Jump to main content.

Summary Report of Air Quality Modeling Research Activities for 2007

Print Version (PDF, 38 pp, 3.07 MB) About PDF

Past Reports

CONTENTS
NOTICE
ABSTRACT
ACKNOWLEDGEMENTS
1 INTRODUCTION
2 Providing Scientifically-Advanced Models and Tools to Support Environemtnal Policy Decision
3 Evaluating the Impact of Regulatory Policies on Air Quality and Ecosystems
4 Linking Sources to Human Exposure
5 Linking Soucrces to Ecosystem Exposure
6 Providing Air Quality FOrecast Guidance for Health Advisories
7 Understanding the Relationships Between Climate Change and Air Quality
APPENDIX A: Division Staff Roster
APPENDIX B: Division and Branch Descriptions
APPENDIX C: Awards and Recognition
APPENDIX D: Publications
APPENDIX E: Abbreviations

Notice
The research presented here was performed under the Memorandum of Understanding and Memorandum of Agreement between the U.S. Environmental Protection Agency (EPA) and the U.S. Department of Commerce’s (DOC’s) National Oceanic and Atmospheric Administration (NOAA). It has been subjected to EPA 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
Through a Memorandum of Understanding (MOU) and Memorandum of Agreement (MOA) between the U.S. Department of Commerce (DOC) and the U.S. Environmental Protection Agency (EPA), the Atmospheric Sciences Modeling Division (ASMD) of the National Oceanic and Atmospheric Administration’s (NOAA’s) Air Resources Laboratory (ARL) develops advanced modeling and decision support systems for effective forecasting and management of the Nation’s air quality. As a division within the EPA organizational structure, ASMD is known as the Atmospheric Modeling Division (AMD). The Division is responsible for providing a sound scientific and technical basis for regulatory policies to improve ambient air quality. The models developed by the Division are being used by EPA, NOAA, and the air quality community not only to understand and forecast the magnitude of the air pollution problem, but also to develop emission control policies and regulations. This report summarizes the research and operational activities of the Division for fiscal year 2007.

Acknowledgements
The authors acknowledge the support of Patricia McGhee of the Division for technical editing and manuscript preparation.

Chapter 1. Introduction

The National Oceanic and Atmospheric Administration (NOAA) Atmospheric Sciences Modeling Division (ASMD) works within the frameworks of the Memorandum of Understanding and Memorandum of Agreement between the U.S. Department of Commerce (DOC) and the U.S. Environ­mental Protection Agency (EPA). These agreements are implemented through long-term Interagency Agreements DW13938483 and DW13948634 between EPA and NOAA.

The Division is organized into four research branches:

The first three branches above constitute the Atmospheric Modeling Division (AMD) of the National Exposure Research Laboratory (NERL) of the Office of Research and Devel­opment (ORD) within EPA’s organizational structure. The fourth branch listed is part of the Air Quality Assessment Division of the Office of Air Quality Planning and Standards (OAQPS) within EPA’s organizational structure. Throughout this report, these NOAA-EPA branches will be collectively referred to as “the Division.” The appendices to this report contain a list of Division employees (Appendix A), descrip­tions of the Division and its branches (Appendix B), a list of awards earned by Division personnel (Appendix C), and a list of Division publications (Appendix D).

The Division’s role within the 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 OAQPS 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 contracts and cooperative agreements.

The Division has completed a major strategic planning process begun in 2002. We identified six outcome-oriented Theme Areas:

Research tasks were developed within each Theme Area, by considering these questions:

The result is a research strategy for meeting user needs that is built around the six major Theme Areas and supported by the four branches of the Division, as depicted in Figure 1-2. The Division’s Applied Modeling Branch also supports the three research- and development-focused branches by facilitating the transition of atmospheric modeling systems and other research tools to regulatory applications.

This report summarizes the research and operational activities of the Division for fiscal year 2007. 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. The rest of this report (Chapters 2 through 7) is organized according to the six major program themes listed above, also shown in Figure 1‑2.


Adapted from “A Conceptual Framework for U.S. EPA’s National Exposure Research Laboratory,” November 2007 Draft by EPA/NERL.

The Division’s role in the Source-Exposure-Dose-Effects Continuum
Figure 1-1. The Division’s role in the Source-Exposure-Dose-Effects Continuum


The Division's strategy to meet user needs.
Figure 1-2. Strategy to meet user's needs.


Back to Top

Chapter 2
Providing Scientifically-Advanced Models and Tools to Support Environmental Policy Decisions

Introduction
The Clean Air Act (CAA) requires that EPA set National Ambient Air Quality Standards (NAAQS) for air pollutants considered harmful to public health and the environment. Thresholds for six criteria pollutants have been established: carbon monoxide (CO), lead (Pb), nitrogen oxides (NOx), fine particulate matter (PM2.5) and coarse particulate matter (PM10), tropospheric ozone (O3), and sulfur oxides (SOx). EPA reviews each NAAQS every five years, and proposes changes if the most current science on health and ecological effects suggests changing the standards. For example, in 2006 EPA revised the standards for daily average PM2.5 from 65 to 35 μg/m3, and eliminated the annual average standard for PM10, leaving only the daily standard of 150 μg/m3.

When a geographic area exceeds the NAAQS for a criteria pollutant, EPA may designate that area as being in “nonattainment.” In response, the state containing that area must develop a State Implementation Plan (SIP) that explains how the state will achieve compliance with the NAAQS. The principal tools that EPA and the states use to demonstrate this compliance are air quality simulation models. Each SIP must include a modeling demonstration illustrating how the state intends to mitigate emissions (usually through additional emission controls) to achieve compliance with the standard.

In addition to the NAAQS for the criteria pollutants, EPA and the states also study mitigation strategies for other types of pollutants, such as hazardous air pollutants (HAPS, or air toxics) and global pollutants (mercury, for example, falls into both those categories). While there is a range of air quality policy-related issues that are tracked separately for individual pollutants, the pollutants’ chemistry and the sources involved in producing harmful air quality conditions are interrelated. Therefore, a multipollutant model is needed that can simulate the atmospheric processes and emission source inputs that contribute to all of these chemical species and conditions. The Division develops, evaluates, applies, and refines such models. These models represent, in as much detail as possible, the various dynamical, physical, and chemical processes regulating the atmospheric transport and fate of pollutants. The principal modeling platform, the Community Multiscale Air Quality (CMAQ) modeling system, includes components for meteorology, emissions, air quality, and analysis with visualization (see Figure 2-1).

Research Description
Within this Theme Area, the principal elements of the modeling program are Model Development and Model Evaluation. These elements are interrelated and form an iterative process: model evaluation provides information for improving the models; models are then improved through research and development; the improved models are re-evaluated; and (assuming successful re-evaluation) the improved models are then available for regulatory application.

Through the Model Development program element, the Division develops and improves the CMAQ air quality model for a variety of spatial scales (urban through continental) and temporal scales (days to years) and for a variety of pollutants (O3, PM, mercury and other air toxics, visibility, acid depo­sition). The multipollutant model approach permits the testing of emission control strategy impacts on the target pollutant, as well as collateral impacts on other pollutants.

Focus areas of model development include the following:

Integrating meteorology and chemistry modeling is a new program priority designed to provide feedback from air quality parameters (e.g., aerosols) that affect meteorological parameters (e.g., radiation). Developmental areas are guided by the model evaluation results and by model sensitivity and uncertainty tests. New CMAQ model versions are released for public access roughly every one to two years. Workgroups have been formed to focus on these research topics:

Through the Model Evaluation program element, the Division evaluates the models to characterize the accuracy of model predictions and to identify improvements needed in model processes or model inputs. This requires comparisons against observational data. We compare different CMAQ sim­ulations (e.g., different model versions, different chemical mechanisms, different vertical layer structuring) to identify the impact of model changes or options on model performance. Uncertainties in meteorological predictions and emission estimates are considered to help identify where improvements are needed. Regulatory applications of CMAQ are evaluated by comparing model-predicted changes in ozone and aerosols to changes in emission precursors. We conduct model evalu­ation through workgroups dealing with these issues:

Through these efforts, the Division facilitates the transition of research to the regulatory community.

Accomplishments
In the area of model development, a multipollutant version of the CMAQ modeling system was developed to predict ozone, PM, and mercury and 38 other HAPs in a single model configuration. We created this model version in response to increasing interest in modeling multiple pollutants, including criteria and hazardous air pollutants, within a single modeling framework for air quality management. The new model will support regional and urban studies that assess the potential co-benefits and effectiveness of various emission control pro­grams, such as the Clean Air Interstate Rule (CAIR), Clean Air Mercury Rule (CAMR), Clean Air Visibility Rule (CAVR), and various onroad and nonroad mobile source rules. It will also support future assessment studies based on integrated national emission inventories containing both HAPs and criteria pollutants. The multipollutant model was devel­oped by modifying and merging algorithms for gas-phase chemistry, aerosols, clouds, and emissions used in the mercury and HAPs versions of CMAQ. The Carbon Bond 05 (CB05) chemical mechanism was combined with the chemical reactions for chlorine, mercury, and HAPs, and implemented into the CMAQ modeling system. A normalization process was performed to test the model and to ensure that the multipollutant model is consistent with the original versions. Results suggest that consistency is achieved by including the emissions and chemistry of molecular chlorine (Cl2) and hydrochloric acid (HCl) in each model version. The multi­pollutant model will be included in the 2008 release of CMAQ.

During 2007, in collaboration with a variety of private and governmental research organizations, the Division completed the analysis of results from the North American Mercury Model Intercomparison Study (NAMMIS). The NAMMIS employed global-scale modeling of atmospheric mercury to define initial and boundary conditions for three regional-scale mercury models that were the primary subjects of the study: the CMAQ model, developed and applied by the Division; the Regional Modeling System for Aerosols and Deposition (REMSAD), developed and applied by ICF International; and the Trace Element Analysis Model (TEAM), developed by Atmospheric and Environmental Research, Inc. The CMAQ, REMSAD, and TEAM simulations of the air concentration and wet and dry deposition of various mercury species during the 2001 test period were compared on time scales from weekly to annual. The simulations of wet deposition of mercury from CMAQ, REMSAD, and TEAM were also com­pared against observations from the Mercury Deposition Network on time scales from weekly to annual. Considerable model-to-model differences were found for air concentration, dry deposition, and wet deposition. Statistical agreement between simulated annual wet deposition and the corre­sponding observation was found to be largely scaled to the statistical accuracy of the precipitation data input to all three models; these data were derived from prior meteor­o­logical modeling. On shorter time scales, the statistical agree­ment for mercury wet deposition was weaker than for the input precipitation data, indicating that the physicochemical processes controlling the wet deposition of mercury may still not be accurately treated in any of the models tested. At the end of 2007, results from the NAMMIS were being described in two manuscripts intended for publication in peer-reviewed scientific journals.

We also worked to improve the representation of reactive nitrogen chemistry in CMAQ. Similar to other air quality models, the CMAQ model currently accounts for only the homogeneous chemical reactions of nitrous acid (HONO). Studies have indicated that air quality models that take into account only the homogeneous reactions are not adequate to explain the observed ambient HONO. Recent evidence suggests that direct emissions and a heterogeneous reaction involving nitrogen dioxide and water vapor may play an important role in HONO chemistry. To improve the model performance for HONO, these additional sources have been included in CMAQ. The inclusion of these sources does indeed improve the model performance for HONO.

During 2007, several advances were made in the simulation of aerosol chemistry and physics. We focused on heterogeneous nitrogen chemistry, coarse-particle chemistry, trace-elemental composition of particles, and aerosol thermo­dynamics. The heterogeneous reaction probability of N2O5 on wetted particle surfaces (YN2O5) is an influential parameter affecting wintertime predictions of fine-particulate nitrate (NO3-). Division scientists discovered a typographical error in the published parameterization of YN2O5 that had been translated into the CMAQ v4.6 code. Correcting that error led to a degradation in the model predictions of wintertime NO3-. Therefore, a detailed a study of the underlying laboratory data was conducted and a new YN2O5 parameterization was devel­oped. This parameterization is the first to include the effects of temperature, humidity, particle composition, and phase state on YN2O5. When incorporated into the next version of CMAQ, the new YN2O5 parameterization is expected to mitigate current overpredictions of wintertime NO3- under conditions prevalent in the midwestern U.S. In a separate effort during 2007, we made considerable progress in simulating the dynamic interaction between gaseous species (e.g., nitric acid) and coarse particles. These thermo­dynam­ically driven interactions are currently neglected in the CMAQ model, resulting in a gross underprediction of particulate nitrate and an over­estimation of particulate chloride in coastal urban areas. The interactions have been successfully simu­lated in a stand-alone box model of the CMAQ aerosol module and will be incorporated into the full modeling system next year. Progress was also made in modeling the source origin of various trace elements in fine particulate matter. This development will introduce a number of new ways to evaluate model perform­ance for primary PM in urban areas. Finally, we resumed efforts to improve the numerical stability of the gas/particle thermodynamic calculations in CMAQ. This work is being conducted in collaboration with researchers at the Georgia Institute of Technology.

Efforts were also devoted toward transitioning to the Weather Research and Forecasting (WRF) model as the meteorological driver for CMAQ. The Pleim-Xiu land-surface model (PX LSM), the Asymmetric Convective Model version 2 (ACM2) boundary layer model and surface layer scheme, historically used as physics options in the Fifth-Generation Mesoscale Model (MM5), have been added to the WRF model. We provided the codes for these models to the National Center for Atmospheric Research (NCAR) for inclusion in the next release of WRF (version 3.0), due to be released in the spring of 2008. Evaluation of WRF simulations using these new physics components and comparisons to other LSM and PBL options in the WRF system have shown generally comparable or better results for temperature, humidity, and winds. This work has also led to some other significant improvements, including a new indirect nudging scheme for soil temperature, improved treatment of seasonal changes in vegetation, and improved parameterizations for soil, vegetation, and snow heat capacity. To facilitate the linkage between the WRF and CMAQ modeling systems, version 3.3 of the Meteorology-Chemistry Interface Processor (MCIPV3.3) was prepared and delivered to the Community Modeling and Analysis System (CMAS) Center for release to the CMAQ user community; major changes included updates for WRF fields, improve­ments to dry deposition, removal of outdated science options, and addition of metadata to MCIP output files. The Division also completed a systematic investigation of the impacts of data assimilation in meteorological models on air quality predictions from CMAQ. Analyzing MM5 and CMAQ simulations confirmed that the use of nudging throughout the simulation period leads to improved prediction of ozone. MM5 simulations maintain nearly constant statistical performance on average when nudging is used throughout the simulation period; however, CMAQ pre­dictions of ozone tend to degrade as the run time in MM5 increases. A two-part paper summarizing the findings from this investigation was accepted for publication in the Journal of Applied Meteorology and Climatology. Additional invest­igation into this phenomenon will continue in 2008 using WRF rather than MM5.

Finally, model development efforts were also devoted toward developing and testing an on-line integrated meteorology–atmospheric-chemistry modeling system. Integrating meteor­ology and chemistry modeling is a new program priority designed to provide feedback from air quality parameters (e.g., aerosols) that affect meteorological parameters (e.g., radia­tion). A coupled WRF-CMAQ system capable of simulta­neous integration of meteorology and chemistry with two-way data exchange has been developed and tested. Development efforts on the feedback effects of aerosols on solar radiation are progressing.

In the area of model evaluation, the Division continued to probe the performance of the CMAQ system using oper­a­tional, diagnostic, dynamic, and probabilistic evaluation tech­niques. A number of publications examined performance under various synoptic regimes, with alternative chemical mechanisms, and with varying degrees of vertical resolution.

Using operational and diagnostic evaluation techniques, an examination of CMAQ’s PM2.5 predictions showed that the PM “other” component was a major contributor to CMAQ’s overprediction of total PM2.5 in the fall and winter. Work is continuing to reduce the uncertainty of PM “other,” by looking at possible biases in emissions. Another thorny issue that was tackled during 2007 was how to improve the comparison between observed and predicted PM2.5 species to account for artifacts in the observations that are not accounted for in the CMAQ predictions. These components include nitrate volatilization and retained water mass, which can significantly impact the measured PM2.5 mass. The efforts to account for these artifacts will ultimately provide more accurate compar­isons between observed and predicted PM2.5 mass.

We made significant progress with probabilistic evaluation techniques, with an initial focus on CMAQ ozone predictions. Because all models are a simplification of the phenomena they aim to represent, it is often more useful to estimate a model result as a probabilistic range rather than as a single “best” result. A key challenge is that ensemble approaches require a large number of expensive simulations of independent modeling systems. We implemented a computationally effi­cient method to generate ensembles with hundreds of members based on several structural configurations of a single air quality modeling system and using the Decoupled Direct Method (DDM) to directly calculate how ozone concentrations change as a result of changes in input parameters. The modeled probabilistic range was compared to observations and was shown to perform better than more ad hoc estimates of the uncertainty in ozone predictions. Because this technique can generate large ensembles efficiently, it is well suited for diagnosing structural errors in the air quality modeling system.

Exploration into new statistical methods for evaluating comparisons of monitoring data with model predictions also took place. 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. While a variety of approaches could reasonably be utilized, the focus has been on methods that allow one to better understand and utilize the spatial correlation of pollutant fields, such as kriging-based methods. One example is Hierarchical Bayesian Modeling which is used to investigate the relationship between ammonium wet deposition and precipitation, and kriging with adjustments for anisotropy, used to better understand ozone and PM2.5 concentrations in the northeastern U.S. In addition, we have recently assessed the impact on model evaluation of incommensurability—that is, the mismatch between point-based measurements and areal averages (model output). Ideas for improving regional air quality model evaluation techniques were explored at an American Meteorological Society (AMS)- and Division-sponsored workshop during the summer of 2007.

Lastly, the Atmospheric Model Evaluation Tool (AMET) was made publically available. AMET is a combination of open-source software that includes a relational database to store paired observed-predicted values and a statistical program to create various plots and calculate statistics. AMET is a valu­able tool that can aid in the evaluation of both meteorological and air quality simulations. Because AMET utilizes a relational database, the user can query data in the database based on any number of criteria, making it ideal for identifying any specific problems that may exist in the model predictions. Work to improve AMET and extend its capabil­i­ties will continue in the future.

Next Steps
Over the next several years, science and technology advancements planned for the CMAQ modeling system include enhanced emissions modeling and additional model system evaluation. Some of the planned milestones under this Theme Area are the following:

FY-2008

FY-2009

FY-2010

Impacts and Transition of Research to Applications
The Division releases versions of the CMAQ model and associated programs to the public through the ORD-supported CMAS Center; the Center also provides user support and training. The community air quality modeling concept, especially the CMAQ model, has seen growing acceptance since the model was first released in 1998. An annual CMAQ model users conference now attracts over 200 people each year from North and South America, Europe, and Asia.

EPA/OAQPS and the states use CMAQ for assessments conducted during national air quality rulemaking and in their SIPs, respectively. OAQPS has used the model to assess the potential effectiveness of the CAIR and the CAMR. The states, through their Regional Planning Organizations (RPOs), are using CMAQ for visibility assessments in support of the Regional Haze Rule (RHR) and for upcoming SIP assessments for O3 and PM2.5. The CMAQ model is also being used in Canada, the U.K., Spain, Eastern European countries, China, Korea, and many other nations in programs to improve regional air quality management. NOAA’s National Weather Service (NWS), in a collaborative project with EPA, is using CMAQ to make publicly available short-term (next-day) forecasts of ozone air quality across the eastern United States (see Chapter 6).

The end result of all of these efforts will be the ability to better inform (1) the public on current air quality conditions (from forecasting applications), to help them make decisions on health-related exposures to air pollution, and (2) policy makers (from air quality model assessments) to guide them in making the best long-term emission control decisions to reduce air pollution.

The part of the Division organizationally associated with OAQPS oversees and facilitates the process of transitioning the tools we develop and evaluate to regulatory applications, thus providing the foundation for scientifically sound regulatory decisions.

29141 – USB cable – 6.6 FT

Schematic of CMAQ modeling system, including meteorology, emissions, and air quality models.
Figure 2-1. Schematic of CMAQ modeling system, including meteorology, emissions, and air quality models.

Back to Top

Chapter 3
Evaluating the Impact of Regulatory Policies on Air Quality and Ecosystems

Introduction
As discussed in the introduction to Chapter 2, air quality management in the United States is implemented for criteria pollutants through NAAQS. States with nonattainment areas (areas that do not meet the NAAQS for one or more pollutants) must submit SIPs that demonstrate how the state will reduce emissions to achieve attainment. Most criteria pollutants are transported across state boundaries, which complicates the nonattainment issue. Recent rulemakings have recognized that this transport must be considered, requiring that a regional perspective be used when developing strategies for air pollution nonattainment.

In 1998, EPA finalized a rule known as the NOx Budget Trading Program (NBP), requiring 22 states and the District of Columbia to submit SIPs that address the regional transport of ground-level ozone. The actions directed by these plans include reducing emissions of nitrogen oxides (which are a precursor to ozone formation), thereby decreasing the formation and transport of ozone across state boundaries.

The Clean Air Rules of 2004 are a suite of actions designed to improve air quality. Three of the rules specifically address the transport of pollution across state borders. The CAIR will per­manently cap emissions of sulfur dioxide (SO2) and NOx from utilities in the eastern United States. When fully implemented in 2015, CAIR will reduce SO2 emissions in these states by over 70% and NOx emissions by over 60% from 2003 levels. CAMR will build on CAIR to reduce mercury emissions from coal-fired power plants. The Non-Road Diesel Rule will reduce emissions from future non-road diesel engines by changing the way diesel engines function (to reduce emissions) and the way diesel fuel is refined (to remove sulfur).

Deposition of atmospheric nitrogen, sulfur, and mercury to land and water surfaces contributes significant loadings to receiving water bodies, affecting the health of ecosystems. For example, atmospheric deposition of nitrogen accounts for about 30% of the nitrogen coming into the Chesapeake Bay. CAA regulations, including the NBP, CAIR, and CAMR, are expected to reduce the atmospheric deposition of these pollutants.

Research Description
Given the significant costs associated with these rules and control measures, it is important to demonstrate their effectiveness. The Division has demonstrated reductions in observed and modeled ozone concentrations resulting from actions of the NBP. Research will continue to develop ways to systematically track and periodically assess our progress in attaining national, state, and local air quality goals—particularly those related to criteria pollutants regulated under the NAAQS and the Clean Air Rules.

Research under this Theme Area falls into two categories:

The major research questions addressed by this research include the following:

The CMAQ modeling system is used to characterize air quality before and after the implementation of a target regulation and to evaluate correlations between changes in emissions and changes in pollutant concentrations or atmos­pheric deposition. Various scenarios are modeled to estimate the anthropogenic contribution to total ambient concentrations and the impact of not implementing the regulation. Methods have been developed to differentiate changes attributable to emission reductions from those that result from other factors, such as weather and annual and seasonal variations in emissions. Trajectory models, such as NOAA’s Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model, are used to investigate the transport of primary and secondary pollutants from their sources to downwind regions.

Research is initially focusing on regulations affecting NOx and SO2, for which emissions monitoring data are available (e.g., NBP and CAIR regulations). Later research will investigate using other sources of information, such as remote sensing, to evaluate regulations that impact pollutants such as PM and mercury, for which emissions data are sparse or uncertain.

Specifically, we are developing indicators to assess changes in emissions and air quality associated with regulatory actions, and modeling approaches to characterize the processes that impact the relationships between these indicators (process linkages). Figure 3-1 indentifies the full suite of indicators and process linkages associated with the evaluation of the NBP rules. Previous efforts performed under this Theme Area developed the indicators characterizing changes in emissions and ambient NOx and ozone concentration levels. Models and data analyses were used to relate the changes in emissions to the changes in ambient NOx (emitted precursor pollutant) and ozone (secondary pollutant) concentrations by directly relating the fate and transport of these pollutants to levels downwind of their sources.

Accomplishments
The results of research under this Theme Area have indicated that when major point sources of NOx were reduced by the NBP, this decreased ozone concentration levels by 5-8 parts per billion (ppb) at downwind locations. In 2007, evalu­ations of the chemical and physical processes further indicated that, while a dramatic reduction in maximum ozone chemical production rates occurred downwind of major point sources affected by the NBP, net ozone production efficiency actually increased due to the greater decrease in reactive nitrogen product species (NOz). This and other results indicate that the chemical regime has shifted toward more NOx-limited conditions in the plume-impacted areas downwind of the sources, meaning that relatively small increases in NOx emissions (e.g., from the transport corridors in the eastern U.S.) can result in a relatively large increase in ozone, due to changes in production rate efficiencies. Overall, our research has shown that emission control programs implemented under the NBP have been effective in meeting the objective of reducing interstate ozone transport, and have helped improve ozone air quality in source areas of the eastern United States. These results contributed to the annual assessment of the NBP in Report EPA-430-R-07-009, NOx Budget Trading Program 2006: Program Compliance and Environmental Results.

Next Steps
Research over the next five years will

  1. continue the assessment of the NBP by applying ambient concentration and exposure indicators to a health and risk assessment in the greater New York State area; and
  2. assess the impact of the phased implementation of CAIR through the application and further develop­ment of indicators and process linkage methodologies developed for assessing the NBP.

The process of developing indicators and process linkages for assessing the NBP will not only establish an approach for assessing CAIR, but will also establish a baseline description of the state of the environment before the implementation of CAIR. Major deliverables anticipated from this research include the following:

FY-2008

FY-2009

FY-2010

FY-2012

Impacts and Transition of Research to Applications
Quantifying the improvement in air quality and human and ecological health brought about by costly regulations is critical in evaluating whether these actions are making the difference originally anticipated. Research under this Theme Area evaluates the effectiveness of specific regulatory actions. Methods developed for these evaluations will also provide a framework for assessing future regulatory actions. These methods will include

This effort transitions research results to applications by demonstrating the use of the CMAQ and HYSPLIT models and statistical techniques to evaluate the impact of regulations implemented to improve air quality.

Assessing the impact of regulaltions on ecosystems and human health endpoints showing the indicators and process linkages associated with the NOx BUdget Trading Program
Figure 3-1. Assessing the impact of regulations on exosystems and human health endpoints showing the indicators (boxes) and process linkages (arrows) associated with the NOx Budget Trading Program.

Back to Top

Chapter 4
Linking Sources to Human Exposure

Introduction
The goal of this research theme is to reduce uncertainties in quantifying the link between sources of atmospheric pollution and human exposure. The CAA requires EPA to assess which HAPs pose the greatest risk to humans in the United States, and to develop strategies for controlling harmful concen­trations of these compounds. These assessments typically involve the application of different models, depending on program objectives—global, regional, urban, or local scale (Figure 4-1). Performing these assessments often requires a linkage between ambient air quality and human exposure models. The Division conducts research to build this linkage by combining the features of grid-based, regional-scale chemistry-transport models and urban-scale dispersion models. This research facilitates the use of air quality model concentrations in human exposure modeling and health risk assessments, which historically have been limited by their need to rely upon monitored concentrations at a central site.

For exposure assessments, air quality modeling should include local-scale features, long-range transport, photochemistry, and deposition to provide the best estimates of air concentrations. Generally speaking, the two major types of air quality models are source-based Gaussian dispersion models and grid-based chemistry-transport models. Chemistry-transport models, such as CMAQ, can provide estimates of photochemically formed pollutants typically at a 36- to 4-km grid scale, but not local-level details. CMAQ provides volume-average concen­tration values for each grid cell in the modeling domain for given conditions. Emissions are assumed to be instan­taneously well-mixed within the grid cell in which they are emitted. While grid-based models are preferred for simulation of chemically reactive airborne pollutants, dispersion models (such as the AMS/EPA Regulatory Model Improvement Committee [AERMIC] Model [AERMOD] have been developed to simulate the near-field fate of airborne pollutants that are relatively chemically inert.

For multipollutant assessments, a suite of toxic compounds needs to be included in the CMAQ modeling system, and model results should be evaluated with ambient observational data. This research need is closely linked to other research themes within the Division that involve the development and evaluation of the modeling system, improvements in chemical and physical characterization of air toxics, and the measure­ment of ambient air toxics concentrations.

Because exposure assessments are primarily for urban areas, air quality simulation models should accurately depict the physical-chemical processes that occur in these areas. Concentration fields derived from models run at grid resolu­tions on the order of 4 km or larger (such as CMAQ) do not account for the variability of high emission gradients typical in urban areas. Several approaches are available that may yield a better characterization of urban “hot spots,” including brute-force simulations with finer-scale grid models, hybrid modeling that combines chemistry-transport models with dispersion models, and sub-grid variability distribution estimates of concentrations. Meteorological models such as the MM5 and WRF modeling systems now include the capability to assimilate advanced urban canopy descriptions, including building, vegetation, and street canyon character­istics. Databases of high-resolution urban morphological features are needed to support these advanced models for future urban evaluation and application.

A growing number of health studies have identified adverse effects—including respiratory disease, cancer, and death—for populations exposed to air pollution near major roadways, thus raising concerns about building schools near roadways and the general health of people living near roads. Performing near-roadway risk assessments requires characterizing atmos­pheric processes in complex urban settings, especially near major roadways. Near-road air pollution has been selected as a central theme in EPA/ORD’s multiyear clean air research plan, because it is a problem that is of pressing importance (as identified by EPA’s stakeholders), and it requires an integrated, multidisciplinary field and laboratory scientific approach.

Research Description
The Division’s work in this Theme Area is broken into the following two research tasks:

Within the first task, multiscale modeling of air toxics involves

  1. including chemistry and physics for additional toxic air pollutants in CMAQ;
  2. applying CMAQ with toxics for problems of interest to Program and Regional Offices, and the transfer of information to these stakeholders;
  3. reformatting the results from air quality model simulations for use in the Stochastic Human Exposure and Dose Simulation (SHEDS) model; and
  4. developing methods and tools that can be used to predict air pollutant concentrations at urban (or neighborhood) scales, and using these tools to assess the magnitude and variability of concentrations to which urban populations are exposed.

To incorporate the salient features of both grid-based and plume-dispersion approaches, we have been testing a hybrid approach that combines results from CMAQ with the AERMOD model. The CMAQ grid model provides the regional background concentrations and urban-scale photo­chemistry, and the AERMOD local plume dispersion model provides the air concentrations that are due to local emission sources. The results of both model simulations are combined to provide ambient air concentrations for use in exposure models. The advantage of this modeling approach is that researchers can incorporate the spatial and temporal variation of air pollution within a study area without having to rely on dense ambient monitoring networks. This hybrid approach is currently being explored in several studies, including an air quality and exposure study in Detroit, MI, and an accountability study in New Haven, CT.

As a complement to hybrid modeling, we are exploring other methods to obtain model concentration fields at spatial scales needed for improved exposure and risk assessments. This entails running CMAQ with higher-resolution grid meshes (smaller grid cells) than is the normal practice. We are also investigating the use of urbanized versions of the MM5 and WRF models to drive the CMAQ model at 1-km grid resolutions. In addition, partnerships with external collabora­tors are being leveraged to study ways to parameterize concentration distribution statistics to augment CMAQ outputs at 12- or 4-km grid resolutions, based on outputs of fine-scale grid models and/or use of hybrid modeling approaches.

To support improved urban-scale meteorological modeling, the Division is leading the creation of the National Urban Database and Access Portal Tool (NUDAPT). As part of this effort, we are conducting collaborative studies with NCAR on the urbanized version of the WRF model.

Regarding near-roadway modeling, the second task within this Theme Area: Before 2007, the Division was engaged in a number of loosely coordinated research projects involving the near-road environment, including research to support homeland security efforts. In 2007, EPA/ORD initiated a cross-laboratory coordinated near-road research program. The Division is meeting the physical and numerical dispersion modeling needs of this program, by assisting in the design and analysis of field experiments, by conducting laboratory dispersion studies, and by developing improved numerical algorithms for modeling near-road dispersion of emissions from major roadways. Our focus is to examine the signi­fi­cance of near-road emissions from varied roadway conditions on human exposure and related health risks, and to develop tools for addressing this issue.

Accomplishments
The CMAQ modeling system has been modified to include HAPs, and results from the revised model have been coupled with the near-field dispersion model AERMOD to account for urban-scale gradients of air toxics. In addition, outputs from this coupled system have been successfully linked to the SHEDS model and the Hazardous Air Pollutant Exposure Model (HAPEM). This research has been performed in collaboration with scientists from EPA/ORD/NERL’s Human Exposure and Atmospheric Sciences Division (HEASD) and EPA/OAQPS.

During the past two years, we have developed the hybrid approach to estimate concentrations for multiple pollutants that reflects both local features (hot spots) and regional transport. The local impacts from mobile sources and signi­ficant stationary sources are estimated using AERMOD, and the combined concentrations are used for subsequent human exposure analysis. During 2007, we demonstrated an application of this linkage for New Haven, CT. This project is a collaborative effort with state and local agencies, including government, academia, and the New Haven community, to apply and evaluate air quality and human exposure models that can be used with health data. The project goal is to assess the feasibility of using this information to conduct an air accountability study (i.e., to trace the impact of air quality changes through to human health impacts).

Efforts have continued on methods to derive sub-grid variability (SGV) distributions from a combination of ~1-km-grid-resolution CMAQ model simulations and hybrid model results. With this approach, each 4- or 12-km CMAQ grid cell is assigned SGV characteristics, such as the type of distri­bution or the range of concentrations corresponding to user-prescribed percentile values. Initial efforts have focused on SGV distribution functions derived from the Wilmington, DE, and Houston, TX, modeling results. The use of Weibull distributions seems promising. The SGV approach may prove useful for applications that can incorporate estimates of SGV on an a priori basis, such as with population exposure studies.

For urban-scale meteorological modeling, a prototype version of NUDAPT was completed for the Houston area. NUDAPT’s portal features allow users to adapt processed fields of urban canopy parameters and other gridded fields for use with different grid resolutions and map projections. NUDAPT also includes sets of advanced urban canopy parameter implementations for the MM5 and WRF modeling systems. A workshop of the federal, state, academic, and private collaborators was conducted in Boulder, CO, during spring 2007 as a means to perform the initial implementation of NUDAPT. Sensitivity studies using urbanized MM5 and WRF to drive CMAQ for urban applications were begun. For Houston, sensitivity runs of MM5 at 1-km resolution are being performed using both the standard MM5 model and an urbanized MM5 based on an urban canopy approach. Similarly, collaborations are underway with NCAR regarding their urbanized WRF model. Preliminary CMAQ-Toxics (CMAQ-TX) simulations have been made at 1-km grid resolution for the Houston and Wilmington domains; the Delaware modeling is being performed in collaboration with the State of Delaware. A preliminary survey of the distribu­tion functions has been conducted from the Wilmington and Houston CMAQ results. A special session at the 2007 annual CMAS conference showcased the NUDAPT efforts, including a demonstration of the NUDAPT initial prototype.

During 2007, the Near-Roadway and School Infiltration Research Initiative project continued. Fourteen roadway configurations were identified, and a physical model was created for each configuration for performing modeling in the Division’s Meteorological Wind Tunnel. The configurations included a flat roadway with no surrounding obstacles (base case), noise barriers of varied heights and distances from roadway, two different porous barriers intended to simulate rows of vegetation, and depressed and elevated roadways. The experiments generated three-dimensional data sets comprising winds, turbulence, and tracer-gas concentrations. Preliminary results show that the solid noise barriers have a substantial effect on downwind concentrations. When winds blow across the roadway, the barriers increase vertical turbulence. This causes the plume from the roadway to mix more vigorously in the vertical direction, which results in decreased ground-level concentrations immediately downwind of the road. For a single upwind barrier, the downwind concentration (near the edge of the roadway) decreases by a factor of four compared to the base case. By adding a second solid barrier on the downwind edge of the roadway, the downwind concentration decreases by a factor of six compared to the base case. With an upwind “vegetation” barrier of 58% porosity, only minor differences are seen in the downwind concentrations compared to the base case. Although the simulated upwind vegetation causes a modest increase in the vertical extent of the plume, downwind concentrations decrease less than 7% over the base case. A vegetation barrier with less porosity (23%, repre­senting more dense vegetation) shows a decrease in near-road downwind concentrations of about a factor of two. Denser or taller vegetation would be expected to produce greater differences in the concentration field. Finally, depressed roadways are found to affect the downwind concentration fields in a way similar to the case with noise barriers on both sides of the road.

A number of journal articles are being prepared using the results from the wind tunnel, and the data obtained are being used to verify numerical algorithms and to improve the line-source algorithm used in near-roadway dispersion models.

Several members of the Division participated in the Raleigh 2006 Pilot Field Study, which was an EPA/ORD multi­laboratory collaborative effort involving EPA/NERL and EPA’s National Health and Environmental Effects Research Laboratory (NHEERL) and National Risk Management Research Laboratory (NRMRL). The Raleigh field study, conducted in the summer of 2006, was designed to provide data to characterize the influence of traffic-generated emissions in the near-road environment, especially to help assess their impact on air quality and particle toxicity near a heavily traveled highway. The study included several real-time and time-integrated sampling devices that measured air quality concentrations at multiple distances and heights from the road. Pollutants analyzed included EPA-regulated gases, particulate matter (coarse, fine, and ultrafine), and air toxics. Pollutant measurements were synchronized with real-time traffic and meteorological monitoring devices to provide continuous and integrated assessments of the variation of near-road pollutant concentrations and particle toxicity with changing traffic and environmental conditions, as well as distance from the road. This research task helped provide the analysis used to demonstrate the temporal and spatial impact of traffic emissions on near-road air quality.

Using funding from the EPA/ORD near-roadway research initiative, a support contract provided a comprehensive review of available/operational air quality and emission models. A final draft report was completed, and work is underway to transform this report into a review article, possibly for the Journal of the Air Waste and Management Association. This review will provide the air quality modeling community with a convenient summary of the current state-of-science and will serve as a guide for the research needed to develop improved near-roadway air quality and emission models.

Next Steps
During the next few years, the Division is expected to increase emphasis in the areas of near-roadway modeling and linkage of air quality models with human exposure models to assess human health. A major effort during 2008 will be to create and implement a version of the SAPRC07-TX chemical mechanism within the CMAQ system. This mechanism will go directly to a “multipollutant” form. We also plan to add cloud chemistry for chromium compounds and address adding arsenic compounds into the CMAQ-TX model. If time permits, we will start studying the inclusion of polycyclic aromatic hydrocarbons (PAHs) into CMAQ-TX. We hope to simulate the seven to sixteen PAHs that are suspected of causing damage to human health, based on laboratory studies. Besides requiring revision of model algorithms, PAHs present an additional difficulty because parts of the emissions inven­tory for HAPs lump several PAHs into one emissions rate.

Also during 2008, simulations with CMAQ-TX for Baltimore will be performed, as well as simulations to examine the effect of alternative mobile-source fuel compo­si­tion on concentrations of toxic aldehydes. In 2009, we will analyze output from the Baltimore simulations and provide the ambient concentration predictions for input to Environmental Benefits Mapping and Analysis Program (BenMAP) benefits model analyses being performed by EPA/OAQPS.

To support urban modeling, a “fine-scale” Division work­group has been established to develop a more detailed research plan for investigating the efficacy of adapting WRF/CMAQ to a grid mesh of less than 4 km for various test beds. This plan will be used to guide and perhaps redirect research within the Division during 2008 through 2011.

Via an existing collaboration with the State of Delaware, we will continue examining and refining the characteristics of the SGV distributions using both fine-scale and hybrid modeling approaches. Methods to utilize these distributions will be investigated for developing parameterization of SGV varia­tions for coarse (4‑ and 12-km) grid resolutions. We may also investigate parameterizations of SGV distributions with either off-the-shelf or alternative software specifically developed for deriving parametric forms of the distribution function.

Further, we plan to explore developing an easy-to-use method to create modeling input files for on-road mobile sources at a link-based level, and to assess the impact of more spatially resolved emissions on modeled ambient air pollutant con­centrations. We will continue conducting uncer­tainty analyses to evaluate model results and explore an “ensemble-based” approach for generating probabilistic concentration fields.

Follow-on collaboration with NCAR will investigate use of the urbanized version of WRF for driving CMAQ for urban applications. We will also perform sensitivity studies using advanced urbanized versions of MM5 and WRF with CMAQ to examine the impact of using the NUDAPT data in urban areas.

Follow-on wind tunnel dispersion studies will be conducted over several years to expand the near-roadway scenarios to include the influences of wind direction variations, nearby buildings, and upwind and downwind vegetation. The Division and NOAA’s Air Resources Laboratory (ARL) Field Research Division (FRD) are planning to perform a tracer-gas dispersion experiment to enhance a field study that will be conducted in Las Vegas during 2008-2009. The overall purpose of the EPA field study is to characterize the spatial gradient of pollution within ~200 m of a major highway. The research plan envisions the release of SF6 from a 100-m perforated pipe under various flow regimes. Plans have been proposed to deploy sonic anemometers near I-15 in Las Vegas to characterize the decay of vehicular-induced turbulence as a function of distance from the roadway. Further discussions between the Division and FRD will occur during 2008. Design of the field study will be supported by a wind tunnel study of the flow and dispersion in a 1:200 scale model of the selected Las Vegas field site, which will be performed during summer 2008. Data collected in the field and the laboratory, as well as detailed numerical modeling studies from models such as the Quick Urban and Industrial Complex (QUIC) model, will be analyzed and a refined line-source algorithm will be proposed for inclusion in the AERMOD model.

In future years (2009-2011), the Division plans to return to wind tunnel studies of urban street canyon flows within a scale model of a large urban center (Midtown Manhattan). We will examine the general structure of complex urban boundary layers. The Midtown Manhattan model provides a “target of opportunity” because a physical model already exists from an earlier homeland security project. These data, in combination with those from similar wind tunnel studies of urban centers (Lower Manhattan, Oklahoma City, etc.) and from computa­tional fluid dynamics modeling (existing, from studies performed outside of the Division) of the same Midtown area, can be used to characterize the influence of the urban landscape on grid-average concentrations within regional-scale modeling analyses, and could provide the basis for improved urban dispersion algorithms within near-field and hybrid modeling approaches.

Research under this Theme Area is expected to contribute to the following research milestones:

FY-2008

FY-2009

FY-2010

FY-2011

FY-2012

Impacts and Transition of Research to Applications
Development and application of linked models of ambient air quality and human exposure will help epidemiologists reduce uncertainty as they assess the risk of air pollutants to human health, and will help policy-makers reduce uncertainty as they develop control strategies that target air pollutants identified as posing the greatest risk to humans. These uncertainty reductions should result in more accurate risk assessment results and in policies that are more likely to protect human health.

Multiple scales in air quality modeling
Figure 4-1. Multiple scales in air quality modeling


Chapter 5
Linking Sources to Ecosystem Exposure

Back to Top

Introduction
Ecosystems provide us with many life-sustaining benefits—resources and services that contribute to our physical, social, and economic welfare; examples of ecosystem services include clean air and water, fertile soil for crop production, pollination, and flood control. A long-term goal of envi­ron­mental management is to achieve sustainable ecologi­cal resources through a comprehensive assessment of current and projected ecosystem health and services. Such an assess­ment must include identification of the major threats (the specific stressors) to ecosystem health, the sources of those stressors, and how they move through the environment. This is fundamentally a problem of multimedia pollution.

The overall objective of this Theme Area is to develop the atmospheric components of multimedia modeling and assessment tools to allow better management and protection of ecosystems and their associated resources and services. The Division is developing a suite of linked models, tools, and technology to provide long-range ecological forecasts and a scientific basis for decision-making to protect terrestrial and aquatic ecosystems. This research supports EPA’s expanded definition of air quality management that includes ecosystem protection in assessments of air pollution regulations, i.e., in the setting of secondary NAAQS. It also supports the new emphasis of EPA’s Ecosystem Research Program (ERP) on linking sources to exposure in a multipollutant context and developing capabilities for ecosystem services assessments.

The interaction between the atmosphere and the underlying surface is increasingly being recognized as a significant factor in multimedia issues. Atmospheric deposition is an important source of ecosystem stressors, in particular for acidification, eutrophication of coastal estuaries due to excess nitrogen, and bioaccumulation of mercury. Managing the nitrogen cycle is a central issue of the ESRP. Critical-load is the amount of deposition above which natural resources can be negatively affected and is intended as a protective threshold. The Nation­al Academy of Sciences (NAS) has recommended that EPA consider a deposition-based approach such as critical loads to ecosystem management. In support of the ESRP thrust and the NAS recommendation, the Division conducts research to provide the most accurate atmospheric deposition estimates possible.

The Clean Water Act administered by the EPA requires states to develop Total Maximum Daily Load limits (TMDLs), the maximum amount of a pollutant that a body of water can receive while still meeting water quality standards. The atmosphere is an important contributor to stressors such as excess nutrients, but atmospheric deposition is seldom considered in the development of TMDLs. The Division is conducting research to improve the understanding of the atmospheric contribution of stressors to TMDLs.

Research Description
For this research theme, the Division has identified research areas that have the most potential to reduce critical uncer­tain­ties in atmospheric deposition, to assess program account­a­bility, and to link atmospheric deposition to ecosystem resources and services.

Specific research tasks are grouped under the following research program elements:

Through the Air-Surface Research and Development program element, the Division develops and enhances air-surface exchange modules for CMAQ, and advances the link­age between CMAQ and the underlying land-use categories to facilitate improved interactions with ecosystem models. We also develop and enhance air-surface exchange modules for monitoring network operations using an inferential method for dry deposition, focusing primarily on sulfur, nitrogen, and mercury species. The bidirectional air-surface exchange pro­cess is a new feature of this program element.

Focus areas of the Air-Surface Research and Development program element include the following:

Through the second program element, Multimedia Applica­tions, the Division develops and improves linkages between air and water models, and develops and maintains connections to ecosystem resources and services through participation with partners in multimedia assessments. Simulation of deposition estimates at a National scale is an important output from these efforts.

Focus areas of Multimedia Applications include the following:

Through the Multimedia Tool Development program element, the Division develops tools for specialized multimedia analyses and applications involving atmospheric models. The need for specialized tools is especially pertinent to bringing atmospheric components together with watershed components for multimedia management analyses. Most off-the-shelf tools do not address the specialized needs encoun­tered in analyzing data from a multimedia perspective. Significant effort is often required to analyze observations and model results and provide them in forms that are required for supporting management decisions.

Focus areas of Multimedia Tool Development include the following:

Local Navigation

Jump to main content.