{"Bibliographic":{"Title":"Fiscal year 2004 summary report of the NOAA Atmospheric Sciences Modeling Division to the U.S. Environmental Protection Agency","Authors":"","Publication date":"2005","Publisher":""},"Administrative":{"Date created":"08-17-2023","Language":"English","Rights":"CC 0","Size":"0000245300"},"Pages":["QC807. . 5\n.U6\nA7\nno. 255\n2005\nA\nOF\nNOAA Technical Memorandum OAR ARL-\nSTATES\nof\nFISCAL YEAR 2004 SUMMARY REPORT OF THE NOAA ATMOSPHERIC SCIENCES\nMODELING DIVISION TO THE U.S. ENVIRONMENTAL PROTECTION AGENCY\nEvelyn M. Poole-Kober\nHerbert J. Viebrock\n(Editors)\nAtmospheric Sciences Modeling Division\nResearch Triangle Park, North Carolina\nAir Resources Laboratory\nSilver Spring, Maryland\nJune 2005\nNOAA/EPA Golden Jubilee\nATMOSPIC\nAND\nINSURANCE\nNOAA\nNOAA - EPA Partnership\nThis year marks the 50th Anniversary of the collaboration between the U.S. Department of\nCommerce's National Oceanic and Atmospheric Administration (NOAA) and the U.S.\nEnvironmental Protection Agency (EPA), and their predecessor agencies on air quality\nmodeling.\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION\nOFFICE OF OCEANIC AND ATMOSPHERIC RESEARCH","S\nNOAA\nBetty Pet\n103\nWorld\nOF\nCOMMUNITY\nCamp\nNOAA Technical Memorandum OAR ARL-255\nSTATES\nFISCAL YEAR 2004 SUMMARY REPORT OF THE NOAA ATMOSPHERIC SCIENCES\nMODELING DIVISION TO THE U.S. ENVIRONMENTAL PROTECTION AGENCY\nEvelyn M. Poole-Kober\nCenter\nLibrary\nHerbert J. Viebrock\n(Editors)\nNOAA\nAtmospheric Sciences Modeling Division\nResearch Triangle Park, North Carolina\nAir Resources Laboratory\nSilver Spring, Maryland\nJune 2005\nATMOSPIC\nNOAA\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION\nOFFICE OF OCEANIC AND ATMOSPHERIC RESEARCH","NOTICE\nMention of a commercial company or product does not constitute an endorsement by NOAA.\nUse for publicity or advertising purposes of information from this publication concerning\nproprietary products or the tests of such products is not authorized.\nDISCLAIMER\nThe research presented here was performed under the Memorandum of Understanding between\nthe U.S. Environmental Protection Agency (EPA) and the U.S. Department of Commerce's\nNational Oceanic and Atmospheric Administration (NOAA). This work constitutes a\ncontribution to the NOAA Air Quality Program.\nFor sale by the National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161.\nii","PREFACE\nThis year marks the 50th Anniversary of the collaboration between the U.S. Department of\nCommerce's National Oceanic and Atmospheric Administration (NOAA) and the U.S.\nEnvironmental Protection Agency (EPA), and their predecessor agencies on air quality modeling.\nThe relationship between the NOAA and EPA began when the Air Pollution Unit of the Public\nHealth Service, which later became part of the EPA, requested the Weather Bureau to provide it\nwith meteorological expertise. Thus, in 1955, a special Weather Bureau air pollution unit was\nformed and located in Cincinnati, Ohio, until it moved in 1969 to Raleigh, North Carolina, and\nintegrated with the Public Health Service. The unit is now the NOAA Atmospheric Sciences\nModeling Division (ASMD), working within the framework of the Memorandum of\nUnderstanding and Memorandum of Agreement between the U.S. Department of Commerce and\nEPA. These agreements are implemented through long-term Interagency Agreements\nDW13938483 and DW13948634 between EPA and NOAA.\nThis report summarizes the Fiscal Year 2004 research and operational activities of the\nDivision. The summary includes descriptions of research and operational efforts in air pollution\nmeteorology, meteorology and air quality model development, evaluation and application, and air\npollution abatement and compliance programs. ASMD serves as the vehicle for implementing\nthe interagency collaboration on atmospheric research efforts. ASMD conducts research\nactivities in-house and through contracts and cooperative agreements, and provides atmospheric\nsciences expertise, air quality forecasting support, and consultation and guidance on the\nmeteorological and air quality modeling aspects of air quality management to various EPA\noffices, including the Office of Air Quality Planning and Standards. To provide these services,\nthe Division is organized into four research Branches: Atmospheric Model Development Branch,\nModel Evaluation and Applications Research Branch, Air-Surface Processes Modeling Branch,\nand Applied Modeling Branch. This report is organized by major program themes reflecting the\nDivision strategic plan, consistent with NOAA's mission and strategic goals.\nAll inquiries on the research or support activities outlined in this report should be sent to\nthe Director, NOAA Atmospheric Sciences Modeling Division (E243-02), U.S. Environmental\nProtection Agency, 109 T.W. Alexander Drive, Research Triangle Drive, NC 27711.\nNOAA Science Center\nBetty Petersen Memorial Library\nWorld Weather Bldg., Room 103\n5200 Auth Road\nCamp Springs, MD 20746\n9-13-05\niii","iv","CONTENTS\nPage\niii\nPREFACE\nviii\nFIGURES\nTABLES\nX\nABSTRACT\n1\n1. INTRODUCTION\n1\n2. PROGRAM REVIEW\n2\n2.1 Atmospheric Model Development\n2\n2.1.1 Meteorological Modeling for CMAQ Applications\n3\n2.1.2 Linking Meteorology and Chemistry Models for Research Applications\n3\n2.1.3 Planetary Boundary Layer and Land Surface Modeling\n4\n2.1.4 High-Resolution Sea Surface Temperature Initialization for Meteorological\nModels\n5\n2.1.5 Anthropogenic Emissions\n6\n2.1.6 Biogenic Emissions\n6\n2.1.7 Modeling Smoke Emissions from Fires\n7\n2.1.8 Implementation and Testing of New and Refined Chemical Mechanisms\nand Chemical Solvers in CMAQ\n8\n2.1.9 Aerosol Mechanism Improvements in CMAQ\n9\n2.1.10 Plume-in-Grid Modeling\n11\n2.1.11 CMAQ Mercury Model Refinements and Evaluation\n13\n2.1.12 CMAQ Code Integration and 2004 Release\n14\n2.1.13 Development and Testing of an Air Quality Forecast Model\n16\n2.1.14 Linking the Eta Model with CMAQ for Air Quality Forecasting\n17\n2.1.15 Evaluation of Eta-CMAQ Forecast Predictions for Summer 2004\n17\n2.1.16 Assessment of the Real-Time Application of CMAQ\n18\n2.2 Atmospheric Model Evaluation and Application Activities\n19\n2.2.1 Meteorological Model Evaluation\n19\n2.2.2 Diagnostic Metrics for Ozone and Inorganic Particulate Matter\n20\n2.2.3 Diagnostic Evaluation for Carbonaceous Aerosol Components\n21\n2.2.4 CMAQ Model Evaluation to Assess Model Readiness for Application\n22\n2.2.5 Spectral Analysis of the Observed and Predicted Meteorology and Air\nQuality\n24\n2.2.6 Model Evaluation Using Advanced Spatial Statistical Models\n25\nV","2.2.7 Temporal Signatures of Model Output and Observations\n27\n2.2.8 Reduction of the Space-Time Domain Dimensionality for Evaluation of\nModel Performance\n28\n2.2.9 Objective Comparison of CMAQ and REMSAD Performances\n31\n2.2.10 Sensitivity of CMAQ Control Strategy Predictions to Model Input\nUncertainties\n32\n2.2.11 Inverse Modeling for Ammonia: A New Emission Inventory and an\nAnnual Simulation\n33\n2.2.12 Bay Regional Atmospheric Chemistry Experiment Model Evaluation\n34\n2.2.13 Model Evaluation Tool Development\n35\n2.2.14 Evaluating the Effect of NO Reductions\n35\n2.3 Toxic Air Pollutant Modeling\n36\n2.3.1. Extending CMAQ to New Species of Air Toxics\n37\n2.3.2. Comprehensive Version of CMAQ for the National Air Toxics\nAssessment\n39\n2.3.3 Linking CMAQ to a Human Exposure Model in an Urban Area\n43\n2.3.4 Parameterizing the Urban Canopy\n44\n2.3.5 Advancing the Neighborhood-Scale Version of CMAQ\n48\n2.3.6 Modeling Subgrid Concentration Variability\n50\n2.3.7 Developing and Applying CFD Simulations of Pollutant Transport and\nDispersion\n50\n2.4 Multimedia Modeling and Application Studies\n59\n2.4.1 Multimedia Integrated Modeling System Spatial Allocator\n59\n2.4.2 Multi-Layer BioChemical Model for Calculating Dry Deposition\n59\n2.4.3 Chesapeake Bay 2007 Re-Evaluation\n60\n2.4.4 Ammonia Budgets for Coastal Systems\n61\n2.4.5 Tampa Bay Study\n62\n2.4.6 Multimedia Research for CMAQ-Hg\n62\n2.4.7 Multimedia Tool Development\n63\n2.5 Climate Change Impacts on Regional Air Quality\n64\n2.5.1 Regional Climate Downscaling of Meteorology\n64\n2.5.2 Chemical Emissions Processing\n64\n2.5.3 Global Climate and Chemical Transport Simulations\n65\n2.5.4 Methods Developed for Analysis and Evaluation of Regional Climate\nSimulations\n65\n2.6 Specialized Client Support\n66\n2.6.1 European Monitoring and Evaluation Program\n66\n2.6.2 Support of the 1999 National Air Toxics Assessment\n67\n2.6.3 The Philadelphia Air Toxics Project\n67\n2.6.4 Support Center for Regulatory Air Models\n68\nvi","2.6.5 Eta Data Assimilation System Review\n68\n2.6.6 Development of a Response Surface Model for Ozone\n69\n70\nREFERENCES\nAPPENDIX A: ACRONYMS, ABBREVIATIONS, AND DEFINITIONS\n76\n80\nAPPENDIX B: PUBLICATIONS\n87\nAPPENDIX C: PRESENTATIONS\nAPPENDIX D: WORKSHOPS AND MEETINGS\n92\nAPPENDIX E: VISITING SCIENTISTS\n102\nAPPENDIX F: POSTDOCTORAL RESEARCHERS\n105\nAPPENDIX G: ATMOSPHERIC SCIENCES MODELING DIVISION STAFF AND\n106\nAWARDS\nvii","Page\nFIGURES\nFigure 1 -- Comparison of NO emissions estimated using the nonlinear least squares\napproximation in the ETA-CMAQ air quality forecast system with actual SMOKE%/MOBILE6\nmodel results for the 12 UTC forecast of July 19, 2004\n7\nFigure 2 -- Fractions of ground-level aerosol sulfate produced via aqueous-phase oxidation (left)\nand via gas-phase oxidation (right) during January 2001, as calculated using the CMAQ sulfate-\ntracking diagnostic tool.\n11\nFigure 3. -- Ambient concentration of primary carbon originating from biomass combustion (left)\nand food cooking operations (right) during the June 15 - August 31, 1999, time period, as\ncalculated using the CMAQ carbon-apportionment diagnostic tool.\n12\nFigure 4. -- Total mercury wet deposition for 2001 (ug/m²) as simulated by the CMAQ mercury\nmodel\n15\nFigure 5. -- Total mercury dry deposition for 2001 (ug/m²) as simulated by the CMAQ mercury\nmodel\n15\nFigure 6. -- Forecast surface level O3 distributions at 2000 GMT over the 1x domain on July 21,\n2004 (left) and August 12, 2004 (right). Color-coded diamonds indicate observed values.\n19\nFigure 7. -- Ratios of CMAQ model results to ambient measurements of EC, OC, and individual\norganic compounds at eight southeastern United States sites in July 1999. Horizontal lines\nbound the region in which model-observation agreement is within a factor of two. Vertical\ndashed lines distinguish molecular markers specific to different source categories\n23\nFigure 8. -- Pie chart showing the relative contributions of the different temporal components to\nthe total variance of hourly PM2 2.5 concentrations, averaged over all AQS monitors in the analysis\ndomain. a) Observations from TEOM monitors, b) CMAQ predictions, and c) REMSAD\npredictions. ID stands for the intra-day component, DU the diurnal component, SY the synoptic\ncomponent, and BL the baseline (longer-term) component.\n26\nFigure 9. -- Average observed sulfate.\n29\nFigure 10. -- Statistical estimates based on monitoring data.\n29\nFigure 11 -- Average sulfate simulated by CMAQ\n30\nFigure 12. -- Significant differences (CMAQ - statistical estimates).\n30\nviii","Figure 13. Summer 3-month average Formaldehyde concentrations predicted by CMAQ\n40\nFigure 14. -- Winter 3-month average Formaldehyde concentrations predicted by CMAQ.\n40\nFigure 15. -- Comparison of monthly-averaged observed concentrations with model predictions\nfor formaldehyde.\n41\nFigure 16. -- Comparison of monthly-averaged observed concentrations with model predictions\nfor benzene.\n41\nFigure 17. -- Time series of 24-hr averaged formaldehyde concentrations measured at the\nProvidence, Rhode Island site for all monitors falling within the single CMAQ grid cell (135,72)\nand the model predictions at this grid cell.\n42\nFigure 18. -- Time series of 24-hr averaged benzene concentrations measured at the Providence,\nRhode Island site for all monitors falling within the single CMAQ grid cell (135,72) and the\nmodel predictions at this grid cell.\n42\nFigure 19. -- Scheme of the new MM5 canopy parameterization, DA-SM2-U, using the\ndrag-force approach with the soil model SM2-U(3D), compared with the roughness approach. 45\nFigure 20. -- Sensible heat flux (20 UTC, August 2000) DA-SM2-U (left hand side); Standard\nroughness approach (right hand side).\n47\nFigure 21. -- PBL height (m) (20 UTC, August 2000) DA-SM2-U (left hand side); Standard\nroughness approach (right hand side).\n48\nFigure 22. -- Formaldehyde (ppb) simulations for August 30, 2000 at 2100 GMT. Left panel is\nthe 1-km simulation (with UCP), middle panel is native 4-km grid size. Right panel is range of\nvalues of the individual 16 1-km cells in each 4-km grid normalized by the 4 km aggregated cell\n49\nmean\nFigure 23. -- Comparison between vertical profile measurements of wind speed, wind direction,\nand TKE (Turbulent Kinetic Energy) made in the EPA wind tunnel model of lower Manhattan,\nand the corresponding CFD predictions\n52\nFigure 24. -- Predicted wind speed (m/sec). Flow field details for a horizontal slice at the\nsurface\n53\nFigure 25. -- Predicted wind speed (m/sec). Flow field details for a vertical slice\n54\nFigure 26. -- Wind vectors and concentration contours surrounding an emissions source in the\n\"ground zero\" area.\n55\nix","Figure 27. -- Vertical concentration profiles surrounding an emissions source in the \"ground zero\"\narea.\n56\nFigure 28. -- Preliminary predictions of wind speed for a simulation supporting the Department\nof Homeland Security's New York City Urban Dispersion Program.\n57\nFigure 29. - Example of predicted plume dispersion for a simulation supporting the Department\nof Homeland Security's New York City Urban Dispersion Program\n58\nTABLES\nTable 1. -- Toxic air pollutant species modeled explicitly in CMAQ during FY-2004\n38\nTable 2. -- Urban Canopy Parameters (UCP) for Houston, Texas\n46","FISCAL YEAR 2004 SUMMARY REPORT OF THE NOAA\nATMOSPHERIC SCIENCES MODELING DIVISION\nABSTRACT. During Fiscal Year 2004, the Atmospheric Sciences Modeling\nDivision's work on meteorological and air quality modeling, and policy guidance\nwas accomplished in accordance with the memoranda signed by the U.S.\nDepartment of Commerce and the U.S. Environmental Protection Agency (EPA).\nThis ranged from research studies and model applications to the provision of air\nquality forecast, policy advice, and guidance on air quality management. Research\nefforts emphasized the development, evaluation, and application of meteorological\nand air quality models. Among the research studies and results were the release of\nthe September 2004 version of the Community Multiscale Air Quality (CMAQ)\nmodeling system; continued development, evaluation, and improvement of CMAQ\nand its modules; improvement of the SMOKE emission processing system;\nevaluation and improvement of the Eta-CMAQ modeling system for use in air\nquality forecasting; development of model evaluation tools; development of an air\ntoxins version of CMAQ (CMAQ-AT); and the development of techniques for data\nanalysis and interpretation.\n1. INTRODUCTION\nIn Fiscal Year 2004, the Atmospheric Sciences Modeling Division (ASMD) continued its\ncommitment for providing goal-oriented, high-quality research and development, and operational\ntransfer of Division products in support of the missions and strategic goals of NOAA and EPA.\nUsing an interdisciplinary approach emphasizing integration and partnership with EPA and\npublic and private research communities, the Division's primary efforts focused on studying\nprocesses affecting the dispersion of atmospheric pollutants through numerical as well as\nphysical modeling; and developing and evaluating meteorological and air quality models on all\ntemporal and spatial scales. The research products developed by the Division are transferred to\nthe public and private national and international communities. Division research is focused on\nfive program areas: new developments in air quality modeling, climate change and its impact on\nregional air quality; multimedia modeling; data management and analysis; and air quality\nforecasting. The Division is organized to respond effectively to these research directions as more\nfully described in the following sections of the report.","2. PROGRAM REVIEW\n2.1 Atmospheric Model Development\nThis research is aimed at providing state-of-science air quality models and guidance for\nuse in the implementation of National Ambient Air Quality Standards (NAAQS) for ozone and\nfine particulate matter (PM25), and air quality forecasting. The principal effort is to develop and\nimprove the Community Multiscale Air Quality (CMAQ) modeling system, a multiscale and\nmulti-pollutant chemistry-transport model (CTM). Specific research components include:\nmeteorological modeling, land-surface and planetary-boundary layer (PBL) modeling, emissions\nmodeling, gas-phase chemical mechanisms and solvers development, aerosol representations in\ngrid-based air quality models, plume-in-grid treatment for large elevated sources of pollution,\nCMAQ code integration and efficiencies improvement, and air quality forecasting.\nThe objectives of this research program are to continuously improve the mesoscale\n(regional through urban scale) air quality simulation models, including CMAQ, as air quality\nmanagement and NAAQS implementation tools. The CMAQ CTM includes the necessary\ncritical science process modules for handling atmospheric transport, deposition, cloud mixing,\nemissions, gas- and aqueous-phase chemical transformation processes, and aerosol dynamics,\nand atmospheric chemistry. Research is conducted to develop and test appropriate chemical and\nphysical mechanisms, improve the accuracy of emissions and dry deposition algorithms, and to\ndevelop and advance state-of-science meteorological models via improved process\nparameterizations.\nBy design, CMAQ is expected to be used by both scientists and policy makers for various\nassessment activities, research module developments, and detailed model evaluation studies.\nScientists can thus incorporate additional air quality science process modules into the system. A\ngeneralized coordinate approach used in CMAQ allows the CMAQ CTM to be configured\ndynamically consistent with the driver meteorological model. Tested model configurations can\nbe established for use by the policy community to develop and analyze implementation strategies\nfor air quality management. CMAQ utilizes the \"one atmosphere\" approach to air quality\nmodeling. It is capable of concurrently simulating concentrations of oxidants and fine particles,\nvisibility degradation, air toxins, and acidic and nutrient deposition and loadings to ecosystems at\nurban and regional scales. As the understanding of the atmospheric processes, input data, and\nmodel formulations and parameterizations improves, it will be essential to continue to upgrade or\nprovide science options through future releases of CMAQ. Therefore, activities that facilitate the\nmaintenance and science process evolution within CMAQ will be required. The work described\nbelow includes additional model development and testing that led to the September 2004 release\nof the CMAQ modeling system.\n2","2.1.1 Meteorological Modeling for CMAQ Applications\nThe Fifth-Generation Pennsylvania State University (PSU)/National Center for\nAtmospheric Research (NCAR) Mesoscale Model (MM5) is the primary tool for providing the\nmeteorological fields for CMAQ. MM5 is widely used to generate meteorological\ncharacterizations of the atmosphere throughout the air-quality modeling community. For\nCMAQ, MM5 is applied to case studies (episodic, seasonal, and annual) at a variety of spatial\nscales using a series of one-way nested domains. Typically, MM5 is run retrospectively using\nfour-dimensional data assimilation for a dynamic analysis of the simulation period. The output\nrepresents a dynamically-consistent multiscale meteorology simulation for various horizontal\ngrid spacings ranging from continental to urban scales. The MM5 output is ultimately used in\nthe Sparse Matrix Operator Kernel Emission (SMOKE) (emissions) and CMAQ (chemistry)\nmodules to describe the atmospheric state variables and characteristics of the planetary boundary\nlayer.\nDuring FY-2004, MM5 was used to apply CMAQ for an annual simulation of 2001. This\nyear-long MM5 simulation was used to evaluate CMAQv4.4, and it provided meteorological\nfields for several internal development projects. Follow-on modeling with MM5 will extend the\nannual simulations to include 2002, 2003, and 2004. In addition, the proof-of-concept research\nto implement urban canopy parameterizations in MM5 for modeling the effects of urban areas at\nhorizontal grid spacings of ~1 km was developed (Dupont et al. 2004; Otte et al. 2004a).\nAlso during FY-2004, work began to transition toward using the Weather Research and\nForecast (WRF) model for meteorological simulations. WRF is the next-generation\nmeteorological model that is intended to ultimately replace MM5, and it includes many of the\nfeatures that are currently in MM5. It is also attractive for air-quality modeling applications\nbecause, unlike MM5, it contains mass-conserving equations. Collaborative work has begun\ntoward implementing a nudging-based four-dimensional data assimilation capability in WRF, as\nwell as developing a version of the Pleim-Xiu land-surface model (PX LSM) for WRF. It is\nanticipated that the transition to WRF will intensify as those capabilities, which are available in\nMM5 and typically used for air quality modeling simulations, become more fully integrated into\nWRF\n2.1.2 Linking Meteorology and Chemistry Models for Research Applications\nThe Meteorology-Chemistry Interface Processor (MCIP) creates the off-line linkage\nbetween meteorological models and CMAQ for research and regulatory applications. MCIP is\ncompatible with upgrades to the meteorological models that are used by CMAQ to preserve\nnumerical and physical consistency between the meteorology and chemistry models. In\nFY-2004, MCIP was upgraded to version 2.3 and included several new features. MCIP can now\n1 Copyright 1999 MCNC-North Carolina Supercomputing Center\n3","process polar stereographic and Mercator projected meteorological data. Several new fields are\nnow available in the MCIP output, including graupel, 10-m wind speed and direction, full-level\nJacobian, and latitude and longitude coordinates for the CMAQ lateral boundaries, the latter for\ndownscaling global simulations for CMAQ application. MCIP will now only generate certain\nhydrometeor species and fields associated with the PX LSM if those fields are available in the\nMM5 output. Several other features and code adjustments were also made. The details\npertaining to the code changes can be found in Otte (2004).\n2.1.3 Planetary Boundary Layer and Land Surface Modeling\nRealistic simulation of land-surface and planetary-boundary layer (PBL) processes is\nimportant for both meteorology and air quality modeling. Interactions between surface\ncharacterization, surface fluxes, and PBL processes are very tightly coupled. In addition, surface\nfluxes and PBL mixing of chemical constituents closely follow the meteorological processes.\nTherefore, efforts in this area involve both meteorology and chemical transport models to\ndevelop realistic and consistent modeling of the surface and PBL processes.\nParameterizations of the vertical transport due to boundary-layer turbulence are among\nthe most important components of meteorology and air quality models. However, the PBL\nschemes employed in meteorological models and those used in air quality models are often quite\ndifferent. Part of the reason for this is simply different model development histories, but the kind\nof scheme that works well in meteorological models have not worked SO well in air quality\nmodels or vice versa. Clearly, the vertical mixing of trace chemical species should be similar to\nthe vertical mixing of heat and water vapor.\nMesoscale meteorological models typically include either simple non-local closure\nschemes or higher-order schemes that involve prognostic equations for turbulent kinetic energy\n(TKE) and sometimes other higher order terms such as turbulent dissipation or potential\ntemperature variance. The non-local schemes, in particular, have been developed to address the\ninadequacies of local schemes that cannot produce realistic profiles of both first order quantities\nand their fluxes in convective conditions. Air quality models typically use simple local closure\n(eddy diffusivity), although both non-local and higher-order schemes have also been used. A\ndifficulty for air quality models is that chemical profile data for evaluation, including\nozonesondes and aircraft measurements, are very sparse. Therefore, ground-level concentration\ndata are often used for evaluation, which may be affected by many other processes. Without\nextensive comparisons to PBL profiles of chemical measurements, it is difficult to know whether\nan accurate simulation of ground-level concentrations represents realistic PBL mixing or results\nfrom compensating errors.\nDuring the past year, we have been experimenting with various PBL formulations as well\nas continuing development of new PBL schemes. Most notably, a new more advanced version of\nthe Asymmetric Convective Model (ACM) has been developed and is currently being tested and\nevaluated. The new model (ACM2) is a combination of local- and non-local closure.\n4","Specifically, ACM2 is a combination of the original ACM (Pleim and Chang, 1992) and eddy\ndiffusion. The challenge is to match the two schemes at a certain height, in this case the top of\nthe lowest model layer, and apportion the mixing rate between the two schemes SO that the\nresultant flux is identical to that produced by either scheme running alone.\nThe ACM2 has been implemented in MM5 and CMAQ. Evaluation is proceeding along\ntwo tracks. A series of MM5 and CMAQ simulations has been made for the summer of 2004.\nOperational evaluations of these runs have shown preliminary statistics for temperature,\nhumidity, and winds, that are similar to previous MM5 simulations using ACM. Preliminary\nevaluation of CMAQ runs show improved simulation of ozone compared to the 2004 air quality\nforecasts. A more rigorous evaluation of the PBL model will be made involving comparisons of\nPBL heights and profiles of meteorological and chemical parameters to ICARTT (International\nConsortium for Atmospheric Research on Transport and Transformation) 2004 field\nmeasurements. The other evaluation track involves idealized case study simulations for various\nstability and wind conditions. Profiles of temperature, winds, and chemical tracers as well as\ntheir flux profiles are being compared to large eddy simulation results for identical conditions.\nThe goal of this effort is to develop, test, and evaluate a PBL model that accurately simulates the\nvertical mixing of both meteorological and chemical parameters at the ground and throughout the\nPBL.\nThe data-assimilation scheme for the Pleim Xiu Land Surface Model (PX-LSM) was\nanalyzed and evaluated against Southern Oxidant Study (SOS) 1999 field experiment data. This\nscheme involves Newtonian nudging of surface and root-zone soil moisture according to model\nbiases in 2-m air temperature and relative humidity. Sensitivity tests confirm the value of this\nscheme in improving temperature and surface flux simulations (Pleim and Xiu, 2003). The\nPX-LSM has been an available option in the MM5 for several years. Efforts are underway to\nimplement the PX-LSM in the WRF model.\n2.1.4 High-Resolution Sea Surface Temperature Initialization for Meteorological Models\nThe initial version of a sea surface temperature (SST) processing utility for MM5 was\ndeveloped in FY-2003. The main rationale behind this development is that MM5 typically uses\ncoarse (32 km or greater) gridded SST data interpolated from larger scale models. Within coastal\nareas, the sea temperature is one of the most dominating factors that influence the boundary-layer\nmeteorology; thus, it is important to resolve SST to the horizontal grid scale of the model. The\nfirst MM5 simulation that used the high resolution SST information was focused over Tampa\nBay, Florida (April 20-June 7, 2002). Initial evaluation results indicate that the sea surface\ntemperature did improve the timing of the spring Florida sea-breeze, and resulting temperature\nvariations. This project has expanded to consider not only the spring season, but also the\nsummer and fall of 2002. High resolution SST grids have been prepared for the simulation of\nthese other seasons. A similar approach has been applied to a 4-km and 1-km scale simulation\nover the Houston area. An advancement of the technique was developed in FY-2004 for these\nsimulations because the Galveston Bay diurnal water temperature varied by as much as 5-10°C,\n5","and the prior method did not have the temporal data resolution to replicate this diurnal variation.\nThe procedure involved finding the minimum (at night) and maximum (middle of day) daily\ntemperature at each grid point and fitting a sine curve to match the amplitude and phase. Hourly\nSST files were generated using this approach, and ingested into a 10 day simulation (August\n2000). Once these simulations have been completed, a thorough evaluation on the benefit of the\nSST to the simulation accuracy will be documented and published.\n2.1.5 Anthropogenic Emissions\nDuring FY-2004, the Sparse Matrix Operator Kernel Emission (SMOKE) modeling\nsystem (www.cep.unc.edu/empd/products/smoke/index.shtml) was enhanced to version 2.1. The\nDivision initiated development of SMOKE, which has evolved into a key emission processing\ntool for air quality management modeling. SMOKE v2.1 allows humidity data to be extracted\nalong with temperature data for use in modeling mobile source emissions. Both temperature and\nhumidity data may be averaged over time, rather than using hourly data, to increase model\nexecution speed. The mobile source model in SMOKE was upgraded to MOBILE6.2.0.3. The\nbiogenic emission model within SMOKE was upgraded from the Biogenic Emission Inventory\nSystem (BEIS) version 3.09 to version 3.12. BEIS3.12 contains updated biogenic emission\nfactors and allows use of the Biogenic Emissions Land cover Database version 3 (BELD3) 1-km\nresolution gridded land-use data set. SMOKE will also model emissions using the polar\nstereographic projection as well as Lambert conformal projections. In addition, a variety of\nsoftware bugs were fixed. Development of SMOKE® is continuing with the gradual\ndevelopment by the Division of an episodic biomass burning emission algorithm (based on the\nU.S. Forest Service's BlueSky model). The Spatial Allocator tool of the Multimedia Integrated\nModeling System (MIMS) is being upgraded to provided the capability to grid input files needed\nby SMOKE®. It is expected to be completed by fall 2005. The Spatial Allocator requires only\ngrid definitions and (for spatial surrogates) GIS (Geographic Information System) shape files.\nAs part of NOAA's air quality forecasting program, an efficient method of estimating\nmobile source emissions was developed and implemented for real-time forecasting. A nonlinear\nleast-squares approach was used to create a relationship between Eta-modeled temperature and\nemission rates for each species from SMOKE/MOBILE6 for each grid cell and for each hour of\nthe week. This relationship was used to create mobile source emissions in real time using\nforecast temperatures to make forecast emission estimates. As shown in Figure 1, the nonlinear\nleast squares approximation compared to the actual SMOKEo/MOBILE6 estimate is nearly\nidentical for the 12 Universal Time Coordinate (UTC) forecast of July 19, 2004, for NOx\n(nitrogen dioxide and nitric oxide).\n2.1.6 Biogenic Emissions\nIntroduced in 1988, the Biogenic Emissions Inventory System (BEIS) provides hourly,\ngridded estimates of biogenic volatile organic compounds and soil NO emissions to such\n6","regional air quality models as CMAQ. During FY-2004, BEIS version 3.12 (Pierce et al., 2003)\nwas tested further. BEIS3.12 includes emission factors for 34 chemical species, including 14\nspecific monoterpenes, methanol, and methyl-butenol. BEIS3.12 was included as part of\nSMOKE v2.1 release.\nr =\n0.99988\nn =\n420453\nnmb =\n-1.02755\n10\n0\n5\n10\n15\n20\n25\nNONLINEAR LEAST SQUARES APPROXIMATION\nFigure 1. Comparison of NO emissions estimated using the\nnonlinear least squares approximation in the Eta-CMAQ air quality\nforecast system with actual SMOKEo/MOBILE6 model results for the\n12 UTC forecast of July 19, 2004.\n2.1.7 Modeling Smoke Emissions from Fires\nIn the past, emission inventories of forest fires, agricultural burning, and range fires were\nestimated crudely, at the county-level on a monthly basis. The Division has begun to assimilate\nresearch to improve spatial and temporal descriptions of biomass burning. The U.S. Forest\nService has developed a tool known as BlueSky to predict air quality impacts of smoke from\nforest, agricultural, and range fires. The BlueSky modeling framework combines state-of-the-art\nemissions, meteorology, and dispersion models. A prototype to combine BlueSky's emission\n7","processing algorithm with the SMOKE emission processing system was completed during FY-\n2004. The combination of these tools will allow a more accurate characterization of fuel loading,\ntemporal and spatial distribution of fire emissions, and a more accurate representation of plume\nrise and vertical distribution. During FY-2005, the Division will continue to improve on this\nnew approach to distribute emissions from forest fires to specific grid cells on an hourly basis,\nand the Division will begin to evaluate the effectiveness of this approach in improving aerosol\npredictions in CMAQ.\n2.1.8 Implementation and Testing of New and Refined Chemical Mechanisms and\nChemical Solvers in CMAQ\nAtmospheric gas-phase chemistry is a critical component of the CMAQ modeling system.\nThe ability of CMAQ to accurately predict ambient concentrations of trace gases in the\natmosphere is dependent upon the validity of the gas-phase chemical interactions and\ntransformations contained in the chemical mechanism that is used in CMAQ. Accurate\nrepresentation of gas-phase chemistry is also vital for the simulation of such other important\natmospheric processes as the formation of aerosols, the chemical transformations taking place in\nthe liquid phase, and the deposition of air contaminants to land and water surfaces.\nCommensurate with the need for an accurate chemistry representation is the need for gas-phase\nchemistry solution techniques that are both highly accurate and computationally efficient. Since\nnumerical solution techniques that have been used historically consume about 50 to 75 percent of\nthe computer time required for model simulations, any substantial computational efficiencies that\ncan be gained will significantly lower the computational time of CMAQ. Therefore, the\nunderlying objectives of this research effort are twofold: (1) to improve and enhance the\nrepresentation of atmospheric gas-phase chemistry in CMAQ by refining existing chemical\nmechanisms, by adding new chemical mechanisms, and by investigating new approaches for\nincreasing chemical information in the model, and (2) to reduce computer time required to\nsimulate gas-phase chemistry by enhancing the computational efficiency of existing solvers, by\ninvestigating new approaches that can be used in conjunction with existing solvers to lower\ncomputational requirements without sacrificing numerical accuracy, and by testing and\nevaluating new chemistry solver algorithms. The results of this work will help improve the\nscientific integrity of CMAQ by incorporating new scientific knowledge in atmospheric\nchemistry, while ensuring the practicality of using CMAQ as a modeling tool in\nregulatory/operational modeling applications by lowering the computational burden.\nDuring FY-2004, two new gas-phase chemistry solvers were added to the array of solvers\navailable in the CMAQ modeling system. One new solver, ROS3, is a particular version of the\nRosenbrock class of solvers (Sandu et al., 1997). It is a general solver and, hence, does not\nrequire any modifications when an existing chemical mechanism is changed or when a new\nmechanism is introduced. The ROS3 solver can achieve high accuracy at a relatively low\ncomputational cost and is the fastest of the existing CMAQ generalized solvers. The other new\nsolver addition to CMAQ is an Euler Backward Iterative (EBI) solver (Hertel et al., 1993)\ndeveloped specifically for the SAPRC99 chemical mechanism (an EBI solver for the CMAQ\n8","Carbon-Bond-IV (CB-IV) mechanism was developed in the previous year). Unlike ROS3, EBI\nsolvers are not generalized and must be adapted to individual mechanisms. Although they are\nsomewhat less accurate than several of the generalized solvers, they are the fastest of all the\nsolvers available in CMAQ. Both EBI and ROS3 are also very robust and relatively easy to use,\nmaking them particularly well-suited for solving the set of stiff ordinary differential equations\nthat arise when simulating the chemical transformations of trace gases in the atmosphere. Both\nsolvers were included in the latest public release of CMAQ, and are, thus, available for use by the\nCMAQ modeling community.\nDuring FY-2004, the existing gas-phase CB-IV mechanism in the CMAQ model was\nextended to include atmospheric chlorine chemistry developed at the University of Texas at\nAustin (Chang at el, 2002; Tanaka et al, 2003). The resulting gas-phase mechanism were\nincorporated into the CMAQ model and tested with the following three gas-phase chemistry\nsolvers: Euler Backward Iterative solver, Rosenbrock solver, and Sparse-Matrix Vectorized Gear\nsolver. The combined mechanism was used to evaluate the effect of chlorine emissions on\natmospheric ozone in the continental United States. The study included anthropogenic molecular\nchlorine emissions, natural molecular chlorine released from sea-salt aerosol, and anthropogenic\nhypochlorous acid emissions from cooling towers and swimming pools. When molecular\nchlorine emissions were included in the model, the only significant impact occurred near the\nGreat Salt Lake in Utah. Morning and evening ozone levels at that location increased by a\nmaximum of 14 parts per billion by volume (ppbv) and 4 ppbv, respectively; but the impact on\nthe daily peak ozone mixing ratios was not substantial. When both molecular chlorine and\nhypochlorous acid emissions were included in the model, the impacts were evident at several\nareas, including the Great Salt Lake and the Houston area. Ozone mixing ratios in the Great Salt\nLake did not change compared to the case with molecular chlorine emissions. The morning and\nthe daily peak ozone mixing ratios in the Houston area increased by a maximum of 10 and 7\nppbv, respectively, for the days simulated. In contrast to hydroxyl radicals, chlorine radical\nmixing ratios peaked in the morning and reached up to 15 and 4 percent of the corresponding\nhydroxyl radical mixing ratios at the Great Salt Lake and the Houston area, respectively.\nChlorine emissions appeared to increase the hydroxyl radical mixing ratios compared to the case\nwithout any chlorine emissions. The increases in ozone mixing ratios were accompanied by\ndecreases in volatile organic compounds mixing ratios.\n2.1.9 Aerosol Mechanism Improvements in CMAQ\nThe CMAQ aerosol module was revised to improve computational efficiency, numerical\nstability, and the in-line code documentation. Testing of the 2003 CMAQ release revealed that\n47 percent of the aerosol module computational time is spent on calculating coagulation\ncoefficients and 13 percent on partitioning secondary organic aerosol (SOA) material between\nthe gas and aerosol phases. A new subroutine was developed to calculate coagulation\ncoefficients in a more efficient manner than the approach used previously. In the new subroutine,\ncoagulation coefficients for particle number and aerosol third moment are calculated from\nanalytical expressions reported by Binkowski and Shankar (1995), and the second moment\n9","coagulation coefficients are calculated using a combination of analytical expressions and\ncorrection factors similar to the approach outlined by Whitby et al. (1991). The coagulation\ncoefficients obtained using the new subroutine are within one percent of the results from the\nprevious model version, but the computational efficiency of the new routine is 60 times faster.\nThe SOA partitioning calculation involves the iterative solution of a system of quadratic\nequations. In the revised aerosol module, the initial guess for this iterative solution was revised\nto incorporate information on SOA concentrations from the previous time step. This reduces the\nnumber of iterations required for convergence, and hence, the computational time, by\napproximately 60 percent. The net effect of both improvements is a factor of two increase in the\ncomputational efficiency of the CMAQ aerosol module.\nDuring developmental testing of previous CMAQ versions, small numerical perturbations\ncaused by the use of different Fortran compilers were found to produce large (~10 ug m ³\ntransient impacts on aerosol-phase nitrate concentrations over the arid southwestern United\nStates. These effects were attributed to numerical instabilities in the ISORROPIA\nthermodynamics module. In collaboration with the developer of ISORROPIA at Georgia\nInstitute of Technology, two problems were identified and corrected in the thermodynamics\nmodule. As a result, the large numerical instabilities observed in previous CMAQ versions are\nnow removed. Recently, a second family of numerical instabilities was identified in the\nISORROPIA module, which results in small (~1 ug m²3 transient impacts on aerosol-phase\nnitrate concentrations in low-humidity environments. A solution to this numerical error is under\ninvestigation, with an anticipated release in FY-2005. A significant effort was made to\nstrengthen the in-line documentation in the CMAQ aerosol code. Every major scientific formula\nin the aerosol chemistry and dynamics routines was annotated with comments that permit users\nto trace the formula back to an equation in a journal article or published report where it is\ndescribed thoroughly. These in-line documentation enhancements should aid the CMAQ users\nseeking to modify scientific algorithms within the model. The above revisions are described in\nBhave et al. (2004b) and are incorporated into the 2004 CMAQ public release.\nIn addition to the aerosol module revisions for the 2004 public release, a set of diagnostic\ntools are under development that will enable users to probe the sources of emissions for the\nmodeled aerosol concentrations. With one tool, the ambient sulfate concentrations formed via\ndifferent pathways can be determined quantitatively. For example, Figure 2 illustrates model\ncalculations of the fractions of ground-level aerosol sulfate produced by aqueous-phase oxidation\nand gas-phase oxidation, when averaged over January 2001. With a second diagnostic tool, one\nmay calculate the contributions from individual emission source categories and/or geographic\nregions to the ambient primary carbonaceous aerosol burden. An example of the results obtained\nusing this tool are shown in Figure 3, where the modeled concentrations of primary carbon\noriginating from biomass combustion and food cooking are calculated and averaged over the\nJune 15-August 31, 1999, time period. Both of these diagnostic tools are anticipated for the\nCMAQ release in FY-2005.\n10","Aqueous Production\nGas-Phase Production\n1.0\n0.9\n0.8\n0.7\n0.6\n0.5\n0.4\n0.3\n0.2\n0.1\n0.0\nFigure 2. Fractions of ground-level aerosol sulfate produced via aqueous-phase oxidation\n(left) and via gas-phase oxidation (right) during January 2001, as calculated using the CMAQ\nsulfate-tracking diagnostic tool.\n2.1.10 Plume-in-Grid Modeling\nThe plume-in-grid (PinG) technique provides a subgrid scale treatment of the dynamic\nand chemical processes governing gas-phase and aerosol species concentrations in isolated,\nmajor point-source plumes within the CMAQ Eulerian grid modeling system. The CMAQ/PinG\napplies a Lagrangian approach and simulates plume growth in a gradual, real-world manner due\nto turbulence and wind shear processes. This treatment is in contrast to the traditional Eulerian\ngrid modeling method which is instantaneous dilution of point-source emissions into an entire\ngrid cell volume. The overdilution effect becomes more pronounced with increasing horizontal\ngrid sizes generally specified in regional- or continental-scale modeling domains. The key\nalgorithms are a plume dynamics model (PDM) processor and a Lagrangian reactive plume\nmodel (PinG module), which are designed to simulate the relevant plume processes at the proper\nspatial and temporal scales for CMAQ model domains with grid cell sizes greater than 10 km.\nThe PinG treatment is able to simulate plumes from multiple point sources. A continuous plume\nis represented by a series of plume sections, each of which has been released at a 1-hour interval.\nThe horizontal dimension of each plume cross-section is internally resolved by a set of attached\nplume cells. The PinG module is fully integrated into the CMAQ grid model. It is exercised\nconcurrently during a CMAQ chemical transport model (CTM) simulation and takes advantage\nof grid cell concentrations as boundary conditions for each plume section edge. An important\nfeedback occurs when a plume section reaches the model grid cell size. At that time, the subgrid\nplume treatment ceases for the particular plume section and plume concentrations are\n11","incorporated into the Eulerian grid framework. A full description of the capabilities of the\nCMAQ/PinG modeling treatment and its technical formulation are described in Gillani and\nGodowitch (1999).\nBiomass Combustion\nFood Cooking\n0.25\n0.20\n0.15\n0.10\n0.05\n0.00\nug/m3\nFigure 3. Ambient concentration of primary carbon originating from biomass combustion\n(left) and food cooking operations (right) during the June 15-August 31, 1999, time period, as\ncalculated using the CMAQ carbon-apportionment diagnostic tool.\nThe updated aerosol algorithm version 3 (AE3) employed in the CMAQ /CTM was also\nincorporated and successfully tested in the PinG module. The 2004 public release of the CMAQ\nmodeling system contained the PinG module with the capability to perform aerosol formation in\nsubgrid plumes. Therefore, PinG currently simulates aerosol species and PM25 along with\ngas-\nphase photochemistry in the subgrid plumes. CMAQ/PinG test simulations were successfully\ncompleted on various computational platforms with a single processor and multi-processors in a\nparallel processing mode. Preliminary results of modeling aerosols in PinG revealed differences\nin aerosol sulfate (SO4) concentrations in the vicinity of high NOx and SO2 point sources. For\npoint sources with comparable SO2 emissions, greater sulfate formation occurred in those plumes\nexhibiting a lower NOX emission rate. These PinG results appeared to be supported by recent\nexperimental aerosol data obtained in plumes. Additionally, CMAQ model simulations were\nconducted with the PinG approach and without it using existing emissions and meteorology data\nsets from a summer 2001 period for a continental domain with a 36-km grid cell size. There\nwere 47 high emission NOX and SO2 point sources in the modeling domain simulated with the\nPinG approach. Comparisons of modeled gaseous and aerosol species against surface monitoring\n12","network data were underway. So far, analyses of peak and hourly ozone concentrations for the\ndays modeled reveal that the CTM/PinG results displayed better agreement and less bias than the\nCTM/NoPinG results, particularly in areas where numerous large point sources exist. Additional\nmodel runs are planned for a winter period and another summer experimental period in order for\na comparison of model results against observed plume data collected by various airborne\nplatforms. Further sensitivity test runs will be performed to assess computational times and to\ninvestigate the impact on oxidant and aerosol species concentrations with different chemical\nmechanisms (e.g., CB-IV, SAPRC) and different chemistry solvers (e.g., Gear or Rosenbrock\nsolver) available in the CMAQ modeling system.\n2.1.11 CMAQ Mercury Model Refinements and Evaluation\nDuring FY-2004, the third and final phase of a European intercomparison study of\nnumerical models for long-range atmospheric transport of mercury was completed. This study\nwas organized by the Cooperative Programme for Monitoring and Evaluation of the Long-Range\nTransmission of Air Pollutants in Europe, previously known as the European Monitoring and\nEvaluation Programme (EMEP). Operational organization of the study was provided by EMEP's\nMeteorological Synthesizing Center - East in Moscow, Russia. The first two phases of the\nintercomparison involved rather simple simulations of mercury chemistry in a hypothetical cloud\nvolume and full-scale model simulations of the emission, transport, transformation, and\ndeposition of mercury over Europe during two short episodes in 1995 and 1999. Comparisons\nwere made of the full-scale modeling results to field measurements of elemental mercury gas,\nreactive gaseous mercury, and particulate mercury in air. The third and final phase of the\nintercomparison study involved full-scale simulations over Europe spanning longer time periods\nfollowed by comparisons to observations of both air concentrations and wet deposition fluxes of\nmercury. While some models simulated the entire calendar year of 1999, the CMAQ mercury\nmodel could only simulate the months of February and August as these were the only time\nperiods for which MM5-derived meteorological inputs were available for the European test\ndomain because of limited resources available for this European study.\nDuring FY-2004, the CMAQ mercury model was updated to simulate atmospheric\nmercury based on CMAQv4.3 with additional corrections and enhancements conforming to\nversion 06FEB04 as defined by the EPA Office of Air Quality Planning and Standards (OAQPS)\nfor regulatory applications. In cooperation with OAQPS, the CMAQ mercury model was used to\nperform a number of simulations of the entire calendar year of 2001 over most of North America\nto test a number of mercury chemistry sub-model configurations. Figures 4 and 5 display some\nexample results of these simulations. All of the full-year simulations used initial\ncondition/boundary condition (IC/BC) inputs derived from a global-scale simulation of the\nGEOS-CHEM model developed and applied by the Harvard University Department of Earth and\nPlanetary Sciences (Bey et al., 2001). Further CMAQ mercury model testing will continue into\nFY-2005.\n13","During FY-2004, organization and planning for a mercury model intercomparison study\nfor North America continued. As described above, atmospheric mercury model developers have\nbegun to use model inter-comparison studies to investigate differences regarding their modeling\nof various atmospheric processes, to gauge the level of modeling uncertainty with respect to\nspecific parameters and variables, and to collect evidence of the most important knowledge gaps\nleading to these uncertainties. Wet deposition of mercury is measured on a regular basis at many\nlocations in North America through the Mercury Deposition Network (MDN). While the MDN\nmay not provide sufficient observational closure for comprehensive model evaluation, it can\nprovide information leading to a better understanding of why models differ in their simulations\nof wet deposition and various related atmospheric processes. Modeling assessments conducted\nby the EPA and by the Electric Power Research Institute regarding the importance of mercury\nemissions outside of the United States to the deposition of mercury within its boundaries have\nled to very different conclusions. This model inter-comparison study over North America would\nhelp explain these conflicting results.\n2.1.12 CMAQ Code Integration and 2004 Release\nSeveral computational efficiency improvements were incorporated into the 2004 release\nof CMAQ (version 4.4). In collaboration with the Department of Energy's Sandia National\nLaboratory, a known bottleneck in the parallel I/O implementation was removed, which provided\na significant reduction in model runtimes on Linux computer clusters. Efficiencies were also\nincorporated into the aerosol module of CMAQ, including a new method for calculating\ncoagulation coefficients and updates to the gas-particle equilibrium solver for SOA. Also, a fast\nchemical solver, known as the Euler Backward Iterative (EBI) scheme, was developed for the\nSAPRC99 mechanism.\nIn addition to efficiency improvements, other updates in CMAQv4.4 include: (1) aerosol\ntreatment in the plume-in-grid module (PinG); (2) a new generalized chemical solver\n(Rosenbrock); (3) in-line documentation; (4) code structure simplifications; (5) removal of\noutdated modules; and (6) several bug fixes. Note that another component of the CMAQ system,\nMCIP (version 2.3), was also revised and released in conjunction with the FY-2004 CMAQ\nrelease. CMAQv4.4 is publicly available from the internet site of the Community Modeling and\nAnalysis System center at www.cmascenter.org\n14","30.000112\n25.000\n20.000\n15.000\n10.000\n5.000\n0.000\n1\n148\n1\nFigure 4. Total mercury wet deposition for 2001 (ug/m²) as simulated by the\nCMAQ mercury model.\n30.000112\n25.000\n20.000\n15.000\n10.000\n5.000\n0.000\n1\n148\n1\nFigure 5. Total mercury dry deposition for 2001 (ug/m²) as simulated by the\nCMAQ mercury model.\n15","2.1.13 Development and Testing of an Air Quality Forecast Model\nIn FY-2003, NOAA and EPA signed a Memorandum of Agreement to collaborate on the\ndesign and implementation of a capability to produce daily air quality modeling forecast\ninformation for the United States. The Eta meteorological model and CMAQ were linked\ntogether to form the core of this forecast system. Testing of the system was conducted during the\nFY-2003 and FY-2004 ozone seasons for the northeastern United States, and the system became\nfully operational in September 2004. Over the next five years the model domain will be\nexpanded to the continental United States, and PM25 will be added to the model forecast\ncapability.\nDuring the summer of 2004, the air quality forecast (AQF) system was exercised in three\nstreams: (1) experimental production of O3 forecasts over the northeast United States. (1x\ndomain) for dissemination to the general public; (2) developmental forecasts of O3 over an\nexpanded eastern U.S. domain (3x domain) for dissemination to a focus group of forecasters; and\n(3) developmental forecasts of both O3 and particulate matter (PM) concentrations over the 3x\ndomain for initial assessments of PM forecast capabilities. In the first two applications, aerosols\nwere not simulated by CMAQ. In all applications, the CB-IV mechanism was used, the\nhorizontal grid-cell size was 12 km, while the vertical extent from the surface to 100 mb was\ndiscretized using 22 layers of variable thickness. The emission inventories used by the AQF\nsystem were updated to represent the 2004 forecast period. NO emissions from point sources\nwere projected to 2004 (relative to a 2001 base inventory) using estimates derived from the\nannual energy outlook by the Department of Energy. Area source emissions were based on the\n2001 National Emissions Inventory, version 3, while BEIS3.12 was used to estimate the biogenic\nemissions. Mobile emissions were estimated using a computationally-efficient, least-square\nregression-based approximation to the MOBILE6 model.\nThe turbulent mixing scheme in CMAQ was enhanced to allow the minimum value of the\nsurface layer vertical-eddy diffusivity (Kz) to vary spatially depending on the fraction of urban\narea in each grid cell. The approach allows for Kz in rural regions to fall off to a lower value\nthan predominantly urban regions. This allows increased nighttime O3 titration in rural areas,\nreducing modeled O3 overpredictions. The approach also allows simulated nighttime precursor\nconcentrations in urban areas from becoming too large. Two approaches for specifying lateral\nboundary conditions (LBCs) to CMAQ were investigated based on typical \"clean\" tropospheric\nbackground values; and based on seasonal averages derived from prior CMAQ simulations over\nthe continental United States for the summer of 2001. The default clean profiles were used for\nthe 1x domain, while the seasonal LBC profiles were used in the developmental 3x forecast\napplications. To improve the representation of O3 in the free troposphere, additional\nmodifications of the O3 LBCs using forecast results from the National Centers for Environmental\nPrediction's (NCEP) Global Forecast System (GFS) were explored. In the GFS, O3 is initialized\nusing the Solar Backscatter Ultra Violet (SBUV-2) satellite observations. The evolution of the\n3-dimensional O3 fields in the GFS is then simulated by its transport schemes and a zonally\naveraged production and depletion scheme.\n16","2.1.14 Linking the Eta Model with CMAQ for Air Quality Forecasting\nA key ingredient in the linkage between the NCEP's Eta model and CMAQ is a new pre-\nprocessor for CMAQ (PREMAQ) is largely equivalent to MCIP and parts of SMOKE in the\ncommunity version of the CMAQ modeling system. PREMAQ places the post-processed Eta\nmodel output into the required horizontal and vertical grids for CMAQ. Like MCIP, PREMAQ\ncomputes state variables and other derived variables (e.g., air density, Jacobian, dry deposition\nvelocities for chemical species) that are required by CMAQ. Unlike MCIP, PREMAQ also\nincludes calculations of the meteorology-dependent emissions (i.e., biogenic and mobile sources)\nadapted from SMOKE®. The output from PREMAQ includes the full set of meteorology and\nemissions files that are used by CMAQ. A description of PREMAQ can be found in Otte et al.\n(2004b).\nIn FY-2004, PREMAQ underwent several key modifications for the 2004 ozone\nforecasting season. PREMAQ was modified to process the vertical eddy diffusivity field directly\nfrom the Eta model for experimental simulations to improve the coupling between the Eta model\nand CMAQ. Ozone analyses from NCEP's GFS were also added to PREMAQ and used at\nheights greater than 6 km to provide time-varying chemical boundary conditions. PREMAQ was\nmodified to process hourly precipitation increments allowing for the Eta model's accumulation\n\"bucket\" to periodically empty, as is typically done during the operational forecasts, which\ncorrected the wet deposition calculations in CMAQ and the precipitation-based soil moisture\ncalculations in the biogenic emissions. PREMAQ was further modified to define its input fields,\nwhich are post-processed from the Eta model on scalar points rather than dot points, which\neliminates an interpolation step and ensures a more precise alignment of the meteorological\nfields. PREMAQ now processes percentage of urban land use, which is obtained from the\nemissions system and used by CMAQ to improve the nighttime vertical mixing in urban areas.\nFinally, PREMAQ outputs the 10-m wind speed and direction from the Eta model for assistance\nin model verification activities.\nThe linkage of the Eta model and CMAQ in the national air quality forecasting system\nwas declared operational in September 2004. A manuscript describes the process and software\ncomponents that were used to link the Eta Model and CMAQ, discusses several technical and\nlogistical issues that were considered, and provides examples of ozone forecasts from the air\nquality forecasting system and relates them to the forecast meteorological fields (Otte et al.,\naccepted for publication).\n2.1.15 Evaluation of Eta-CMAQ Forecast Predictions for Summer 2004\nNOAA, in partnership with the EPA, has been developing an operational, nationwide air\nquality forecasting (AQF) system. An experimental phase of the AQF program, which couples\nthe Eta meteorological model with CMAQ, began operations in June of 2004 and is providing\nforecasts of ozone (O3) concentrations over the northeastern United States. These forecasts were\nfirst made public via NOAA's website in September 2004.\n17","An important component of the AQF system was the development and implementation of\nan evaluation protocol. Accordingly, a suite of statistical metrics that facilitates evaluation of\nboth discrete-type forecasts and categorical-type forecasts of O3 concentrations was developed\nand applied to the system to characterize its performance. The results reveal that the AQF system\nperformed reasonably well in this inaugural season (mean domain wide correlation coefficient =\n0.59), despite anomalously cool and wet meteorological conditions that were not conducive to\nthe formation of O3 (mean, domain-wide peak 8-hour concentrations = 44.6 ppb). Due in part to\nthese anomalous conditions, the AQF system overpredicted concentrations (mean, domain-wide\nforecast peak 8-hour O3 = 54.8 ppb) resulting in a mean bias of +10.2 ppb (normalized mean bias\n= +22.8 percent). In terms of error, the domain-wide root mean square error averaged 15.7 ppb\n(normalized mean error = 28.1 percent) for the season. The systematic overprediction was also\nevident from a categorical perspective, as most of these metrics (i.e. false alarm rates) indicated\nan excessive number of exceedances during the four-month period.\nA closer examination of the AQF system's performance over time revealed a systematic\npattern of varied accuracy that was attributed to the synoptic-scale meteorology impacting the\ndomain. Figure 6 illustrates regional distributions of model forecasts for surface O3 on two\ntypical days characterized by high and low O3 levels, respectively. The model performed very\nwell during periods when anticyclones, characterized by clear skies, dominated the domain (e.g.,\nleft panel of Figure 6). Conversely, periods characterized by extensive cloud associated with\nfronts and/or cyclones resulted in poor model performance (e.g., right panel of Figure 6).\nSubsequent analysis revealed two main factors contributing to this overprediction. The first\ninvolved the excessive downward transport of O3 rich air via CMAQ's convective cloud scheme\nin conjunction with O3 profiles derived from the GFS model. The second factor involved too\nlittle attenuation of actinic flux by CMAQ's simulated cloud cover, resulting in too much\nphotolysis and subsequently too much production of O3. In combination, these factors resulted in\nthe AQF system's systematic overprediction of O3 in and around areas of cloud cover. Changes\nto CMAQ's cloud schemes are underway that are expected to significantly improve the AQF\nsystem's performance.\n2.1.16 Assessment of the Real-Time Application of CMAQ\nIn collaboration with NOAA and EPA, the New York State Department of\nEnvironmental Conservation (NYSDEC) has begun to perform near real-time simulations of O3\nand PMs over the northeastern United States. This pilot project builds upon the operational air\nquality forecasting program of National Weather Service (NWS), NOAA, and EPA by utilizing\nthe same Eta/PREMAQ/CMAQ modeling system and the same meteorological and emission\ninputs.\nThe pilot study is aimed at assessing the feasibility of applying the CMAQ modeling\nsystem in a near real-time mode for prototyping continuous PM2. modeling. The pollutants of\ninterest are ozone, PM2.5 total mass, and PM2. species composition in New York State with a\nparticular emphasis on the greater New York City Metropolitan Area. The study examines both\n18","operational aspects such as data transfers, computing power, and data storage as well as scientific\nquestions such as determining CMAQ performance for the prediction of speciated PM2 2.5 over\nNew York State and assessing the merits of CMAQ based forecasts compared to traditional\nforecasting approaches. The collaboration with EPA and NOAA includes NYSDEC's\nparticipation in the beta-testing of future CMAQ releases. Throughout the duration of the\nproject, periods of model simulations alternate with phases of model evaluation, analysis, and\npossible refinements to the modeling setup. Finally, NYSDEC will apply and evaluate tools to\ncombine CMAQ model outputs with observations in an attempt to better characterize ambient air\nquality over New York. Such combined fields may be of value when studying linkages between\nair pollution and public health.\n120\n105\n90\n75\n60\n45\n30\n15\n0\nppb\nFigure 6. Forecast surface level O3 distributions at 2000 GMT over the 1x domain on July 21,\n2004 (left) and August 12, 2004 (right). Color-coded diamonds indicate observed values.\n2.2 Atmospheric Model Evaluation and Application Activities\n2.2.1 Meteorological Model Evaluation\nAir quality modeling simulations are strongly dependent on the meteorology. Thus, the\nperformance of meteorological simulations must be assessed in conjunction with the\nchemical-transport predictions. Traditional verification methods do not take advantage of the\nincreasing amounts of non-standard meteorological observations (e.g., wind profilers, aircraft,\nsatellite winds, gridded radar derived precipitation, etc.). During FY-2003, the initial\ndevelopment on a meteorological evaluation system was performed. In FY-2004, the tool was\n19","used for several evaluations that served as a test bed to improve the performance/speed, find and\neliminate bugs in the system, develop new analysis options, and an interactive web-based\ninterface. The evaluation system was also packaged for easier portability to other platforms.\nSeveral major evaluation efforts were conducted using the meteorological evaluation tool.\nThe main evaluation was an annual (2001) MM5 simulation over the continental United States at\na grid size of 36 km, and a similar simulation over the eastern United States with 12-km grid\ncells. Surface meteorological variables (2 m temperature, 10 m wind speed and direction, and\nmixing ratio), tropospheric wind profiles, and precipitation were evaluated, and results presented\nin various formats. The tool was also used in several other experiments, including an 8-km and\n2-km meteorological simulation (MM5) for the Bay and Regional Atmospheric Chemistry\nExperiment (BRACE), a 12-km WRF and 12-km NCEP Eta model comparison during the\nsummer of 2004, and a 12-km MM5 simulation during the summer of 2004 to aid in the\ndevelopment of a more accurate boundary-layer parameterization.\nPromising experiments were conducted to directly link errors in the meteorology with\nthose of the air quality model within the framework of the evaluation tool. In one particular\nexperiment, observed and simulated aerosol-nitrate data was loaded into the model evaluation\ndatabase. Observation-model (both meteorology and air quality data) pairs were extracted from\nthe database for situation where the relative humidity was overpredicted and for times when the\nrelative humidity was underpredicted. Model performance metrics were calculated for these two\ndata sets. Statistics indicated that the aerosol nitrate was overpredicted by an average of 50\npercent when the relative humidity is overpredicted, and there was little to no over or\nunderprediction when the relative humidity was either accurately simulated or underestimated.\nThis approach is actively being applied to other simulations and chemical species. The\nmeteorological evaluation tool is being improved to suit air quality applications.\n2.2.2 Diagnostic Metrics for Ozone and Inorganic Particulate Matter\nDiagnostic metrics enable the examination of the process side of a model to better study\nthe degree of reliability of control strategy predictions. Diagnostic metrics require a special set of\nnon-routine measurements, because they typically involve ratios of species involved in\nphotochemical production or aerosol equilibrium processes. Earlier work (Tonnesen and Dennis,\n2000a; 2000b) had identified a set of metrics designed to assess the photochemical state of the\natmosphere relative to ozone production and to the expected magnitude and direction of a change\nin O3 concentrations due to a change in hydrocarbon (VOC) or nitrogen oxide (NOX) emissions.\nThe metrics are based on measurement of O3, NO+true NO2 = NO, NO, and NO-NO = NO.\nDiagnostic tests using these metrics were applied to CMAQ for Nashville, Tennessee, for the\n1995 SOS field measurements (Arnold et al., 2003). The need for true-NO and NO\nmeasurements was passed to EPA, spawning instrument development and additional guidance\nwithin its monitoring strategy.\n20","Examination of the metrics is underway for use in assessing the physical and chemical\nstate of inorganic fine particles in the atmosphere. The inorganic fine particle system is a priority\nbecause the inorganic fine particles represent a majority of the total fine mass in the eastern\nUnited States and the inorganic system has some important nonlinearities. The gas ratio (GR)\ndefined by Ansari and Pandis (1998) was identified as the leading inorganic aerosol metric to\nsupport diagnostic testing. The GR is equal to the free ammonia divided by total-nitrate, where\nfree ammonia is defined as the moles of gaseous ammonia plus aerosol-phase ammonium minus\ntwice the moles of aerosol-sulfate, and total-nitrate equals gas- + aerosol-nitrate.\nThe GR is being assessed for its ability to provide insight into the nonlinear responses of\nthe inorganic fine aerosols in CMAQ predictions. The first nonlinear response being studied is\nthe degree to which aerosol nitrate will replace sulfate as SO2 emissions are reduced. The\nconceptual model represented in the GR and articulated through thermodynamic calculations\n(Ansari and Pandis, 1998) indicates that the degree of nonlinearity in the PM response to sulfate\nreductions will depend on the value of the GR. For GR values much greater than 1 (i.e.,\nammonia-rich regime), the decrease in PM is expected to be proportional to the sulfate reduction\nbecause two moles of ammonium are removed along with each mole of sulfate. At GR values\nmuch less than 1 (i.e., nitrate-rich regime), the increase in PM is expected to be proportional to\nthe sulfate reduction because two moles of nitrate will replace each mole of sulfate that is\nremoved. At GR values close to 1, the PM-mass response to a sulfate reduction is expected to be\nnonlinear.\nPreliminary results indicate that in a full 3-D model like CMAQ, the nonlinear transitions\nare less well-defined than in the thermodynamic equilibrium box models. Nonetheless, it\nappears the GR provides valuable insights into the nonlinear responses in CMAQ. Examination\nof the CMAQ results for winter 2002 suggest that almost the entire eastern United States will be\naffected by these nonlinearities in the inorganic aerosol system. Several inorganic aerosol\nsensitivity studies are underway.\n2.2.3 Diagnostic Evaluation for Carbonaceous Aerosol Components\nA substantial fraction of fine particulate matter across the United States is composed of\ncarbon (Malm et al., 2004). At routine monitoring sites, carbonaceous aerosol is segregated into\norganic carbon (OC) and elemental carbon (EC) based on its thermal and optical properties. The\nOC fraction may be subdivided further into primary organic carbon (OCpri), which is emitted\ndirectly to the atmosphere in particulate form, and secondary organic carbon (OCsec), which is\nformed in the atmosphere through oxidation of reactive organic gases and subsequent\ngas-to-particle conversion processes. It is important to determine the relative contributions of\nOCpri and OCsec to the ambient aerosol burden, SO that policymakers may decide which portion\nof the organic aerosol complex to target in their control strategy selection process. In FY-2004,\nYu et al. (2004) devised a method to estimate the ambient concentrations of OCpri and OCsec\nusing routine OC and EC measurements in combination with model calculations of the primary\nOC to EC concentration ratio. This methodology was applied to data spanning the June 15-\n21","August 31, 1999, time period. Results indicate that during summer months, the fractional\ncontribution of OCsec to total OC ranges from 48 percent in the western United States to 77\npercent in the Northeast. In FY-2005, the methodology will be extended to a full year of data to\nassess the seasonal cycle of OCpri and OCsec across the continental United States.\nIn the southeastern United States, carbonaceous aerosol is the largest component of fine\nparticulate mass (Hansen et al., 2003) and a significant portion of it is OCpri (Zheng et al.,\n2002). OCpri is emitted from numerous sources, including motor vehicle exhaust, residential\nwood combustion, coal combustion, forest fires, agricultural burning, solid waste incineration,\nfood cooking operations, and road dust.\nA diagnostic tool was developed within the CMAQ modeling framework that allows\nusers to calculate the contributions from individual emission source categories to the ambient\nprimary carbonaceous aerosol burden. This tool was exercised in a CMAQ simulation of the\n1999 summer season. The model results were converted into concentrations of individual\norganic compounds and evaluated against organic-tracer measurements collected in July 1999 at\neight monitoring sites across the southeastern United States, as shown in Figure 7.\nModeled-to-observed concentration ratios are displayed along the vertical axis.\nSeventeen organic tracers along with bulk EC and OC are arranged in groups along the\nhorizontal axis, separated by vertical dashed lines that delineate conserved tracers emitted from\ndifferent source categories. Model results for total OC are in reasonable agreement with\nobservations at the Atlanta, Georgia, site, but are low by a factor of three at the remaining sites.\nThis indicates that total OC is underestimated across the southeastern United States, but it is\nimpossible to determine which source contributions have been underestimated based on bulk OC\nand EC measurements alone. By evaluating model results against the individual organic tracer\nmeasurements, it is found that modeled OCpri contributions from motor vehicle exhaust and\nnatural gas combustion are reasonably accurate whereas the contributions from biomass\ncombustion and food cooking are underestimated by a factor of 4 or more. Complete details of\nthis evaluation are reported by Bhave et al. (2004a).\n2.2.4 CMAQ Model Evaluation to Assess Model Readiness for Application\nAn operational evaluation of the 2004 release of the CMAQv4.4 was performed that\ncompares an annual simulation (2001) covering the contiguous United States against monitored\ndata from four nationwide networks. This effort represents one of the most spatially and\ntemporally comprehensive performance evaluations of CMAQ, and reveals a continuation of\nimprovement in the model's ability to accurately simulate ambient air concentrations of critical\ngas and particulate matter species.\n22","1\n10\nBulk Comp.\nMotor Vehicle Exhaust\nBiomass Comb\nCooking\nOil & NG\n0\n10\nCentreville, AL\n-1\n10\nN. Birmingham, AL\nOak Grove, MS\nGulfport, MS\nYorkville, GA\nAtlanta, GA\nOLF#8, FL\nPensacola, FL\n-2\n< 10\nFigure 7. Ratios of CMAQ model results to ambient measurements of EC, OC, and\nindividual organic compounds at eight southeastern United States sites in July 1999.\nHorizontal lines bound the region in which model-observation agreement is within a factor of\ntwo. Vertical dashed lines distinguish molecular markers specific to different source\ncategories.\nSimulations of the peak 1 - and 8-hour O3 concentrations during the \"O3 season\" (April\nthrough September) were good (r = 0.68,0.69; Normalized Mean Bias (NMB) = 4.0 percent, 8.1\npercent; and Normalized Mean error (NME) = 18.3 percent, 19.6 percent), respectively. The\nNMB statistics associated with the peak 1-hour O3 concentrations are better because the peak\n8-hour concentrations usually include evening hours that CMAQ, historically, has overpredicted\ndue to difficulties simulating the collapse of the boundary layer. CMAQ did display a tendency\nto\noverpredict (NMB often > 30 percent) along coastal regions, which may be tied to poor\ncharacterization of coastal boundary layers and their interaction with land/sea breezes in the\n36-hour grid-cell size used for the annual simulation.\nThe annual simulations of SO 2- were also good (0.77 r < 0.92, depending upon network)\nthough slightly negatively biased (-2.0 percent NMB -10.0 percent) with relatively small error\n23","(25.0 percent NME < 42.0 percent). Spatially, CMAQ's performance was better over the\neastern half of the United States than in the western region, while temporally, the performance\nwas somewhat degraded during the winter months.\nThe performance of CMAQ's NO3 simulations, though still lagging that of O3 and SO42-\nhas shown marked improvement over previous releases. The correlations reflect this progress\n(0.37 0.62, depending upon network) as do the measures of error (80.0 percent NME\n94.0 percent) and measures of bias (-16.0 percent NMB 4.0 percent), which have shown the\nlargest, though somewhat misleading, improvement. Misleading in that when examined over\nspace and time, the NO3 simulations exhibit large, though often compensating NMBs, which are\nthought to be attributable to an incomplete understanding of ammonia emissions. The quality of\nNH4+ simulations is similar to, though somewhat better than that for NO Correlations range\nfrom 0.56 to 0.79, depending on the network. The model produces relatively modest amounts of\nerror (35.0 percent NME < 63.0 percent) and even less, though once again somewhat\nmisleading (for the same reason as seen with NO), (bias -4.0 percent NMB 14.0 percent).\nThe evaluation of CMAQ's performance in simulating NO is influenced both by the large\nuncertainties associated with characterization of ammonia emissions and also with the current\nscientific state of uncertainty surrounding the heterogeneous N205 production pathway for HNO\nformation.\nThe quality of simulations of EC and OC, while similar, are both fairly poor, which is not\nsurprising given the level of uncertainties associate with emissions and the current state-of-the-\nscience. Correlations range from 0.35 (OC) to 0.47 (EC). The model produces fairly large,\nthough not unreasonable amounts of error (NME = 68.0 percent for OC, 58.0 percent for EC) and\nencouragingly small amounts of bias (NMB = -6.0 percent for EC and 12 percent for OC). The\nquality of CMAQ simulations of PM2 much like PM2.5 itself, represents a compilation of the\nquality of all of the simulated particulate species. Overall, the performance of this release\nsignifies a marked improvement over the earlier version of CMAQ as correlations range from 0.51\nto 0.70, depending on the network. The annual bias is very small and identical for each network\n(NMB = -3.0 percent) and the error, though still considerable (NME = 45 percent for both\nnetworks), is greatly improved.\nPotential areas of research into the sources of the deficiencies identified in this evaluation\ninclude uncertainties in the emissions inventories (especially the temporal allocation of NH3\nemissions and emissions associated with the carbonaceous species), imperfect representation of\nthe meteorological fields, because of coarse-grid simulations as well as an incomplete\nunderstanding of the aerosol dynamics.\n2.2.5 Spectral Analysis of the Observed and Predicted Meteorology and Air Quality\nEvaluation of regulatory simulations of fine particulate matter (PM25) is complex because\nsimulations need to be performed for an entire annual cycle while taking into account the\nchemical speciation of PM2.5 to identify its origin and develop meaningful air quality management\n24","strategies. Moreover, annual simulations with air quality modeling systems were performed only\nfairly recently, making the investigation of methodologies to evaluate such extended simulations a\ntopic of increasing interest. In this study, we applied temporal scale analysis as a technique to\nevaluate an annual simulation of PM2. and its chemical components over the eastern United\nStates. The concept of scale analysis is widely used for research in physical sciences, including\nmeteorology, climatology, and air pollution, and also was applied for several air quality model\nevaluation studies during the past several years. The technique was applied to identify the\ntemporal scales that are the largest contributors to the temporal variability in general and to\nperiods of elevated PM2 concentrations in particular. Next, the ability of two air quality models\n(CMAQ and REMSAD) to reproduce the variability and temporal evolution of total and speciated\nPM2.5 fluctuations on different time scales extracted from observations was evaluated.\nThe spectral decomposition of total PM2.5 mass from hourly observations and CMAQ and\nREMSAD (Regional Modeling System for Aerosols and Deposition) model predictions revealed\nthat on days of high PM2.5, concentrations are generally characterized by positive forcings from\nfluctuations having periods equal to or greater than a day (i.e., the diurnal, synoptic, and\nlonger-term components), while the magnitude of intra-day fluctuations showed only small\ndifferences between average and episodic conditions. Furthermore, both modeling systems did not\ncapture most of the variability of the high-frequency variations (i.e.,, intra-day component) for all\nvariables for which hourly measurements were available (Figure 8). It was also illustrated through\nthe use of correlation analysis that correlations were insignificant on the intra-day time scale for\nall variables, suggesting that these models in the setup used for this study were not skillful in\nsimulating the higher-frequency variations in meteorological variables and concentrations of all\npollutants. The models exhibited greatest skills at capturing longer-term (seasonal) fluctuations\nfor temperature, wind speed, O3, sulfate, and nitrate. Correlations for total PM2.5 ammonium,\nelemental carbon, organic carbon and crustal PM2.5 correlations were highest for the synoptic time\nscale implying problems with factors other than meteorology, such as emissions or boundary\nconditions, in capturing the baseline fluctuations. This indicates that capturing the meteorological\nfluctuations on all scales is a necessary, but not sufficient, prerequisite for capturing pollutant\nfluctuations on all time scales, especially for the simulation of PM 2.5\n2.2.6 Model Evaluation Using Advanced Spatial Statistical Models\nA typical model evaluation for CMAQ includes the comparison of each monitoring value\nwith the value simulated by CMAQ for the grid cell in which the monitor lies. Based on these\npaired values, various analyses can be performed based on simple scatterplots, measures of\ncorrelation, and estimates of bias. Such methods allow large amounts of data to be processed\nquickly and produce easily understandable summary plots and statistics. However, for a detailed\nstudy of a particular pollutant and/or region, these traditional methods can be inadequate,\nespecially when monitoring data is relatively scarce.\n25","a) Observations\nDU 19%\nID 10%\nBL 13%\nSY 58%\ntotal variance = 89.2 (ug/m ²²\nb) CMAQ\nc) REMSAD\nDU 23%\nDU 27%\nID 3%\nID\n1%\nSY 60%\nBL 12%\nBL 14%\nSY 60%\ntotal variance = 134.2 (ug/m3)\ntotal variance = 208.4 (ug/m3)\nFigure 8. Pie chart showing the relative contributions of the different temporal\ncomponents to the total variance of hourly PM2 2.5 concentrations, averaged over all\nAQS monitors in the analysis domain. a) Observations from TEOM monitors, b)\nCMAQ predictions, and c) REMSAD predictions. ID stands for the intra-day\ncomponent, DU the diurnal component, SY the synoptic component, and BL the\nbaseline (longer-term) component.\nAdvanced statistical methods can be used to account for the spatial correlation structure\ninherent in the atmospheric process. For instance, Bayesian spatial modeling techniques can be\nused to produce estimates of pollution for each grid cell and the likely errors associated with these\nestimates based on the data collected by sparsely located monitors. These estimates can then be\ncompared with the actual simulated values provided by CMAQ. If the difference between an\nestimated grid cell value and the value simulated by CMAQ is substantially larger than the error\nassociated with the statistical estimate, then that grid cell is identified for further inspection.\n26","In conjunction with the 2001 CMAQ evaluation, this method was used to assess CMAQ's\nability to simulate aerosol sulfate in a portion of the southeastern United States during selected\ntime periods. A pictorial summary of results for the period January 2-January 29, 2001, is\ncontained in Figures 9-12; all sulfate values are given in units of micrograms per cubic meter\nug/m3). Figure 9 displays the average sulfate observed during this period at monitoring stations\nin the southeastern United States. The CMAQ study area is outlined in gray. Note that observed\nvalues from outside the study region's borders are incorporated into the statistical model to\nimprove the sulfate estimates at the edges of the region of interest. Figure 10 displays statistical\nestimates of the sulfate concentrations for each grid cells, based on this observed data.\nFigure 11 reveals the average sulfate values as they were actually simulated by CMAQ. A\ncomparison of Figures 10 and 11 reveals that the differences lie mainly in the northern and\nwestern portions of the region, with an additional smaller area in North Carolina. It is important\nto take the likely errors of the statistical estimates into account when identifying grid cells for\nfurther study. Figure 12 shows only the differences (CMAQ less statistical estimates) for which\nthe respective CMAQ simulated values are outside the 95 percent credible intervals associated\nwith the respective statistical estimates. From Figure 12, primary candidate areas for further\ninvestigation include the southern Indiana-Kentucky, southwestern Tennessee, and northern\nVirginia regions. Further development and applications of Bayesian statistical techniques for\nspatially-correlated processes are continuing.\n2.2.7 Temporal Signatures of Model Output and Observations\nTime series decomposition methods were applied to meteorological and air quality data\nand their numerical model estimates. Decomposition techniques express a time series as the sum\nof a small number of independent modes which hypothetically represent identifiable forcings,\nthereby helping to untangle complex processes. The objectives were to (1) compare the\nperformance of decomposition techniques in characterizing time scales in meteorological and air\nquality variables, (2) use these methods to identify temporal characteristics of observations and\nmodel outputs, and (3) compare outputs expressed as temporal modes from two different\nmodeling systems operated under nearly identical conditions. The results of this effort are briefly\nsummarized here.\nThe decomposition methods included empirical orthogonal functions (EOF), empirical\nmode decomposition (EMD), and wavelet filters (WF). EOF, a linear method designed for\nstationary time series, is principal component analysis (PCA) applied to time-lagged copies of a\ngiven time series. EMD is a relatively new nonlinear method that operates locally in time and is\nsuitable for nonstationary and nonlinear processes. Wavelet filters are linear and band-width\nguided with the number of modes set by the analyst.\nThe time series for these studies included modeled and observed temperature, PM2.5, and\nozone. Temperature should represent a relatively easy test for the decomposition methods as\ncompared to the air quality variables. Since modeled estimates of temperature are forced to\n27","closely track observations from a dense observation network, temporal modes of observations and\nmodel temperature time series should, therefore, be in close agreement. Comparison of modeled\nand observed ozone and PM2.5, on the other hand, are more difficult tests for the decomposition\ntechniques.\nThe aims of observation and model output comparisons involving temporal modes are\nslightly less ambitious than those of signal detection. For example, metrics can be created from\nmodal amplitudes and periodograms without speculation about possible forcings. It is planned to\ncontinue these analyses using observed and modeled ozone. The abilities of the meteorological\nmodel (MCIP) and air quality models REMSAD and CMAQ to capture time scales of temperature\nand ozone were also compared. It was found that models and observations differed in the number\nof significant time scales present (with observations containing one to two more modes than\nmodel outputs). Individual modes differ greatly between observation and model until modes,\noccupying similar frequency bands are added together. As been reported elsewhere, there was a\ntendency for model performance to improve as mode frequency decreased. With respect to model\ndevelopment, displays of temporal modes may be more informative than raw outputs or\ncorrelation-like summary statistics. Literature studies of this nature used much larger signal/noise\nratios than encountered in the hourly PM2.s observations that were analyzed. One should also\nattempt to identify forcings by analyzing nearby observations and climatological time series\neffected by similar forcings.\n2.2.8 Reduction of the Space-Time Domain Dimensionality for Evaluation of Model\nPerformance\nDepending on the spatial extent of the domain, the grid cell size and the length of the\nperiod modeled, models generate enormous amounts of information. Such is the case for the\nannual (year 2001) model simulations using CMAQ and the REMSAD photochemical recently\nexecuted by the EPA. Although the huge amount of information generated by the models is\nvaluable for evaluating their performance, failure to organize this information properly may lead\nto confusion and hamper the evaluation procedure. Therefore, the challenge is to identify a\ntechnique that, utilizing all relevant data available, indicates, in a clear and concise manner, which\nspatial and temporal features of the observations are well simulated and which other ones are\nblurred or not captured by the model.\n28","Figure 9. Average observed sulfate.\nFigure 10. Statistical estimates based on monitoring data.\n29","Figure 11. Average sulfate simulated by CMAQ.\nFigure 12. Significant differences (CMAQ - statistical estimates).\n30","This challenge was addressed by organizing the spatial and temporal observational domain\ninto a limited number of spatially and temporally homogeneous categories and assessing model\nperformance in each category. The technique utilized to compartmentalize the domain is rotated\nprincipal component analysis (RPCA). The procedure was applied to the observed time series of\nthe variable evaluated at multiple observation sites, leading to delineation of regions responding to\ndistinct modes of variations. RPCA was also applied to maps of daily mean sea level pressure,\nleading to identification of the numbers of distinct synoptic patterns observed during the\nsimulated period and the classification of each simulated day in the relevant synoptic pattern. An\nillustration of the use of the technique is being published in FY-2005. The modeled variable\nutilized for the purpose is the 10 m wind speed estimated for the eastern United States (East of the\nRocky mountains) by MCIPv 2.2 applied to MM5 fields. These models generated the\nmeteorological inputs utilized in the annual CMAQ and REMSAD runs previously mentioned.\nIn summary, the quality of daily mean wind speed estimates was not even throughout the\ndomain, as significantly different evaluation metrics were calculated for the various regions\nidentified by RPCA. More specifically, wind speed was generally underestimated (negative bias)\nin the western portion of the domain (from the Dakotas southward to Texas), overestimated along\nthe Atlantic Ocean shores (from New England to Florida), and apparently adequately simulated in\nthe center of the domain. For the 2001 simulation considered in this study, neither the seasonal\nnor the synoptic-based temporal classification allowed detection of time periods with systematic\nweaknesses in wind speed estimates. It is speculated that the data assimilation scheme used to\nnudge MM5 (and, therefore, MCIP) fields towards existing observations may be the reason for\nthis absence of contrast.\n2.2.9 Objective Comparison of CMAQ and REMSAD Performances\nTwo of the most prominent photochemical air quality modeling systems that can be used\nto assess the impact of emission reduction strategies are CMAQ and REMSAD. To promote\nmodel-to-model comparison of these two modeling systems, the EPA performed simulations of\nair quality over the contiguous United States during year 2001 (horizontal grid cell size of 36 X 36\nkm) with CMAQ and REMSAD driven by identical emission and meteorological fields. The\nresults of these simulations were used to compare the abilities of CMAQ and REMSAD to\nreproduce measured aerosol nitrate and sulfate concentrations. Model estimates were compared\nto observations reported by the Interagency Monitoring of PROtected Visual Environment\n(IMPROVE) and Clean Air Status and Trend Network (CASTNet) networks. Root mean squared\nerrors (RMSEs) were calculated for simulation-observation pairs within 10 geographic regions\nand for 12 seasons (months). The Wilcoxon matched-pair signed rank test was used to assess the\nsignificance of the differences between corresponding RMSEs characterizing CMAQ and\nREMSAD skills, respectively. These conclusions were reached:\nCMAQ is more skillful than REMSAD for simulations of aerosol sulfate. The model was\nshown better at reproducing months of high concentrations when compared with\nCASTNet data. CMAQ superiority was not as prevalent although existent when assessed\n31","with IMPROVE observations, leading us to speculate that the strength of the CMAQ\nmodel does not reside in its ability to simulate the shorter term (one day to the next), but\non the longer-term (weekly and longer-term) fluctuations.\nCMAQ and REMSAD performances could seldom be differentiated for nitrate. In the rare\ncases where differentiation was possible, REMSAD and CMAQ were alternatively found\nmore appropriate. As a result, the only fair statement is that both models seem to perform\nabout equally.\n2.2.10 Sensitivity of CMAQ Control Strategy Predictions to Model Input Uncertainties\nSensitivity analyses are important adjuncts to model-data comparisons. The prediction of\nthe effects of emission controls on air concentrations is a key use of the air quality models, which\nare used in the State Implementation Plan (SIP) process to assess impacts of potential emissions\nreduction strategies for the criteria pollutants, in particular ozone and PM25. These predictions\ncan be affected by model input uncertainties, model parameter uncertainties, and structure of the\nmodel itself. One area of concern addressed in FY-2003 is the choice of vertical mixing\nalgorithms because they alter the species concentration mixing histories and, hence, the\nphotochemical processing, potentially altering the control strategy response from CMAQ. The\narea of concern for a chemistry-sensitivity analysis in FY-2004 is the chemical mechanism\nchoices in CMAQ relative to ozone control strategy predictions. Chemical mechanisms in\nEulerian models are uncertain approximations of variables and reactions in the mixed layer;\nhence, their mathematical implementations are varied. Differing implementation of these\nuncertainties in different chemical mechanisms affect photochemical dynamics. Newer chemical\nmechanisms that have better scientific justification can sometimes degrade the chemical transport\nmodel's performance against observed O3. Such degraded performance against observations\nunder current conditions lowers confidence in the control cases using the new science.\nSensitivities were designed to test the performance of O3 base case predictions and also the\nchange in O3 due to emissions reductions for the three photochemical mechanisms implemented\nin CMAQ.\nCMAQ was run to simulate urban and regional tropospheric conditions in the southeastern\nUnited States over 14 days in July 1999 at 32-km, 8-km, and 2-km grid spacing. Runs were\ncompleted with either of the two older mechanisms, CB-IV and RADM2, and with the more\nrecent and complete SAPRC99. The sensitivity matrix included the base case and separate 50\npercent spatially uniform reductions for anthropogenic emissions of nitrogen oxides (NO) and\nvolatile organic compounds (VOC). Comparisons to observations for the base case were\nperformed for the grid cells containing the SOS Cornelia Fort Airpark (CFA) site downwind of\nNashville, Tennessee, and the Southeastern Aerosol Research and Characterization Study\n(SEARCH) site at Jefferson Street (JST) in Atlanta, Georgia. Nashville (CFA) represents a small-\nto-moderate sized city and Atlanta (JST) represents a moderately large urban area.\n32","For the base case, SAPRC99 predicted higher O3 than either CB-IV or RADM2 at JST and\nCFA, especially for afternoon maxima. The 8-km simulation was better than the 32-km grid\nresolution at JST, but the performance for O3 at the 2-km grid resolution was not better than the\n8 km at either site. This is a result found quite often for O3, performance at the finest resolution is\nnot better, and often worse, than at intermediate resolutions of 8 or 12 km.\nFor the emission reduction cases, the differential O3 response to the emission reductions\nshowed differences among the mechanisms, most interestingly at JST. For the 50 percent VOC\nemission reduction cases, the range of O3 response at JST is smaller than that for NO reductions,\nand the O3 benefits are mostly smaller from CB-IV than from SAPRC99. RADM2 benefits were\nmore similar to SAPRC99. For the 50 percent NO emission reductions, both O3 benefits and\nnon-benefits were predicted at JST by all photochemical mechanisms at all grid spacings though\nthere are differences. At JST, CB-IV and RADM2 predict less O3 benefit and more non-benefit\nthan SAPRC99, as well as non-benefits during some hours when SAPRC99 predicts benefits.\nThus, there are differences between O3 responses to emission reductions from CB-IV VS.\nSAPRC99 for the days simulated. The differential sensitivity varied by more than 15 percent\nwith CB-IV and by more than +10 percent with RADM2 for some hours during the days\nsimulated.\n2.2.11 Inverse Modeling for Ammonia: A New Emission Inventory and an Annual\nSimulation\nIn a previous study by Gilliland et al. (2003), a top-down inverse modeling method was\nused to estimate seasonally-varying ammonia (NH3) emissions. As the first published estimates of\nseasonal NH3 emissions, the results were heavily relied upon in air quality modeling applications;\nhowever, uncertainties are inherent because of model prediction errors for meteorology and air\nquality and because of interannual variability in agricultural practices and meteorological\nconditions, both of which can influence the strategies of seasonal forcing in the of NH3\nemissions.\nIn the newest series of 2001 annual air quality simulations using the Community\nMultiscale Air Quality (CMAQ) model, prior estimates of seasonal NH3 emissions were produced\nbased on the seasonal estimates from Gilliland et al. (2003) and newer bottom-up NH3 inventory\nestimates for dairy cattle and fertilizer (Pinder et al., 2004; Goebes et al., 2003). An inverse\nmodeling study was conducted to evaluate the confidence in the prior use of NH3 emission\nestimates for 2001. Advantages to this analysis over the previous study include having a full\nannual simulation including all months and more speciated aerosol data than in the previous study\nthat focused on the year 1990. Disadvantages include having no total ammonium (NH3+NH4) air\nconcentration data available as a cross-check to the inverse modeling results using NH4\nprecipitation chemistry. Several refinements were made to the inverse modeling application in\nGilliland et al (2003) that were used in this study including the incorporation of precipitation\nchemistry uncertainty and removal of uncertainty biases at monitors with extremely low wet\ndeposition or concentration values. Since CMAQ simulations are more computationally efficient\n33","than in the past, inverse modeling results were produced using both NH4 wet deposition and wet\nconcentration data as the optimizing indicators for the inverse modeling application.\nInverse modeling results suggest that the summertime emissions should be higher and\nwintertime emissions should be lower than those originally estimated. While increasing the\noverall amplitude of the seasonality in this way, the results do not suggest any bias in the annual\ntotal emission inventory for NH3. This is a notable result because a high bias was detected in\nprevious versions of the NH3 inventory where the annual total NH3 emissions were approximately\n25 to 30 percent too high. This new inverse modeling study suggests that corrections to the\ninventory have been successful in correcting that bias.\n2.2.12 Bay Regional Atmospheric Chemistry Experiment Model Evaluation\nThe Tampa Bay Estuary Program and the Florida Department of Environment asked the\nEPA and NOAA to enter into a partnership to apply CMAQ to understand the sources of nitrogen\ndeposition affecting Tampa Bay. The majority (60 percent) of the nitrogen deposition to the\nestuary and watershed is estimated to come from sources local to Tampa Bay, which is unusually\nhigh, due to Tampa's isolation from other large source regions. Tampa Bay provides an important\ncoastal atmospheric problem involving coarse particles and sea salt. CMAQ was selected as the\nmodel for the Tampa Bay Assessment, in part because CMAQ will incorporate sea salt in its\naerosol module in FY-2005 and a sectional model incorporating sea salt, CMAQ-University of\nCalifornia, Davis, was under development. Prior to any Tampa Bay assessment, it was agreed that\nCMAQ, starting with CMAQ-UCD, needs to be evaluated against high-quality local data and that\nthe nitrogen budget around Tampa Bay needs to be more carefully characterized.\nThe Bay Regional Air Chemistry Experiment (BRACE), designed for the above two\npurposes, was conducted during May 2002. ASMD scientists, along with ARL colleagues, were\ninvolved in the planning of BRACE. ASMD scientists worked on deployment of true-NO\nmonitors and with Hillsborough and Pinellas Counties' air quality professionals on deployment of\nNO instruments. NOAA scientists took the lead on siting three wind profilers around the Bay,\nand helped define the complete chemistry package of instruments for the NOAA Twin Otter\naircraft flown by ARL. Analysis of the May 2002 data showed a large discrepancy in the\nmeasurement of HNO Hence, ASMD scientists took the lead in organizing and conducting a\nHNO intercomparison study during October 2003. The preliminary analyses from\nintercomparison showed that indeed there was most likely an inlet problem during May 2002, and\nthe October 2003 results will provide the modelers with crucial guidance in how best to create a\nbest estimate of the nitrogen budget for comparisons against CMAQ.\nThe Wexler sectional aerosol model, Aerosol Inorganic Model (AIM), was adapted to\nincorporate sea salt in its calculations. During FY-2004, this sectional model was implemented\ninto the 2004 release version of CMAQ, named CMAQ-UCD. CMAQ-UCD was tested and\nsuccessfully adapted to run on multiple IBM eServer CPU's (EPA's supercomputer).\nCalculations for the period of August 1-14,1999, and comparisons to sectional data from Tampa\n34","Bay allowed the sea salt emissions algorithm to be tested and improved. The MM5 simulations for\nthe May 2002 period at 8-km and 2-km resolutions were evaluated. The stratification of days with\nstrong, weak, and no sea breeze was accurately replicated by MM5. The fidelity of MM5's\nsimulation of the sea breeze swings appeared to be quite good. This was gratifying because\nmodelers in Houston, Texas, had problems accurately replicating the sea breeze. However, there\nwas little difference in the quality of the MM5 simulations between the 8-km and 2-km\nresolutions. At the close of FY-2004, the MM5 simulations were deemed to have passed this\nphase of the evaluation, allowing the CMAQ-UCD simulations to proceed.\n2.2.13 Model Evaluation Tool Development\nSignificant effort is often required to compare observations and model results. Most off-\nthe-shelf tools do not address the specialized needs encountered in model evaluation. The R\nModel Evaluation Toolkit, RMET, was commissioned and developed at the prototype level in the\nR statistical system to address Division and broader modeling community needs. The statistical\nroutines and graphics use R, an open source statistical package that is free. A preliminary version\nof the RMET tool was delivered in December 2003. RMET provides model evaluation capability\nfor working with CMAQ simulations and measurement data, particularly graphical displays of the\ncomparisons. The tool handles CMAQ output in Models-3 I/O-API format. Division-wide\ntraining on R and on the RMET tool will take place in October 2004. RMET will be extended to\nincorporate more advanced model and data comparison capabilities.\n2.2.14 Evaluating the Effect of NOX Reductions\nThis is the first phase of a long-term project aimed at identifying scientific methodologies\nthat will lead to the development of innovative analytical tools to evaluation the effectiveness of\nthe emission control strategies implemented in achieving the intended benefits, namely,\naccountability. Significant reductions in NO emissions from stationary sources have occurred in\nthe eastern United States, and additional reductions are anticipated in the future. These emission\ncontrol programs have been required to reduce the tropospheric ozone levels based primarily on\ncomputer modeling. The NO Reasonable Available Control Technology rule in the NorthEast\nOzone Transport Region, CAAA Title IV, NO SIP call, and Section 126 rule all require NO\nemission reductions throughout the eastern United States from major sources. In addition,\nsignificant future reductions are expected from mobile sources, diesel engine rules, and the Clean\nAir Interstate Rule. Given the significant costs associated with these emission control measures, it\nis important to demonstrate the effectiveness of these rules through analysis of model outputs and\nobservations and to track progress on improving air quality.\nThe conceptual framework of accountability is based on measuring environmental\noutcomes using an integrated environmental assessment model that enables assessing and\ndocumenting relationships between emissions, air quality, atmospheric deposition, and effects to\npublic health and ecosystems. This initiative will begin by reviewing:\n35","1.\nEmission reductions observed in ambient air and atmospheric deposition. Since the 1990\nClean Air Act Amendments, a greater number of stationary sources of SO2 and NO\nemissions have installed continuous emission monitoring systems. Improved systems for\ntracking emission from mobile sources have also been developed. At this level, an\naccountability framework provides a bridge between measured emission reductions and\nchanges in the ambient environment. Resources under this initiative would be applied to\nanalyze specific primary and transformed emission products in ambient air and in\natmospheric deposition (e.g., nitrogen oxide, particle nitrate) over relevant geographic\nareas.\n2.\nPredicted air quality and atmospheric deposition improvements. Resources would be\napplied to enhance the predictive capability to address whether emission reductions have\nresulted in the expected improvements in air quality and deposition, for example:\nReduced ozone, PM s concentrations;\nReduced deposition of NO transformations (e.g., wet and dry deposition of\nnitrate); and\nDiagnostic species (e.g., peroxides, nitric acid, ammonia) useful for model\nevaluations and interpreting dynamic-changes in the atmosphere associated\nemission reductions.\nIn addition to assessing whether the ozone improvements have occurred, this would also\nentail assessing whether these improvements can be attributed to specific emission control\nstrategies that have been implemented. The objective is to research and develop analytical tools\nthat will quantify the effect of regional NOX emission reductions on ambient ozone air quality,\nthus providing a measure of control strategy accountability. The Division is coordinating and\nsharing research plans and products with the EPA Office of Air Quality Planning and Standards\nand the Office of Atmospheric Programs.\n2.3 Toxic Air Pollutant Modeling\nThe Clean Air Act Amendments (CAAA) of 1990 identified almost 200 individual\ncompounds or mixtures of compounds as toxic air pollutants or hazardous air pollutants (HAPs)\nwith the potential for causing adverse health effects. Air quality models for predicting ambient\nconcentrations of these toxic compounds are needed to provide human exposure estimates for\nboth risk assessment and risk management. To obtain accurate estimates of the ambient\nconcentrations of these compounds, there must be a proper accounting of the important processes\nthat control their fate. Since each compound, or mixture of compounds, has unique physical and\nchemical properties that affect the relative importance of those processes, each compound must be\nconsidered individually. The objective of this work is to develop the capability to model toxic\ncompounds at urban and regional scales using CMAQ, and at finer scales using both probabilistic,\nwith probability distributions embedded within CMAQ, and deterministic, using computational\nfluid dynamics models approaches. These models are used to develop spatially and temporally\n36","variable estimates of concentrations of important air toxins at the appropriate resolutions, and to\nevaluate the model predictions. This task is closely linked to other tasks that involve the\ndevelopment and evaluation of the modeling system, improvements in chemical and physical\ncharacterization of air toxins, and the measurement of ambient air toxics concentrations.\n2.3.1. Extending CMAQ to New Species of Air Toxics\nTo assess and manage the risk from HAPs to human health and ecosystems, it is important\nto know how their ambient concentrations and atmospheric deposition vary over location and\ntime. An efficient and perhaps the best way to obtain this information over a national domain at a\nhigh spatial and temporal resolution is the use of air quality models to simulate the chemical and\nphysical processes that control the fate of emitted HAPs. Historically, Gaussian plume models\nhave been used to compute concentrations of HAPs to assess risks to human health reported in the\nNational Air Toxics Assessment (NATA). An EPA Science Advisory Board concluded that the\nNATA's modeling approach inaccurately predicted effects from long-range transport of less\nreactive HAPs and atmospheric photochemistry, because its Gaussian model was unable to\naccount for complex chemical reactions and long-range (>50 km) transport. This neglect can\ncause significant errors in predictions, despite corrections that added \"background\" concentrations\nto model output. The corrections seemed inappropriate, because the background value sometimes\naccounted for most of the total risk. The resultant predictions also did not provide a good way to\nassess contributors to HAP concentrations or to develop control strategies. In response to these\nproblems, CMAQ was modified and applied to simulate HAP concentrations across the\ncontinental United States.\nDuring FY-2004, applications were completed that simulated 20 HAPs within the CMAQ\n(Table 1). The CB-IV mechanism and the Statewide Air Policy Research Center (SAPRC99)\nmechanism were modified and tested to determine whether the choice of mechanism affects the\npredictions. A version of the fast numerical solver, the Euler Backward Iterative (EBI) solver,\nwas developed for both. Four HAPs- formaldehyde, acetaldehyde, 1,3-butadiene and acrolein-\nwere included within the EBI solver because they affect key compounds in such atmospheric\nphotochemistry as ozone and hydroxyl radical. The remaining HAPs were not included in the\nsolver, and they undergo losses from atmospheric photochemistry based on reaction rates and air\nconcentrations after the solver converges to a solution. This treatment was rewritten into a\ngeneralized algorithm to easily incorporate additional HAPs into CMAQ. Reactive tracers were\nincluded to track the emissions of formaldehyde, acetaldehyde, and acrolein. The tracers allow\nthe determination of how production from photochemistry determines the overall concentrations,\na key question in developing HAPs control strategies. The modified mechanism maintains\ncomputational efficiency and does not compromise accuracy in predicted HAP concentrations.\nFor these applications, HAP concentrations across the continental United States during 2001 were\nsimulated, and the results were evaluated against observations. The CMAQ output for HAP\nconcentrations were processed into a format that supports the risk assessment model used in\nNATA.\n37","Table 1. Toxic air pollutant species modeled explicitly\nin CMAQ during FY-2004.\nCompound Name\nCAS number\nformaldehyde\n50-00-0\n1,3-butadiene\n106-99-0\nnapthalene\n91-20-3\nacrolein\n107-02-8\nacetaldehyde\n75-07-0\n1,3-dichloropropene\n542-75-6\nquinoline\n91-22-5\nvinyl chloride\n75-01-4\nacrylonitrile\n107-13-1\ntrichloroethylene\n79-01-6\nbenzene\n71-43-2\n1,2-dichloropropane\n78-87-5\nethylene oxide\n75-21-8\n1,2-dibromoethane\n106-93-4\n1,2-dichloroethane\n107-06-2\ntetrachloroethylene\n127-18-4\ncarbon tetrachloride\n56-23-5\ndichloromethane\n75-09-2\n1,1,2,2-tetrachloroethane\n79-34-5\nchloroform\n67-66-3\nAlso in FY-2004, the modified CB-IV mechanism was used to evaluate the role of\nbiogenic sources in the concentrations of formaldehyde and acetaldehyde. The results showed a\nlarge increase in the role from winter to summer conditions. The increase occurs because emitted\nvolatile organic compounds (VOCs) increase and photochemically produce formaldehyde and\nacetaldehyde. Results indicate the type of VOCs that contribute to most of the photochemical\nproduction based on location and time of year.\n38","2.3.2. Comprehensive Version of CMAQ for the National Air Toxics Assessment\nNATA is designed to help EPA, state, local and tribal governments, and the public better\nunderstand the air toxics problem in the United States. The national-scale assessment includes\nfour steps:\n1.\nCompiling an inventory of air toxic emissions;\nEstimating the annual average outdoor air toxic concentrations;\n2.\nEstimating the exposure (what people are estimated to breathe); and\n3.\n4.\nCharacterizing potential public health risks.\nIn general, larger urban areas appear to carry greater risk than smaller urban and rural areas,\nbecause the air toxic emissions tend to be higher in areas having more people, but this trend is not\nuniversal, and can vary from pollutant to pollutant, according to their sources. Although large\nuncertainties (e.g., emission levels, exposure, toxicity) are inherent in this analysis, EPA uses\nthese results to answer such questions as which pollutants or source sectors may be associated\nwith higher risks than others (e.g., priority setting for data collection).\nThe Division's contribution to NATA is to provide an estimate of the outdoor air toxin\nconcentrations that includes an accurate, state-of-the-science description of all relevant chemical\nand physical processes that can affect the concentrations of toxic pollutants. Model simulations\nwere completed of air toxin concentrations during 2001 over the continental United States. The\nspecies simulated included 20 gas phase toxins. Using CMAQ, spatially and temporally variable\nestimates of concentrations of important air toxins were developed, and the model predictions\nwere evaluated against observations. With these resolved concentration fields, several\nobservations were made about the continental-scale characteristics of HAPs in the atmosphere.\nFigure 13 shows an example of the formaldehyde concentrations predicted by CMAQ for 2001 in\nsummer and Figure 14 shows the concentrations in winter.\nThe results have been compared with observations taken during the 2001 air toxins pilot\nstudy, which consists of 35 monitors in 8 cities (San Jacinto, California; Grand Junction,\nColorado; Tampa, Florida; Cedar Rapids, Iowa; Detroit, Michigan; Rio Rancho, New Mexico;\nProvidence, Rhode Island; and Seattle, Washington). This data set was collected using consistent\nprotocols and was extensively analyzed and quality-assured. Data were taken at 1, 3, or 6 day\nintervals although some monitors did not start collecting data until spring of 2001. Explanation of\nthis data set and its use is given in Luecken and Hutzell (2004). Figure 15 shows the monthly\naverages of formaldehyde concentrations measured at all sites in the air toxins pilot study and\ntheir predicted values from CMAQ. Figure 16 shows the same comparisons for benzene. The\n1:1, 1:2 and 2:1 lines are shown. The majority of the model match the observations within a\nfactor of 2. Figures 17 and 18 show that CMAQ also matches the temporal trends very well for\nformaldehyde and benzene. For both of these pollutants, the day-to-day variability is predicted\nwith skill, although at some sites, the predicted concentrations are lower than the observations.\n39","Formaldehyde - summer\n3-mo average concentations, SAPRC99\n6.00\n86\n4.50\n3.00\n1.50\n0.00\n1\nug/m^3\n1\n131\nFigure 13. Summer 3-month average Formaldehyde concentrations\npredicted by CMAQ.\nFormaldehyde - winter\n3-mo average concentrations, SAPRC99\n1.50\n86\n1.12\n0.75\n0.38\n0.00\n1\nug/m^3\n1\n131\nFigure 14. Winter 3-month average Formaldehyde concentrations\npredicted by CMAQ.\n40","12\n10\n8\n6\n4\n2\n0\n0\n2\n4\n6\n8\n10\n12\nObserved conc (ug/m³)\nFigure 15. Comparison of monthly-\naveraged observed concentrations with\nmodel predictions for formaldehyde.\n8\n7\n6\n5\n4\n3\n2\n1\n0\n0\n1\n2\n3\n4\n5\n6\n7\n8\nObserved conc (u.g/m³)\nFigure 16. Comparison of monthly-averaged\nobserved concentrations with model\npredictions for benzene.\n41","12\n10\n8\n6\n4\n2\n0\nDate\nJWRI\nWERI\nVERI\nEPRI\nULRI\nPredicted\nFigure 17. Time series of 24-hr averaged formaldehyde\nconcentrations measured at the Providence, Rhode Island site\nfor all monitors falling within the single CMAQ grid cell\n(135,72) and the model predictions at this grid cell.\n5\n4\n3\n2\n1\n0\nDate\nJWRI\nWERI\nVERI\nEPRI\nULRI\nPredicted\nFigure 18. Time series of 24-hr averaged benzene\nconcentrations measured at the Providence, Rhode Island site\nfor all monitors falling within the single CMAQ grid cell\n(135,72) and the model predictions at this grid cell.\n42","The CMAQ model results were compared with predictions from the Gaussian plume\nmodel, ASPEN, used to support NATA. It was found that CMAQ produced the same values for\nthe general statistics used to compare the ASPEN model to observations, although with slightly\nmore of a bias. The results were post-processed into a format that supports the risk assessment\nmodel used in the NATA. Hourly values for acrolein were extracted and delivered to other EPA\nlaboratories for consideration of the diurnal variation of acrolein in an upcoming mobile source\nrule. This work has demonstrated that a numerical chemical transport model, CMAQ-CTM, is\na\nuseful and appropriate tool for predicting concentrations of HAPs across the United States for a\nyear-long period. Overall, key findings from this analysis include the following:\nFormaldehyde concentrations are largest in the Southeast and central California, due\nmainly to reaction products of biogenic isoprene.\nThere is a large degree of seasonal, daily, and hourly variability in the concentrations.\nFormaldehyde, for example, can vary by a factor of six from winter to summer, and a\nfactor of two from mid-morning to noon within a typical day.\nAtmospheric chemical formation plays a critical role in the concentrations of some HAPs.\nFormaldehyde and acetaldehyde, for example, are mostly chemically-formed in the\natmosphere, with only a small portion due to direct emissions.\nA comparison of the modeled concentrations with a limited set of observations shows that\nCMAQ is generally able to reproduce the temporal variability in the data.\nCMAQ concentrations can differ substantially from ASPEN predictions for species whose\nconcentration depends substantially on atmospheric chemistry.\nTo follow up this work, plans for the 2002 NATA include improving on the previous analysis by\nincluding additional toxic species and developing ways to account for local \"hot spots\" of high\nconcentration.\n2.3.3 Linking CMAQ to a Human Exposure Model in an Urban Area\nThe Division completed a FY-2004 pilot study to develop the capability to provide\nadvanced photochemical grid-model air-toxic concentrations to a human exposure model. The\npilot study also began to assess whether a grid-based chemical transport model could successfully\nreplace and/or augment traditional Gaussian plume modeling approaches to providing annual\nambient concentration estimates of air toxics for human exposure assessments in urban settings.\nThe study consisted of: (1) extending an air toxics version of the CMAQ modeling system\nto a modeling domain centered over Philadelphia at 12- and 4-km grid meshes; (2) performing\nmodel simulations for the year 2001; (3) comparing the modeling results limited with\nobservational data collected at Camden, New Jersey, under the Urban Air Toxics Monitoring\n43","Programwww.epa.gov/ttn/amtic/files/ambient/airtox/main-2a.pdf; (4) reformatting the modeling\nresults into the 3-hour annual averages needed for input to HAPEM5; and, (5) assessing the\npracticality of using CMAQ for estimating air toxics for human exposure assessments by\nexamining the computational requirements needed for this exercise.\nThe pilot study demonstrated that CMAQ can be an useful tool to simulate the air toxic\nconcentration fields needed to drive a human exposure model. A comparison of the modeled\nconcentrations with a limited set of observations suggests that the CMAQ model was able to\nreproduce the temporal features embedded in the data. For this pilot study, air toxic\nconcentrations generated by the CMAQ modeling system for a 4-km grid mesh overlaying\nPhiladelphia were successfully formatted for direct use in the HAPEM5. Based on these results,\nCMAQ is being considered for application for the EPA's NATA program.\nThe next phase of this research effort will focus on extending the CMAQ modeling system\nfor simulating air toxics with finer grid cell sizes (~1 km) and to examine the practicality of\nintermingling Gaussian dispersion model estimates with CMAQ results. In addition, CMAQ\napplications are being performed for Houston, which is an excellent urban test bed for further\ndevelopment because it has a detailed building morphology database to test the urban\nparameterizations for meteorological modeling and it has detailed air toxic concentration data\nfrom such field studies as the Texas 2000 Air Quality Study that can be used for extensive model\nevaluation.\n2.3.4 Parameterizing the Urban Canopy\nThe difficulty of predicting air pollutant dispersion at high spatial resolution is exacerbated\nby the need for high quality, high definition of the meteorological fields that govern transport and\nturbulence in urban areas. Air quality fields are now being modeled at finer spatial resolution to\nreveal \"pollutant hot spots\" in urban areas. These fine resolution mesh simulations will need to be\ndriven by meteorology at commensurate mesh sizes. Since most of the primary atmospheric\npollutants are emitted inside the roughness sub-layer (RSL), and consequently, the first chemical\nreactions and dispersion occur in this layer, it is necessary to generate detailed meteorological\nfields inside the RSL to perform air quality modeling at high-spatial resolutions.\nAt neighborhood scale (on order of 1-km horizontal grid spacing), the meteorological\nfields are strongly influenced by the presence of the vegetation and building morphology of\nvarying complexity, which requires developing more detailed treatment of the influence of canopy\nstructures in the models and using additional morphological databases as input. The assumptions\nof the roughness approach, used by most of the mesoscale models, are unsatisfactory at this scale.\nHence, a detailed urban and rural canopy parameterization (Dupont et al., 2004), called\nDA-SM2-U, was developed and incorporated inside MM5 to simulate the meteorological fields\nwithin and above the urban and rural canopies. DA-SM2-U uses the drag-force approach to\nrepresent the dynamic and turbulent effects of the buildings and vegetation, and a modified\nversion of the soil model SM2-U (Dupont et al., a and b accepted for publication), called\n44","SM2-U(3D), to represent the thermodynamic effects (e.g., estimates the heat and humidity fluxes)\nof the canopy elements at different levels within the canopy.\nRoughne ss\nDrag-Force\nNet radiation\na aproa ch\napproach\nSee. sible\nLater\nSamage\n-\nheat floor\nbeat fax\nheat four\nheat floor\nR\nH\nLE\nProfessional\nT\nLS\n1\nnoof\n-\nQ\nnatural\nsoil\nT\nFaxed\nbase\nsurface\nsoil\nCream\nwillayer\nRunner supar\n-\n3\" will layer\nFigure 19. Scheme of the new MM5 canopy parameterization, DA-SM2-U, using the\ndrag-force approach with the soil model SM2-U(3D), compared with the roughness approach.\nThe drag-force approach transmits directly to the atmosphere the dynamic, thermodynamic\nand turbulent effects of the canopy elements (vegetation and buildings) by changing the\nconservation equations of the mesoscale model. The lower level of the computational domain\ncorresponds to the real level of the ground, and additional vertical layers are included within the\ncanopy to allow more detailed meteorological fields within the RSL (Figure 19). Inside the\ncanopy, the effects of buildings and vegetation are represented by adding in the dynamic equation\na friction force induced by horizontal surfaces of buildings, and a pressure and viscous drag force\ninduced by the presence of buildings and vegetation. In the temperature equation, the sensible\nheat fluxes due to buildings and vegetation, and the anthropogenic heat flux parameterization\nfollowed Taha (1999). In the specific humidity equation, the humidity sources coming from the\nevapotranspiration of the vegetation and the evaporation of the water intercepted by buildings\nwere included. A shear production term induced by horizontal surfaces of buildings, turbulent\nkinetic energy sources induced by the presence of buildings and vegetation, and buoyant\nproduction terms from the sensible heat fluxes emitted by buildings and vegetation were included\nin the turbulent kinetic energy equation. The turbulence length scale was also modified inside the\nurban canopy, as proposed by Martilli et al. (2002). The volume of buildings is considered in\neach cell, whereas the volume of the vegetation is neglected. The turbulent transport in the\nvertical is also modified to consider the real volume of air in the cell.\n45","DA-SM2-U is thus a multi-layer canopy and soil model with few layers of a couple meters\nwithin the canopy depending on the mesh of the mesoscale model domain, and three layers within\nthe ground; a surface soil layer for the natural surfaces, a root zone layer, and a deep soil layer. A\nfirst evaluation of DA-SM2-U on the City of Philadelphia (Dupont et al., 2004) with a simple\nurban morphology representation showed that the model is capable of simulating the important\nfeatures observed in the urban and rural areas.\nIn FY-2004, an effort was undertaken to apply the DA-SM2-U version of MM5 to\nHouston, Texas. A detailed GIS database of urban canopy parameters (UCP) gridded at 1 km was\ncreated for the entire MM5 computational domain. To provide the most accurate representation\nof the morphological parameters for the entire MM5 computational domain on Houston, a GIS\nurban database was created (Burian et al., 2004). This database includes multiple surface\ntopography and surface cover digital data sets, including land use, bare earth elevation,\nfull-feature digital elevation model, and roadway locations. The parameters were then correlated\nto the underlying land-use type using area-weighted averages. The average values for each\nparameter for each land use type were then extrapolated to each 1-km grid cell in the MM5\nmodeling domain using an area-weighting scheme based on land use fraction with the grid cell.\nTable 2 lists the 23 different UCPs that were developed for this project.\nTable 2. Urban Canopy Parameters (UCP) for Houston, Texas.\nCanopy UCPs:\nBuilding UCPs:\nVegetation, Other UCPs:\nMean Canopy Height\nMean building height\nMean vegetation height\nCanopy plan area density\nStandard deviation of\nVegetation plan area density\nCanopy top area density\nbuilding height\nVegetation top area density\nCanopy frontal area\nBuilding height histograms\nVegetation frontal area density\ndensity\nBuilding wall-to-plan area\nRoughness length\nratio\nMean orientation of streets\nDisplacement height\nBuilding height-to-width\nPlan area fraction surface covers\nSky view factor\nratio\nPercent directly connected\nBuilding plan area density\nimpervious area\nBuilding rooftop area density\nBuilding material fraction\nBuilding frontal area density\nThis effort provides the first implementation of this detailed set of gridded UCPs into the\nDA-SM2-U/MM5 system. A case study for August 30, 2000, was selected for a domain\nencompassing the greater Houston-Galveston area (Ching et al., 2004a). The period of interest\nalso corresponds to the Texas 2000 Air Quality Study. Simulations were made at 36-, 12-, and\n4-km grid sizes using 30 sigma layers in the vertical.\n46","Sensitivity studies of work in progress have indicated a need for accurate input boundary\nconditions of flow fields from the coarser 4-km grid mesh influenced by the mesoscale circulation\ninduced by Gulf of Mexico and Galveston Bay. The inclusion of high resolution, diurnally\nvarying sea surface temperatures, was necessary for accurate fine scale flow simulations and was\napplied in the comparative results for Figures 20 and 21, which display significant differences in\npredicted dispersion parameters between standard MM5 (using Reynolds averaging) and the\nurbanized version of MM5 (using DA-SM2-U-UCP) at a 1 km grid mesh.\nThe DA-SM2-U/MM5 system was successfully implemented using a sophisticated set of\ngridded UCPs based on high resolution building and vegetation data. Modeling tools to help\nresolve physically, flows in urban areas that are impacted by the presence of canopy features at\n1-km grid sizes were developed. This method reduces the problems or uncertainties associated\nwith simple interpolation schemes that cannot be expected to accurately represent the flow in\nurban areas. Thus, the combination of UCP-driven meteorology for fine scale modeling and\nmore-accurately modeled lake-land breeze circulations will provide stronger scientific basis for\nadvancing the simulations of the flow and air quality for Houston and other urban areas with\nsimilar climatic features.\nA city-specific database of urban morphological data and UCP daughter products are\nneeded for running these advanced UCP for all urban areas. Initial discussions are underway to\nexplore the creation of such a database (Williams and Ching, 2004). If such a database can be\nachieved, the utility of advanced approaches as described in this section may become attractive.\nFigure 20. Sensible heat flux (20 UTC, August 2000) DA-SM2-U (left hand side); Standard\nroughness approach (right hand side).\n47","Figure 21. PBL height (m) (20 UTC, August 2000) DA-SM2-U (left hand side); Standard\nroughness approach (right hand side).\n2.3.5 Advancing the Neighborhood-Scale Version of CMAQ\nAir quality simulation models need a more advanced capability for application at finer\nscales and to serve as tools for performing exposure and risk assessments in urban areas (Ching et\nal., 2004a; 2004b; 2004c). Also, air pollutants need to be modeled at relatively-fine spatial\nresolutions to reveal \"hot spots\" in urban areas. Fine resolution mesh simulations require that\nemissions and meteorology be characterized at commensurate mesh sizes and that a suitable\ndatabase be available to evaluate the model results. One of the challenges in extending CMAQ to\na relatively fine mesh is to account for the presence of urban streets and tree canopies.\nPreliminary results by Ching et al. (2003) at 1.3 km mesh size, using a simplified set of urban\ncanopy parameters for Philadelphia based on surveys of urban building geometries (Otte et al.,\n2004), have shown that the resulting MM5 and CMAQ fields are significantly impacted by the\nintroduction of urban canopy parameters (UCPs) of buildings. To examine the impact of a finer\ngrid mesh, the use of a more sophisticated urban canopy parameterization (Dupont et al., 2004)\nwith a detailed urban morphology data set, and an improved depiction of the nearby water\ntemperatures, the CMAQ system is being tested over the Houston, Texas, area.\nBefore embarking on a comprehensive suite of model simulations and evaluations, the\nsensitivity of the CMAQ system was examined using a one-day test case on August 30, 2000. For\nthis case, sensitivity tests were designed to examine the differences in concentration fields\nresulting from the urbanized MM5 system and the standard version of MM5 run at 1-km\nresolution. Before setting up the 1-km simulations, 4-km grid-mesh simulations were improved\nusing high-resolution sea-surface temperature (SST) observations taken from the Polar-orbiting\nOperational Environmental Satellites Advanced Very-High Resolution Radiometer. The use of\n48","the SST observations in MM5 improved the accuracy of the near-surface land-bay breeze\ncirculation simulations, by clearly reproducing the observed wind directions and the wind shift at\nthe time of the Bay breeze passage not captured in the control run. Subsequently, it was decided\nto adopt the use of the more accurate temporally-resolved SSTs for Galveston Bay in the Houston\nstudy.\nThe sensitivity of CMAQ to different grid meshes is shown in Figure 22. The left panel\nshows a one-hour snapshot of formaldehyde from the 1-km grid simulations driven by the\nDA-SM2-U version of MM5, and the center panel shows results from the 4-km standard version\nof CMAQ. While the general pattern is similar for the two panels, \"hot spots\" with the 1-km\nversion of the model are clearly visible. The right most panel illustrates the sub-grid variability\nassociated with the 4-km resolution concentration variations that are possible using the 1-km\npredictions. Variability in this panel is indicated by computing in each 4-km grid the ratio of the\nrange of the 1-km results by the mean of the 1-km results. These results were obtained by\nsampling the 16 1-km grid values in each 4 km cell. Areas with large normalized range-to-mean\nvalues are visible throughout the model domain.\nDuring FY-2005, additional studies are planned to examine the use of the UCP for 1-km\nCMAQ simulations over Houston. Working in collaboration with scientists from the University\nof Houston, the Division plans to simulate several weeks of the Texas 2000 air quality study\nperiod and to evaluate model results with observations collected during the study. The\ncombination of UCP-driven meteorology for fine-scale modeling and more accurately modeled\nlake-land breeze circulations may in the future improve the simulations of atmospheric flow and\nair toxic concentrations for Houston and other urban areas with similar climatic features.\n2.0\n10\n8\n1.5\nS\n1.0\n2\n0.5\no\nppbV\n0.0\nFigure 22. Formaldehyde (ppb) simulations for August 30, 2000 at 2100 GMT. Left panel is\nthe 1 km simulation (with UCP), middle panel is native 4 km grid size. Right panel is range\nof values of the individual 16 1-km cells in each 4 km grid normalized by the 4 km\naggregated cell mean.\n49","2.3.6 Modeling Subgrid Concentration Variability\nWhen air quality (AQ) models are used for human exposure and risk assessments they\nneed to provide detailed information on the location and magnitude of hazardous air pollutant, or\nair toxics, concentrations, with particular interest in capturing extreme values, or hot spots.\nRegional-scale AQ models are typically limited to relatively coarse resolutions when simulating\nmean pollutant concentrations for each grid cell volume. Computational fluid dynamics and\ncoupled large-eddy simulation with photochemistry techniques allow AQ simulations with much\nfiner grid spacings, but these types of simulations are impractical for long-time integrations or\noperational use. Thus, procedures are needed for representing the subgrid pollutant concentration\nextremes in regional models without requiring concurrent fine resolution simulations.\nDuring FY-2004, an initial effort to formulate distribution functions to represent subgrid\nvariability (SGV) fields was begun by the Division and the ARL Atmospheric Turbulence and\nDiffusion Division, Oak Ridge, Tennessee. Initial description and results are provided in\nHerwehe et al. 2004). The approach is to improve air quality simulation capability for human\nexposure and risk assessment modeling tools in urban settings by providing within-grid\nconcentration variability distribution functions in the form of probability density functions (pdf)\nas a complement to regional or urban scale grid simulations (Ching and Byun, 1999). This effort\ncontinues an exploratory phase begun in FY-2003, using results from a CMAQ simulation for\nPhiladelphia at 1.3-km grid size to determine the pollutant pdf characteristics and parameters\nrepresenting the subgrid scale variability for 4-km and 12-km grid meshes. Preliminary analyses\nindicate that the within-grid spatial variability in concentration fields arising from the distribution\nof sources in each grid can be very important, and in many instances, perhaps even as or more\nimportant than the grid resolved fields for exposure analyses. Further refinements will be possible\nwhen modeling results from the source distributions and from subgrid chemistry become\navailable.\nWork continues to describe and characterize properties of distribution functions in terms\nof their shape, location and scale parameters as a means to derive parametric fields of SGV\nfunctions. The Dataplot command script under development for research was named\nConcentration Distribution Functionware, or CDFware. Desired CDFware improvements include\nthe ability to detect multi-modal, particularly the relatively common bimodal, data and to fit such\nmixed multiple distributions as a mixture of two Weibull distributions in the bimodal case to the\ndata set. CDFware will also be applied to higher resolution outputs from neighborhood-scale\nmodels in order to determine whether more coherent parameter fields can be detected at the finer\nresolutions.\n2.3.7 Developing and Applying CFD Simulations of Pollutant Transport and Dispersion\nDevelopments and applications of Computational Fluid Dynamics (CFD) are ongoing for\nsupport of urban air toxics assessments and homeland security issues. CFD modeling has\nemerged as a promising technology for simulating wind flow and pollutant dispersion in urban\n50","microenvironments. Development and applications are linked closely with the advancing\ncapabilities of both software and hardware. In addition to using EPA computing resources,\ncooperation has been established with the Department of Energy's Argonne National Laboratory\nfor use of their large Linux cluster. Much is being learned about how best to set up CFD\nsimulations to support environmental simulations and the issues that most affect comparability\nwith physical model studies and field measurement studies. The choice of boundary conditions,\ngrid resolution and structure, and turbulence models affect the outcome of a solution significantly.\nTransport and dispersion can be well simulated for flat plate like atmospheric boundary layers.\nNo work has been done for stable stratified flows. Transport and dispersion simulations are more\ncomplicated for atmospheric flows due to the complex temporal-spatial wind fluctuations. The\nproject has focused on RANS (Reynolds-Averaged Navier-Stokes) steady-state solutions and the\nstandard k-e (turbulent kinetic energy and turbulent energy dissipation rate) turbulence models.\nThis study is being extended to include unsteady solutions and higher order turbulence models.\nProgress has been made on CFD development of wide ranging atmospheric boundary layers\n(Huber et al., 2004a).\nWhile setting up a working model of the extremely complex building environment for\nLower Manhattan was a challenging exercise, there were many lessons learned that should make it\neasier to set up similarly complex urban environments in the future. An overview was published\non numerical modeling done in support of the studies following the event of September 11, 2001,\nin Lower Manhattan (Huber et al., 2004b). A CFD simulation for the Lower Manhattan building\nmodel has not been done. Cooperation from other agency is being arranged to support a full\nsimulation during FY-2005. Nonetheless, the current results are valid for winds from the West\n(negative X direction) since all upwind buildings to the Hudson River are included. Figure 23\npresents a comparison with a vertical profile of measurements from the EPA wind tunnel model.\nFigures 24 and 25 show an example of the flow field details for a vertical and horizontal slice.\nFigures 26 and 27 show example flow and concentration detail surrounding an emissions source\nin the \"ground zero\" area. Understanding the pathway of toxic air pollutants from source to\nhuman exposure in urban areas finds immediate application for both routine air pollution\nassessments and in support of homeland security (Huber, 2004).\nThe collapse of the New York World Trade Center towers on September 11, 2001,\ndemonstrated some of the shortcomings in conducting rapid exposure and risk analyses in urban\nareas where the understanding of airflow around large buildings is poor. While problem-specific\napplications of CFD may not be feasible in \"real-time\" support, there is a major role for CFD\nsimulations to be run for developing archives that could be tabularized for supporting real-time\napplications. Also, CFD simulations could have a significant role in supporting field studies in\nurban environments, which could be used to develop performance verification. Future research\nand development, including CFD simulations, could lead to the development of reliable simplified\nmodels (or databases) as needed to support emergency responders. CFD simulations can be used\nto support necessary post-event analyses as is being done in support of post 9/11 studies. CFD\nmodeling is being extended as part of multi-agency support for the Department of Homeland\nSecurity's New York City Urban Dispersion Program (UDP). Preliminary results supporting this\nprogram are presented in Figures 28 and 29. Participation with the UDP program provided a good\nopportunity to demonstrate potential uses for CFD simulations and major field measurement\ncampaigns and physical modeling will help evaluate the CFD model results.\n51","Comparisons at Port 206: Free-stream wind along x direction\nHoriz. Wind Speed\nHoriz. Wind Direction\n3\n120\n206 Hmog\nCFD hmag\n100\n2\nEE\n80\n1\n206 Hold\n60\nCFD hdir\n0\n40\n0\n50\n100\n150\n200\n0\n50\n100\n150\n200\nHeight [m]\nHeight [m]\nVert. Wind Speed\n0.2\n0\nPort 206\n-0.2\n206 Vmag\nCFD vmag\n-0.4\n0\n50\n100\n150\n200\nHeight [m]\nTKE\n0.6\n0.5\n206 THE\n0.4\nCFD TKE\n0.3\n0.2\n0.1\n0\n0\n50\n100\n150\n200\nHeight [m]\nFigure 23. Comparison between vertical profile measurements of wind speed, wind direction,\nand TKE (Turbulent Kinetic Energy) made in the EPA wind tunnel model of lower\nManhattan, and the corresponding CFD predictions.\n52","Example: Surface Winds (10% sample)\nVisualization\n0\n2 4 6 8 10 12\nwind speed [m/s]\nFigure 24. Predicted wind speed (m/sec). Flow field details for a horizontal slice at the\nsurface.","Example: Winds of Vertical Plane (10% sample)\n0\n2 4 6 8 10 12\nwind speed [m/s]\nFigure 25. Predicted wind speed (m/sec). Flow field details for a vertical slice.","Winds and Concentration: Z = 10 m\n100 m\nVisualization\n-2\n-3\n-4\n-5\n-6\n-7\n-8\nFigure 26. Wind vectors and concentration contours surrounding an emissions source in the\n\"ground zero\" area.\n55","Cross-stream Profile: Free-stream wind along x direction\nVisianalization\nLogo(Tracer Mass Fraction)\n-8\n-7\n-6\n-5\n-4\n-3\n-2\nFigure 27. Vertical concentration profiles surrounding an emissions source in the \"ground\nzero\" area.\n56","Winds from Southwest: Madison Square Garden\nWindSpeed Tm/s\nFigure 28. Preliminary predictions of wind speed for a simulation supporting the Department\nof Homeland Security's New York City Urban Dispersion Program.\n57","Example Plume: Madison Square Garden\nFigure 29. Example of predicted plume dispersion for a simulation supporting the\nDepartment of Homeland Security's New York City Urban Dispersion Program.\n58","2.4 Multimedia Modeling and Application Studies\n2.4.1 Multimedia Integrated Modeling System Spatial Allocator\nASMD continued development of the Multimedia Integrated Modeling System (MIMS),\nwhich includes the MIMS framework and the Spatial Allocator2. The MIMS framework provides\na software infrastructure to support configuring, applying, and evaluating environmental models.\nASMD enhanced the framework to make it easier to set up complex simulations, for instance,\nwhere models are invoked repetitively or where there are many input or output parameters. A\nJava scientific plotting library, based on the open source statistical package R, was also\nsignificantly enhanced. OAQPS used MIMS as the basis for a prototype decision support system.\nThe decision support system uses nonlinear optimization to find promising alternatives that\nbalance economic and environmental objectives. In that system, MIMS provides the\ninfrastructure for user interaction and managing multiple model executions, which are used to\nexplore the decision space.\nThe Spatial Allocator tool of MIMS is in the process of being upgraded to provide the\ncapability to grid input files needed by SMOKE, including emission inventory data, surrogate files\nfor spatial allocation of emission sources and land cover data needed for modeling of biogenic\nemissions. It is expected to be completed by fall 2005. The Spatial Allocator requires only grid\ndefinitions and for spatial surrogates GIS shape files.\n2.4.2 Multi-Layer BioChemical Model for Calculating Dry Deposition\nThe Clean Air Status and Trends Network (CASTNet) is operated by EPA's Clean Air\nMarkets Division and the National Park Service to monitor concentration and dry deposition at\nsites across the country to assess long-term trends in air quality and environmental protection\nresulting from regulatory policies and emission reductions required under the Clean Air Act.\nCASTNet estimates dry deposition flux by combining measured concentrations of pollutants with\nmodeled deposition velocities. The Multi-Layer Biochemical Model (MLBC) (Wu et al., 2003a;\n2003b) is being examined for use in the network operations for predicting the deposition velocity.\nIn its original design, MLBC treats all canopies as a mixture of the plant species. For CASTNet,\nthe canopy at a site is treated as spatially distinct species where the deposition velocity is\ndetermined from area weighting the deposition velocities for each of the local species.\nMLBCv1.0 was modified to allow for this latter approach to develop a new version of the model,\nMLBC-AW (Area Weighted). Development of MLBC-AW is continuing to assure that the model\ncan be run with network meteorological data. One area of focus will be determining a better\n2 Developed by the Carolina Environmental Program at the University of North Carolina\nat Chapel Hill, the Spatial Allocator is a free tool for generating spatial surrogates for emissions\nand performing other spatial allocation without requiring a geographic information system.\n59","approximation for the leaf temperature since this is not measured at the network sites, but is an\nimportant input to MLBC-AW.\nLeaf area index (LAI) and canopy height are not routinely measured at CASTNet sites and\nare currently modeled using a step-function that is based on measurements made in 1991-1992,\nand 1997 and depend only on the day of the year. Thus, the same annual leaf-out profile is used\neach year, whereas actual LAI values respond to interannual variability in rainfall, temperature,\nradiation, etc. Given the sensitivity of deposition models to LAI, obtaining better estimates of\nLAI should provide more realistic estimates of deposition flux. Plant growth models offer one\noption for predicting the response of plants to interannual variability in meteorological conditions.\nThere are many different approaches to plant growth modeling ranging from simple\nparameterizations to more complicated photosynthetically- based models. During FY-2004, the\nplant growth algorithm was extracted from the Erosion Productivity-Impact Calculator and\nembedded within MLBC-AW. This version of MLBC is denoted as MLBC-PG. In MLBC-PG,\nthe meteorology used as input for the deposition model is also used for the plant growth\nalgorithm. Water and temperature stress factors calculated by the deposition model for\ndetermining canopy resistance were used in the plant growth algorithm. The LAI and canopy\nheight determined from the plant growth algorithm are used in the deposition model for\ndetermining pollutant deposition velocities. Comparisons of model derived LAI with the\nCASTNet step-function showed potentially important differences in LAI. However, the\nCASTNet sites have numerous extended periods with missing meteorological data, which make\nthe use of the plant growth model unfeasible at this time. As a result of this modeling work,\nconsideration is being given to having site operators routinely measure LAI for input to the\ndeposition model.\n2.4.3 Chesapeake Bay 2007 Re-Evaluation\nASMD has established a long-term relationship with the EPA and NOAA Chesapeake Bay\nPrograms to address multi-media environmental problems where the atmosphere is an important\nsource of reduced and oxidized nitrogen through deposition. Chesapeake Bay is a leader in using\nmulti-media modeling approaches. Two major Chesapeake Bay re-evaluations or assessments of\nrequired nitrogen load reductions to the Bay have already occurred.\nChesapeake Bay has been placed on EPA's list of impaired waters. The Chesapeake 2000\nagreement calls for pre-empting the need for a TMDL (Total Maximum Daily Load) plan by\ncleaning up the Bay by 2010. The Bay 2007 re-evaluation is a critical step in this process towards\nthe 2010 cleanup and delisting, and ASMD is participating in the re-evaluation process. The best\nscience is desired for the re-evaluations, and during the period between major re-evaluations,\nASMD is changing its multi-media modeling of nitrogen from the Extended RADM to its new\nmodel, CMAQ. The CMAQ dry deposition algorithms were revised in FY-2003, improving\ndeposition parameterizations for NH3, HNO and other nitrogen containing species. CMAQ has\nbeen sufficiently evaluated for deposition to show that it is an improvement over the Extended\nRADM. A newly designed aggregation data set with 40 cases was developed for CMAQ that can\n60","directly address seasonal deposition. The outer, continental grid resolution is 36-km, a significant\nreduction over the 80-km resolution used with the Extended RADM, and now covers the entire\nlower 48 States. For Chesapeake Bay multi-media simulations, a 12-km nest over the Mid-\nAtlantic region covering most of the Chesapeake Bay airshed was developed. This inner nest\nbetter resolves the Bay surface compared to the 20-km nest used with the Extended RADM. In\nFY-2004, the adaptation of the aggregation method to the new aggregation data set was completed\nand the base year simulation was updated to use 1999 and 2001 emissions to better match the new\nbase year of 2000 used by the water quality model. Also, a new 2010 NO SIP call futures case\nwas simulated to compare the changes with previous estimates made using the Extended RADM,\nwhich used a 1990 base year. The CMAQ-predicted relative change was consistent with the\nExtended RADM-predicted relative. As expected, the relative change from 2001 to 2010 was\nsmaller than the change from 1990 to 2010 because NO emission began trending downward after\n1996. However, the absolute changes in nitrogen deposition were not consistent. Comparisons\nagainst the 2001 annual run of CMAQ also showed inconsistencies with the aggregation results.\nThis will be investigated in FY-2005.\n2.4.4 Ammonia Budgets for Coastal Systems\nAn important fraction of atmospheric nitrogen deposition is reduced nitrogen\n(ammonia/ammonium). With successful implementation of the EPA regulations on NO\nemissions for control of ozone and increases in animal operations in the eastern seaboard states,\nreduced nitrogen is expected to become a majority of the nitrogen deposited from the atmosphere.\nHowever, ammonia is not receiving the attention it deserves, in part, because many ecologists\ndealing with marine estuaries and watersheds believe ammonia deposits instantly SO that none\nleaves the immediate area. Long-range transport of ammonia is ignored. ASMD has an\nopportunity to correct this misinterpretation of data through modeling and model-data\ninterpretation studies using the regional models. Model atmospheric budget analyses were\nperformed in FY-2002 with MAQSIP (Multiscale Air Quality Simulation Platform), a\ndevelopment predecessor to CMAQ, for North Carolina ammonia emissions associated with the\nlarge increase in the hog population. The analysis, covering a short summer period and reported\nat the International N2001 Conference, show that only 5 to 10 percent of the NH budget dry-\ndeposits locally, while most of the ammonia emissions are involved in long-range transport,\ncontrary to conventional wisdom. The model results are consistent with spatial and temporal\ntrends in the ammonia wet deposition data. Nonetheless, the conventional wisdom persists and\ndistorts studies of nitrogen-cycling in coastal estuaries, introducing significant errors in them.\nThe 1999 summer period was chosen for the next phase of analysis due to availability of\nspecial 12-hour integrated gas and particle measurements at the Clinton site in the middle of the\nhog farm area in North Carolina. The ammonia inverse modeling was re-applied to July and\nAugust 1999. Simulations were carried out at 32-km and 8-km grid-cell sizes with the newer\nCMAQ with the new M3Dry deposition algorithms. For the 8-km simulation, process analysis\nwas turned on in CMAQ. Preliminary comparisons showed very reasonable agreement between\nmodeled and measured NHx levels at Clinton and Atlanta, Georgia. However, with the updated\n61","M3Dry deposition algorithms, the process analysis now shows that dry deposition of ammonia in\nCMAQ is 4-5 times higher than previously estimated. The new estimates may be too high\nbecause the bi-directional air-surface exchange of ammonia is ignored. A new simulation period\nis being selected for which dry deposition measurements in North Carolina exist to better assess\nand bound the model results. The goal now is to provide some bounds on the ammonia budget\ncalculations.\n2.4.5 Tampa Bay Study\nThe Tampa Bay Estuary Program and the Florida Department of Environment asked EPA\nand NOAA to enter into a partnership to apply CMAQ to understand the sources of nitrogen\ndeposition affecting Tampa Bay. The majority (60 percent) of the nitrogen deposition to the\nestuary and watershed is estimated to come from sources local to Tampa Bay, which is unusually\nhigh, due to Tampa's isolation from other large source regions. ASMD was asked to work with\nthe Tampa Bay National Estuary Program to assess the atmospheric contribution of nitrogen to\nTampa Bay. Tampa Bay provides an important atmospheric multi-media problem involving\ncoarse particles and sea salt. Two of the largest power plants in the nation, in terms of NOx\nemissions, are on the shores of the Bay and there are serious questions as to how much of the\natmospheric deposition is due to the power plants versus mobile sources in the area surrounding\nthe Bay. CMAQ was selected as the model for the Tampa Bay Assessment, in part because\nCMAQ will incorporate sea salt in its aerosol module in FY-2005. Prior to any Tampa Bay\nassessment, it was agreed that CMAQ needs to be evaluated against high-quality local data.\nThe Tampa Bay study needs to have an annual average deposition as its basis to be able to\nbe\nused by the Tampa Bay National Estuary Program. The use of the aggregation was\ncontemplated. Because the sea breeze has an important influence on transport over Tampa Bay, a\nwind hodogram analysis was conducted to ascertain whether the aggregation set, with 5-day long\nsequences, could support a credible analysis of transport over Tampa Bay. The conclusion was\nno, the aggregation set would not adequately capture coastal sea breeze effects. A straight\nsimulation of CMAQ at 32-, 8-, and 2-km cell sizes to create an annual average would be more\ndefensible. Retaining the 2-km resolution would also provide better simulation of the chemistry\nin the power plant plumes and the differential in productivity due to the mandated reductions in\nNOX emissions. Precipitation records for sites in and around Tampa Bay were compared against\n15-year and 40-year rainfall averages. Except for December 2002, the period of April 2002 to\nMarch 2003 would have close to average rainfall on every month except June 2002, which had\nrainfall be 50 percent higher than average. April 2002 through March 2003, excluding December\n2002, will be used for the Tampa Bay study. The meteorology for this period will be generated\nduring FY-2005 while CMAQ-UCD is being evaluated.\n62","2.4.6 Multimedia Research for CMAQ-Hg\nDuring FY-2004, the CMAQ-Hg code as described in Bullock and Brehme (2002) was\nupdated to the September 2004 release of CMAQ. After confirmation of correct implementation,\nthe code was moved to an ASMD Linux machine. Proper implementation was again confirmed.\nUpon completion, the code was shared with OAQPS. Following initial Linux implementation, a\nsingle column version of CMAQ-Hg was defined, which will be used to test proposed CMAQ\nmodifications to facilitate dynamic surface/atmosphere flux estimation within the CMAQ system.\n2.4.7 Multimedia Tool Development\nSignificant effort is often required to analyze observations and model results and provide\nthem in a form required to support management decisions. Most off-the-shelf tools do not address\nthe specialized needs or applications encountered in analyzing data from a multimedia\nperspective, making it more difficult than is necessary to link elements of the multimedia\ncomponents together. The need for specialized tools is especially pertinent to bringing\natmospheric components together with watershed components for multimedia management\nanalyses. For many air-water linkages, multi-year averaged deposition is desired. The\nAggregation method was developed and embodied in a software tool to create climatological\naverage fields of atmospheric deposition of sulfur and nitrogen across the United States to support\nair-water linkages, using outputs from a regional air quality model. This method essentially uses a\nset of air quality modeling simulations based on an archive of atmospheric transport cases to\ncreate climatological average deposition estimates. The Aggregation method/air-water linkage\nwas used extensively by the Chesapeake Bay Project for its nitrogen management planning,\ndemonstrating its usefulness. However, ASMD is the only group using the Aggregation method.\nWider use was limited because it was developed in SAS.\nThe objective of this project in Multimedia Tools Development is to make the\nAggregation method available to the wider client community through an easy to use software tool.\nTo meet the objective, the upgraded Aggregation software program was converted from SAS to R,\na widely used and supported open-source statistical package. The output was converted to the\nModels-3/CMAQ format, an easy to understand front end to the package was developed, and\ndocumentation was written. The R package is easier to use than the SAS approach, allowing less\nexperienced users to be able to carry out the aggregation methodology. The output files can be\nviewed with the Package for Analysis and Visualization of Environmental DataTM3 (PAVET), the\nModels-3 visualization tool that is publically available through CMAS, and is accessible to other\nModels-3 tools. The ability to use PAVETM is very attractive and easy compared to mapping the\n3 Copyright 1997-2000 MCNC-North Carolina Supercomputing Center, Research Triangle\nPark, NC.\n63","results using SASGraphTM4 and enhances the users ability to display results. The R Aggregation\nPackage, RAGG, documentation, and a test data set with inputs and the correct output answer are\non EPA's anonymous ftp site for download. The ftp site can be reached from the ASMD website\npage on Multimedia Modeling Tools (www.epa.gov/asmdnerl/Multimedia/software.html) under\nthe heading R Aggregation Tool or RAGG. A URL is also provided for the R website in case a\nuser needs to install R.\n2.5 Climate Change Impacts on Regional Air Quality\nThe Climate Impacts on Regional Air Quality (CIRAQ) project was initiated in FY-2002\nand will directly contribute to the EPA Global Change Research Program's (EPA GCRP)\nassessment reports of global climate change impacts on air quality. The Division's role in the\nassessment is to simulate air quality on a national domain under current and future climate change\nconditions. The planned products for this effort are designed to provide results and analysis in a\ntimely manner for the EPA GCRP 2007 air quality assessment report. Current and future (2050)\n10-year regional climate simulations were developed during FY-2003-04. During FY-2004, a\nQuality Assurance Project Plan was developed and approved for CIRAQ and baseline model-\nready meteorology and emissions files were processed. During FY-2005-06, these baseline\nscenarios will be followed by processing of model-ready future (2050) meteorology and emissions\nscenarios and CMAQ air quality simulations. The primary goal of these simulations is to develop\nfuture air quality modeling scenarios to compare against current conditions to test the sensitivity\nof air quality to potential climate change.\n2.5.1 Regional Climate Downscaling of Meteorology\nTo support this project and ultimately the air quality assessment, the EPA GCRP is\nfunding the Department of Energy's (DOE) Pacific Northwest National Laboratory (PNNL) to\ndevelop current and future regional climate simulations. These simulations rely on MM5 with\ninitial and boundary conditions from global climate model (GCM) simulations, and the future\nGCM simulations rely on Intergovernmental Panel on Climate Change future greenhouse gas\nscenarios. During FY-2003, 10-year MM5 simulations using NCEP reanalysis fields as boundary\nconditions were completed (1990-2000) and transferred to ASMD for archiving. During\nFY-2004, PNNL completed two additional 10-year MM5 simulations with boundary condition\nlinks to the National Aeronautics and Space Administration (NASA) Goddard Institute for Space\nStudies (GISS) GCM. MM5-GISS simulations were performed for a reference period (e.g., 2000\n+ 5 years) and a future period under climate change conditions (e.g., 2050 I 5 years). As regional\nclimate modeling simulations become available, ASMD regularly archives the data and develops\nthe model-ready meteorology fields needed for CMAQ and the emissions processor SMOKE®.\nSAS is a registered trademark of SAS Institute, Inc., Cary, NC, USA\n64","2.5.2 Chemical Emissions Processing\nThis effort was formally initiated during FY-2004. Sample model-ready regional climate\nscenarios were processed through SMOKE to better determine resource requirements and to\nidentify potential quality assurance issues. The base inventory used is the EPA 2001 modeling\ninventory, projected from the 1999 National Emission Inventory (NEI) version 3. Preliminary\nresults indicate successful processing and expected levels of regional emission response to\nclimate variability. Emissions processing for the remaining reference period will be completed in\nearly FY-2005, with future scenario processing to be completed as model-ready meteorology\nbecome available. The same EPA 2001 modeling inventory will be used with the future climate\nscenarios for input to the 2007 air quality assessment product. Collaboration is on-going with the\nEPA National Risk Management Research Laboratory regarding the development of emission\ninventories that could be used in future (2010) analyses products.\n2.5.3 Global Climate and Chemical Transport Simulations\nDuring FY-2003 and 2004, global climate and chemical transport simulations were\ncompleted and analyzed by collaborators at Harvard University under direct project support.\nResults from this study suggest that under potential future climate change conditions, an increase\nin stagnation events and pollution episodes would occur. These findings are particularly relevant\nfor the ASMD portion of this study because these global climate and chemical boundary\nconditions are used in CIRAQ regional downscaling simulations of MM5 and CMAQ. These\nglobal chemical transport modeling results have been archived at ASMD, and software code has\nbeen developed and tested to produce CMAQ boundary conditions from these modeling results.\nSince this series of global simulations only included ozone chemistry, the boundary condition\nestimates for aerosol species is currently limited to default estimates. A second series of global\nCTM simulations has been offered by Carnegie Mellon University (CMU) that includes aerosols.\nWhile the CMU simulations are driven by the same global climate model, full consistency\nbetween the global climate and global chemistry drivers cannot be guaranteed. During 2005,\nchemical boundary conditions from both sets of global CTM will be considered for the CMAQ\nsimulations.\n2.5.4 Methods Developed for Analysis and Evaluation of Regional Climate Simulations\nAnalysis tools were developed during FY-2004 to evaluate historical MM5 simulations\nagainst a suite of observations and to generate graphs using GrADS (Grid Analysis and Display\nSystem)5. A second tool was developed to extract meteorological data, corresponding to weather\nCopyright 1988-2004 by Brian Doty, Center for Ocean-Land-Atmosphere\nStudies(COLA), Institute of Global Environment and Society, Calverton, MD. All Rights\nReserved. Permission is granted for an individual or institution to use this software (in any\n65","observation sites, from the MM5 simulation and insert into a CIRAQ database table. Additionally,\nthe tool decodes observed meteorology and inserts it into the same database. A program was\ndeveloped to extract simulated and observed meteorological variables from the database and plot\ndistributions for comparison between model and observations. During FY-2005, these tools will\nbe combined with time series and spatial analysis methods developed during FY-2004 to analyze,\nevaluate and compare present-day downscaled climate scenarios to observations.\nSpatial analysis during FY-2004 focused on preliminary evaluation of two global\nreanalysis data sets as a method development test and as a means to estimate uncertainty in our\nestimates of present climate condition. A cluster analysis method was implemented in which\ndominant patterns of 700 mb atmospheric transport derived from u and V component wind are\nidentified from reanalysis meteorological data. Distributions of dominant pattern frequency, a\ncombination of event frequency and persistence, were developed and applied seasonally to 1800\nUTC 700 mb reanalysis data from 1985 through 1994 across the continental United States.\nResults of this analysis indicate that, although the dominant 700 mb transport patterns appear very\nsimilar (very good agreement regarding direction and +/-15 percent windspeed agreement), the\nfrequency of these patterns often occur differently within their respective data sets. During\nFY-2005, these estimates of baseline uncertainty will be compared with similar GISS/MM5\nbaseline results to construct statements regarding baseline and future model pattern similarities\nand differences as well as bias implications for the air quality simulations to be initiated during\nFY-2005.\nTime series analysis methods were tested to separate out temporal variability in the\nregional climate modeling results. With a 10-year simulation, it will be possible to consider daily,\nsynoptic, seasonal, and interannual variations. Use of time series analysis may help to identify\ndifferences between periodic variations and actual trends in climate. The same general methods\ncan also be used to analyze air quality predictions once CMAQ simulations are completed.\nA summary report regarding analysis and evaluation of the regional climate simulations\nwill be produced during FY-2005. The report comprises a description of the ASMD regional\nclimate scenario data management and quality assurance tool along with quality assurance\nsummaries. The statistical methods that have been developed for the regional climate analysis\napplication will be described and analysis results for point, time series and spatial analyses will be\npresented. Scenario evaluation for the baseline period will include comparison to surface\nobservations and global reanalysis data sets. Such comparison will be used to identify regional\nclimate model strengths, weakness and bias' that could potentially impact the air quality model\nsimulations. If future climate scenario results are available, the analysis will also include\ncomparison of baseline climate scenarios to projected future conditions. These analysis will be\nperformed on the full 20 year (10 year baseline and 10 year future) climate simulation.\nform); however, certain copyright restrictions do apply.\n66","2.6 Specialized Client Support\n2.6.1 European Monitoring and Evaluation Program\nA Division scientist serves as the United States representative to the European Monitoring\nand Evaluation Program (EMEP) that oversees the cooperative program for monitoring and\nevaluation of the long-range transmission of air pollutants in Europe. The primary goal of EMEP\nis to use regional air quality models to produce assessments evaluating the influence of one\ncountry's emissions on another country's air concentrations or deposition. The emphasis has\nshifted from acidic deposition to ozone and there is now interest in fine particulates and toxic\nchemicals. The United States and Canadian representatives report on North American activities\nrelated to long-range transport. The Division scientist also evaluates European studies of special\nrelevance to the program, providing technical critiques of the EMEP work during formal and\ninformal interactions, and develops and coordinates such programs with EMEP as the modeling\nstudies of the Modeling Synthesizing Center West at the Norwegian Meteorological Institute in\nOslo, Norway.\nIn FY-2004, the United States and Canadian representatives hosted a workshop on fine\nparticulate matter measurement and modeling to summarize the experiences of the EPA Supersite\nprogram for the benefit of the EMEP program. The workshop was held in April 2004 in New\nOrleans, Louisiana. The last EMEP workshop sponsored by Canada and the United States was\nover a decade ago in Nova Scotia. The purposes of this workshop were to exchange information\nwith European colleagues and arrive at a sense of the community regarding the current state of\nmeasurement and regional modeling for fine particulate matter in the atmosphere and recommend\nwhere we need advances in tools and techniques. Approximately 60 participants attended, 18\nfrom Europe, 15 from Canada, and 28 from the United States. A United Nations UNECE (United\nNations Economic Commission for Europe) EMEP summary report with key, high-level\nworkshop conclusions is on the EMEP website in Geneva. The workshop discussions had such a\nwealth of information and insights of use to both EPA and Environment Canada that it was\ndecided to produce a longer workshop report.\n2.6.2 Support of the 1999 National Air Toxics Assessment\nSpatial allocations were improved for aircraft emissions in support of the 1999 National\nAir Toxics Assessment. This improvement is implemented in an EMS-HAP (Emissions\nModeling System for Hazardous Pollutants) User's Guide, Version 3 (U.S. Environmental\nProtection Agency, 2004). Specifically, the improved spatial allocations for aircraft involve\nincluding itinerant operations for general aviation, commercial aircraft, military, and air taxi as a\nbasis for allocating associated county-level emissions to corresponding airport locations. This\ncapability was added to the SMOKE 2.0 emission processor. As a result of this work,\ninconsistencies in the 1999 National Emission Inventory were exposed.\n67","Data structures for NATA involving concentration, exposure, and risk model estimates\nwere synthesized for application within the EPA Data Warehouse. This synthesis allows greater\nutilization of the single data set, opens the data set to more improvement and refinement by the\nuser community, and allows rapid implementation of last-minute program changes. This data\nformat will be used as a base for the NATA Explorer, a visualization tool for all 1999 NATA\ndata.\n2.6.3 The Philadelphia Air Toxics Project\nAn extensive air toxics modeling project was conducted in Philadelphia to assist\ncommunities in reducing air toxics emissions and risk through voluntary air toxics emissions\nreduction efforts. This project modeled voluntary emission reduction efforts to address the air\ntoxics contributions from diesel powered trucks, buses, and other mobile sources, which were\nidentified through NATA as major contributors to air toxics emissions in the region.\nThe Air Toxics Project included modeling of emission sources in Philadelphia and the\nsurrounding counties in Pennsylvania, New Jersey, and Delaware. Point, area, mobile, and\nemissions from such large sources as airports and landfills were processed through the EMS-HAP\nfor input in the Industrial Source Complex dispersion model. The concentrations estimated by\nISC were then adjusted to take into account air toxics concentrations resulting from secondary\nreactions in the atmosphere and from background sources. These results are a component of\nhuman exposure modeling and risk analysis. In addition to focusing on current air toxics\nemissions and risks, this detailed modeling study includes an assessment of expected air toxics\nemissions and concentrations levels in 2010. This information will assist in determining the likely\neffect of required control efforts.\n2.6.4 Support Center for Regulatory Air Models\nSCRAM (Support Center for Regulatory Air Models) website continued to be updated to\nreflect additions of models and data sets. Some of these changes included modifications to the 7th\nModeling Conference for Air Quality Modeling (http://www.epa.gov/scram001/ttn26.htm) files\nspecifically related to the promulgation of the AERMOD (AMS EPA Regulatory MODel) system.\nModifications to AERMOD include new dry and wet deposition algorithms. The SCRAM\nwebsite is currently undergoing a structural change to provide more information to users and\nutilize a more intuitive logic flow for navigation. Each major area of SCRAM is being analyzed:\nmodels, guidance, meteorological data, etc. The Model Clearinghouse area, which provides\nscientific and technical support to the Regional Offices, will be restructured to provide full access\nand search capabilities to the general public. Responsibilities and technical support of the Model\nClearinghouse were provided on an on-going basis to the Regional Offices on a variety of issues\nrelated to PSD (Prevention of Significant Deterioration) and NSR (New Source Review).\n68","2.6.5 Eta Data Assimilation System Review\nSix- and seven-year EDAS (Eta Data Assimilation System) 700mb wind cluster data were\nanalyzed to study the effects of global change. The time series output were stratified by season (3\nmonths/season) using the SAS FASTCLUS®6 routine to generate clusters of days with the most\nsimilar spatial patterns through the entire national domain. These preliminary clusters were then\nre-analyzed to compare the EDAS (1997-2002) analysis results to those presented in Cohn et al.\n(2001) for the period of 1984-1992. The resulting array of data consisted of 1,935 observations.\nThese two output analyses were compared to see if the methodologies employed were consistent.\nHalf of the two analyses compared very well and the other half compared moderately well, while\nvery few of the analyses did not match.\n2.6.6 Development of a Response Surface Model for Ozone\nOne of the most time-consuming tasks in identifying effective strategies to improve air\nquality is running the air quality model for all possible control options. More air quality control\noptions can be considered by using a response surface model that accurately replicates the\ncomplex interactions of an individual air quality simulation. ASMD staff, with various EPA and\nnon-governmental partners, applied techniques used in meteorological and other model response\nanalyses to develop an ozone response surface based on CAMx modeling that will be used to\nestimate the effects of a series of planned mobile source emission reductions. Over 140 CAMx\nruns were analyzed and interpreted to develop the surface that considered 14 separate dimensions.\nAn additional 10 runs were analyzed for validation purposes. The \"model of the model\" was\ndetermined to accurately reproduce the results from an individual simulation, typically within 0.2\nppb depending upon the output metric.\n6\nCopyright 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.\n7\nComprehensive Air Quality Model with Extensions\n69","REFERENCES\nAnsari, A.S., and S.N. Pandis. Response of inorganic PM to precursor concentrations.\nEnvironmental Science & Technology 32:2706-2714 (1998).\nArnold, J.R., R.L. Dennis, and G.S. Tonnesen. Diagnostic evaluation of numerical air quality\nmodels with specialized ambient observations: Testing the Community Multiscale Air\nQuality (CMAQ) modeling system at selected SOS 95 ground sites. Atmospheric\nEnvironment 37:1185-1198 (2003).\nBey, I., D.J. Jacob, R.M. Yantosca, J.A. Logan, B.D. Field, A.M. Fiore, Q. Li, H.Y. Liu, L.J.\nMickley, and M.G. Schultz. Global modeling of tropospheric chemistry with assimilated\nmeteorology: Model description and evaluation. Journal of Geophysical Research-\nAtmospheres 106(D19), 23073-23096, 10.1029/2001JD000807 (2001).\nBhave, P.V., G.A. Pouliot, and M. Zheng. Source apportionment of primary carbonaceous\naerosol using the Community Multiscale Air Quality model. Proceedings of the 27th\nNATO/CCMS International Technical Meeting on Air Pollution Modeling and Its\nApplications, Banff, Canada, October 24-29, 2004. NATO/CCMS International\nTechnical Meeting of Air Pollution Modeling and Its Applications, 25-32 (2004b).\nBhave, P.V., S.J. Roselle, F.S. Binkowski, C.G. Nolte, S. Yu, G.L. Gipson, and K.L. Schere.\nCMAQ aerosol module development: recent enhancements and future plans. CMAS\nModels-3 Users' Workshop, One Atmosphere, One Community, One Modeling System:\nModels-3, October 27-29, 2003, Research Triangle Park, North Carolina. Community\nModeling and Analysis System and the UNC at Chapel Hill Carolina Environmental\nProgram, Chapel Hill, NC (2004a).\nBinkowski, F.S., and U. Shankar. The Regional Particulate Matter Model 1. Model description\nand preliminary results. Journal of Geophysical Research-Atmospheres\n100(D12):26191-26209 (1995).\nBullock, O.R. Jr., and K.A. Brehme. Atmospheric mercury simulation using the CMAQ model:\nFormulation description and analysis of wet deposition results. Atmospheric Environment\n36:2135-2146 (2003).\nBurian, S., S. Stetson. W. Han, J. Ching, and D. Byun. High resolution data set of urban canopy\nparameters for Houston, Texas. Preprints, Fifth Symposium on the Urban Environment,\nAugust 23-26, 2004, Vancouver, BC, Canada. American Meteorological Society, Boston,\nPaper No. 9.3 (2004).\nChang, S., E.C. McDonald-Buller, Y. Kimura, G. Yarwood, J. Neece, M. Russel, P. Tanaka, D.\nAllen. Atmospheric Environment 36:4991-5003 (2002).\n70","Ching, J.K.S., S. Dupont, R. Gilliam, S. Burian, and R. Tang. Neighborhood scale air quality\nmodeling in Houston using urban canopy parameters im MM5 and CMAQ with improved\ncharacterization of mesoscale lake-land breeze circulation. Preprints, Fifth Symposium on\nthe Urban Environment, August 23-26, 2004, Vancouver, BC, Canada. American\nMeteorological Society, Boston, Paper No. 9.2. (2004a).\nChing, J., S. Dupont, J.A. Herwehe, T. Otte, A. Lacser, D.W. Byun, and R. Tang. Air quality\nmodeling at coarse-to-fine scales in urban areas. Preprints, Sixth Conference on\nAtmospheric Chemistry: Air Quality in Megacities (Joint with the Symposium on\nPlanning, Nowcasting, and Forecasting in the Urban Zone), January 11-15, 2004,\nSeattle, Washington. American Meteorological Society, Boston, Paper No. J2.18 (2004).\nChing, J.K.S., T.E. Pierce, T. Palma, W.T. Hutzell, R. Tang, A. Cimorelli, and J. Herwehe.\nLinking air toxic concentrations from CMAQ to the HAPEM5 exposure model at\nneighborhood scales for the Philadelphia area. Preprints, Fifth Symposium on the Urban\nEnvironment, Vancouver, BC, Canada, August 23-26, 2004. Joint Session J4: Human\nBiometeorology: Air Quality (Joint between the 16th Conference on Biometeorology and\nAerobiology and the Fifth Symposium on the Urban Environment). American\nMeteorological Society, Boston, Paper No. J4.4 (2004).\nChing, J., S. Dupont, J. Herwehe, and R, Tang. Community scale air toxics modeling with\nCMAQ. CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One\nModeling System: Models-3, October 27-29, 2003, Research Triangle Park, North\nCarolina. Community Modeling and Analysis System and the UNC at Chapel Hill\nCarolina Environmental Program, Chapel Hill, NC, CD-ROM, Poster Session (2003).\nChing, J.K.S., and D.W. Byun. Introduction to the Models-3 framework and the Community\nMultiscale Air Quality Model (CMAQ). In Science Algorithms of the EPA Models-3\nCommunity Multiscale Air Quality (CMAQ) modeling system. Chapter 1. D.W. Byun and\nJ.K.S. Ching (Eds.). EPA-600/R-99/030, National Exposure Research Laboratory, U.S.\nEnvironmental Protection Agency, Research Triangle Park, NC, 1-17 (1999).\nCohn, R., B.K. Eder, S.K. LeDuc, and R.L. Dennis. Development of an aggregation and episode\nselection scheme to support the Models-3 Community Multiscale Air Quality Model.\nJournal of Applied Meteorology 40:210-228 (2001).\nDupont, S., T.L. Otte, and J.K.S. Ching. Simulations of meteorological fields within and above\nurban and rural canopies with a mesoscale model (MM5). Boundary-Layer Meteorology\n113:111-158 (2004).\n71","Gillani, N.V., and J.M. Godowitch. Plume-in-grid treatment of major point source emissions. In\nScience algorithms of the EPA Models-3 Community Multiscale Air Quality modeling\nsystem. Chapter 9. D.W. Byun, and J.K.S. Ching (Eds.). EPA-600/R-99/030, National\nExposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle\nPark, NC, 1-40 (1999).\nGilliland, A.B., R.L. Dennis, S.J. Roselle, and T.E. Pierce. Seasonal NH3 emission estimates for\nthe eastern United States based on ammonium wet concentrations and an inverse modeling\nmethod. Journal of Geophysical Research-Atmospheres 108:doi:10.1029/2002JD003063\n(2003).\nGoebes, M.D., R. Strader, and C. Davidson. An ammonia emission inventory for fertilizer\napplication in the United States. Atmospheric Environment 37:2539-2550 (2003).\nHansen, D.A., E.S. Edgerton, B.E. Hartsell, J.J. Jansen, N. Kandasamy, G.M. Hidy, and C.L.\nBlanchard. The southeastern aerosol research and characterization study: Part 1 -\nOverview. Journal of the Air & Waste Management 53:1460-1471 (2003).\nHertel O., R. Berkowicz, J. Christensen, and O. Hov. Test of two numerical schemes for use in\natmospheric transport-chemistry models. Atmospheric Environment 27A:2591-2611\n(1993).\nHerwehe, J.A., J.K.S. Ching, and J. Swall. Quantifying subgrid pollution variability in Eulerian\nair quality models. Preprints, 13th Conference on the Applications of Air Pollution\nMeteorology with the Air & Waste Management Association, August 23-26, 2004,\nVancouver BC, Canada. Session 7: Variability and Uncertainty. American\nMeteorological Society, Boston, Paper No. 7.5 (2004).\nHuber, A. Program in support of CFD simulations of micrometeorology and contaminant\ntransport within exterior urban building environments. Preprints, Eighth Annual George\nMason University Conference on Transport and Dispersion Modeling, July 13-15, 2004,\nFairfax, Virginia. George Mason University, Fairfax, Virginia. (2004).\nHuber, A., P. Georgopoulos, R. Gilliam, G. Stenchikov, S. Wang, B. Kelly, and H. Feingersh.\nModeling air pollution from the collapse of the World Trade Center and assessing the\npotential impacts on human exposures. EM. February 2004:25-40. (2004b).\nHuber, A., W. Tang, A. Flowe, B. Bell, K. Kuehlert, and W. Schwartz. Development and\napplications of CFD simulations in support of air quality studies involving buildings.\nPreprints, Fifth Symposium on the Urban Environment, Vancouver, BC, Canada, August\n23-26, 2004. Joint Session J2: Urban Scale Dispersion and Air Quality (Joint between\nthe Fifth Symposium on the Urban Environment and the 13th Conference on the\nApplications of Air Pollution Meteorology with the A&WMA). American Meteorological\nSociety, Boston, Paper No. J2.2 (2004a).\n72","Luecken, D.J., and W.T. Hutzell. Use of pilot study monitoring data to help evaluate an air\nquality model (CMAQ) for toxic air pollutants. Workshop on Air Toxics Data Analysis,\nJune 2-3 2004, Chicago, Illinois. Available at:\nwww.ladco.org/toxics/AT%20presentations%20pdf/CMAQ pilot%20_studies.pdf.\nMalm, W.C., B.A. Schichtel, M.L. Pitchford, L.L. Ashbaugh, and R.A. Eldred. Spatial and\nmonthly trends in speciated fine particle concentration in the United States. Journal of\nGeophysical Research-Atmospheres 109:doi:10.1029.2003JD003739 (2004).\nMartilli, A., A. Clappier, and M.W. Rotach. An urban surface exchange parameterization for\nmesoscale models. Boundary-Layer Meteorology 104:261-304 (2002).\nOtte, T.L. What's new in MCIP? CMAS Models-3 Users' Workshop, One Atmosphere, One\nCommunity, One Modeling System: Models-3, October 27-29, 2003, Research Triangle\nPark, North Carolina. Community Modeling and Analysis System and the UNC at\nChapel Hill Carolina Environmental Program, Chapel Hill, NC, CD-ROM 7.2 (2004).\nOtte, T.L., A. Lacser, S. Dupont, and J.K.S. Ching. Implementation of an urban canopy\nparameterization in a mesoscale meteorological model. Journal of Applied Meteorology\n43:1648-1665 (2004a).\nOtte, T. L., G. Pouliot, and J. E. Pleim. PREMAQ: A new pre-processor to CMAQ for air quality\nforecasting. CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One\nModeling System: Models-3, October 27-29, 2003, Research Triangle Park, North\nCarolina. Community Modeling and Analysis System and the UNC at Chapel Hill\nCarolina Environmental Program, Chapel Hill, NC, CD-ROM 7.1 (2004b).\nOtte, T.L., G. Pouliot, J.E. Pleim, J.O. Young, K.L. Schere, D.C. Wong, P.C.S. Lee, M. Tsidulko,\nJ.T. McQueen, P. Davidson, R. Mathur, H.-Y. Chuang, G. DiMego, and N.L. Seaman.\nLinking the Eta model with the Community Multiscale Air Quality (CMAQ) modeling\nsystem to build a national air quality forecasting system. Weather Forecasting (in press).\nPierce, T., W. Benjey, J. Ching, D. Gillette, A. Gilliland, S. He, M. Mebust, and G. Pouliot.\nAdvances in emissions modeling of airborne substances. Emission Inventories -\nApplying New Technologies, 12th International Emission Inventory Conference, April\n29-May 1, 2003, San Diego, California. Session 3 - Data Management. Office of Air\nQuality Planning and Standards, Research Triangle Park, NC (2003). Available at\nhttp://www.epa.gov/ttn/chief/conference/ei12/index.html#ses-9\nPinder, R.W., R. Strader, C.I. Davidson, and P.J. Adams. A temporally and spatially resolved\nammonia emission inventory for dairy COWS in the United States. Atmospheric\nEnvironment 38:3747-3756 (2004).\n73","Pleim, J.E., and J.S. Chang. A nonlocal closure model for vertical mixing in the convective\nboundary layer. Atmospheric Environment 26A:965-981 (1992).\nPleim, J.E., and A. Xiu. Development of a land surface model. Part II: Data assimilation.\nJournal of Applied Meteorology 42:1811-1822 (2003).\nSandu, A., J.G. Verwer, J.G. Blom, E.J. Spee, G.R. Carmichael, and F.A. Potra. Benchmarking\nstiff ODE solvers for atmospheric chemistry problems II: Rosenbrock solvers.\nAtmospheric Environment 31:3459-3472 (1997).\nTaha, H. Modifying a mesoscale model to better incorporate urban heat storage: A bulk\nparameterization approach. Journal of Applied Meteorology 38:466-473 (1999).\nTanaka, P.L., D.T. Allen, E.C. McDonald-Buller, S. Chang, Y. Kimura, C.B. Mullins, G.\nYarwood, and J.D. Neece. Development of a chlorine mechanism for use in the carbon\nbond IV chemistry model. Journal of Geophysical Research-Atmospheres, 108 (D4),\n4145, loi:10.1029/2002JD002432 (2003).\nTonnesen, G.S., and R.L. Dennis. Analysis of radical propagation efficiency to assess ozone\nsensitivity to hydrocarbons and NO. Part 1: Local indicators of instantaneous odd oxygen\nproduction sensitivity. Journal of Geophysical Research-Atmospheres 105:9213-9225\n(2000a).\nTonnesen, G.S., and R.L. Dennis. Analysis of radical propagation efficiency to assess ozone\nsensitivity to hydrocarbons and NO. Part 2: Long-lived species as indicators of ozone\nconcentration sensitivity. Journal of Geophysical Research-Atmospheres 105:9227-9241\n(2000b).\nU.S. Environmental Protection Agency. User's guide for the Emissions Modeling System for\nHazardous Air Pollutants (EMS-HAP, Version 3.0). EPA-454/B-03-006, Office of Air\nQuality Planning and Standards, Research Triangle Park, NC (2004).\nWhitby, E.R., P.H. McMurry, U. Shankar, and F.S. Binkowski. Modal aerosol dynamics\nmodeling. EPA/600/3-91/020, Atmospheric Research and Exposure Assessment\nLaboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC (1991).\nWilliams, D.J., and J.K.S. Ching. A federated partnership for urban meteorological and air\nquality modeling. Preprints, Fifth Symposium on the Urban Environment, August 23-26,\n2004, Vancouver, BC, Canada. American Meteorological Society, Boston, Paper No. 9.6\n(2004).\nWu\nY.,\nB. Brashers, P.L. Finkelstein, and J.E. Pleim. A multilayer biochemical dry deposition\nmodel, 1, Model formulation. Journal of Geophysical Research-Atmospheres, 108(D1),\n4013, doi: 10.1029/2002JD002293 (2003a).\n74","Wu Y., B. Brashers, P.L. Finkelstein, and J.E. Pleim. A multilayer biochemical dry deposition\nmodel, 2, Model evaluation. Journal of Geophysical Research-Atmospheres, 108(D1),\n4014, doi:10.1029/2002JD002306 (2003b).\nYu, S., R.L. Dennis, P.V. Bhave, and B.K. Eder. Primary and secondary organic aerosols over the\nUnited States: Estimates on the basis of observed organic carbon (OC) and elemental\ncarbon (EC), and air quality modeled primary OC/EC ratios. Atmospheric Environment\n38:5257-5268 (2004).\nZheng, M., G.R. Cass, J.J. Schauer, and E.S. Edgerton. Source apportionment of PM2.5 in the\nsoutheastern United States using solvent-extractable organic compounds as tracers.\nEnvironmental Science & Technology 36:2361-2371 (2002).\n75","APPENDIX A: ACRONYMS, ABBREVIATIONS, AND DEFINITIONS\nACM\nAsymmetric Convective Model\nACM2\nAsymmetric Convective Model and eddy diffusion\nAE3\nAerosols algorithm version 3\nAERMOD\nAir dispersion model\nAIM\nAerosol Inorganic Model\nAQ\nAir quality models\nAQF\nAir quality forecast\nARL\nAir Resources Laboratory\nASMD\nAtmospheric Sciences Modeling Division\nASPEN\nGaussian plume model\nAT\nAir toxins\nAVIRIS\nAirborne Visible/Infrared Imaging Spectrometer\nBEIS3.12\nBiogenic Emissions Inventory System version 3.12\nBRACE\nBay Regional Atmospheric Chemistry Experiment\nCAA\nClean Air Act\nCAAA\nClean Air Act Amendments\nCAMx\nComprehensive Air Quality Model with extensions\nCB-IV\nCarbon-Bond-IV\nCASTNet\nClean Air Status and Trend Network\nCDFware\nConcentration Distribution Functionware\nCFA\nSOS Cornelia Fort Airpark\nCFD\nComputational fluid dynamics\nCIRAQ\nClimate Impact on Regional Air Quality\nCLEANER\nCollaborative Large-Scale Engineering Analysis Network for\nEnvironmental Research\nCMAQ\nCommunity Multiscale Air Quality modeling system\nCMAQ-AT\nAir toxins version of CMAQ\nCMAQ CTM\nCommunity Multiscale Air Quality modeling system chemistry-\ntransport model\nCMAQ-Hg\nCommunity Multiscale Air Quality - Mercury model\nCMAQ PinG\nCommunity Multiscale Air Quality modeling system Plume in Grid\nmod\nCMAQ-UCD\nThe Wexler sectional aerosol model, Aerosol Inorganic Model\n(AIM), was adapted to incorporate sea salt in its calculations,\nimplemented into the September 2004 release CMAQ, and named\nCMAQ-UCD.\nCMAS\nCommunity Modeling and Analysis System plume-in-grid model\nCTM\nChemistry-Transport Model\nCMU\nCarnegie Mellon University\n76","A multi-layer canopy and soil model with few layers of a couple\nDA-SM2-U\nmeters within the canopy depending on the mesh of the mesoscale\nmodel domain, and three layers within the ground; a surface soil\nlayer for the natural surfaces, a root zone layer, and a deep soil layer\nDOE\nDepartment of Energy\nEuler Backward Iterative\nEBI\nEC\nElemental Carbon\nEDAS\nEta Data Assimilation System\nEMD\nEmpirical Mode Decomposition\nEuropean Monitoring and Evaluation Program\nEMEP\nEmissions Modeling System for Hazardous Pollutants\nEMS-HAP\nEOF\nEmpirical Orthogonal Functions\nEPA\nEnvironmental Protection Agency\nEPA GCRP\nEPA Global Change Research Program\nNational Center for Environmental Prediction Mesoscale Model\nEta\nEtaDry\ndry deposition routine\nRegional Acid Deposition Model with full dynamics of secondary\nExtended RADM\ninorganic fine particle formation taken from the RPM\nGCM\nGlobal Climate Models\nGlobal Change Research Program\nGCRP\nA global three-dimensional model of atmospheric composition\nGEOS-CHEM\ndriven by assimilated meteorological observations from the\nGoddard Earth Observing System\nGFS\nGlobal Forecast System\nGIS\nGeographic Information System\nGISS\nGoddard Institute for Space Studies\nGR\nGas Ratio\nGrADS\nGrid Analysis and Display System\nHAPEM5\nHazardous Air Pollutant Exposure Model version 5\nICARTT\nInternational Consortium for Atmospheric Research on Transport\nand Transformation\nIC/BC\nInitial Condition/Boundary Condition\nIMPROVE\nInteragency Monitoring of PROtected Visual Environment Network\nISORROPIA\nThermodynamics module\nJST\nSOS Jefferson Street\nMinimum value of the surface layer vertical-eddy diffusivity\nKz\nLAI\nLeaf Area Index\nLateral boundary conditions\nLBC\nModels-3 Dry Deposition Scheme\nM3Dry\nMultiscale Air Quality Simulation Platform\nMAQSIP\nMCIP\nMeteorology-Chemistry Interface Processor\n77","MCIP2.2\nMeteorology-Chemistry Interface Processor version 2\nMDN\nMercury Deposition Network\nMIMS\nMultimedia Integrated Modeling System\nMLBC\nMultiLayer Biochemical Model\nMLBC-AW\nMultiLayer Biochemical Model-Area Weighted\nMLBC_PG\nMultiLayer Biochemical Model- - Calculator (EPIC) model was\nembedded in MLBC and denoted as MLBC-PG\nMM5\nMesoscale Model - version 5\nMobile6\nMobile Source Emission\nNAAQS\nNational Ambient Air Quality Standards\nNASA\nNational Aeronautics and Space Administration\nNATA\nNational Air Toxics Assessment\nNATA Explorer\nA visualization tool for all 1999 NATA data\nNCAR\nNational Center for Atmospheric Research\nNCEP\nNational Centers for Environmental Prediction\nNEI\nNational Emission Inventory\nNMB\nNormalized Mean Biases\nNME\nNormalized Mean Errors\nNMSE\nNormalized Mean Square Error\nNOAA\nNational Oceanic and Atmospheric Administration\nNSR\nNew Source Review\nNWS\nNational Weather Service\nNYSDEC\nNew York State Department of Environmental Conservation\nOAQPS\nOffice of Air Quality Planning and Standards\nOC\nOrganic Carbon\nOCpri\nPrimary Organic Carbon\nOCsec\nSecondary Organic Carbon\nODE\nOrdinary Differential Equation\nPAVE\nPackage for Analysis and Visualization of Environmental Data\nPBL\nPlanetary Boundary Layer\nPDFs\nProbability Density Functions\nPDM\nPlume Dynamics Model\nPinG\nPlume-in-Grid\nPinG Module\nPlume-in-Grid Model\nPM\nParticulate Matter\nPNNL\nPacific Northwest National Laboratory\nppbv\nParts per billion by volume\nPREMAQ\nPre-processor for CMAQ\nPSD\nPrevention of Significant Deterioration\nPSU\nPennsylvania State University\nPX LSM\nPleim Xiu Land-Surface Model\n78","RADM2\nRegional Acid Deposition Model version 2\nRANS\nReynolds-Averaged Navier-Stokes\nRegional Modeling System for Aerosols and Deposition\nREMSAD\nRMET\nThe R Model Evaluation Toolkit,\nRMSEs\nRoot mean squared errors\nROS3\nA particular version of the Rosenbrock class of solvers\nRSL\nRoughness sub-layer\nA gas-phase chemical mechanism (Statewide Air Policy Research\nSAPRC99\nCenter)\nSCRAM\nSupport Center for Regulatory Air Models\nSGV\nSubgrid variability\nSIP\nState Implementation Plan\nSparse Matrix Operator Kernel Emission model\nSMOKE®\nSIP\nState Implementation Plan\nRepresents the thermodynamic effects (e.g.,estimates the heat and\nSM2-U(3D)\nhumidity fluxes) of the canopy elements at different levels within\nthe canopy\nSOA\nSecondary Organic Aerosol\nSOS\nSouthern Oxidants Study\nSST\nSea Surface Temperature\nTEOM\nTapered element oscillating microbalance\nTMDL\nTotal Maximum Daily Loan\nUCPs\nUrban Canopy Parameterizations\nUDP\nDepartment of Homeland Security's New York City Urban\nDispersion Program\nUNECE-EMEP\nUnited Nations Economic Commission for Europe-European\nMonitoring and Evaluation Program\nUTC\nUniversal Time Coordinate\nVOC\nVolatile Organic Compounds\nWF\nWavelet Filters\nWRF\nWeather Research and Forecast\nWTC\nWorld Trade Center\n79","APPENDIX B: PUBLICATIONS\nAlapaty, K., R. Mathur, and S. Arunachalam. Preliminary results on the development of a\nvariable-grid-resolution air quality model. CMAS Models-3 Users' Workshop, One\nAtmosphere, One Community, One Modeling System: Models-3, October 27-29, 2003,\nResearch Triangle Park, North Carolina. Community Modeling and Analysis System and\nthe UNC at Chapel Hill Carolina Environmental Program, Chapel Hill, NC, CD-ROM,\nSession 4 (2003).\nBenjey, W.G., and T.E. Pierce. An approach to an unified process-based regional emission flux\nmodeling platform. 13th Emissions Inventory Conference \"Working for Clean Air in\nClearwater\", Clearwater, Florida, June 7-10, 2004. Available at\nhttp://www.epa.gov/ttn/chief/conference/eil3/index.html\nBhave, P.V., S.J. Roselle, F.S. Binkowski, C.G. Nolte, S. Yu, G.L. Gipson, and K.L. Schere.\nCMAQ aerosol module development: recent enhancements and future plans. CMAS\nModels-3 Users' Workshop, One Atmosphere, One Community, One Modeling System:\nModels-3, October 27-29, 2003, Research Triangle Park, North Carolina. Community\nModeling and Analysis System and the UNC at Chapel Hill Carolina Environmental\nProgram, Chapel Hill, NC (2004).\nBowker, G.E., S.G. Perry, and D.K. Heist. A comparison of airflow patterns from the QUIC\nmodel and an atmospheric wind tunnel for a two-dimensional building array and a multi-\ncity block region near the World Trade Center site. Preprints, Fifth Symposium on the\nUrban Environment, Vancouver, BC, Canada, August 23-26, 2004. Joint Session J5: Fast\nResponse Urban Dispersion Models (Joint between the Fifth Symposium on the Urban\nEnvironment and the 13th Conference on the Applications of Air Pollution Meteorology\nwith the A&WMA). American Meteorological Society, Boston, Paper No. J5.4 (2004).\nBurian, S.J., M.J. Brown, J.K.S. Ching, M.L. Cheuk, M. Yuan. W.S. Han, and A.T. McKinnon.\nUrban morphological analysis for mesoscale meteorological and dispersion modeling\napplications: current issues. Preprints, Fifth Symposium on the Urban Environment,\nVancouver, BC, Canada, August 23-26, 2004. American Meteorological Society, Boston,\nPaper No. 9.1 (2004).\nBurian, S.J., S.R. Stetson, W.S. Han, J.K.S. Ching, and D.W. Byun. High-resolution data set of\nurban canopy parameters for Houston, Texas. Preprints, Fifth Symposium on the Urban\nEnvironment, Vancouver, BC, Canada, August 23-26, 2004. American Meteorological\nSociety, Boston, Paper No. 9.3 (2004).\n80","Ching, J., S. Dupont, J.A. Herwehe, T. Otte, A. Lacser, D.W. Byun, and R. Tang. Air quality\nmodeling at coarse-to-fine scales in urban areas. Preprints, Sixth Conference on\nAtmospheric Chemistry: Air Quality in Megacities (Joint with the Symposium on\nPlanning, Nowcasting, and Forecasting in the Urban Zone), January 11-15, 2004,\nSeattle, Washington. American Meteorological Society, Boston, Paper No. J2.18 (2004).\nChing, J., S. Dupont, J. Herwehe, and R. Tang. Community scale air toxics modeling with\nCMAQ. CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One\nModeling System: Models-3, October 27-29, 2003, Research Triangle Park, North\nCarolina. Community Modeling and Analysis System and the UNC at Chapel Hill\nCarolina Environmental Program, Chapel Hill, NC, CD-ROM, Poster Session (2003).\nChing, J.K.S., T.E. Pierce, T. Palma, W.T. Hutzell, R. Tang, A. Cimorelli, and J. Herwehe.\nLinking air toxic concentrations from CMAQ to the HAPEM5 exposure model at\nneighborhood scales for the Philadelphia area. Preprints, Fifth Symposium on the Urban\nEnvironment, Vancouver, BC, Canada, August 23-26, 2004. Joint Session J4: Human\nBiometeorology: Air Quality (Joint between the 16th Conference on Biometeorology and\nAerobiology and the Fifth Symposium on the Urban Environment). American\nMeteorological Society, Boston, Paper No. J4.4 (2004).\nChing, J.K.S., S. Dupont, R. Gilliam, S. Burian, and R. Tang. Neighborhood scale air quality\nmodeling in Houston using urban canopy parameters in MM5 and CMAQ with improved\ncharacterization of mesoscale lake-land breeze circulation. Preprints, Fifth Symposium on\nthe Urban Environment, Vancouver, BC, Canada, August 23-26, 2004. American\nMeteorological Society, Boston, Paper No. 9.2 (2004).\nDavidson, P.M., N. Seaman, K. Schere, C. Wayland, and K. Carey. National air quality\nforecasting capability: First steps toward implementation. CMAS Models-3 Users\nWorkshop, One Atmosphere, One Community, One Modeling System: Models-3, October\n27-29, 2003, Research Triangle Park, North Carolina. Community Modeling and\nAnalysis System and the UNC at Chapel Hill Carolina Environmental Program, Chapel\nHill, NC, CD-ROM, Session 4 (2003).\nDupont, S., J. Ching, and S. Burian. Introduction of urban canopy parameterizations into MM5 to\nsimulate urban meteorology at neighorhood scales. Symposium on Planning, Nowcasting,\nand Forecasting in the Urban Zone, January 11-15, 2004, Seattle, Washington. American\nMeteorological Society, Boston, Paper No. 4.4 (2004).\nDupont, S., T.L. Otte, and J.K.S. Ching. Simulation of meteorological fields within and above\nurban and rural canopies with a mesoscale model (MM5). Boundary-Layer Meteorology\n113:111-158 (2004).\n81","Eder, B.K., D. Kang, and R.C. Gilliam. An evaluation of the Eta-CMAQ air quality forecast\nmodel as part of NOAA's national program. CMAS Models-3 Users' Workshop, One\nAtmosphere, One Community, One Modeling System: Models-3, October 27-29, 2003,\nResearch Triangle Park, North Carolina. Community Modeling and Analysis System and\nthe UNC at Chapel Hill Carolina Environmental Program, Chapel Hill, NC, CD-ROM,\nSession 3 (2003).\nEder, B.K., and S. Yu. An evaluation of the 2003 release of Models-3 CMAQ. CMAS Models-3\nUsers' Workshop, One Atmosphere, One Community, One Modeling System: Models-3,\nOctober 27-29, 2003, Research Triangle Park, North Carolina. Community Modeling\nand Analysis System and the UNC at Chapel Hill Carolina Environmental Program,\nChapel Hill, NC, CD-ROM, Session 3 (2003).\nEtyemezian, V., S. Ahonen, D. Nickolic, J. Gillies, H. Kuhns, D. Gillette, and J. Veranth.\nDeposition and removal of fugitive dust in the arid Southwest: Measurements and model\nresults. Journal of the Air & Waste Management Association 54:1099-1111 (2004).\nFinkelstein, P.L., A.W. Davison, H.S. Neufeld, T.P. Meyers, and A.H. Chappelka. Sub-canopy\ndeposition of ozone in a stand of cutleaf coneflower. Environmental Pollution\n131:295-303 (2004).\nGillette, D.A., and A.M. Pitchford. Sand flux in the northern Chihuahuan desert, New Mexico,\nUSA, and the influence of mesquite-dominated landscapes. Journal of Geophysical\nResearch-Earth Surfaces, 109, F04003, doi: 10. .1029/2003JF000031 (2004).\nGillette, D.A., R.E. Lawson, Jr., and R.S. Thompson. A \"test of concept\" comparison of\naerodynamic and mechanical resuspension mechanisms for particles deposited on field rye\ngrass (Secale cercele). Part 1. Relative particle flux rates. Atmospheric Environment\n38:4789-4797 (2004).\nGillette, D.A., R.E. Lawson, Jr., and R.S. Thompson. A \"test of concept\" comparison of\naerodynamic and mechanical resuspension mechanisms for particles deposited on field rye\ngrass (Secale cercele). Part 2. Threshold mechanical energies for resuspension particle\nfluxes. Atmospheric Environment 38:4799-4803 (2004).\nGillette, D.A., D. Ono, and K. Richmond. A combined modeling and measurement technique for\nestimating windblown dust emissions at Owens (dry) Lake, California. Journal of\nGeophysical Research-Earth Surface 109, F01003, doi: 10.1029/2003JF000025 (2004).\n82","Heist, D.K. S.G. Perry, and G.E. Bowker. Evidence of enhanced vertical dispersion in the wakes\nof tall buildings in wind tunnel simulations of lower Manhattan. Preprints, Fifth\nSymposium on the Urban Environment, Vancouver, BC, Canada, August 23-26, 2004.\nJoint Session 7: Special Session in Honor of Dr. E. Plate. American Meteorological\nSociety, Boston, Paper No. 7.5 (2004).\nHerwehe, J.A., J.K.S. Ching, and J. Swall. Quantifying subgrid pollution variability in Eulerian\nair quality models. Preprints, 13th Conference on the Applications of Air Pollution\nMeteorology with the Air & Waste Management Association, Vancouver, BC, Canada,\nAugust 23-26, 2004. Session 7: Variability and Uncertainty. American Meteorological\nSociety, Boston, Paper No. 7.5 (2004).\nHicks, B., and J.S. Irwin. Atmospheric transport and diffusion modeling systems for effective\nemergency response. EM February:41-42 (2004).\nHuber, A.H., P. Georgopoulos, R.C. Gilliam, G. Stenchikov, S. Wang, B. Kelly, and H.\nFeingersh. Modeling air pollution from the collapse of the World Trade Center and\nassessing the potential impacts on human exposures. EM February:35-40 (2004).\nHuber, A., W. Tang, A. Flowe, B. Bell, K. Kuehlert, and W. Schwartz. Development and\napplications of CFD simulations in support of air quality studies involving buildings.\nPreprints, Fifth Symposium on the Urban Environment, Vancouver, BC, Canada, August\n23-26, 2004. Joint Session J2: Urban Scale Dispersion and Air Quality (Joint between\nthe Fifth Symposium on the Urban Environment and the 13th Conference on the\nApplications of Air Pollution Meteorology with the A&WMA). American Meteorological\nSociety, Boston, Paper No. J2.2 (2004).\nIrwin, J.S., E. Gego, C. Hogrefe, J.M. Jones, and S.T. Rao. Comparison of sulfate concentrations\nsimulated by two regional-scale models with measurements from the IMPROVE network.\n9th International Conference on Harmonization within Atmospheric Dispersion Modelling\nfor Regulatory Purposes, June 1-4, 2004, Garmisch Parkenkirchen, Germany. Institut für\nMeteorologie und Klimaforschung (IMK-IFU), Germany (2004). Available on-line\nhttp://www.harmo.org/conferences/Proceedings/_Garmisch/Garmisch_proceedings.asp\nIrwin, J.S., and S.R. Hanna. Characterizing uncertainty in plume dispersion models. 9th\nInternational Conference on Harmonization within Atmospheric Dispersion Modelling for\nRegulatory Purposes, June 1-4, 2004, Garmisch Parkenkirchen, Germany. Institut für\nMeteorologie und Klimaforschung (IMK-IFU), Germany (2004). Available on-line at\nhttp://www.harmo.org/conferences/Proceedings/_Garmisch/Garmisch_proceedings.asp\n83","Kinnee, E.J., J.S. Touma, R. Mason, J. Thurman, A. Beidler, C. Bailey, and R. Cook. Allocation\nof onroad mobile emissions to road segments for air toxics modeling in an urban area.\nTransportation Research Part D 9:139-150 (2004).\nLee, S.-M., H.J.S. Fernando, D.W. Byun, and J.K.S. Ching. CFD modeling of fine scale flow and\ntransport in the Houston Metropolitan Area, Texas. Preprints, Fifth Symposium on the\nUrban Environment, Vancouver, BC, Canada, August 23-26, 2004. American\nMeteorological Society, Boston, Paper No. 9.12 (2004).\nLintner, B.R., A.B. Gilliland, and I.Y. Fung. Mechanisms of convection-induced modulation of\npassive tracer interhemispheric transport Interannual variability. Journal of Geophysical\nResearch 109, D13102, doi: 10.1029/2003JD004306 (2004).\nOkin, G.S., and D.A. Gillette. Modeling wind erosion and dust emission on vegetated surfaces.\nIn Spatial Modeling of the Terrestrial Environment. R.E. Kelly, N.A. Drake, and S.L.\nBarr, (Eds.). John Wiley & Sons Ltd., Chichester, West Sussex, England, 127-156\n(2004).\nOtte, T. L., G. Pouliot, and J. E. Pleim. PREMAQ: A new pre-processor to CMAQ for air quality\nforecasting. CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One\nModeling System: Models-3, October 27-29, 2003, Research Triangle Park, North\nCarolina. Community Modeling and Analysis System and the UNC at Chapel Hill\nCarolina Environmental Program, Chapel Hill, NC, CD-ROM 7.1 (2004b).\nPerry, S.G., D.K. Heist, R.S. Thompson, W.H. Snyder, and R.E. Lawson, Jr. Wind tunnel\nsimulation of flow and pollutant dispersal around the World Trade Center site. EM\nFebruary:31-34 (2004).\nPleim, J., G. Gipson, S. Roselle, and J. Young. New features of the 2003 release of the CMAQ\nmodel. CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One\nModeling System: Models-3, October 27-29, 2003, Research Triangle Park, North\nCarolina. Community Modeling and Analysis System and the UNC at Chapel Hill\nCarolina Environmental Program, Chapel Hill, NC, CD-ROM, Session 1 (2003).\nPoole-Kober, E.M., and H.J. Viebrock (Eds.). Fiscal year 2003 summary report of the NOAA\nAtmospheric Sciences Modeling Division to the U.S. Environmental Protection Agency.\nNOAA Technical Memorandum ERL ARL-252, Atmospheric Sciences Modeling\nDivision, Research Triangle Park, NC, 100 pp. (2004).\nRao, S.T. Homeland security: Managing the risks. EM February: 12 (2004).\n84","Roy, B., J.B. Halverson, and J. Wang. The influence of radiosonde \"age\" on TRMM field\nsoundings humidity correction. Journal of Atmospheric and Oceanic Technology\n21:470-480 (2004).\nStreicher, J.J., W.C. Culverhouse, Jr., M.S. Dulbert, and R.J. Fornaro. Modeling the anatomical\ndistribution of sunlight. Photochemistry and Photobiology 79(1):40-47 (2004).\nVette, A., S.G. Perry, D.K. Heist, A.H. Huber, M. Lorber, P. Lioy, P. Georgopoulos, S.T. Rao,\nW.B. Petersen, B. Hicks, J.S. Irwin, and G. Foley. Environmental research in response to\n9/11 and homeland security. EM February: 14-22 (2004).\nVette, A., R. Seila, E. Swartz, J. Pleil, L. Webb, M. Landis, A. Huber, and D. Vallero. Air\npollution measurements in the vicinity of the World Trade Center: Summary of\nmeasurements conducted by EPA-ORD. EM February:23-26 (2004).\nWilliam, D.J., and J.K.S. Ching. A federated partnership for urban meteorological and air quality\nmodeling. Preprints, Fifth Symposium on the Urban Environment, Vancouver, BC,\nCanada, August 23-26, 2004. American Meteorological Society, Boston, Paper No. 9.6\n(2004).\nYoung, J.O. and D.C. Wong. Recent developments for parallel CMAQ. CMAS Models-3 Users\nWorkshop, One Atmosphere, One Community, One Modeling System: Models-3, October\n27-29, 2003, Research Triangle Park, North Carolina. Community Modeling and\nAnalysis System and the UNC at Chapel Hill Carolina Environmental Program, Chapel\nHill, NC, CD-ROM, Session 4 (2003).\nYu, S., R.L. Dennis, P.V. Bhave, and B.K. Eder. Primary and secondary organic aerosols over the\nUnited States: Estimates on the basis of observed organic carbon (OC) and elemental\ncarbon (EC), and air quality modeled primary OC/EC rations. Atmospheric Environment\n38:5257-5268 (2004).\nYu, S., R.L. Dennis, P.V. Bhave, and B.K. Eder. Primary and secondary organic aerosols over the\nUnited States: Estimates on the basis of observed organic carbon (OC) and elemental\ncarbon (EC), and air quality modeled primary OC/EC ratios. CMAS Models-3 Users ,\nWorkshop, One Atmosphere, One Community, One Modeling System: Models-3, October\n27-29, 2003, Research Triangle Park, North Carolina. Community Modeling and\nAnalysis System and the UNC at Chapel Hill Carolina Environmental Program, Chapel\nHill, NC, CD-ROM, (2003).\n85","Yu, S., R.L. Dennis, B.K. Eder, S.J. Roselle, A. Nenes, and J. Walker. Can the thermodynamic\nmodel and 3-D air quality model predict the aerosol NO3 reasonably? CMAS Models-3\nUsers' Workshop, One Atmosphere, One Community, One Modeling System: Models-3,\nOctober 27-29, 2003, Research Triangle Park, North Carolina. Community Modeling\nand Analysis System and the UNC at Chapel Hill Carolina Environmental Program,\nChapel Hill, NC, CD-ROM, (2003).\nYu, S., B.K. Eder, R.L. Dennis, S. Chu, and S. Schwartz. New unbiased symmetric metrics for\nevaluation of the air quality model. CMAS Models-3 Users' Workshop, One Atmosphere,\nOne Community, One Modeling System: Models-3, October 27-29, 2003, Research\nTriangle Park, North Carolina. Community Modeling and Analysis System and the UNC\nat Chapel Hill Carolina Environmental Program, Chapel Hill, NC, CD-ROM, (2003).\n86","APPENDIX C: PRESENTATIONS\nBenjey, W.G. An approach for a processed-based unified emission flux based modeling platform.\nPresentation at the 13th International Emission Inventory Conference, \"Working for Clean\nAir in Clearwater,\" Clearwater, FL, June 10, 2004.\nBhave, P.V. Postprocessing of model output for comparison to ambient data. Invited\npresentation at the Particulate Matter Model Performance Evaluation Workshop, Chapel\nHill, NC, February 10, 2004.\nBhave, P.V. Measurement needs for evaluating model calculations of carbonaceous aerosol.\nPresentation at the European Monitoring and Evaluation Program (EMEP) Workshop,\nParticulate Matter Measurement & Modeling, New Orleans, LA, April 22, 2004.\nBowker, G.E. Wind tunnel projects in EPA's Fluid Modeling Facility. Presentation at the EPA\nScience Forum 2004: Healthy Communities and Ecosystems, Washington, DC, June 2,\n2004.\nBullock, O.R., Jr. Modeling atmospheric mercury deposition to the sounds and other water\nbodies. Presentation at the Mercury and CO2 Workshop, sponsored by the North Carolina\nDepartment of Environment and Natural Resources, Raleigh, NC, April 19, 2004.\nBullock, O.R., Jr. Modeling transport and transformation of mercury and its compounds in\ncontinental air masses. Presentation at the International Workshop on Harmonization of\nMercury Measurements, Methods and Models to Assess Source-Receptor Impact on Air\nQuality and Human Health, Maratea, Italy, May 24, 2004.\nBullock, O.R., Jr. Aqueous reduction of Hg2+ to Hg° by HO2 in the CMAQ-Hg model.\nPresentation at the 7th International Conference on Mercury as a Global Pollutant,\nLjubljana, Slovenia, July 1, 2004.\nBullock, O.R., Jr. Atmospheric chemistry and the relative importance of mercury sources.\nPresentation at the 2004 Mercury Workshop sponsored by the U.S. Geological Survey -\nEastern Region, Reston, VA, August 17, 2004.\nChing, J.K.S. Community modeling with CMAQ. Presentation at the 2003 Models-3 Users'\nWorkshop, One Atmosphere, One Modeling System: Models-3, Research Triangle Park,\nNC, October 27, 2004.\nCooter, E.J. Detecting effect of climate/downscaling issues (scenario development and testing).\nPresentation at the U.S. EPA Consequences of Global Change for Air Quality STAR\nGrant Progress Review Workshop, Research Triangle Park, NC, May 24, 2004.\n87","Dennis, R.L. Air quality modeling of ammonia: A regional modeling perspective. Presentation at\nthe NADP Ammonia Workshop, Washington, DC, October 24, 2003.\nDennis, R.L. Time-resolved & in-depth evaluation of PM and PM precursors using CMAQ.\nPresentation at the Particulate Matter Model Performance Workshop, Chapel Hill, NC,\nFebruary 10, 2004.\nDennis, R.L. In-depth/diagnostic model evaluation. Presentation at the EPA-NOAA Air Quality\nScientist-to-Scientist Meeting, Research Triangle Park, NC, March 2, 2004.\nDennis, R.L. Evaluation of 3-D regional particulate models: measurement needs for inorganic\nspecies. Presentation at the US-Canada EMEP Workshop on Monitoring and Modeling,\nNew Orleans, LA, April 22, 2004.\nDupont, S. Application of the urbanized version of MM5 for Houston. Presentation at the 2003\nModels-3 Users Workshop, Research Triangle Park, NC, October 27, 2003.\nEder, B.K. An evaluation of the 2003 release of CMAQ. Presentation at the 2003 Models-3\nUsers' Workshop, One Atmosphere, One Modeling System: Models-3, Research Triangle\nPark, NC, October 28, 2004.\nEder, B.K. An evaluation of the Eta-CMAQ air quality forecast model as part of NOAA's\nnational program. Presentation at the 2003 Models-3 Users' Workshop, One Atmosphere,\nOne Modeling System: Models-3, Research Triangle Park, NC, October 29, 2004.\nEder, B.K. An evaluation of the Eta-CMAQ air quality forecast model. Presentation at the EPA\n2004 National Air Quality Conference: Your Forecast to Breathe By, Baltimore, MD,\nFebruary 24, 2004.\nEder, B.K. An annual evaluation of Models-3 CMAQ using a 2001 simulation. Presentation at\nthe US-Canada EMEP Workshop on Particulate Matter Measurement and Modeling, New\nOrleans, LA, April 21, 2004.\nEder, B.K. An operational evaluation of the Eta-CMAQ air quality forecast model for the\nsummer of 2004. Presentation to NWS at Silver Spring, MD, September 8, 2004.\nFinkelstein, P.L. Recent research at Purchase Knob, NC. Presentation at the National Park\nService Meeting on Sub-Canopy Research, Great Smoky Mountains National Park, March\n25, 2004.\nFinkelstein, P.L. CMAQ evaluation for acid rain. Presentation at the NADP annual meeting,\nHalifax, NS, Canada, September 22, 2004.\n88","Gillette, D.A. Wind characteristics during dust storms at Jornada del Muerto. Seminar given to\nthe Department of Range Science, New Mexico State University, Las Cruces, NM, April\n10, 2004.\nGilliland, A.B. The Climate Impacts on Air Quality (CIRAQ) project. Presentation to the\nCommittee on Environment and Natural Resources (CENR) Air Quality Subcommittee via\nteleconference, July 17, 2004.\nHuber, A.H. Development of CFD simulation applications for local-scale urban areas and\npotential interface with mesoscale models. Invited presentation at the American\nMeteorological Society Workshop on Merging Mesoscale and Computational Fluid\nDynamics Modeling Capabilities, Seattle, WA, January 11, 2004.\nHuber, A.H. Development, application and evaluation of CFD simulations for local-scale\npollutant dispersion. Presentation at the EPA-NOAA Air Quality Scientist-to-Scientist\nMeeting, Research Triangle Park, NC, March 2, 2004.\nHuber, A.H. Program in support of CFD simulations of micrometeorology and contaminant\ntransport within exterior urban building environments. Presentation at the Eighth Annual\nGeorge Mason University Conference on Transport and Dispersion Modeling, Fairfax,\nVA, July 13, 2004.\nHuber, A.H. CFD simulations within complex urban building environments. Presentation at the\nFifth Symposium on the Urban Environment, British Columbia, Canada, August 23, 2004.\nHutzell, W.T. A biogenic role in exposure to two toxic compounds. Poster presentation at the\nFall Meeting of the American Geophysical Union, San Francisco, CA, December 8, 2003.\nLuecken, D.J. Use of pilot study monitoring data to help evaluate an air quality model (CMAQ)\nfor toxic air pollutants. Presentation at the Workshop on Air Toxics Data Analysis,\nChicago, IL, June 3, 2004.\nMathur, R. The Community Multiscale Air Quality (CMAQ) model: Model configuration and\nenhancements for air quality forecasting. Presentation at the 2004 Air Quality Forecasting\nFocus Group Meeting, Silver Spring, MD, September 8, 2004.\nPierce, T.E. The importance of lightning NO for regional air quality modeling. Presentation to\nthe Joint Action Group for Lightning Detection System, Office of the Federal Coordinator\nfor Meteorology, Silver Spring, MD, December 3, 2004.\n89","Pierce, T.E. Emissions modeling research in the Atmospheric Sciences Modeling Division.\nPresentation to representatives from the Mid-Atlantic Regional Air Management\nAssociation (MARAMA), Silver Springs, MD, February 18, 2004.\nPleim, J.E. A new non-local boundary layer model. Presentation at the WRF-MM5 User's\nWorkshop, Boulder, CO, June 22, 2004.\nPleim, J.E. New features of the 2003 release of the CMAQ model. Presentation at the 2003\nModels-3 Users' Workshop, One Atmosphere, One Modeling System: Models-3, Research\nTriangle Park, NC, October 27, 2004.\nPleim, J.E. NOAA's Eta-CMAQ modeling system for air quality forecasting. Lecture provided at\nthe Air Quality Forecast Training Workshop, Environment Canada-Eastern Region\nHeadquarters, Halifax, Nova Scotia, February 11, 2004.\nPleim, J.E. Updates and evaluation of the Community Multiscale Air Quality (CMAQ) model.\nPresented at the 13th Joint Conference on the Applications of Air Pollution Meteorology\nwith the Air & Waste Management Association, Vancouver, BC, Canada, August 23, 2004\nPouliot G.A. Recent advances in the modeling of airborne substances. Presentation at the 2003\nModels-3 Users' Workshop, One Atmosphere, One Modeling System: Models-3, Research\nTriangle Park, NC, October 27, 2003.\nRao, S.T. Using air quality models for emissions management decisions. Presentation at the\nNRC Committee Meeting, Washington, DC, March 18, 2004.\nRao, S.T. Integrated use of observations and model outputs in air quality management.\nPresentation at the 3rd Canadian Workshop on Air Quality, Quebec, Canada, March 24,\n2004.\nRao, S.T. Climate-air quality interactions. Presentation at the NOAA Climate Board,\nWashington, DC, April 9, 2004.\nRao, S.T. Using CMAQ in air quality policy decisions: Research to applications. Presentation at\nthe Office of Management and Budget Meeting, Washington, DC, April 21, 2004.\nRao, S.T. Role of satellite observations in understanding the link between air quality and health.\nPresentation at the EPA-NOAA Meeting on Satellite Data Assimilation, Washington, DC,\nMay 4, 2004.\n90","Rao, S.T. Federal-State partnerships for enhanced understanding of air quality and health\nrelationships. Presentation at the EPA Science Forum 2004: Healthy Communities and\nEcosystems, Washington, DC, June 3, 2004.\nRao, S.T. Role of models in air quality management. Presentation at the EPA-FACA's Science\n& Technology Subgroup Meeting, Research Triangle Park, NC, September 9, 2004.\nSchere, K.L. NOAA's Eta-CMAQ modeling system for air quality forecasting. Lecture provided\nat the Air Quality Forecast Training Workshop, Environment Canada-Eastern Region\nHeadquarters, Halifax, Nova Scotia, March 10, 2004.\nSchere, K.L. Atmospheric modeling of air pollutants with the Community Multiscale Air Quality\n(CMAQ) model. Poster presentation at the EPA Science Forum 2004: Healthy\nCommunities and Ecosystems, Washington, DC, June 1, 2004.\nSwall, J.L. Nonstationary spatial modeling of environmental data using a process convolution\napproach. Presentation at the Joint Statistical Meeting 2004, Toronto, Canada, August 11,\n2004.\nYu, S. Statistics Definitions and issues: Deriving \"unbiased symmetric\" metrics. Invited\npresentation at the Particulate Matter Model Performance Evaluation Workshop, Chapel\nHill, NC, February 10, 2004.\n91","APPENDIX D: WORKSHOPS AND MEETINGS\nTexas Association of Regional Councils Proposal Review Panel, Houston, TX, October 4-7,\n2003.\nR.L. Dennis\nNOAA/NCEP Air Quality Forecast Model Meeting, Camp Springs, MD, October 14-15, 2003.\nT.L. Otte\nJ.E. Pleim\nK.L. Schere\nNARSTO Emissions Inventory Workshop, Austin, TX, October 14-17, 2003.\nR.L. Dennis (Session Co-Chair)\nW.G. Benjey\nT.E. Pierce\nJ.L. West\nStanding Air Emissions Working Group (last meeting), San Francisco, CA, October 18, 2003.\nW.G. Benjey\nNADP Ammonia Workshop, Washington, DC, October 22-24, 2003.\nR.L. Dennis\nOffice of the Federal Coordinator for Meteorology, Committee for Climate Analysis, Monitoring,\nand Services, Silver Springs, MD, October 24, 2003.\nE.J. Cooter\nAmerican Association of Aerosol Research 21st Annual Conference, Anaheim, CA, October\n20-24, 2003.\nP.V. Bhave\nNOAA Design Review Meeting for Air Quality Forecast Capability, Silver Spring, MD, October\n23, 2003.\nK.L. Schere\n92","CMAS Models-3 Users' Workshop, One Atmosphere, One Community, One Modeling System:\nModels-3, Research Triangle Park, NC, October 27-29, 2003.\nW.G. Benjey\nG.L. Gipson\nS.T. Rao\nP.V. Bhave\nJ.M. Godowitch\nS.J. Roselle\nJ.K.S. Ching\nD.J. Luecken\nK.L. Schere\nR.L. Dennis\nC.G. Nolte\nD.B. Schwede\nB.K. Eder\nT.L. Otte\nJ.O. Young\nP.L. Finkelstein\nT.E. Pierce\nS. Yu\nR.C. Gilliam\nJ.E. Pleim\nA.B. Gilliland\nG.A. Pouliot\nMeeting of External Advisory Committee of the CMAS Center, Research Triangle Park, NC,\nOctober 30, 2003.\nK.L. Schere\nNational Exposure Research Laboratory Leadership Team Meeting, Athens, GA, November 3-6,\n2003.\nS.T. Rao\nEPA Joint TMDL (Total Maximum Daily Loads) Coordinators' Meeting, Chicago, IL, November\n3-7, 2003.\nR.L. Dennis\nNational Science Foundation, Collaborative Large-scale Engineering Analysis Network for\nEnvironmental Research Workshop, Duke University, Durham, NC, November 10-11, 2003.\nR.L. Dennis\nEPA-DOE Sandia National Laboratory Scientist-to-Scientist Workshop, Albuquerque, NM,\nNovember 12-13, 2003.\nJ.E. Pleim\nS.T. Rao\nK.L. Schere\nAmerican Geophysical Union Fall Meeting, San Francisco, CA, December 8-12, 2003.\nW.T. Hutzell\n93","EPA-DOE Lawrence Livermore National Laboratory Scientist-to-Scientist Workshop, Research\nTriangle Park, NC, December 16-17, 2003.\nE.J. Cooter\nEPA Scientist-to-Scientist Air/Water Meeting on Effects of Nitrogen in Estuarine and Coastal\nMarine Systems, Washington, DC, December 17, 2003.\nR.L. Dennis\nAmerican Meteorological Society Workshop on Merging Mesoscale and Computational Fluid\nDynamics Modeling Capabilities, Seattle, WA, January 11, 2004.\nA.H. Huber\n84th American Meteorological Society Annual Meeting, Seattle, WA, January 11-15, 2004.\nA.H. Huber\nE.M. Poole-Kober\nAmerican Meteorological Society's MGA Advisory Board, Seattle, WA, January 12, 2004.\nE.M. Poole-Kober\nSeventh Annual Atmospheric Science Librarians International Conference, Seattle, WA, January\n14-16, 2004.\nE.M. Poole-Kober\nWorkshop on Carbonaceous Particulate Matter: The State of the Science - Part I, EPA's National\nCenter for Environmental Assessment, Research Triangle Park, NC, January 14, 2004.\nP.V. Bhave\nC.G. Nolte\nJ.E. Pleim\nK.L. Schere\n7th Annual Electric Utilities Environmental Conference, Tucson, AZ, January 18-22, 2004.\nS.T. Rao\n94","Particulate Matter Model Performance Evaluation Workshop, Chapel Hill, NC, February 10-11,\n2004.\nP.V. Bhave\nR. Mathur\nR.L. Dennis\nS. Yu\nB.K. Eder\nAir Quality Forecast Training Workshop, Environment Canada-Eastern Region Headquarters,\nHalifax, Nova Scotia, Canada, February 11, 2004.\nJ.E. Pleim\nArctic Council Action Plan Steering Group Meeting on Mercury Inventory Development,\nMoscow, Russia, February 17-20, 2004.\nO.R. Bullock, Jr.\nBRACE Data Analysis Workshop, Tampa Bay, FL, February 18-19, 2004.\nR.L. Dennis\nUrban Atmospheric Observatory Planning Meeting, New York, NY, February 23-24, 2004.\nS.T. Rao\nEPA 2004 National Air Quality Conference, Baltimore, MD, February 23-25, 2004.\nB.K. Eder\nEPA-NOAA Scientist-to-Scientist Meeting on Air Quality Research to Guide National Policy and\nPrograms, Research Triangle Park, NC, March 2-3, 2004.\nJ.K.S. Ching\nA.H. Huber\nW.B. Petersen\nE.J. Cooter\nD.J. Luecken\nJ.E. Pleim\nR.L. Dennis\nR. Mathur\nS.T. Rao\nB.K. Eder\nC.G. Nolte\nK.L. Schere\nP.L. Finkelstein\nT.L. Otte\nJ.L. Swall\nA.B. Gilliland\nT.E. Pierce\n95","Earth Observations Systems Expert Panel Workshop, Research Triangle Park, NC, March 9-10,\n2004.\nJ.K.S. Ching\nE.J. Cooter\nJ.E. Pleim\nAir Quality Forecast Training Workshop, Environment Canada-Eastern Region Headquarters,\nHalifax, Nova Scotia, Canada, March 10, 2004.\nK.L. Schere\nWatershed Analysis Risk Management Framework Model Peer Review Panel Meeting convened\nby the Minnesota Sea Grant, Minneapolis, MN, March 10-12, 2004.\nO.R. Bullock, Jr.\nNational Research Council Committee on Models in the Regulatory Decision Process,\nWashington, DC, March 18-19, 2004.\nS.T. Rao\n3rd Canadian Workshop on Air Quality, Quebec, Canada, March 23-25, 2004.\nS.T. Rao\nNASA 2004 AVIRIS (Airborne Visible/Infrared Imaging Spectrometer) Earth Science Workshop,\nNational Aeronautics and Space Administration, Pasadena, CA, March 31-April 2, 2004.\nJ.J. Streicher\nNOAA Climate Board Meeting, Washington, DC, April 9, 2004.\nS.T. Rao\nWorkshop on Carbonaceous Particulate Matter: The State of the Science - Part II, EPA's National\nCenter for Environmental Assessment, Research Triangle Park, NC, April 12, 2004.\nP.V. Bhave\nS.T. Rao\nC.G. Nolte\nK.L. Schere\nJ.E. Pleim\n96","Mercury and CO2 Workshop sponsored by the North Carolina Department of Environment and\nNatural Resources, Raleigh, NC, April 19-21, 2004.\nO.R. Bullock, Jr.\nMercury Model Intercomparison Meeting, Meteorological Synthesizing Center-East, Moscow,\nRussia, April 20-25, 2004.\nO.R. Bullock, Jr.\nEuropean Monitoring and Evaluation Program (EMEP) Workshop, Particulate Matter\nMeasurement and Modeling, New Orleans, LA, April 20-23, 2004.\nR.L. Dennis (Co-organizer of the workshop)\nP.V. Bhave\nB.K. Eder\nMeeting of the ARL/RTP, NCEP AQ Forecast Team, Research Triangle Park, NC, April 23,\n2004.\nT.L. Otte\nJ.E. Pleim\nK.L. Schere\nWorkshop on Air Quality Applications of Satellite Data: Using Satellite Data to Monitor and\nImprove Air Quality Forecast, NOAA/NESDIS/Center for Satellite Applications and Research\n(STAR), Camp Springs, MD, May 4, 2004.\nJ.E. Pleim\nS.T. Rao\nNational Wildland Fire Emissions Technical Workshop, New Orleans, LA, May 4-6, 2004.\nJ.M. Godowitch\nT.E. Pierce\n14th Annual SAIL (Southeast Affiliate of IAMSLIC Libraries) Annual Meeting, University of\nTexas-Marine Sciences Institute, Port Aransas, TX, May 11-14, 2004.\nE.M. Poole-Kober\n97","Air Pilot Project Group for Environmental Public Health Tracking Meeting, New York, NY, May\n12, 2004.\nS.T. Rao\nInterdepartmental Committee for Meteorological Services and Support Research Meeting,\nWashington, DC, May 14, 2004.\nS.T. Rao\nU.S. EPA Consequences of Global Change for Air Quality Science-to-Achieve Results (STAR)\nGrant Progress Review Workshop, Research Triangle Park, NC, May 24-25, 2004.\nE.J. Cooter\nU.S. EPA Science Forum 2004: Healthy Communities and Ecosystems, Washington, DC, June\n1-3, 2004.\nC.G. Nolte\nS.T. Rao\nG. Sarwar\nK.L. Schere\nWorkshop on Air Toxics Data Analysis, Chicago, IL, June 3, 2004.\nD.J. Luecken\nAnnual Review of the EPA Science-to-Achieve Results (STAR) Grant Program on Carbonaceous\nParticulate Matter, Research Triangle Park, North Carolina, June 7-8, 2004.\nP.V. Bhave\nC.G. Nolte\nU.S. EPA International Emission Inventory Conference, Working for Clean Air in Clearwater,\nClearwater, FL, June 8-10,2004\nW.G. Benjey\nT.E. Pierce\n98","WRF/MM5 User's Workshop, Boulder, CO, June 22-25, 2004.\nR.C. Gilliam\nJ.E. Pleim\nMeeting of the External Advisory Committee to the Columbia University Project on New York\nClimate and Health, New York City, NY, June 25, 2004.\nK.L. Schere\n7th International Conference on Mercury as a Global Pollutant, Ljubljana, Slovenia, June 27-July\n2, 2004.\nO.R. Bullock, Jr.\nWorkshop on CFD Models Applied to Atmospheric Transport and Dispersion, George Mason\nUniversity, Fairfax, VA, July 12, 2004.\nA.H. Huber\nEight Annual George Mason University Transport and Dispersion Modeling Conference, Fairfax,\nVA, July 13-15, 2004.\nA.H. Huber\nS.T. Rao\nAtmospheric Impacts Committee Meeting of the Coordinating Research Council, Detroit, MI,\nJuly 20-21, 2004.\nK.L. Schere\nICARTT (International Consortium for Atmospheric Research on Transport and Transformation)\nScience Meeting and visit to the NOAA Ship Ronald H. Brown, Durham and Portsmouth, NH,\nJuly 23-27, 2004.\nR.L. Dennis\nAlbemarle-Pamlico National Estuary Program, Science & Technology Advisory Committee\nMeeting, Greenville, NC, July 28, 2004.\nR.L. Dennis\n99","2004 Joint Statistical Meeting, Statistics as a Unified Discipline, Toronto, Ontario, Canada,\nAugust 8-12, 2004.\nJ.L. Swall\n2004 Mercury Workshop sponsored by the U.S. Geological Survey, Eastern Region, Reston, VA,\nAugust 17-18, 2004.\nO.R. Bullock\nFifth Symposium on the Urban Environment, American Meteorological Society, Vancouver,\nBritish Columbia, Canada, August 23-26, 2004.\nJ.K.S. Ching\nT.E. Pierce\nG.E. Bowker\nJ.E. Pleim\nD.K. Heist\nW. Tang\nA.H. Huber\n13th Joint AMS/AWMA Conference on the Applications of Air Pollution Meteorology,\nVancouver, British Columbia, Canada, August 23-26, 2004.\nA. H. Huber\nW. Tang\nU.S. EPA National Center for Environmental Research Relevancy Review Meeting on Source\nApportionment of Particulate Matter, Research Triangle Park, NC, August 26, 2004.\nP.V. Bhave\nTexas Association of Regional Councils Proposal Review Panel, Houston, TX, September 1-2,\n2004.\nR.L. Dennis\nEuropean Monitoring and Evaluation Programme (EMEP) Steering Body Meeting, Geneva,\nSwitzerland, September 6-8, 2004.\nR.L. Dennis\n100","Air Quality Forecast Focus Group Meeting, Silver Spring, MD, September 8-9, 2004.\nD. Kang\nJ.E. Pleim\nH.-M. Lin\nK.L. Schere\nR. Mathur\nJ.O. Young\nT.L. Otte\nS. Yu\nNOAA/NCEP Air Quality Forecast Model Meeting, Silver Spring and Camp Springs, MD,\nSeptember 9-10, 2004.\nR. Mathur\nT.L. Otte\nJ.E. Pleim\nK.L. Schere\nNASA-NOAA-EPA Workshop on Air Quality and Related Climate Change Issues, Research\nTriangle Park, NC, September 14-15, 2004.\nW.G. Benjey\nT.E. Pierce\nJ.K.S. Ching\nK.L. Schere\nR. Mathur\nOffice of the Federal Coordinator for Meteorology, Workshop on Urban Meteorology, Rockville,\nMD, September 20-23, 2004.\nJ.K.S. Ching\nT.E. Pierce\nS.T. Rao\nChallenges in Urban Meteorology: A Forum for Users and Providers, Rockville, MD, September\n21-23, 2004.\nJ.K.S. Ching\nT.E. Pierce\nS.T. Rao\nNADP Technical Committee Meeting, Halifax, Nova Scotia, Canada, September 21-24, 2004.\nP.L. Finkelstein\n101","APPENDIX E: VISITING SCIENTISTS\nDr. Georg Grell\nNOAA/Forecast Systems Laboratory\nBoulder, Colorado\nDr. Grell visited the Division on January 29, 2004, to discuss the WRF/Chem model, and gave a\npresentation entitled, A next generation air quality prediction model based on the Weather\nResearch and Forecast (WRF) model.\nDr. Mohammed Majeed\nDelaware Department of Natural Resources and Environmental Control\nState of Delaware\nNew Castle, Delaware\nDr. Majeed visited the Division on February 12, 2004, to discuss his collaboration on fine-scale\nMM5/CMAQ modeling.\nJeff McQueen\nDrs. Pius Lee, and Marina Tsidulko\nNOAA/NCEP\nScientific Applications International Corporation\nCamp Springs, Maryland\nNCEP\nCamp Springs, Maryland\nMr. McQueen, and Drs. Lee and Tsidulko visited the Division on April 23, 2004, to participate in\nan Air Quality Forecast Team meeting with ASMD staff.\n102","Dr. Michael Reynolds\nDr. Petra M. Kastner-Klein\nBrookhaven National Laboratory\nAssistant Professor\nEnvironmental Sciences Department\nSchool of Meteorology\nUpton, New York\nUniversity of Oklahoma\nDr. Pablo Huq\nAssociate Professor\nCollege of Marine Studies\nUniversity of Delaware\nNewark, Delaware\nDrs. Reynolds, Kastner-Klein, and Huq visited the Division during March 31-April 2, 2004, to\ndiscuss modeling urban pollution transport and dispersion related to potential future studies as\npart of the developing New York City Urban Atmospheric Observatory (UAO). Dr. Reynolds\nwas the program manager developing the UAO, which is now the Department of Homeland\nSecurity, New York City Urban Dispersion Program (UDP). Dr. Petra M. Kastner-Klein\npresented a seminar entitled, Mean flow, turbulence and dispersion characteristics in urban\nareas. Dr. Huq gave a seminar entitled, Modeling the atmospheric boundary layer and flow\naround buildings in a water tunnel.\nDr. Jerry Davis\nMarine, Earth, and Atmospheric Sciences Department\nNorth Carolina State University\n5134 Jordan Hall, Box 8208\nRaleigh, North Carolina\nDr. Davis has been working on-site at the Division as a visiting scientist since May 2004.\nMr. Jim Tuccillo\nIBM Corporation\nPeachtree City, Georgia\nMr. Tuccillo visited the Division on June 28, 2004, to discuss the computational efficiency\naspects of the Eta-CMAQ Air Quality Forecasting system.\n103","Dr. Daewon Byun\nDepartment of Geosciences\nUniversity of Houston\nHouston, Texas\nDr. Byun visited the Division during July 28-29, 2004. He presented a seminar on entitled,\nSimulation of Houston-Galveston Metropolitan airshed episode with CMAQ.\n104","APPENDIX F: POSTDOCTORAL RESEARCHERS\nDr. Daiwen Kang worked on operational model evaluations of ozone for the NOAA air quality\nmodel forecast project.\nDr. Wei Tang conducted research on computational fluid dynamics modeling and assessing the\nmodeling techniques with field study data.\nDr. Shaocai Yu conducted research on CMAQ diagnostic model evaluation for the EPA\nretrospective assessment simulations and the NOAA forecast simulations.\n105","APPENDIX G: ATMOSPHERIC SCIENCES MODELING DIVISION\nSTAFF AND AWARDS\nAll personnel listed are National Oceanic and Atmospheric Administration employees, except\nthose designated EPA, who are employees of the Environmental Protection Agency, or SEEP,\nwho are part of the EPA Senior Environmental Employment Program.\nOffice of the Director\nDr. S.T. Rao, Supervisory Meteorologist, Director\nJ. David Mobley (EPA), Environmental Engineer, Associate Director\nWilliam B. Petersen, Physical Science Administrator, Assistant Director\nJeffrey L. West, Physical Science Administrator\nBarbara R. Hinton (EPA), Secretary (Until December 2003)\nPatricia F. McGhee, Secretary (Since January 2004)\nProgram Operations Staff\nHerbert J. Viebrock, Supervisory Physical Scientist, Chief\nLinda W. Green, Administrative Specialist\nEvelyn M. Poole-Kober, Librarian\nJohn H. Rudisill, III, Equipment Specialist (Until May 2004)\nAtmospheric Model Development Branch\nKenneth L. Schere, Supervisory Meteorologist, Chief\nDr. Prakash V. Bhave, Physical Scientist\nO. Russell Bullock, Meteorologist\nRobert C. Gilliam, Meteorologist\nGerald L. Gipson (EPA), Physical Scientist\nJames M. Godowitch, Meteorologist\nDr. Alan H. Huber, Physical Scientist\nDr. William T. Hutzell (EPA), Physical Scientist\nDeborah Luecken (EPA), Physical Scientist\nDr. Rohit Mathur, Physical Scientist (Since January 2004)\nDr. Christopher G. Nolte (EPA), Physical Scientist\nTanya L. Otte, Meteorologist\nDr. Jonathan E. Pleim, Physical Scientist\nShawn J. Roselle, Meteorologist\nDr. Golam Sarwar (EPA), Physical Scientist\nDr. Jeffrey O. Young, Mathematician\n106","Patricia F. McGhee, Secretary (Until January 2004)\nShirley Long (SEEP), Secretary\nModel Evaluation and Applications Research Branch\nDr. Alice B. Gilliland, Supervisory Physical Scientist, Chief\nDr. Robin L. Dennis, Physical Scientist\nDr. Brian K. Eder, Meteorologist\nDr. Peter L. Finkelstein, Physical Scientist\nSteven C. Howard, IT Specialist\nDr. Biswadev Roy (EPA), Physical Scientist (Since December 2003)\nDr. Jenise L. Swall, Statistician\nAlfreida R. Torian, IT Specialist\nGary L. Walter, Computer Scientist\nSherry A Brown, Secretary\nAir-Surface Processes Modeling Branch\nThomas F. Pierce, Supervisory Physical Scientist, Chief\nDr. William G. Benjey, Physical Scientist\nDr. George E. Bowker (EPA), Physical Scientist\nDr. Jason K.S. Ching, Meteorologist\nDr. Ellen J. Cooter, Meteorologist\nDr. Dale A. Gillette, Physical Scientist\nDr. David K. Heist, Physical Scientist\nDr. Steven G. Perry, Meteorologist\nDr. George A. Pouliot, Physical Scientist\nDonna B. Schwede, Physical Scientist\nJohn J. Streicher, Physical Scientist\nBruce Pagnani (SEEP), Computer Programmer\nAshok Patel (SEEP), Engineer\nJohn Rose (SEEP), Machinist/Modeler\nJane Coleman (SEEP), Secretary (Since March 2003)\nApplied Modeling Branch\nMark L. Evangelista, Supervisory Meteorologist, Chief\nDennis A. Atkinson, Meteorologist\nDr. Desmond T. Bailey, Meteorologist\nPatrick D. Dolwick, Physical Scientist\nJohn S. Irwin, Meteorologist (Until August 2004)\nRichard A. Mason, Physical Scientist (Since January 2004)\n107","Brian L. Orndorff, Meteorologist\nJawad S. Touma, Meteorologist\nAwards\nWilliam B. Petersen received an EPA Gold Medal Award for his contributions as a member of the\nBrentwood Post Office Anthrax Crises Exception Team.\nJohn S. Irwin received a NOAA Distinguished Career Award for his leadership in boundary-layer\ndynamics, transport and diffusion, and model evaluation; and for merging sound science and\nregulatory policy.\nDr. Robin L. Dennis and Jeffrey L. West received an EPA Bronze Medal for serving as\ncontributing authors to the NARSTO PM Assessment document.\nDr. Alan H. Huber received an EPA Bronze Medal for the promotion of strong science in making\nagency decisions.\nShawn J. Roselle received an EPA Bronze Medal for his work with the Five-Year PM\nAccomplishment Team.\n108"]}