United States Environmental Protection Agency Office of Research and Development National Exposure Research Laboratory Research Abstract Government Performance Results Act (GPRA) Goal: Air Significant Research Findings: Community Multiscale Air Quality (CMAQ) Model - September 2003 Public Release Version Scientific The U.S. EPA and the states have cooperated in implementing the provisions of Problem and the 1990 Clean Air Act Amendments to attain National Ambient Air Quality Policy Issues Standards of criteria pollutants, including ozone and fine particulate matter. Air quality simulation models are being used in the implementation process to assess the impacts of envisioned emissions mitigation strategies on the criteria pollutants. Our research has led to the latest version of EPA's Community Multiscale Air Quality (CMAQ) modeling system, a numerical grid model capable of simulating regional through urban patterns of ozone and photochemical oxidants, fine and coarse particulate matter, visibility, and acid deposition. Research The CMAQ air quality model, which is based on the surface hydrology and Approach meteorological model (MM5) and the Sparse Matrix Operator Kernel Emissions (SMOKE) model, was extensively revised in 2003. Changes include updated chemistry and physics, upgrades to earlier implementations, and computational efficiency enhancements. A preliminary evaluation of the new release of the CMAQ model, which is located on the EPA Models-3 website (http://www.epa.gov/asmdnerl/models3/). compares model results to observations of gas and aerosol species. There are also model-to-model comparisons between the 2002 and 2003 CMAQ model releases. The most significant changes involve aerosol modeling, particularly nitrate aerosols and secondary organic aerosols (SOA). Nitrate aerosol modeling was updated to be consistent with the most recent literature and the SOA implementation was modified to allow for reversible semi volatility. These changes resulted in substantially lower predictions of concentrations of both aerosol nitrates and SOA when compared with the 2002 CMAQ version. Minor changes have also been made to aqueous processes and dry deposition. There have been major modifications to improve the model's computational efficiency. A new fast gas-phase chemistry solver, known as the Euler Backward Iterative (EBI) scheme, has been developed for the Carbon Bond Version 4 (CB4) chemical mechanism used in the CMAQ model. Also, some of the fastest reacting species have been dropped from the transport processors. The time step for operator splitting has been revised to allow different advective time steps by vertical layer. Note that other components of the CMAQ system such as SMOKE and Meteorology-Chemistry Interface Processor (MCIP) have also been revised recently. Revisions to MCIP are especially important because issues in ------- Research Approach Continued the processing of the horizontal and vertical wind fields have been addressed for version 2.1. Version 2.2 includes additional updates to the treatment of dry deposition, particularly for winter conditions. Results and The new version of the CMAQ modeling system has been tested for the Impact continental United States. An evaluation was performed involving two simulation periods (4 January - 19 February, 2002 and 15 June - 16 July, 1999) using two chemical mechanisms, the Statewide Air Pollution Research Center (SAPRC) mechanism and the CB4. Initial results from the test simulations indicated that the new version of the CMAQ model reproduced patterns of major pollutants reasonably well during this time period. At the national scale, air quality data from the Aerometric Information Retrieval System (AIRS) (ozone); the Interagency Monitoring of Protected Visual Environments (IMPROVE); the Clean Air Status and Trends Network (CASTNet); and the Scientific and Technical Information Network (STN) (particulate matter) monitoring networks were used to check model simulations. The mean bias of maximum estimated 1-hr and 8-hr ozone concentrations was within 10 percent of observed values, and PM2.5 component observations were within a factor of 2. Results from these simulations were also compared with results using the previous version of the model. The introduction of new science and numerical routines in the CMAQ system over the past year has led to considerable improvements in some of the model's pollutant concentration estimates. These preliminary results indicate that the model is ready for applications and testing by the air quality research and management community. The research leading to the latest version of the CMAQ modeling system was aided with collaborative work on meteorological modeling by the Carolina Environmental Programs (CEP) at the University of North Carolina and the National Center for Atmospheric Research (NCAR). Collaboration on emissions modeling is ongoing with NCAR and EPA's National Risk Management Research Laboratory (NRMRL). CEP and the University of California-Riverside also assisted with this version of the CMAQ model. The science within the CMAQ model is being documented in a journal article currently in preparation. The modeling system is available for downloading at htto ://www. cmascenter.org/html/models .html. Future Research Intensive model evaluation exercises are ongoing with the September 2003 version of the CMAQ model being used by NERL. Special field data, including those collected in the Nashville and Atlanta regions, and routine network data (i.e., CASTNet, IMPROVE, and the National Atmospheric Deposition Program (NADP)) will be used to evaluate model performance for ozone and fine particulate matter and the precursor trace gases to these pollutants. Research Collaboration and Research Products ------- Contacts for Questions and inquiries can be directed to: Additional Information Jonathan E. Pleim Phone: 919/541-1336 E-mail: pleim.jon@epa.gov Brian K. Eder Phone: 919/541-3994 E-mail: eder.brian@epa.gov Shawn J. Roselle Phone: 919/541-7699 E-mail: roselle.shawn@epa.gov U.S. EPA, Office of Research and Development National Exposure Research Laboratory (E243-03) Research Triangle Park, NC 27711 ------- |