and Model Evaluation Branch
Tom Pierce, Chief
Jane Coleman (SEEP1), Secretary
Wyat Appel
Brian Eder
Kristen Foley
Jim Godowitch
Steve Howard
Sergey Napelenok
George Pouliot
Alfreida Torian
Atmospheric Exposure Integration Branch
Ellen Cooter, Acting Chief
Jesse Bash
Jason Ching
Jim Crooks (Postdoctoral Fellow)
Robin Dennis
Val Garcia
Megan Gore (Contractor)
Vlad Isakov
Donna Schwede
Joe Touma
Myrto Valari (NRC2 Postdoctoral Fellow)
David Heist, Fluid Modeling Facility
Ashok Patel (SEEP), Fluid Modeling Facility
Steve Perry, Fluid Modeling Facility
Bill Peterson (Contractor), Fluid Modeling Facility
John Rose (SEEP1), Fluid Modeling Facility
Atmospheric Model Development Branch
Rohit Mathur, Chief
Shirley Long (SEEP1), Secretary
Prakash Bhave
Ann Marie Carlton
Garnet Erdakos (NRC2 Postdoctoral Fellow)
Rob Gilliam
Bill Hutzell
Deborah Luecken
Martin Otte (Postdoctoral Fellow)
Harshal Parikh (Contractor)
Shawn Roselle
Golam Sarwar
Heather Simon (Postdoctoral Fellow)
John Streicher
David Wong
Jeff Young
Shaocai Yu (Postdoctoral Fellow)
Applied Modeling Branch
Jon Pleim, Acting
Melanie Ratteray (SEEP1), Secretary
Farhan Akhtar (ORISE3 Postdoctoral Fellow)
Bill Benjey
Jared Bowden (NRC2 Postdoctoral Fellow)
Russ Bullock
Barren Henderson (ORISE3)
Jerry Herwehe
Chris Nolte
Tanya Otte
Rob Pinder
Jenise Swall
Ben Wells (Contractor)
Ying Xie (NRC2 Postdoctoral Fellow)
1SEEP - Senior Environmental Employee Program
2NRC - National Research Council
3ORISE - Oak Ridge Science and Education Program
63
-------
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 are being used by EPA,
NOAA, and the air pollution community in understanding
and forecasting not only the magnitude of the air
pollution problem but also in developing emission control
policies and regulations for air quality improvements.
AMAD applies air quality models to support key
integrated, multidisciplinary 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 analytical,
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 are applicable to spatial scales ranging from
local/urban and mesoscale 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 physical 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 areas 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 are consistent with the
observations. By collaborating with other EPA offices
that provide data and algorithms on emissions
characterization 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 health and
ecosystems exposure models and studies. The three
major focus areas of this Branch are (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
analysis 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
strategies 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 atmospheric modeling tools to address
emerging issues related to air quality and atmospheric
influences on ecosystems. Climate change, growing
demand for biofuels, 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.
64
-------
APPENDIX C
2009 Awards and Recognition
EPA Bronze Medal
Alice Gilliland, William Hutzell, Deborah Luecken, Rohit
Mathur, Sergey Napelenok, Christopher Nolte, Tanya
Otte, Thomas Pierce, Robert Pinder, Jonathan Pleim,
George Pouliot, Shawn Roselle, Golam Sarwar, Kenneth
Schere, Donna Schwede, David Wong, and Jeffrey
Young - CMAQ Multi-Pollutant Model Team
ORD Technical Assistance to the Regions or
Program Offices Award
Scientific and Technological Achievement Awards
Winners
Deborah Luecken and William Hutzell - Development
and analysis of air quality modeling simulations for HAPs
Scientific and Technological Achievement Awards
Honorable Mention
Prakash Bhave - Receptor modeling of ambient PM data
using positive matrix factorization: Review of existing
methods
Golam Sarwar and Prakash Bhave - Modeling the effect
of chlorine emissions on ozone levels over the eastern
United States
Joe Touma - Modeling population exposures to outdoor
sources of HAPs
Joe Touma - Impact of underestimating the effects of
cold temperature on motor vehicle start emissions of air
toxics in the United States
NERL Special Achievement Award
ST. Rao - Goal 3: Leader in the Environmental
Research Community
Ellen Cooter, Robin Dennis, Vlad Isakov, Thomas
Pierce, Donna Schwede, and Joe Touma - Goal 4:
Integrate Environmental Science and Technology to
Solve Environmental Problems
Robert Gilliam, Alice Gilliland, Rohit Mathur, Christopher
Nolte, Tanya Otte, Jonathan Pleim, Shawn Roselle,
David Wong, and Jeffrey Young - Goal 5: Anticipate
Future Environmental Issues
AMAD Awards
Besf Paper: Alice Gilliland, Kristen Foley, Robert Pinder,
ST. Rao, and Jim Godowitch - Dynamic evaluation of
regional air quality models: Assessing the changes in
ozone stemming from changes in emissions and
meteorology
Second Best Paper: Russ Bullock - The North American
Mercury Intercomparison Study (NAMMIS): Comparisons
of OC predications with measurements
Third Best Paper: AnnMarie Carlton, Rohit Mathur, and
Shawn Roselle - CMAQ model performance enhanced
when in-cloud secondary organic aerosol is included:
Comparisons of OC predications with measurements
Teamwork Award: Steve Howard - Demonstrating the
quality of unselfish teamwork within the Division to
promote scientific research, as well as external and
internal collaborations
Leadership Award: Prakash Bhave - Demonstrating
leadership abilities in scientific research, external and
internal collaborations, mentorship, and project
management
65
-------
APPENDIX D
2009 Publications
(Division authors are in bold.)
Journal Articles
Brixley, L, J. Richmond-Bryant, D. Heist, G.E. Bowker,
S.G. Perry, and R.W. Wiener. The Effect of a Tall Tower
on Flow and Dispersion Throught a Model Urban
Neighborhood: Part 2. Pollutant Dispersion. Journal of
Environmental Monitoring. Royal Society of Chemistry,
Cambridge, UK, 11(12):2171-2179 (2009).
Bullock, R., D. Atkinson, T. Braverman, K. Civerolo,
A. Dastoor, D. Davignon, J. Ku, K. Lohman, T. Myers,
R. J. Park, C. Seigneur, N.E. Selin, G. Sistla, and
K. Vijayaraghavan. An Analysis of Simulated Wet
Deposition of Mercury from the North American Mercury
Model Intercomparison Study. Journal of Geophysical
Research-Atmospheres. American Geophysical Union,
Washington, DC, 114(D08301):1-12 (2009).
Carlton, A.G., C. Wiedinmyer, and J.H. Kroll. A Review
of Secondary Organic Aerosol (SOA) Formation from
Isoprene. Atmospheric Chemistry and Physics,
Copernicus Publications, Katlenburg-Lindau, Germany,
9(14):4987-5005 (2009).
Ching, J., M. Brown, S. Burian, F. Chen, R. Cionco,
A. Hanna, T. Hultgren, T. McPherson, D. Sailor, H. Taha,
and D. Williams. National urban database and access
portal tool. Bulletin of the American Met. Soc., 90(8),
1157-1168(2009).
Denby, B., V. Garcia, D.M. Holland, and C. Hogrefe.
Integration of Air Quality Modeling and Monitoring Data
for Enhanced Health Exposure Assessment. EM: Air and
Waste Management Association's Magazine for
Environmental Managers. Air & Waste Management
Association, Pittsburgh, PA, (10/2009):46-49 (2009).
Eder, B.K., D. Kang, R. Mathur, J.E. Pleim, S. Yu, T. L.
Otte, and G. Pouliot. A Performance Evaluation of the
National Air Quality Forecast Capability for the Summer
of 2007. Atmospheric Environment, Elsevier Science
Ltd., New York, NY, 43(14):2312-2320 (2009).
Fairlie, T., J. Szykman, A. Gilliland, R. Pierce,
C. Kittaka, S. Weber, J. Engel-CoxX, R.R. Rogers,
J. Tikvart, R. Scheffe, and F. Dimmick. Lagrangian
Sampling of 3-D Air Quality Model Results for Regional
Transport Contributions to Sulfate Aerosol
Concentrations at Baltimore, MD in Summer of 2004.
Atmospheric Environment, Elsevier Science Ltd., New
York, NY, 43(20):3275-3288 (2009).
Georgopoulos, P.G., S. Isukapalli, J.M. Burke,
S. Napelenok, T. Palma, J. Langstaff, M. Majeed, S. He,
D.W. Byun, M. Cohen, and R. Vautard. Air Quality
Modeling Needs for Exposure Assessment from the
Source-To-Outcome Perspective. EM: Air and Waste
Management Association Magazine for Environmental
Managers. Air & Waste Management Association,
Pittsburgh, PA, (10/2009):26-34 (2009).
Heist, D., L. Brixey, J. Richmond-Bryant, S.G. Perry,
and R.W. Wiener. The Effect of a Tall Tower on Flow and
Dispersion Through a Model Urban Neighborhood:
Part 1. Flow Characteristics. Journal of Environmental
Monitoring. Royal Society of Chemistry, Cambridge, UK,
11(12):2163-2170 (2009).
Heist, D., S.G. Perry, and L. Brixey. A Wind Tunnel
Study of the Effect of Roadway Configurations on the
Dispersion of Traffic-Related Pollution. Atmospheric
Environment, Elsevier Science Ltd., New York, NY,
43(32):5101-5111 (2009).
Hu, Y., S. Napelenok, M.T. Odman, and A.G. Russell.
Sensitivity of Inverse Estimation of 2004 Elemental
Carbon Emissions Inventory in the United States to the
Choice of Observational Networks. Geophysical
Research Letters, American Geophysical Union,
Washington, DC, 36(L15806):1-5 (2009).
Isakov, V., J.S. Touma, J. Burke, D. Lobdell, T. Palma,
A. Rosenbaum, and H. Ozkaynak. Combining Regional-
and Local-Scale Air Quality Models with Exposure
Models for Use in Environmental Health Studies. J. Air &
Waste Management Association, 59:461-472 (2009).
MacArthur, R., D. Mobley, L. Levin, I.E. Pierce,
H. Feldman, T. Moore, J. Koupal, and M. Janssen.
Emission Characterization and Emission Inventories for
the 21st Century. EM: Air and Waste Management
Association's Magazine for Environmental Managers. Air
& Waste Management Association, Pittsburgh, PA,
(10/2009): 36-41 (2009).
McKeen, S., G. Grell, S. Peckham, J. Wilczak,
I. Djalalova, E. Hsie, G. Frost, J. Peischl, J. Schwarz,
R. Spackman, A. Middlebrook, J. Holloway, J. de Gouw,
C. Warneke, W. Gong, V. Bouchet, S. Gadreault,
J. Racine, J. McHenry, J. McQueen, P. Lee, Y. Tang,
G. Carmichael, and R. Mathur. An Evaluation of Real-
time Air Quality Forecasts and their Urban Emissions
over Eastern Texas During the Summer of 2006, Second
Texas Air Quality Study Field Study. Journal of
Geophysical Research-Atmospheres, American
Geophysical Union, Washington, DC, 114(DOOF11):1-26
(2009).
66
-------
Papasawa, S., D.J. Luecken, R.L. Waterland,
K. Taddonio, and S. Andersen. Estimated 2017
Refrigerant Emissions of 2,3,3,3-Tetrafluoropropene
(HFC-1234yf) in the United States Resulting from
Automobile Air Conditioning. Environmental Science &
Technology, American Chemical Society, Washington,
DC, 43(24):9252-9059 (2009).
Pleim, J.E., and R.C. Gilliam. An Indirect Data
Assimilation Scheme for Deep Soil Temperature in the
Pleim-Xiu Land Surface Model. Journal of Applied
Meteorology and Climatology, American Meteorological
Society, Boston, MA, 48(7): 1362-1376 (2009).
Pinder, R.W., R.C. Gilliam, W. Appel, S.L. Napelenok,
K. Foley, and A. Gilliland. Efficient Probabilistic
Estimates of Surface Ozone Concentration Using an
Ensemble of Model Configurations and Direct Sensitivity
Calculations. Environmental Science & Technology,
American Chemical Society, Washington, DC,
43(7):2388-2393 (2009).
Rao, S.T. Environmental Monitoring and Modeling
Needs in the 21st Century. EM: Air and Waste
Management Association's Magazine for Environmental
Managers, Air & Waste Management Association,
Pittsburgh, PA, (10/2009):3-4 (2009).
Reff, A.M., P. Bhave, H. Simon, T. Pace, G. Pouliot,
D. Mobley, and M. Houyoux. Emissions Inventory of
PM2.5 Trace Elements across the United States.
Environmental Science & Technology, American
Chemical Society, Washington, DC, 43(15):5790-5796
(2009).
Reis, S., R.W. Pinder, M. Zhang, G. Lijie, and M.A.
Sutton. Reactive Nitrogen in Atmospheric Emission
Inventories. Atmospheric Chemistry and Physics,
Copernicus Publications, Katlenburg-Lindau, Germany,
9(19):7257-7677 (2009).
Sarwar, G., R.W. Pinder, W. Appel, R. Mathur, and
A.G. Carlton. Examination of the Impact of Photoexcited
NO2Chemistry on Regional Air Quality. Atmospheric
Environment, Elsevier Science Ltd., New York, NY,
43(40):6383-6387 (2009).
Scheffe, R., R. Philbrick, C. MacDonald, T. Dye,
M. Gilroy, and A.G. Carlton. Observational Needs for
Four-Dimensional Air Quality Characterization. EM: Air
and Waste Management Association's Magazine for
Environmental Managers, Air & Waste Management
Association, Pittsburgh, PA, (10/2009):5-12 (2009).
Schwede, D.B., R.L. Dennis, and M.A. Bitz. The
Watershed Deposition Tool: A Tool for Incorporating
Atmospheric Deposition in Watershed Analysis. Journal
of American Water Resources Association, American
Water Resources Association, Middleburg, VA,
45(4):973-985 (2009).
Simon, H., Y. Kimura, G. McGaughey, D. Allen, S.S.
Brown, H.D. Osthoff, J.M. Roberts, D.W. Byun, and
D. Lee. Modeling the Impact of CINO2 on Ozone
Formation in the Houston Area. Journal of Geophysical
Research-Atmospheres, American Geophysical Union,
Washington, DC, 114(DOOF03):1-17 (2009).
Soja, A.J., J. Al-Saadi, L. Giglio, D. Randall, C. Kittaka,
G. Pouliot, J. Kordzi, S. Raffuse, T.G. Pace, I.E.
Pierce, T. Moore, B. Roy, R. Pierce, and J. Szykman.
Assessing Satellite-based Fire Data for use in the
National Emissions Inventory. Journal of Applied Remote
Sensing, SPIE/lnternational Society for Optical
Engineering, Bellingham, WA, 3(031504):1-28 (2009).
Swall, J. and K. Foley. The Impact of Spatial Correlation
and Incommensurability on Model Evaluation.
Atmospheric Environment, Elsevier Science Ltd., New
York, NY, 43(6): 1159-1376 (2009).
Tang, Y., P. Lee, M. Tsidulko, H. Huang, J.T. McQueen,
G.J. DiMego, L.K. Emmons, R.B. Pierce, A.M.
Thompson, H. Lin, D. Kang, D. Tong, S. Yu, R. Mathur,
J.E. Pleim, T.L. Otte, G. Pouliot, J.O. Young, K.L.
Schere, P.M. Davidson, and I. Stajner. The Impact of
Chemical Lateral Boundary Conditions on CMAQ
Predictions of Tropospheric Ozone over the Continental
United States. Environmental Fluid Mechanics, Springer,
New York, NY, 9(1):43-58 (2008).
Tong, D.Q., R. Mathur, D. Kang, S. Yu, K.L. Schere,
and G. Pouliot. Vegetation Exposure to Ozone over the
Continental United States: Assessment of Exposure
Indices by the Eta-CMAQ Air Quality Forecast Model.
Atmospheric Environment, Elsevier Science Ltd., New
York, NY, 43(3):724-733 (2009).
Venkatram, A., V. Isakov, R.L. Seila, and R.W. Baldauf.
Modeling and Impacts of Traffic Emissions on Air Toxics
Concentrations near Roadways. Atmospheric
Environment, Elsevier Science Ltd., New York, NY,
43(20):3191-3199(2009).
Wang, H., D.J. Jacob, P. Le Sager, D.G. Streets, R.J.
Park, A. Gilliland, and A. van Donkelaar. Surface Ozone
Background in the United States: Canadian and Mexican
Pollution Influences. Atmospheric Environment, Elsevier
Science Ltd., New York, NY, 43(6):1310-1319 (2009).
Weaver, C., X. Liang, J. Zhu, P. Adams, P. Amar, 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, P.L. 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. McKenzie, L.J.
Mickley, S. O'Neill, C.G. Nolte, S.N. Pandis, P.N.
Racherla, C. Rosenzweig, A. Russell, E. Salathe, A. L.
Steiner, E. Tagaris, Z. Tao, S. Tonse, C. Wiedinmyer,
A. Williams, D. Winner, J. Woo, S. Wu, and D.J.
Wuebbles. A Preliminary Synthesis of Modeled Climate
Change Impacts on U.S. Regional Ozone
Concentrations. Bulletin of the American Meteorological
Society, American Meteorological Society, Boston, MA,
90(12):1843-1863 (2009).
67
-------
Wilczak, J.M., I. Djalalova, S. McKeen, L. Bianco, J. Bao,
G. Grell, S. Peckham, R. Mathur, J. McQueen, and
P. Lee. Analysis of Regional Meteorology and Surface
Ozone During the TexAQS II Field Program and an
Evaluation of the NMM-CMAQ and WRF-Chem Air
Quality Models. Journal of Geophysical Research.
American Geophysical Union, Washington, DC,
114(DOOF14):1-22(2009).
Wu, Y., J. Walker, D.B. Schwede, C. Peters-Lidard, R.L.
Dennis, and W. Robarge. A New Model of Bi-Directional
Ammonia Exchange Between the Atmosphere and
Biosphere: Ammonia Stomatal Compensation Point.
Agricultural and Forest Meteorology, Elsevier Science
Ltd., New York, NY, 149(2):263-280 (2009).
Yu, S., R. Mathur, D. Kang, K.L. Schere, and D. long.
A Study of the Ozone Formation by Ensemble Back
Trajectory-process Analysis Using the Eta-CMAQ
Forecast Model over the Northeastern U.S. During the
2004 ICARTT Period. Atmospheric Environment,
Elsevier Science Ltd., New York, NY, 43(2):355-363
(2009).
Book Chapters
Baklanov, A., J.K. Ching, C. Grimmond, and A. Martilli.
Model Urbanization Strategy: Summaries,
Recommendations and Requirements. Chapter 15,
Alexander Baklanov, CSB Grimmond, Sue Grimmond,
(ed.), Meteorological and Air Quality Models for Urban
Areas, Springer-Verlag, Berlin, Germany, 151-162
(2009).
Bullock, R., and L. Jaegle. Importance of a Global
Approach to Using Regional Models in the Assessment
of Source-Receptor Relationships of Mercury. Chapter
16, N. Pirrone, R. Mason (ed.), Mercury Fate and
Transport in the Global Atmosphere: Measurement,
models and policy implications. Springer-Verlag, Berlin,
Germany, Chapter 16:503-517 (2009).
Keeler, G.J., N. Pirrone, R. Bullock, and S. Sillman. The
Need fora Coordinated Global Mercury Monitoring
Network for Global and Regional Models Validation.
Chapter 13, Mercury Fate and Transport in the Global
Atmosphere. Springer, New York, NY, 391-424 (2009).
Published Reports
Burian, S. J. and J.K. Ching. Development of Gridded
Fields of Urban Canopy Parameters for Advanced Urban
Meteorological and Air Quality Models. U.S.
Environmental Protection Agency, Washington, DC,
EPA/600/R-10/007 (2009).
Rao, S.T., R.L. Dennis, V. Garcia, A. Gilliland,
R. Mathur, D. Mobley, I.E. Pierce, and K.L. Schere.
Summary Report of Air Quality Modeling Research
Activities for 2007. U.S. Environmental Protection
Agency, Washington, D.C., EPA/600/R-09/025 (NTIS
PB2009-111394) (2009).
68
-------
APPENDIX E
Acronyms and Abbreviations
ACM
AEIB
AERMOD
AMAD
AMB
AMDB
AMET
AMS
ARM
APMB
AQI
AQMEII
ARL
ASMD
BEIS
BELD3
BOSC
BRACE
CAA
CAAAs
CAIR
CASTNET
CBL
CCSP
CCTM
CDC
CEM
CHERUBS
CIRAQ
CIYA
CMAQ
CMAQ-TX
CMAQ-UCD
Asymmetric Convective Model
Atmospheric Exposure Integration
Branch
American Meteorological Society/EPA
Regulatory Model
Atmospheric Modeling and Analysis
Division
Applied Modeling Branch
Atmospheric Model Development
Branch
Atmospheric Model Evaluation Tool
American Meteorological Society
Annual Performance Measure
Air-Surface Processes Modeling
Branch
air quality index
Air Quality Model Evaluation
International Initiative
Air Resources Laboratory
Atmospheric Sciences and Modeling
Division
Biogenic Emission Inventory System
Biogenic Emissions Land Cover
Database, v3
Board of Scientific Counselors
Bay Regional Atmospheric Chemistry
Experiment
Clean Air Act
Clean Air Act Amendments
Clean Air Interstate Rule
EPA's Clean Air Status and Trends
Network
convective boundary layer
Climate Change Science Program
CMAQ Chemistry-Transport Model
Centers for Disease Control and
Prevention
Continuous Emission Monitoring
Childhood Health Effects from
Roadway and Urban Pollutant Burden
Study
Climate Impacts on Regional Air
Quality
Cash in Your Account
Community Multiscale Air Quality
Model
Community Multiscale Air Quality
Model-Texas
University of California Davis aerosol
module coupled to the Community
Multiscale Air Quality Model
CMAS
CO
CTM
DDM
DDM-3D
DEM
DOC
DTM
EC
ECU
EMEB
EMEP
EPA
EPIC
ESRP
FDDA
FDEP
FEST-C
FHA
FMF
FML
FRD
GBMM
GCM
GFDL
GHG
CIS
GISS
GLIMPSE
GPL
HAP
HAPEM
HEASD
HNO3
HONO
HO2
H2O2
HUC
IC/BC
Community Modeling and Analysis
System
carbon monoxide
Chemical Transport Model for Mercury
Decoupled Direct Method
Decoupled Direct Method-3D
digital evaluation model
U.S. Department of Commerce
digital terrain model
elemental carbon
electric generating units
Emissions and Model Evaluation
Branch
European Monitoring and Evaluation
Programme
U.S. Environmental Protection Agency
Environmental Policy Integrated
Climate Model
Ecological Services Research Program
4D data assimilation
Florida Department of Environmental
Protection
Fertilizer Emissions Scenario Tool for
CMAQ
Federal Highway Administration
Fluid Modeling Facility
Future Midwestern Landscapes
NOAA's Field Research Division
Grid Based Mercury Model
global climate model
Geophysical Fluid Dynamics
Laboratory
greenhouse gas
geographic information system
Goddard Institute for Space Studies
Geos-CHEM LIDORT Integrated with
MARKAL for the Purpose of Scenario
Exploration
Gnu Public License
Hazardous Air Pollutant
Hazardous Air Pollutant Exposure
Model
Human Exposure and Atmospheric
Sciences Division
nitric acid
nitrous acid
hydroperoxyl radical
hydrogen peroxide
hydrologic unit code
initial condition and boundary condition
69
-------
ICARTT
IE
IMPROVE
INTEX
INTEX-NA
IPCC
ISORROPIA
ITM
LBC
LES
LIDAR
LIDORT
LSM
LW
MAE
MCIP
MDA
MDN
MEGAN
MLBC
MM5
MPI
MYSQL
NAAQS
NADP
NAM
NAMMIS
NARSTO
NAS
NASA
NATO
NBP
NCAR
NEI
NERL
NGA
NH3
NLCD
NMM
NO
NO2
NO3
N205
International Consortium for
Atmospheric Research on Transport
and Transformation
Institute for the Environment (UNC-CH)
Interagency Monitoring of Protected
Visual Environment Network
Intercontinental Chemical Transport
Experiment
Intercontinental Chemical Transport
Experiment-North America
International Panel on Climate Change
thermodynamics partitioning module
International Technical Meeting
lateral boundary condition
large-eddy simulations
Light Detection and Ranging
Linearized Discreet Ordinate Radiative
Transfer
land surface model
longwave
mean absolute error
Meteorology-Chemistry Interface
Processor
maximum daily average
Mercury Deposition Network
Model of Emissions of Gases and
Aerosols from Nature
multilayer biochemical model
fifth generation of the Penn
State/UCAR Mesoscale Model
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
formerly the North American Research
Strategy for Tropospheric Ozone
National Academy of Sciences
National Aeronautics and Space
Administration
North Atlantic Treaty Organization
NOX Budget Trading Program
National Center for Atmospheric
Research
National Emission Inventory
National Exposure Research
Laboratory
National Geospatial Agency
ammonia
National Land Cover Data
Nonhydrostatic Mesoscale Model
nitrogen oxide
nitrogen dioxide
nitrate
dinitrogen pentoxide
NOX
NOV
NOAA
NOAH
NPS
Nr
NRMRL
NUDAPT
03
OAP
OAQPS
OC
OH
ORD
PAH
PAN
PEL
PM
PMML
PRISM
PXLSM
QUIC
Qv
RCM
RELMAP
REMSAD
RMSE
SCIAMACHY
SEARCH
SGV
SHEDS
SIP
SMOKE
S02
SO4
SOA
SOAdd
SPS
STAR
STENEX
STN
SW
TBEP
TEAM
TES
oxides of nitrogen
oxidized nitrogen
National Oceanic and Atmospheric
Administration
NOAA's land surface model
National Park Service
reactive nitrogen
National Risk Management Research
Laboratory
National Urban Database and Access
Portal Tool
ozone
Office of Air Programs
Office of Air Quality Planning and
Standards
organic carbon
hydroxy radical
Office of Research and Development
polycyclic aromatic hydrocarbon
Peroxyacyl nitrate
planetary boundary layer
particulate matter
Predictive Model Markup Language
Parameter-Elevation Regressions on
Independent Slopes Model
Pleim-Xiu Land Surface Model
Quick Urban Industrial Complex
Water vapor mixing ratio
Regional Climate Model
Regional Lagranian of Air Pollution
Regional Modeling System for
Aerosols and Deposition
root mean squared error
Scanning Imaging Absorption
Spectrometer for Atmospheric
Cartography
SouthEastern Aerosol Research and
Characterization Study
subgrid variability
Stochastic Human Exposure and Dose
Simulation
State Implementation Plans
Sparse Matrix Operator Kernel
Emissions
sulfur dioxide
sulfate
secondary organic aerosol
secondary organic aerosol formed in
clouds
Science for Peace and Security
Science To Achieve Results
Stencil Exchange
Speciated Trends Network
shortwave
Tampa Bay Estuary Program
Trace Element Analysis Model
Tropospheric Emission Spectrometer
70
-------
TexAQS
TM
TMDL
UCP
UNC-CH
uses
Texas Air Quality Study
Thematic Mapper
total maximum daily load
urban canopy parameter
University of North Carolina at
Chapel Hill
U.S. Geological Survey
VERDI
VOC
WDT
WRF
WSOC
YSU
Visualization Environment for Rich
Data Interpretation
volatile organic compound
Watershed Deposition Tool
weather research and forecasting
water soluble organic compound
Yonsei University
71
------- |