The Community Multiscale Air Quality (CMAQ) Modeling System: Ongoing and Planned Developments
EPA CMAQ Development Team; Contact: Rohit Mathur

The last major release of the CMAQ modeling system (version 4.7 in 2008) included numerous updates to facilitate the simulation of multi-pollutant interactions. Public release of an updated version of the model is planned for October 2011 and is expected to include several significant scientific and structural updates, which have been guided by analysis and evaluation of the existing system as well as needs associated with addressing emerging environmental problems. Model applications to date have clearly demonstrated the continued need to account for and to improve the representation of interactions of atmospheric processes occurring at the various spatial and temporal scales. The tightening of the NAAQS to lower threshold values (e.g., the recent revisions to the O3 NAAQS) places additional requirements on the ability of the model to accurately represent the entire spectrum of ambient concentrations including the background values. Updated gas-phase mechanisms (SAPRC-07 and RACM2) are being tested and included in the model. To better understand discrepancies between model PM2.5 predictions and measurements especially during the cool seasons, current efforts are focusing on improving the characterization of unspeciated PM2.5 (trace element PM2.5 composition, non-carbon organic matter, sampling artifacts associated with gravimetric PM2.5 measurements). To address the needs of emerging assessments for both air quality-climate interactions and for finer scale air quality applications, an integrated WRF-CMAQ system is being developed wherein radiative effects of absorbing and scattering aerosols in the troposphere estimated from the spatially and temporally varying simulated aerosol distribution can be fed-back to the WRF radiation calculations, resulting in a "2-way" coupling between the atmospheric dynamical and chemical modeling components. Additionally, the system provides means for finer scale applications, wherein higher frequency of data exchange between meteorological and chemical calculations is necessary to capture the effects of meteorological variability on modeled concentrations. A summary of these ongoing developments, anticipated features in the 2011 version of CMAQ, and results from their preliminary testing will be presented.

Development of Cost-minimized Integrated Control Strategies for Regional Ozone and PM2.5 Reductions

Kuo-Jen Liao(1), Praveen Amar(2) and Armistead G. Russell(3)

1 Department of Environmental Engineering, Texas A&M University-Kingsville, Kingsville, TX
2 NESCAUM, Boston, MA
3 School of Civil & Environmental Engineering, Georgia Institute of Technology, Atlanta, GA

Identification of regional air quality control strategies is a difficult task because formation of air pollutants is interdependent and air quality at different locations may have different responses to common emission sources. In this study, we develop and demonstrate a mathematical programming model, OPERA, that finds cost-minimized control strategies for simultaneously attaining prescribed ozone and PM_2.5targets at multiple locations using sensitivities of pollutants to emissions and cost functions of emission reductions. Developments of cost-minimized control measures of regional anthropogenic SO₂, NO_xand VOC and local primary PM_2.5emissions for reducing ozone and PM_2.5concentrations in five U.S. cities (Atlanta, Chicago, Houston, Los Angeles and New York) are presented. Results show that emission reductions from distant regions could be cost-effective for decreasing peak ozone and average PM_2.5levels in the cities. Reductions in anthropogenic NOx and VOC as well as local primary PM_2.5emissions are generally more cost-effective than SO₂controls for decreasing a small magnitude of reductions (i.e., 10%) in ozone and PM_2.5concentrations. When targeted reductions in pollutant levels are higher (i.e., 15%), controls of primary PM_2.5emissions become less cost-effective than reductions in the precursors from the six regions because per-ton costs of primary PM_2.5reductions are higher than the other precursors. In some cases (i.e., 20% in this study), prescribed air quality targets may not be achieved because required magnitudes of emission reductions are out of their feasible ranges. As a result, the optimization model has been demonstrated to be efficient for developing cost-minimized control strategies for attaining multi-pollutant targets at multiple locations and could help air quality managers make effective decisions.

Simulating the Annual-Average PM2.5 Mass Concentration & Composition Using CMAQ: A Decade in Review

Prakash Bhave

For 10 years, EPA scientists along with research partners in the academic, regulatory, and commercial sectors have exercised the Community Multiscale Air Quality (CMAQ) modeling system for year-long simulations that span the continental United States. Extensive evaluations of those model results against speciated measurements of PM_2.5 have helped to identify shortcomings in our model formulations. Reflecting on a decade of studies conducted by EPA staff, this presentation will describe the three largest PM_2.5 compositional errors that were identified in past CMAQ model evaluations: nitrate, carbon, and unspeciated mass. For each case, numerous efforts to mitigate model biases will be reviewed and evaluated.

The first full-year CMAQ simulation of air pollutants across the United States was conducted in 2001. The 1996 calendar year was simulated and ground-level results were compared against measurements across two expansive monitoring networks: IMPROVE and CASTNet. Although model performance for total PM_2.5 mass concentrations looked excellent, it was attributed to compensating errors of overestimated nitrate and underestimated organic carbon.

In the next 5 years, the model performance problem for nitrate was solved by enhancing the mathematical representations of NH₃ emissions, soil temperature and moisture, nighttime N₂O₅ hydrolysis, HNO₃ dry deposition, and wet removal of NO₃^. From 2003 - 2008, the model underestimates of organic carbon were greatly mitigated by inventorying the wild fire emissions, enhancing the biogenic emission estimates, and adding numerous pathways for the formation of secondary organic aerosol (SOA).

The deployment of a national network of urban speciated PM_2.5 monitoring stations (CSN) permitted more extensive evaluations of the CMAQ modeling system starting in 2004. These recent evaluations (e.g., Foley et al., Geoscientific Model Development, 3, 205-226, 2010) reveal that the single largest error in PM_2.5 models over the United States is not organic carbon or the inorganic ions (e.g., SO₄^2-, NO₃^-). Rather, it is the unspeciated mass which we refer to as PM_other. This presentation will close with a summary of five ongoing efforts to improve the simulation of PM_other across the United States.

To What Extent Can Biogenic SOA be Controlled

Annmarie G. Carlton^1*, Robert W. Pinder¹, Prakash V. Bhave¹, George A. Pouliot¹

¹ U.S. Environmental Protection Agency, Office of Research and Development, National Exposure Research Laboratory, 109 TW Alexander Drive Research Triangle Park, NC 27711

The implicit assumption that biogenic secondary organic aerosol (SOA) is natural and can not be controlled hinders effective air quality management. Anthropogenic pollution facilitates transformation of naturally emitted volatile organic compounds (VOCs) to the particle phase, enhancing the ambient concentrations of biogenic secondary organic aerosol (SOA). It is therefore conceivable that some portion of ambient biogenic SOA can be removed by controlling emissions of anthropogenic pollutants. Direct measurement of the controllable fraction of biogenic SOA is not possible, but can be estimated through 3-dimensional photochemical air quality modeling. To examine this in detail, 22 CMAQ model simulations were conducted over the continental U.S. (August 15 - September 4, 2003). The relative contributions of five emitted pollution classes (i.e., NO_x, NH₃, SO_x, reactive non methane carbon (RNMC) and primary carbonaceous particulate matter (PCM)) on biogenic SOA were estimated by removing anthropogenic emissions of these pollutants, one at a time and all together. Model results demonstrate a strong influence of anthropogenic emissions on predicted biogenic SOA concentrations, suggesting more than 50% of biogenic SOA in the eastern U.S. can be controlled. Because biogenic SOA is substantially enhanced by controllable emissions, classification of SOA as biogenic or anthropogenic based solely on VOC origin is not sufficient to describe the controllable fraction.

NOx Dependent Organic Aerosol Parameterizations

Havala Pye, California Institute of Technology

Arthur Chan, California Institute of Technology

John Seinfeld, California Institute of Technology

The interaction between biogenic emissions, such as monoterpenes, and pollutants such as NOx, represent a potential anthropogenic control on what has traditionally been considered natural organic aerosol. The chemical transport model, GEOS-Chem, is used to explore the representation of organic aerosol on a global and regional scale. Primary organic aerosol is treated as either nonvolatile or semivolatile. Secondary organic aerosol (SOA) precursors include monoterpenes, sesquiterpenes, isoprene, benzene, toluene, xylene, intermediate volatility organic compounds, and semivolatile organic compounds. NOx-dependent aerosol yields for monoterpenes and sesquiterpenes are used to determine if the lack of NOx-dependent biogenic SOA parameterizations could be responsible for disagreement between models, observations, and other global estimates of organic aerosol.

Variability in the Direct Radiative Forcing of Aerosols From Distinct Source Regions, Seasons and Emissions Sectors.

Modeling the trace-elemental composition of PM2.5 in CMAQ

Heather Simon, Prakash Bhave, George Pouliot, Daniel Tong, Wyat Appel

Much of the fine particulate matter (PM_2.5) in urban areas is directly emitted to the atmosphere in the particle phase (i.e., primary PM). In the SMOKE-CMAQ modeling system, these primary emissions are routinely speciated into sulfate, nitrate, elemental carbon (EC), organic carbon (OC), and other unspeciated PM_2.5 (i.e., PM_other). The CMAQ model systematically overestimates wintertime concentrations of PM_2.5, and several recent evaluations have revealed that PM_other accounts for the bulk this bias. However, the exact cause of the model bias is difficult to diagnose using the current SMOKE-CMAQ modeling system because the chemical identity of PM_other is lost during emissions processing through SMOKE.

EPA scientists recently completed the first step in overcoming this problem by subdividing the anthropogenic inventory of PM_other emissions into its known constituents: trace elements, metal-bound oxygen, non-carbon organic matter (NCOM), etc. (Reff et al., Environ. Sci. Technol. 43:5790-5796, 2009). This was accomplished by assembling a library of speciation profiles from emission studies in which all of these chemical components were measured. Next, we incorporated the profiles into SMOKE to produce gridded model-ready inventories of the standard set of PM_2.5 species (i.e., sulfate, nitrate, EC, OC) plus thirteen additional chemical constituents: Na, Mg, Al, Si, Cl, K, Ca, Ti, Mn, Fe, NCOM, ammonium, and water.

In this presentation, we will describe an updated version of CMAQ which reads the emissions of these new species and tracks their concentrations and deposition patterns throughout the modeling domain. We evaluate the model results from one winter and one summer month (January and July 2002) against routine ambient measurements of each trace element, in hopes of diagnosing the major causes of the CMAQ bias for PM_other. The model updates described in this presentation are slated for the next public release of CMAQ.

Uncertainties influencing health-based prioritization of ozone abatement options

Daniel S. Cohan (Rice University)

Antara Digar (Rice University)

Wei Tang (Rice University)

Michelle L. Bell (Yale University)

In selecting ozone abatement strategies, decision makers may consider the resulting benefits to human health along with the regulatory imperative of attaining ambient standards. However, evaluations of ozone mitigation health benefits are complicated by uncertainties in both photochemical sensitivity modeling and epidemiological concentration-response functions. Recent work has shown that ozone-precursor sensitivities are in certain cases far more susceptible than ozone concentrations to being influenced by parametric uncertainties in photochemical models. There are also substantial uncertainties in the ozone concentration-response functions for mortality and other health impacts. Here, we simulate parametric uncertainties of ozone responsiveness to precursor emission controls and assess population-weighted health impacts in two domains with distinct ozone sensitivity regimes, Texas and Georgia. We then evaluate the relative importance of uncertainties in the photochemical model parameters and in ozone mortality concentration-response functions in influencing the prioritization of ozone abatement options in each region.

Impact of an updated toluene mechanism on air quality in the western US

Golam Sarwar, Wyat Appel, Annmarie Carlton, Rohit Mathur, Kenneth Schere

Toluene is an important aromatic compound that can affect ozone and secondary organic aerosol concentrations in the atmosphere. However, atmospheric chemistry of toluene is poorly understood and a known source of uncertainty in air quality models. Recently, an updated condensed mechanism for toluene has been proposed for use with the Carbon Bond 2005 (CB05) mechanism. The updated toluene mechanism contains 26 chemical reactions involving 13 species for toluene oxidation while the base toluene mechanism contains 10 chemical reactions involving 5 chemical species. The updated toluene mechanism can better explain the results of environmental chamber experiments involving toluene and oxides of nitrogen. In this study, the impact of the updated toluene mechanism on air quality in the western US is examined using the Community Multiscale Air Quality (CMAQ) model. Model simulations were performed using the Carbon Bond 2005 mechanism containing the base and the updated toluene mechanisms for July 2002. Anthropogenic emissions were obtained from the National Emissions Inventory developed by the United States Environmental Protection Agency and biogenic emissions were estimated using the Biogenic Emission Inventory System. Meteorological data for the CMAQ model were obtained from the MM5 model. The updated toluene mechanism increased the monthly mean of daily maximum 8-hr ozone by more than 1.0 ppbv in Los Angeles, Portland, and Seattle and by 0.5 ppbv or more in other urban areas compared to those obtained with the base toluene mechanism. Model bias and model error with updated toluene mechanism for ozone improved by small margins in Los Angeles. The updated toluene chemistry increased the monthly mean secondary organic aerosols from toluene by a maximum of 7%, hydroxyl radicals by a maximum of 11%, and hydroperoxy radicals by a maximum of 18%, respectively, compared to those with the base toluene chemistry.

A New Decision Support System Based on a Service-Oriented Architecture

Neil Wheeler, Tami Funk, Sean Raffuse, Stacy Drury, Paul Nuss, Kevin Unger, Liron Yahdav, Daniel Pryden, Alan Healy, Michael Haderman, and Lyle Chinkin

Sonoma Technology, Inc., 1455 N. McDowell Blvd., Suite D, Petaluma, CA 94954

John Cissel, Joint Fire Science Program, 3833 S. Development Ave., Boise, ID 83705

H. Michael Rauscher, Rauscher Enterprises LLC, 1733 Old NC 20, Leicester, NC 28748

During the past decade, a proliferation of data, software systems, and analysis tools has emerged in various modeling communities. The heterogeneity of available data, data formats, software systems, and ad-hoc tools has fractured the awareness, access, and distribution of data and software tools. Consequently, analysts and decision makers are left with an assortment of analysis and modeling methods, as well as unconnected software systems in various stages of development.

In response to these issues, the Joint Fire Science Program (JFSP), acting in concert with the interagency Fuels Management Committee, initiated the Software Tools and Systems Study in 2007 to address the proliferation of unconnected and unmanaged modeling systems in the fire and fuels domain. A strategic assessment was completed in 2008 that led directly to development of a conceptual design and a software design for a service-oriented, framework architecture for fuels treatment planning. Under the guidance of an interagency team, these designs were developed into the provisionally named Interagency Fuels Treatment Decision Support System (IFT-DSS). During 2009, JFSP funded development of a proof-of-concept version of IFT-DSS. A fully functional version of IFT-DSS, now under development, will be completed, tested, and in use by spring 2012. IFT-DSS provides command and control for pre-existing and newly developed models and datasets within a common framework. It allows users to customize analysis flow paths and store intermediate and final results and analysis flow paths for repeated analysis of alternative scenarios. Scientific model developers will be able to register their models within the system as callable software services as part of the web-based collaborative Service-Oriented Architecture (SOA).

This paper describes the IFT-DSS SOA design and discusses implementation issues and potential uses in other subject matter domains, such as meteorological, emissions, and air quality modeling.

Understanding the impact of isoprene nitrates and OH reformation on regional air quality using recent advances in isoprene photooxidation chemistry

Xie, Ying; Paulot, Fabien; Pinder, Robert W.; Carter, William P. L.; Nolte, Christopher G.; Luecken, Deborah J.; Hutzell, William T.; Wennberg, Paul O.; and Cohen, Ron

Isoprene chemistry has large impacts on ozone, oxidized nitrogen, and second organic aerosol concentrations. Photooxidation of isoprene can produce unsaturated hydroxy nitrates, which serve as an important sink for NO_x. The production efficiency and the fate of these isoprene nitrates significantly affect ambient NO_x and ozone levels. Consequently, the chemistry of isoprene nitrates represents a major uncertainty in determining the response of ozone to future changes in biogenic emissions. A new isoprene photooxidation mechanism has been developed based on recent chamber experiments. These studies provide new constraints on the yield of isoprene nitrates, their atmospheric lifetime, and the amount of NO_x released from the oxidation of these nitrates. New insights have also been obtained for isoprene oxidation under pristine conditions, which suggest efficient OH reformation and large flux of dihydroxyepoxides.

We have incorporated these recent advancements into the SAPRC07 chemical mechanism. The new scheme is evaluated against smog chamber experiments from three different laboratories. The results suggest improved predictions of ozone for the experiments with the lowest NO_x concentrations and better simulation of methyl vinyl ketone. We have also implemented the new scheme into the CMAQ model and have conducted simulations for the INTEX-NA field campaign period in summer 2004. The new scheme generally results in higher NO_x and HNO₃ concentrations due to more efficient NO_x recycling. The new scheme also predicts higher OH levels and shows better agreement with INTEX-NA observations. Additional sensitivity studies are conducted to constrain uncertainties such as the dry deposition rates of isoprene nitrates and biogenic emission rates. The updated SAPRC07 mechanism will be an essential tool for modeling future emission scenarios, so that the sensitivity of ozone to isoprene emissions changes can be better assessed.

Temporal Source Apportionment of Policy-Relevant Air Quality Metrics

Nicole MacDonald, Amir Hakami (Carleton University)

Source apportionment is typically performed on pollutant concentrations at specific locations and times. However, the concept can be equally applied to integrated metrics that depend on concentrations such as visibility, health indices, mortality, regulatory nonattainment, etc. Currently, various methods are used for source apportionment of concentrations across a spatial domain. Adjoint sensitivity analysis is among the most promising methods for source apportionment of integrated metrics. The adjoint method allows for gradients of such metrics to be calculated with respect to perturbations in the system inputs, including emissions. As a receptor-based approach, the adjoint method is advantageous in policy applications because the individual sources that have the most influence on a chosen metric can be efficiently identified. Adjoint methods have been previously used to apportion integrated metrics among spatially distributed sources. These applications have shown that sources at different locations may have significantly varying degrees of influence on integrated metrics of concentration. Likewise, it is probable that emissions at different times would have different impacts on air quality metrics.

We will apply the adjoint of gas-phase CMAQ to a summer of 2007 episode for temporal source apportionment of air quality metrics. Various air quality metrics such as exposure and health indices will be considered. Time- and location-dependent gradients of metrics with respect to anthropogenic emissions of NO_x and VOCs will be calculated. Temporal characteristics of these gradients across various regions will be explored with the objective to use such regional trends in pollution control strategies.

Updates to the Carbon Bond Mechanism for Version 6 (CB6)
Greg Yarwood and Jaegun Jung, Gookyoung Heo, Gary Z. Whitten, Ph.D., Jocelyn Mellberg, and Mark Estes

The Carbon Bond mechanism describes tropospheric oxidant chemistry in a concise manner suitable for use in complex 3-dimensional atmospheric models. Existing versions (CB4 and CB05) are used in photochemical grid models for ozone and particulate matter (PM) such as CMAQ, CAMx and WRF-Chem. This paper describes the development and evaluation of version 6 of the Carbon Bond mechanism (CB6). Updates include: (1) Incorporating new scientific information released since the previous mechanism update in 2005 (CB05) especially as evaluated by IUPAC and NASA review panels. (2) Reviewing and updating reactions for alkanes, alkenes and aromatics with the most changes resulting for isoprene and aromatics. (3) Adding explicitly several long-lived volatile organic compounds (VOCs) that form ozone at regional scales, specifically propane, benzene, acetone and other ketones. (4) Adding explicitly acetylene and benzene because they are precursors to secondary organic aerosol (SOA) formation and useful as anthropogenic emission tracers. (5) Adding explicitly VOC degradation products that can produce SOA via aqueous-phase reactions, specifically glyoxal, glycolaldehyde and methyl glyoxal. The CB6 updates expand chemical detail (15% more species with 35% more reactions than CB05) to take advantage of advances in computer capacity while maintaining a balance with other resource needs (e.g., grid extent and resolution). CB6 has been tested and evaluated against results from ~350 environmental chamber experiments and demonstrates performance gains compared to previous mechanisms. Results of CAMx ozone simulations using CB6 will be presented and discussed.

A Comparison of CMAQ Predicted Contributions to PM2.5 from Aircraft Emissions to CMAQ Results Post-Processed Using the Speciated Modeled Attainment Test

Matthew Woody, Saravanan Arunachalam, J. Jason West, and Uma Shankar

The Speciated Modeled Attainment Test (SMAT) is a policy relevant CMAQ post-processing tool which combines model results with ambient air quality monitoring data to determine the effects of changes in emissions. Here, the results from a CMAQ investigation of the impacts of aviation emissions at 99 airports on fine particulate matter (PM_2.5) in the continental U.S. were post processed through SMAT. The primary objective was to quantify the influence of SMAT on PM_2.5 contributions attributed to aircraft emissions as compared to CMAQ. Within the continental U.S, SMAT results indicated that 2005 aircraft emissions contributed on average 0.0024 μg m^-3 (0.03%) to annual average PM_2.5 concentrations while 2025 aircraft emissions contributed 0.0096 μg m^-3 (0.11%). CMAQ results predicted 2005 aircraft emissions contributed 0.0037 μg m^-3 (0.05%) to annual average PM_2.5 concentrations while aircraft emissions in 2025 increased PM_2.5 concentrations by 0.0127 μg m^-3 (0.21%). The differences in the overall contributions to PM_2.5 from aircraft are primarily attributed to sulfate and nitrate aerosol. For these two species, SMAT predicted 1.67 and 2.46 times higher sulfate aerosol contributions and 0.26 and 0.41 times lower nitrate aerosol contributions from aircraft emissions compared to CMAQ in 2005 and 2025, respectively. This shift in contributions from sulfate and nitrate aerosols is due to differences between ambient air quality observations and CMAQ predicted base concentrations and relates to both algorithms used by SMAT to speciate monitor data as well as model performance. This work has implications when results are used for further analysis, such as health impact studies where speciated and total PM_2.5 impacts could change depending on which set of results are used.

Use of Geostationary Satellite Observations for Dynamical Support of Model Cloud Fields

Arastoo Pour Biazar, Richard T. McNider, Kevin Doty, Yun-Hee Park

Universityof Alabama in Huntsville

Maudood Khan

The Universities Space Research Association

Bright Dornblaser

TexasCommission on Environmental Quality (TCEQ)

One of the major sources of uncertainty in weather and air quality forecast models is the prediction of clouds. Clouds are the key component in radiation energy balance and climate change as well as the Earth's hydrological system. They also greatly impact the chemical composition of the atmosphere through their radiative impact on photochemistry, influencing heterogeneous chemistry, as well as gas/aerosol formation and removal. Thus, the accurate prediction of clouds in space and time is of outmost importance to many applications in atmospheric sciences. Here in this study we present results from our first attempt at assimilating Geostationary Operating Environmental Satellite (GOES) observations within Weather Research and Forecasting (WRF) model to improve model prediction of clouds. The technique cultivates on our previous work in which GOES observations were assimilated within the PSU/NCAR mesoscale model (MM5). By way of multi-variable regression equations, key cloud parameters, including a target vertical velocity, are expressed as a function of satellite observed variables and then using the observations, model fields are adjusted accordingly. The technique insures that model clouds are removed where clear sky is indicated by the observations, and clouds are produced in a sustainable manner where it is cloudy. An overview of the technique and the preliminary results will be presented.

Proof-of-Concept Evaluation of Use of Photochemical Grid Model Source Apportionment Techniques for Prevention of Significant Deterioration of Air Quality Analysis Requirements

Bret Anderson, Ralph Morris, Chris Emery, Kirk Baker, Andy Hawkins, Erik Snyder

This paper will discuss a joint evaluation project between the US Forest Service, US Environmental Protection Agency, and ENVIRON to evaluate the use of photochemical grid model (PGM) source apportionment techniques for air quality analysis requirements under the Prevention of SIgnificant Deterioration of Air Quality (PSD) major source air quality permitting program. Under the PSD permitting program, applicants are required to demonstrate that they do not cause or contribute to violations of applicable national ambient air quality standards (NAAQS), PSD increments, or cause degradation of air quality related values (AQRVs) in federally protected Class I areas.

The Comprehensive Air Quality Model with Extensions (CAMx) with ozone and particulate matter source apportionment technologies (OSAT/PSAT) will be used to examine the efficacy of the use of PGM's for single source ozone analyses and visibility and acid deposition analyses of Class I airsheds. This approach will be compared to current methods employed to conduct PSD Class I air quality impact analyses to help elucidate potential resource requirements and impacts to normal air quality permitting procedures and timeframes.

Incorporation of Dynamic Boundary Conditions into the AIRPACT regional air quality forecast system
Farren L. Herron-Thorpe, Joseph K. Vaughan, George H. Mount, Serena Chung, and Brian K. Lamb

Laboratory for Atmospheric Research
Department of Civil & Environmental Engineering
Washington State University

Louisa Emmons, National Center for Atmospheric Research
Boulder, CO

Dynamic boundary conditions for the Pacific Northwest regional AIRPACT air quality forecast system have been developed from daily forecasts using the global MOZART model. Furthermore, the daily MOZART forecast results include assimilation of MOPITT carbon monoxide observations. The effects of using dynamic vs monthly averaged BC and using BC with assimilated CO are investigated by comparing results for several months during 2010. The new MOPITT CO assimilated BC are shown to have a significant effect in the free troposphere, especially when there is periodic trans-Pacific pollution from Asia that reaches the western US boundary. Timelines and monthly averages are compared to AIRS carbon monoxide retrievals to determine the effect of these changes on AIRPACT forecast accuracy.

UK AIR QUALITY FORECASTING USING WRF AND CMAQ
Andrea Fraser, Trevor Daviesa, Justin Lingarda

AEA, Fermi Avenue, Harwell IBC, Oxon, OX11 0QR, UK.
Andrea.Fraser@aeat.co.uk

An operational air quality forecasting model has been set up based on the Advanced Research - Weather Research and Forecasting (WRF) model used to predict the atmospheric circulation and the Community Multiscalar Air Quality (CMAQ) model used for chemical transformations, transport and deposition. The aim is to produce a two day forecast for AQ (O3, NO2, SO2, CO, PM10, PM2.5) and weather (temperature, precipitation, wind direction and speed). These along with data from a number of different sources are used by AEA to produce the UK operational Air Quality forecast for the UK Department of food and rural affairs (Defra) and the Devolved Authorities (DA). This work forms one of the tools used to generate the forecast and is part of the continued improvements. The air quality information is disseminated to the authorities and public using a variety of different media.

WRF is run with initial and boundary conditions from the NCEP Global Forecast and is used to produce the numerical description of the weather. This along with emissions data from the EMEP and UK national emissions inventories are used in the Community Multiscalar Air Quality (CMAQ) model. The forecasts are produced as two grids the 50km resolution grid covers a large area of Europe, and provides boundary conditions for a second 10km resolution grid covering the UK. WRF is operated with 48 vertical layers and CMAQ with 24. This is a new model system being developed for the UK forecast and is undergoing evaluation and refinements.

The WRF-CMAQ is producing forecast based on the UK AQ Index, and is one of the tools used for the UK operational AQ forecast. The European forecast has been operational since June 2009, and the UK forecast has been produced daily since November 2009. The WRF-CMAQ forecasts are evaluated daily with real-time (provisional) monitoring data from the UK National Automated Urban and Rural Network. The forecast is evaluated retrospectively using ratified measurements.

The AQ forecast is funded by Defra.

Modeling biomass burnings by coupling a sub-grid scale plume model with Adaptive Grid CMAQ

Aika Yano, Fernando Garcia Menendez, Yongtao Hu, and M. Talat Odman, School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA, 30332-0512, USA

D. Scott McRae, Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC 27695-7910, USA

Gary L. Achtemeier, Forestry Sciences Laboratory, USDA Forest Service, Athens, GA, 30602-2044, USA

Biomass burning plumes are not well resolved in current modeling systems due to insufficient grid resolution and/or inadequate sub-grid treatments. We developed a comprehensive modeling system that enhances the ability to predict air quality impacts from biomass burnings. The adaptive grid method that increases grid resolutions has been demonstrated to accurately track the biomass burning plumes at the regional grid scales. However, precise burning emission inputs to the model, such as the spatial spread and vertical profile at the sub-grid scale are also required for accurate solutions. Daysmoke, a Lagrangian plume model used to predict short range dispersion of prescribed burning plumes, is coupled with Adaptive Grid CMAQ to fulfill these requirements for burning emissions.

Our new modeling system consists of the Adaptive Grids CMAQ coupled with Daysmoke as the sub-grid scale plume model. When a biomass burning plume travels downwind and becomes fully developed, the plume concentrations are carried over to CMAQ with vertical and chemical resolutions. Until the plume is fully developed, Daysmoke is responsible for tracking the evolution of all the emissions. The interface where the two models share their information is time dependant and is calculated and updated for every CMAQ time step in an analysis called "handover". The details of this analysis will be discussed. Biomass burning simulations will be evaluated to show the improvement in the results of the newly developed system.

Investigating Differences in Ozone Production from CB05 and CBMIV Versions of the NAQFC

Rick Saylor, Hsin-Mu Lin, Pius Lee, Binyu Wang, Tianfeng Chai, Ariel Stein,Daniel Tong, Hyun-Cheol Kim, Yunsoo Choi, Fantine Ngan, and Daewon Byun

Air Resources Laboratory

National Oceanic and Atmospheric Administration

Silver Spring, MD 20910

Sensitivity studies were performed to investigate the underlying causes of observed differences in ozone distributions produced by the operational and experimental versions of the National Air Quality Forecasting Capability (NAQFC). The operational NAQFC uses a version of CMAQ adapted for daily production of next day ozone forecasts using the NOAA National Center for Environmental Prediction North American Model (NAM) as the meteorological forecast driver. The operational NAQFC utilizes the CBMIV(Carbon Bond Mechanism IV) gas-phase chemical mechanism. An experimental version of the NAQFC is run in parallel with the operational version and uses the same meteorological data and is dependent on the same base emissions inventory, but uses the updated 2005 version of the Carbon Bond Mechanism (CB05) as the gas-phase chemical mechanism. Comparison of the parallel NAQFC forecast systems reveals a consistent bias between ozone distributions generated by the two forecast systems where the experimental track, utilizing CB05, routinely generates higher ozone mixing ratios across the domain and larger biases as compared to surface measurement networks. An investigation was initiated to better understand the reasons for the observed biases of the CB05 version of the NAQFC and a set of sensitivity studies were performed for a two-week period in August 2007. The sensitivity studies investigate possible hypotheses for the observed differences including meteorological (humidity), chemical mechanism (PAN, organic nitrate, terpene), and removal (deposition) processes. Initial results indicate that recycling of reactive nitrogen from organic nitrate concentrations simulated by the CB05 version of the model may account for a large portion of the observed ozone bias. The paper will describe the sensitivity studies performed and summarize the overall findings.

Development of a Plume-in-Grid Version of Global-through-Urban WRF/Chem

Prakash Karamchandani¹, Krish Vijayaraghavan¹, Yang Zhang² and Shu-Yun Chen¹

¹ENVIRON International Corporation, 773 San Marin Drive, Suite 2115, Novato, CA 94998

²North Carolina State University, Raleigh, NC 27695

A unified Global-through-Urban Weather Research and Forecasting Model with Chemistry (GU-WRF/Chem) is being developed to simulate the feedbacks between global climate change and urban/regional air quality to support the development of emission control strategies with potential co-benefits for air quality management and climate mitigation. The model will be applied in nested grid mode for domains covering the globe, the Trans-Pacific region, the continental US (CONUS), and the eastern U.S. The finest grid resolution, for the eastern U.S. domain, is of the order of 12 km. However, even a grid resolution of 12 km (or even 4 km) is insufficient to correctly represent the near-source transport and chemistry of emissions from elevated point sources, since the initial dimensions of stack plumes from these sources are of the order of tens of meters, and the dimensions of these plumes typically approach the grid resolution several grid cells downwind of the sources. Since a significant fraction of emissions in the eastern U.S. can be attributed to elevated point sources, this can be an important limitation of the grid model that can lead to errors in simulating (1) the contribution of elevated point sources to ambient concentrations and deposition fluxes, (2) the model response to emissions changes from elevated point sources or from other sources that may affect the chemistry of the elevated point source emissions, and (3) the feedback from the inner eastern U.S. domain to the outer domains. We describe the development of a Plume-in-Grid (PiG) version of GU-WRF/Chem to incorporate a more realistic treatment of the transport and chemistry of point source plumes in the model. The model is applied in a two-way nested grid simulation for the CONUS and eastern US domains for a summer episode. We present results from this application showing the model performance as well as the impact of using PiG treatment in the inner eastern US grid on model predictions in the outer CONUS grid.

CMAQ PM2.5 forecasts adjusted to errors in model wind fields

Eun-Su Yang, Sundar A. Christopher, Shobha Kondragunta, and Xiaoyang Zhang

The Community Multiscale Air Quality (CMAQ) model often fails to predict high PM2.5 episodes near and downwind of fires. This disagreement between CMAQ predictions and ground-based PM2.5 observations is primarily caused by errors in model-simulated winds. We find a first-order autoregressive process, AR(1), for an error term of simulated wind time series, where the error is shown as part of the previous errors plus random shock. Due to accumulated effect of previous errors, a predicted position of smoke tends to drift farther away from the actual location over time. In order not to miss high PM2.5 episodes during fires, we allow smoke plumes to reach into all grids that are within the range of wind-produced position error with AR(1). This approach eventually increases spatial range of simulated smoke plumes and makes PM2.5 predictions to move from false negative toward false positive. The approach, therefore, may provide more practical PM2.5 forecasts in that a false negative rate can be easily controlled for decision making.

Discretizing the Sphere for Multi-Scale Air Quality Simulations using Variable-Resolution Finite-Volume Techniques

Martin J. Otte, Atmospheric Modeling and Analysis Division, U.S. EPA

Robert L. Walko, Rosenstiel School of Meteorology and Physical Oceanography, University of Miami

The accurate representation atmospheric processes from global down to regional scales is necessary for air quality modeling. Global processes can quickly propagate and influence the local meteorology around your area of interest. Aerosols and other chemical constituents effect the microphysical and radiative structure of the atmosphere, which feeds back to the entire global climate system. Most current air quality models cannot represent these effects because they can only be applied to a small portion of the globe and because each model grid must have constant resolution. In addition, because air quality models are only of limited area the atmospheric and chemical fields at the boundaries is required, which imparts significant errors on the forecasted fields.

To alleviate these shortcomings and represent the entire global system within air quality models, new computational techniques are being developed that discretize the entire globe and are able to seamlessly telescope down to much higher resolution in local areas of interest. These multi-scale modeling systems replace the finite-differencing schemes used in most current atmospheric models with unstructured finite-volume methods. With finite volume solvers, the entire earth can be discretized without the need for spherical coordinate transformations which result in singularities at the poles or transformations to various map projections that restrict the model to be used for limited areas only. In local regions, the computational mesh can be increased to much higher resolution to represent regional and local pollution transport. This presentation will overview the scientific and computation aspects of unstructured global model.

Impact of temporal fluctuations in power plant emissions on air quality forecasts

Prakash Doraiswamy, Christian Hogrefe, Eric Zalewsky, Winston Hao, Ken Demerjian, Jia-Yeong Ku and Gopal Sistla

The New York State Department of Environmental Conservation (NYSDEC) has been performing model-based air quality forecasts since June 2005 using a single model configuration, and from June 2008 using an ensemble-based system. The ability of any modeling system to accurately predict ozone and PM2.5 air quality is dependent in part on the quality of the emissions used and the associated uncertainties. In an air quality forecasting context, the anthropogenic emissions are usually annual average emissions that are then allocated to each hour based on typical temporal profiles for each source category. Some source categories, such as electric generating units (EGUs), may exhibit significant temporal variations in emissions in response to weather conditions. This study examines the sensitivity of model predictions to such changes in the temporal variations of activity of EGUs. We use the archived forecasted meteorological fields (NCEP WRF-NMM UTC 12z) for the summer of 2007 to drive the CMAQ model retrospectively for different emission scenarios. The analysis will examine the actual temporal variability of emissions from select EGUs, how they differ from the mean seasonally and diurnally adjusted activity data and the differences in the model predictions of ozone based on these different profiles. As part of this effort, we also intend to compare model predictions with observations available from the AIRNOW network.

Dynamical Downscaling of NASA/GISS ModelE: Continuous, Multi-Year WRF Runs

Tanya L. Otte, Jared H. Bowden, Christopher G. Nolte, Martin J. Otte, Jonathan E. Pleim, Jerold A. Herwehe, Greg Faluvegi, and Drew T. Shindell

The WRF Model is being used at the U.S. EPA for dynamical downscaling of the NASA/GISS ModelE fields to assess regional impacts of climate change in the United States. The ModelE fields were included in the IPCC Fourth Assessment Report, and updated science in the improved ModelE will contribute toward the IPCC Fifth Assessment Report. The dynamically downscaled climate fields from WRF ultimately will be used to predict the regional impacts of climate change on air quality and other regional environmental concerns. The WRF model has been successfully linked to the ModelE fields in their raw hybrid vertical coordinate, and continuous, multi-year WRF downscaling simulations have been performed.

The use of nudging for downscaled regional climate simulations has been somewhat controversial over the past several years but has been recently attracting attention. Several recent studies that have used reanalysis (i.e., verifiable) fields as a proxy for GCM input have shown that nudging can be beneficial toward achieving the desired downscaled fields. In this study, the value of nudging will be shown using fields from ModelE that are downscaled using WRF. Several different methods of nudging are explored, and it will be shown that the method of nudging and the choices made with respect to how nudging is used in WRF are extremely critical to balance the constraint of ModelE against the freedom of WRF to develop its own fields.

Categorical performance evaluation of air quality forecasting in Georgia

Yongtao Hu1, M. Talat Odman1, Michael E. Chang2 and Armistead G. Russell1

¹School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA, 30332

²Brook Byers Institute of Sustainable Systems, Georgia Institute of Technology, Atlanta, GA, 30332

The high-resolution air quality forecasting system Hi-Res has been operationally serving the metropolitan areas of Georgia for the past five years. We evaluate Hi-Res's air quality forecasting ability by examining ozone and PM_2.5 as well as PM_2.5 components performance during 2006-2010 using a categorical evaluation approach. The spatial synoptic classification (SSC) calendars for Georgian metropolitan areas are used to cluster the forecasting days into typical weather types of dry polar, dry moderate, dry tropical, moist polar, moist moderate, moist tropical, and a transition category. Observational datasets from SLAMS air quality monitors as well as from SEARCH and ASACA networks are utilized. In addition to the air quality indexes, the forecasting performances for meteorological variables are also examined under these weather types.

Evaluation results show that O₃ performance was worse than average on moist days and better than average on dry days. This suggests that extra attention should be paid to moist tropical days since a sizeable number of the National Ambient Air Quality Standard (NAQQS) exceeding days are observed and expected under this weather type in Georgia. It is also found that PM_2.5 performance during 2006-2008 was worse on dry days, especially on dry tropical days, than on moist days, despite the higher level of PM_2.5 concentration on dry days. The performance on dry days has been improved dramatically since 2009 by integrating a new secondary organic aerosol (SOA) module into the forecasting system. This study will provide helpful information that can be used in assisting local air quality forecasting operations.

Downscaling comparison using Weather Research and Forecast model with between spectral nudging and grid nudging techniques

Peng Liu, Alexandra P. Tsimpidi, Yongtao Hu, Armistead G. Russell,

Athanasios Nenes

In order to investigate how future air quality will respond to different emission scenarios under various climate-change responsive control strategies, high resolution regional meteorology fields that are consistent with future climate predictions need to be developed. Downscaling large scale fields from global climate model (GCM) by applying a regional meteorological model equipped with nudging techniques such as Weather Research and Forecast (WRF) model is a scientific sound choice. Grid nudging has been intensively applied to downscaling reanalysis data for regional air quality modeling of historical episodes. However, spectral nudging may have an advantage over grid nudging in capturing large-scale features of GCM climate predictions^1,2.

In this work, we will investigate which of these two techniques performs better in regards of sticking to GCM modeled climate patterns. We will apply WRF model using NAM analysis data to simulate meteorology fields over North America for four months, each of which represents one season of a year. The modeling domain has a 36-km horizontal grid-spacing, and 35 layers in vertical with the top at 70mb. Comparisons will be conducted in terms of simulated winds and temperature at the surface as well as precipitation predictions. In addition, sensitivity tests will also be carried out to investigate the impacts of parameter configurations, such as the wave-numbers for spectral nudging and nudging coefficients for grid nudging.

References

1. Miguez-Macho, G., G.L. Stenchikov, and A. Robock, Regional climate simulations over North America: Interaction of local processes with improved large-scale flow. Journal of Climate, 2005. 18(8): 1227-1246.

2. Miguez-Macho, G., G.L. Stenchikov, and A. Robock, Spectral nudging to eliminate the effects of domain position and geometry in regional climate model simulations. J. Geophys. Res.-Atmos., 2004. 109(D13).

Improvements to Canadian wintertime particulate-matter forecasting with GEM-MACH15

M.D. Moran, C.A.Stroud, P.A. Makar,_,S. Manard, R. Pavlovic, M. Sassi, P.-A. Beaulieu, D. Anselmo, C.J. Mooney, W. Gong, S. Gong, and J. Zhang

High PM_2.5 concentrations have been observed in North America in all seasons, underlining the need for year-round air-quality (AQ) forecasts. AQ forecasting in the winter, however, poses unique challenges given well-known seasonal variations in emissions, meteorology, chemistry, and removal processes. In the case of PM_2.5, both systematic and episodic overpredictions have been noted in the wintertime for current North American AQ forecast models, including Environment Canada's GEM-MACH15 AQ forecast model.

GEM-MACH15 is the operational regional forecasting configuration of the GEM-MACH multi-scale, in-line AQ model. It is run twice per day at the Canadian Meteorological Centre for a 48-hour forecast over a continental-scale domain with 15-km grid spacing and 58 vertical levels. A simplified 2-bin representation of the PM size distribution is employed, and eight PM chemical components are considered. Input anthropogenic emission files are produced on the GEM-MACH15 rotated latitude-longitude grid from the 2006 Canadian and 2005 U.S. national inventories with the SMOKE emissions processing system. Biogenic emissions are estimated on-line using the BEIS v3.09 algorithms.

Performance evaluations of GEM-MACH15 for several recent winter periods have pointed to several sources of forecast error, including the spatial and temporal allocation of primary PM_2.5 emissions and the treatment of vertical diffusion. This paper will present a comparison of recent speciated PM data from the Canadian NAPS and CAPMoN networks and identify which particle components are being overpredicted in the wintertime. This paper will also review recent improvements that have been made to the GEM-MACH15 modelling system and their resulting impacts on forecast performance.

Accounting for the combined effect of the regional and local-scales on urban pollutant concentrations: comparison between two methods on two case studies in Atlanta, Georgia

Myrto Valari¹, Vlad Isakov²

1 National Research Council Research Associate, US Environmental Protection Agency, NERL/AMAD, RTP NC

2 US Environmental Protection Agency, NERL/AMAD, RTP NC

Pollutant concentrations over urban areas are driven by physical and chemical processes relating to both regional and local-scales. We compare two different modeling efforts as for their ability to account simultaneously for the "regional background" and "local" components of concentration over urban sites. The first method combines concentrations from a regional chemistry transport model (CMAQ) with the output of a local-scale dispersion model (AERMOD) [Stein et al., 2007]. The second method modifies the calculation of the concentration in the regional chemistry-transport model to capture local-scale effects on pollutant concentrations due to heterogenous subgrid scale emissions and dispersion over source-specific surfaces [Valari and Menut, in press]. Results over two case study at the city of Atlanta, GA are presented for one summer and one winter pollution episodes.

References:

Stein A., Isakov V. Godowitch J. and Draxler R., A hybrid modeling approach to resolve pollutant concentrations in an urban area, Atmospheric Environment, 2007, 41,9410-9426.

Valari M. and Menut L., Transferring the heterogeneity of surface emissions to variability in pollutant concentrations over urban areas through a chemistry transport model. Atmospheric Environment (in press)

Dynamical Downscaling of CCSM Using WRF/CMAQ

Yang Gao, Joshua S. Fu, Yun-Fat Lam, John Drake, Kate Evans

Improved observational data of global mean air and ocean temperatures, continued observation of significant ice melting, and a variety of other indicators signal that our climate is warming. This increase in temperature coupled with altered precipitation frequency, ice and snow melting, rising ocean levels, andincreasing air pollution, can significantly alter the natural and human environment. The Community Climate System Model (CCSM) has been used extensively to produce IPCC AR4 projections of possible climate change and is being used to simulate IPCC AR5 scenarios. To study climate change on regional and local scales, dynamic downscaling is a technique to link global and regional models using lateral boundary conditions from global climate models to drive regional models. We couple CCSM4, with high resolution (12km by 12km CONUS and 4km by 4km eastern US domain) in the regional model WRF/CMAQ, and downscale the IPCC RCP4.5 scenario to simulate current and future daily temperature, humidity and precipitation levels for defining heat waves, floods and droughts.

There are some existing issues we need to solve before we can trust the downscaled results. First, how nudging impacts regional climate model (WRF) downscaled results and how strongly we should nudge to keep the global climate model spatial pattern without losing the local characteristics in regional model Second, since CMAQ is a tropospheric model, boundary O₃ concentrations above the tropopause from downscaling could have a large impact to regional air quality model, and the downwash of the high O₃ could lead to too high O₃ concentrations on the surface. This study mainly focuses on solving these two issues.

The Adjoint of ISOYYOPIA II: Continued Development and Preliminary Application

Shannon Capps¹, Armistead Russell², and Athanasios Nenes^1,3

¹School of Chemical & Biomolecular Engineering, Georgia Institute of Technology

²School of Civil & Environmental Engineering, Georgia Institute of Technology

³School of Earth & Atmospheric Sciences, Georgia Institute of Technology

Thousands of model parameters affect the concentration of PM_2.5 simulated by CMAQ. The emissions rates as well as initial and boundary conditions of inorganic species related to sea salt, dust, and anthropogenic sources impact air quality predictions and model experiments. The importance of accurate PM_2.5 concentrations to investigations of health effects and predictions of air quality motivates the optimization of these model parameters. Current approaches allow evaluation of the results from specific values against observations (e.g., tracer method) or prediction of changes due to alteration of these parameters (e.g., CMAQ DDM-3D/PM). With other chemical transport models (e.g., GEOS-Chem), assimilation of satellite, field campaign, or observation site measurements via an adjoint has proven an effective means of efficiently adjusting spatially and temporally varying parameters to improve simulations. The model adjoint produces receptor-oriented sensitivities of PM concentrations to the model parameters of interest, which can be used to optimize agreement between observations and simulations by adjusting model parameters.

The adjoint augments CMAQ in a modular fashion; that is, the adjoint of each CMAQ module must be assembled to achieve this aim. An essential component of the CMAQ aerosol module is the inorganic thermodynamic equilibrium model ISOYYOPIA, the update of which (ISOYYOPIA II) is promised in future releases of CMAQ. ISOYYOPIA II treats the partitioning of inorganic species between the gaseous phase and aerosol associated with sea salt, dust, and anthropogenic pollution (K⁺-Ca²⁺-Mg²⁺-NH₄⁺-Na⁺-SO₂⁴-NO₃-Cl-H₂O). The discrete nature of the adjoint requires following the solution algorithm of ISOYYOPIA II; therefore, adjustments to ISOYYOPIA II that were required for application of an automatic differentiation tool are also outlined. This work demonstrates the capabilities of the adjoint of ISOYYOPIA II by evaluating it against finite difference sensitivities as well as applying it to experimental data. The development of the adjoint of ISOYYOPIA II is an important step toward completion of the adjoint of CMAQ with the capability of treating aerosol-related species.

Air Mass Characterization of Air Quality and Health Impacts under Current and Future Climate Scenarios
Adel Hanna1, Aijun Xiu1, Karin Yeatts1, Richard Smith,1 Zhengyuan Zhu1, Neil Davis1, Kevin Talgo1, Zac Adelman1 Sarav Arunachalam1, Gurmeet Arora1,Qingyu Meng2, Scott Sheridan3, and Joseph Pinto2

1 The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599
2 U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711
3 Kent State University, Kent, Ohio 44242

We investigated the associations among climate, air quality and health outcomes under current and future climate conditions. Meteorological conditions at various locations were characterized by the prevailing daily weather type or air mass pattern. We classified the daily air masses for five cities in North Carolina using the Spatial Synoptic Classification (SSC) system (Kalkstein and Greene, 1997; Sheridan 2002; 2003). This system takes in surface weather data for a given station and classifies each day into one of eight air masses; Dry Moderate (DM), Dry Polar (DP), Dry Tropical (DT), Moist Moderate (MM), Moist Polar (MP) and Moist Tropical (MT). The latter is further expanded to its extremes; the Moist Tropical+ (MT+) and the Moist Tropical ++ (MT++). Analysis with SSC under current climate conditions over ten years (1996-2005), reveals statistically significant increases in asthma hospital admissions with increases in ozone for the DT and MT+/MT++ air masses. Conversely, the MP and DM air masses show a statistically significant negative impact on asthma hospitalizations. For myocardial infarction hospital admissions our results show only the MP air mass to be significantly associated with decrease in hospitalization. It should be noted that the air mass characterization includes all prevailing conditions (e.g. temperature, moisture, pollen, etc.) that may contribute to asthma exacerbation.

Future climate and air quality scenarios were simulated for the months of May, June, July, and August for two periods, 2018 to 2020 and 2048 to 2050. A 10-day (April 21-30) spin-up model simulation preceded the first day of the simulation for each year. Large-scale meteorological conditions were based on Community Climate System Model (CCSM) simulations of those months during the specified years. Regional meteorological conditions were simulated by the Weather Research and Forecasting (WRF) model, and emissions were processed using the Sparse Matrix Operator Kernel Emissions (SMOKE) processor. The meteorological and emissions outputs were used by CMAQ version 4.7 to simulate the predicted air quality parameters for each climate scenario.

We review the climatic and air quality features of the future scenarios in terms of the frequency of air masses and their meteorological and air quality characteristics. We also review the corresponding health outcomes of such future scenarios compared to our findings under current climate conditions.

Development of High-Order DDM Sensitivity Analysis for Particulate Matter in Multidimensional Air Quality Models

Wenxian Zhang, Shannon Capps, Athanasios Nenes, Armistead Russell

Sensitivity analysis is widely used in air quality studies to quantify the responses of simulated pollutant concentrations to model inputs and parameters. The decoupled direct method (DDM) is an efficient approach to perform sensitivity analysis and DDM sensitivities have been applied to conduct emission source impact analysis and to develop potential emission control strategies. DDM sensitivities used in these applications are typically first order, though second-order DDM sensitivities of gaseous species are also available. Due to the known interactions between gases and aerosols, the responses of the concentrations of both the gaseous and aerosol species to the changes in source emissions can be more precisely predicted if high-order sensitivities of particulate matter (PM) are included. In this work, DDM has been extended to calculate high-order sensitivities of PM, and this high-order approach has been implemented in the Community Multiscale Air Quality (CMAQ) model.

High-order DDM sensitivity analysis of PM (HO-DDM/PM) has been applied to a week-long winter episode over the US. The corresponding high-order brute force sensitivities are also calculated. Comparison between the two sensitivities shows good agreement. DDM provides reasonable high-order sensitivities in locations where the brute force results are demonstrably inaccurate. Although the magnitude of high-order PM sensitivities are relatively small compared to first-order ones, they do reflect the non-linear responses of the PM concentrations to the changes in source emissions. For example, both sulfate and nitrate aerosols show noticeable second-order sensitivities to their precursor gases. With the high-order DDM sensitivities of PM, a more detailed source apportionment is then performed.

2006-2008 GEOS-Chem Simulations for CMAQ Initial and Boundary Conditions

Yun-Fat Lam, Joshua S. Fu, Daniel J. Jacob, Carey Jang and Pat Dolwick

In recent years, global atmospheric models (such as GEOS-Chem) have been accepted by the air quality modeling community for generating initial condition (IC) and boundary (BC) conditions in the regional models. One of the greatest advantages of using the output from global models is to allow one to take into account both intercontinental long-range transport of air pollutants and the increase of background pollutant concentrations. The IC/BC generated from the global model gives a regional modeler the ability to use a time and geospatial varied IC/BC during regional simulation, which is found to give a better result than the traditional profile IC/BC. In this paper, we have developed a new GEOSChem2CMAQ mapping table for CB05 with AE5 mechanism to address the changes being made in SOA for the CMAQ 4.7. We performed three years of GEOS-Chem simulations from 2006 to 2008 and generated initial and boundary conditions for the CONUS 12km and CONUS 36km domains. To investigate the effects of boundary conditions, we have also performed CMAQ simulations on the CONUS 36km domain in each of the three years (only two months from each year: January and July are simulated) using 2006 based emissions. Summary of chemical background concentrations along with inter-annual and inter-seasonal variation of boundary conditions will all be presented.

Development of a full adjoint for CMAQ

ShunLiu Zhao and Amir Hakami (Carleton University)

Shannon Capps, Athanasios Nenes, and Ted Russell (Georgia Tech)

Jaroslav Resler (ICS Prague)

Tianfeng Chai and Daewon Byun (NOAA)

Matt Turner and Daven Henze (University of Colorado)

Peter Percell (University of Houston)

Jaemeen Baek, Charles Stanier, and Gregory Carmichael (University of Iowa)

Sergey Napelenok and Rob Pinder (USEPA)

Adrian Sandu (Virginia Tech)

A full adjoint model for CMAQ (CMAQ-ADJ) is being developed as part of a collaborative effort between multiple institutions. This work builds upon the previous version of CMAQ-ADJ (adjoint of CMAQ 4.5.1) that was limited to gas-phase processes. The current effort extends the gas-phase adjoint model to include aerosol dynamics and thermodynamics, cloud processes, and heterogeneous chemistry. It will also include important updates and additions such as parallelization, inclusion of CB05 mechanism, nesting in the backward mode, 4D-Var capabilities, enhanced user interface for various sensitivity analysis applications, improved modularity, etc. The model will be updated to the latest release of CMAQ (expected in Fall 2010 but made available earlier to the development team). Upon completion of the adjoint implementation, the model will be publicly released (expected spring 2011). The presentation includes the approach for CMAQ-ADJ development, status of the adjoint code implementation, and brief overview of partial results. Detailed results from individual groups will be presented separately. Implementation strategies and user control parameters will be discussed. CMAQ-ADJ is being developed with the goal of wide applicability in the modeling community and as such feedbacks are sought as to how users can optimally interact with the adjoint model.

Simulation of the indirect aerosol effect by the two-way coupled WRF-CMAQ over the continental United States

Shaocai Yu, Rohit Mathur, Jonathan Pleim, David Wong,
Annmarie G. Carlton, Shawn Roselle, and S.T. Rao

Atmospheric Modeling and Analysis Division
National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC 27711

Atmospheric emissions resulting from consumption of fossil fuels by human activities contribute to global warming and degrade air quality. The IPCC (2007) concludes that the total direct aerosol radiative forcing is estimated to be -0.5 [ 0.4] W m-2, with a medium-low level of scientific understanding, while the radiative forcing due to the cloud albedo effect (also referred to as first indirect), is estimated to be -0.7 [-1.1, +0.4] W m-2, with a low level of scientific understanding. For a given cloud liquid water content, an increase in the cloud droplet number concentration implies a decrease in the effective radius, thus increasing the cloud reflectivity; this is know as the first indirect aerosol effect. The second indirect aerosol effect is based on the idea that decreasing the mean droplet size in the presence of enhanced aerosols decreases the cloud precipitation efficiency, producing clouds with a larger liquid water content and longer lifetime. In this study, the indirect aerosol effect is estimated with the newly developed two-way coupled WRF-CMAQ over the continental United States. The cloud droplet number concentrations are diagnosed from the activation of CMAQ-predicted aerosol particles of CMAQ simulation. The resulting cloud droplet number is used to calculate variations in droplet effective radius, which in turn allows us to estimate aerosol effects on cloud optical depth and microphysical process rates for indirect aerosol forcing by tying a two-moment treatment of cloud water (cloud water mass and cloud droplet number) to precipitation (the Lin cloud microphysics scheme and an existing radiation scheme in the WRF. With the satellite observation data such as MODIS and CALIPSO, we will evaluate the cloud properties such as cloud optical depth, cloud droplet effective radius, and liquid water content and indirect aerosol forcing in the newly-developed coupled WRF-CMAQ.

Impact of meteorological inputs on surface ozone predictions

Jianping, Huang^1*, Jeff McQueen², Binbin Zhou¹, Brad Ferrier¹, Youhua Tang¹, Marina Tsidulko¹, Ho-Chun Huang¹ Sarah Lu¹, Caterina Tassone¹, Bill Lapenta², Geoff DiMego², Daewon Byun³, Pius Lee³, Daniel Tong⁴, Ivanka Stajner⁵, and Paula Davidson⁶

*Corresponding Author: Jianping Huang, NCEP/EMC, SAIC, 5200 Auth Road, Camp Springs, MD 20746-4304; jianping.huang@noaa.gov

¹Science Applications International Corporation, Camp Springs, MD

²NOAA National Centers for Environmental Prediction, Camp Springs, MD

3NOAA Air Resources Laboratory, Silver Spring, MD

⁴Science and Technology Corporation, Hampton, VA

⁵Noblis Inc, Falls Church, VA

⁶Office of Science and Technology, NOAA/National Weather Service, Silver Spring, MD

Ozone is often over-predicted in coastal areas and for the conditions where high ozone would be expected. In addition to emission and chemistry mechanism limitations, uncertainties of meteorological inputs such as cloud cover and planetary boundary layer (PBL) height could contribute to this bias. The National Air Quality Forecasting Capability (NAQFC) links the National Centers for Environmental Prediction (NCEP)'s Weather Research and Forecasting/Non-hydrostatic Mesoscale Model (WRF/NMM) with the Community Multiscale Air Quality (CMAQ) model, which includes CB05 gas-phase mechanism, to produce experimental ozone predictions. In the NAQFC, cloud parameters and PBL are re-calculated by a pre-processor, PreMAQ, rather than taken directly from NCEP's operational WRF/NMM outputs. In this study, WRF/NMM outputs and CMAQ predictions over the continental USA and several sub-regions for the year 2009 ozone season are examined. We use the NCEP Environmental Modeling Centers (EMC) Forecast Verification System (FVS) to perform comprehensive evaluations of cloud (e.g., cloud cover, cloud water content) and short wave radiation parameters. The observational data include National Environmental Satellite, Data, and Information Service (NESDIS) data, the Clouds from Advanced Very High Resolution Radiometer (AVHYY) Extended (CLAVR-x), and Air Force Weather Agency (AFWA) total cloud cover data. We discuss how cloud prediction errors are correlated with ozone prediction bias. In addition, several CMAQ sensitivity studies are conducted to further investigate the possible impacts of clouds on surface ozone predictions.

Modeling volcanic and marine emissions for Hawaii air quality forecast

Daniel Tong*, Pius Lee, Rick Saylor, Mo Dan, Ariel Stein, Daewon Byun

NOAA Air Resources Laboratory (ARL), Silver Spring, MD 20910

A. Jeff Sutton, Tamar Elias, James Kauahikaua

USGS Hawaiian Volcano Observatory, Hawaii National Park, HI 96718

Xiaoming Liu and Kent Hughes

NOAA Center for Satellite Applications and Research (STAR), Camp Spring, MD

* Corresponding author: daniel.tong@noaa.gov

The National Air Quality Forecasting Capability (NAQFC) extended the forecast domain to include Hawaii in 2009. Modeling air quality over marine environments poses new challenges for forecasters, due, in part, to the unique emission characteristics. Using "conventional" emission data, including inventories of anthropogenic emissions and BEIS-derived terrestrial biogenic emissions, the current forecasting system is only able to reproduce 10% of the typical variability in observed PM_2.5concentrations. Our study presents the development and initial application of emission modeling approaches to account for missing volcanic emissions and marine phytoplankton emissions in the Hawai'i domain. SO₂ emissions from the K+lauea volcano are estimated based on USGS in-situ UV optical absorption spectrometermeasurements, which are further processed at ARL to implement a simple plume rise algorithm. Marine isoprene emissions are calculated by combining NOAA satellite chlorophyll-a observations and emission factors derived from laboratory culture studies. Two CMAQ simulations, one with the new emissions and one without, are conducted to study the effect of including these emissions on air quality forecasting performance by comparing model results with AQS observations over this region. Such mechanisms enable us to include "unconventional" emissions from these natural processes in a near real-time fashion to improve air quality forecasting capacity.

Evaluation of the ACM2 Vertical Mixing Scheme for the Simulated Planetary Boundary Layer in Eastern Texas

Jenna Kolling, Jonathan Pleim, William Vizuete, Harvey Jeffries

This study will evaluate the performance of the Weather Research and Forecasting meteorological model using the ACM2 vertical mixing scheme. The modeling episode provided by the US EPA is from August 1 - October 10, 2006, with a 4 km grid spacing centered over Houston, Texas. Observational data sets are available from both the Second Texas Air Quality Study (TexAQS II) and the Moody Tower TexAQS II Radical Measurement Project. These field studies will allow for a comprehensive spatial and temporal evaluation of ACM2. First, mixing heights were visually estimated from a micropulse Lidar deployed on the roof of Moody Tower in downtown Houston in September of 2006. Second, mixing heights identified from radiosonde balloon data (up to 5 launches per day during high ozone events) are available for August and September of 2006. Lastly, hourly radar wind profiler-derived mixing heights are available for August and September of 2006. This observational data will be used to evaluate the ability of the WRF model with the ACM2 mixing scheme to accurately predict the morning rise, afternoon peak, and evening decay of the planetary boundary layer for different geographic locations in eastern Texas.

Photochemical Modeling of the Ozarks Isoprene Experiment (OZIE): Comparison of MEGAN and BEIS to Field Measurements

Kirk Baker, U.S. Environmental Protection Agency

Annmarie Carlton, U.S. Environmental Protection Agency

Volatile organic compounds (VOC) of biogenic origin play an important role in regional and global scale ozone and particulate matter formation. Two widely used biogenic emissions models, The Model of Emissions and Gases and Aerosols from Nature (MEGAN) and Biogenic Emission Inventory System (BEIS) are employed to generate biogenic emissions for application in the 3-D photochemical grid model CMAQ. Model estimates of isoprene, monoterpenes, and formaldehyde are compared to surface and aloft measurements made during a 1998 intensive monitoring study conducted in the Ozarks (OZIE) in southern Missouri, a large isoprene emitting region. MEGAN emissions resulted in higher secondary organic carbon (SOC) at nearby monitors but model estimates are substantially less than observational estimates. CMAQ using MEGAN biogenics resulted in isoprene, monoterpene, and formaldehyde estimates that tended to be higher and much more variable than observations. CMAQ using BEIS biogenics showed less average error for isoprene and formaldehyde and had a tendency to be under-estimates of observations. CMAQ/MEGAN estimates compared much better to observations when a satellite radiation product is used in place of estimates from a prognostic meteorological model. CMAQ/BEIS showed very little photosensitivity. CMAQ predictions using MEGAN overestimate isoprene at the surface and aloft. Surface formaldehyde is overestimated but aloft formaldehyde agrees well with observations. This suggests that inferred isoprene emissions from formaldehyde satellite retrievals are subject to errors/biases inherent in the gas phase chemical mechanisms.

How sensitive are trace gas concentrations to the method used to parameterize clouds within CMAQ

Christopher P. Loughner, Dale J. Allen, Russell R. Dickerson, Da-Lin Zhang (University of Maryland); Kenneth E. Pickering (University of Maryland and NASA Goddard Space Flight Center); Yi-Xuan Shou (National Satellite Meteorological Center, China Meteorological Administration)

High resolution simulations of a July 2007 mid-Atlantic pollution event during which 8-hour maximum ozone and 24-hour average PM_2.5 concentrations reached 125ppb and 40μg/m³ in northeastern MD were performed with CMAQ driven by a version of WRF that includes an urban canopy model. Comparisons of model output with ground measurements and satellite observations were used to evaluate how well the 13.5, 4.5, 1.5, and 0.5 km horizontal resolution WRF and CMAQ simulations resolve the urban heat island and the Chesapeake Bay Breeze. Since cloud parameters differ between WRF and CMAQ and between the aqueous phase chemistry and photolysis modules within CMAQ, sensitivity simulations were also performed to determine the impact of varying the cloud parameterization scheme within CMAQ on the distribution of clouds and trace gas concentrations.

Evaluation of the coupling of CMAQ with a nitrogen geochemical cycling model on NHx wet deposition

Jesse O. Bash, Ellen J. Cooter, Robin Dennis, Jon Pleim, Megan Gore

U.S. EPA/AMAD

Reactive nitrogen in the atmosphere contributes to particle formation and, via wet and dry deposition, to soil acidification and loss in species biodiversity in sensitive ecosystems. The sources, transport and fate of reduced nitrogen, particularly Ammonia, in the atmosphere remains highly uncertain due the complexity of emissions from agricultural practices and difficulties in measuring ambient concentrations and fluxes at background levels. A soil nitrogen geochemical model with crop specific fertilization scenarios has been coupled to CMAQ v4.7.1 with a bidirectional NH₃ exchange model in the future Midwestern landscapes pilot project. The bidirectional case and a base CMAQ v4.7.1 case was run for a 2002 annual simulation for the Eastern Continental United Sates. Modeled NH_x wet deposition results from the bidirectional exchange and a base version of CMAQ were evaluated against NADP observations. CMAQ with bidirectional exchange introduced a positive bias in annual NH_x observations. However, the biases and spatial correlation in the bidirectional case were largely removed when the wet deposition was corrected for model precipitation biases while the spatial correlation improved but increased the biases in the base case. The bidirectional exchange model also reduced regional and seasonal biases in modeled NH_x wet deposition when the precipitation correction was applied. The results indicate that coupling a nitrogen geochemical cycling model to the bidirectional NH₃ exchange version of CMAQ can improve model performance once precipitation biases are accounted for.

Application and Analysis of Kolmogorov-Zurbenko Filter in the Dynamic Evaluation of a Regional Air Quality Model

Daiwen Kang, S. Trivikrama Rao, Rohit Mathur, Sergey Napelenok, and Thomas Pierce

Dynamic model evaluation examines a retrospective case(s) to evaluate whether the model has properly predicted air quality response to known emission and/or meteorological changes. A good dynamic evaluation case study is the time period between 2002 and 2005 because of the large NOx emission changes that occurred as a result of the U.S. Environmental Protection Agency's (USEPA) NOx State Implementation Plan (SIP) call in addition to a more gradual decreasing trend in mobile emissions. The various spectral components representing different scales of forcing in the time series of atmospheric O₃ mixing ratios can be effectively decomposed using the Kolmogorov-Zurbenko (KZ) filter. Through analysis of the different components caused by different forcings in the dynamic evaluation, we can discern in which component(s) or which scale(s) of forcing the model performs well and in which component(s) or which scale(s) of forcing the model needs further improvement. In this study, we have applied the KZ filter to both the observed and Community Multiscale Air Quality (CMAQ) model simulated O₃ time series in 2002 and 2005. Preliminary results suggest that the CMAQ model responds well to the changes in synoptic forcings from 2002 to 2005. However, the model's responses to baseline changes at most locations within the model domain are not enough compared with the changes in baseline values inferred from the observations. The contributing factors to baseline signals include emissions, boundary conditions, and other slow varying parameters. A detailed analysis and evaluation results of this study will be presented.

Evaluation of European emissions created with a modified version of the SMOKE model

Bieser, J., Aulinger, A., Matthias, V., Quante, M.

A modified version of the SMOKE model, called SMOKE for Europe (SMOKE-EU) was used to create emission data for Europe and surrounding countries. SMOKE-EU uses pan-European public domain datasets and meteorological data to disaggregate officially reported emissions for European countries [1]. Emission data for the year 2000 created with SMOKE-EU were compared to three different emission datasets from widely used European emission models. The emission datasets used for comparison are the official emissions from the European Monitoring and Evaluation Program (EMEP) [2], a dataset created by the Dutch Institute TNO called GEMS [3] and a purchased dataset from the German Institute for Rational use of Energy (IER) [4]. It could be shown that differences of SMOKE-EU emissions compared to the other emission datasets are in the same range as differences amongst them.

Further, concentrations of criteria pollutants calculated with CMAQ4.6 using the four different emission datasets were compared against EMEP measurements with hourly and daily resolution. Using SMOKE-EU emission O3, NO2 and SO4 aerosols could be modelled most reliably. The amount of simulated concentrations within a factor of 2 (F2) of the observations for these species are: O3( F2 = 0.79), NO2 (F2 = 0.55) and SO4 (F2 = 0.62). The lowest values were found for NH4 (F2 = 0.34) and No3 (F2 = 0.25). NH4 concentrations were generally overestimated, leading to a fractional bias (FB) averaged over 22 measurement stations of (FB = 0.83 +- 0.41) while better agreements with observations were found for SO4 (FB = 0.06 +- 0.38, 51 stations) and NO3 (FB = 0.13 +- 0.75, 18 stations). CMAQ runs using the other three emission datasets produces similar results.

The main findings of a regional analysis are that over the Spanish peninsula NO2 and SO2 concentrations were generally underestimated in all four CMAQ runs and also show very low correlations. Generally the least agreement with measurements were found for extremely remote, polar stations which have very low concentrations throughout the year.

The results of this comparison conmfirm that it is adequate to use European emissions created by SMOKE-EU as input for CMAQ or other Chemistry Transport Models.

[1] Bieser, J., Aulinger, A., Matthias, V., Quante, M., Builtjes, P., 2010: SMOKE for Europe - Adaptation, modification and evaluation of a comprehensive emission model for Europe. Geoscientific Model Development (submitted).

[2] Simpson, D., Fagerli, H., Jonson, J.E., Tsyro, S., Wind, P., Tuovinen, J., 2007: Transboundary Acidification, Eutrophication and Ground Level Ozone in Europe Part I Unified EMEP Model Description, EMEP Status Report 2003, Norwegian Meteorological Institute.

[3] Visschedijk, A.J.H., Zandveld, P., Denier van der Gon, H.A.C., 2008: A high resolution gridded European emission database for the EU integrated project GEMS, TNO-report 2007-A-R0233/B.

[4] Friedrich, R., Reis, S., 2004: Emissions of air pollutants, Srpinger, Berlin Heidelberg New York.

Ozone Transport Analysis using Back-Trajectories and CAMx Probing Tools
Greg Yarwood, Susan Kemball-Cook, Bonyoung Koo and Jeremiah Johnson, Jim Price and Mark Estes

Lowering the 8-hour ozone standard raises the importance of transport in contributing to ozone nonattainment. Accurate simulation of ozone transport in photochemical grid models will be critical for the development of effective ozone control strategies. This study evaluated modeled ozone transport in the CAMx photochemical model and used CAMx probing tools to assess transport contributions and their response to potential emission changes.

CAMx was applied for three Texas high ozone episodes from 2005/6 using an updated vertical transport algorithm and the Zhang dry deposition scheme, which is newly implemented in CAMx version 5.21. Periods favorable for the transport of ozone and precursors into Texas were identified from HYSPLIT back trajectories based on EDAS meteorological fields. CAMx ozone performance during transport periods was evaluated at rural monitors within Texas, in adjacent states and in potential upwind source regions for long-range ozone transport into Texas. CAMx simulated ozone at rural sites with good accuracy and use of the Zhang dry deposition scheme combined with the new vertical transport algorithm tended to increase surface layer ozone during transport periods.

The representation of transport pathways from ozone source regions into Texas was investigated by comparing HYSPLIT back trajectories based on EDAS meteorology with back trajectories based on the MM5 data supplied as input to CAMx. The CAMx Anthropogenic Culpability Assessment (APCA) tool and the CAMx Higher Order Direct Decoupled Method (HDDM) were used to provide complementary information on upwind source contributions to Texas ozone. HDDM was also used to evaluate the sensitivity of Texas ozone to potential changes in emissions in upwind source regions and to assess the effect of uncertainty in biogenic emissions on those sensitivities.

Modeling Air Quality Impact of the Deepwater Horizon Oil Spill in the Gulf of Mexico

Daewon Byun^1*, Daniel Tong^1,2, Hyun-Cheol Kim^1,2, Yunsoo Choi^1,2, Pius Lee¹, Rick Saylor¹

and Dong Wu³

¹Air Resources Laboratory, NOAA, Silver Spring, MD

²Earth Resources and Technology, Annapolis, MD

³Aerosol and Cloud Group, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA

* Corresponding author: daewon.byun@noaa.gov

The ongoing oil spill, which began on April 20, 2010 from a well at the ocean floor following the explosion and subsequent sinking of the Deepwater Horizon offshore oil platform, has caused serious environmental concerns. The source of the oil spill is approximately 1 mile below sea level and 40 miles southeast of the Louisiana coast in the Gulf of Mexico. Air quality concerns arise once the oil reaches the ocean surface. Smokes from burning oil gathered on the surface of the water and evaporative emissions from oil spread over the ocean surface can affect air quality. The amount of surface oil and emissions factors for controlled oil burning and evaporative hydrocarbon emissions are quite uncertain. Evaporative emissions are approximated into the area source emissions of hydrocarbons utilizing the NOAA surface oil forecast. The emission factors for crude oil burning over water are compiled by reanalyzing data from previous in-situ ocean oil burning measurements, EPA emission factor documents, and EIA fossil fuel emission report. The simulation results with the bottom-up emissions will be compared with available observations and plume height estimates by the MISR satellite. CMAQ simulations will be performed to assess the impacts of the oil spill on regional air quality with specific focus on ozone, PM_2.5 (particulate matter of aerodynamic diameter less than 2.5 micrometers) and benzene. Simulations with a wide range of emission scenarios will be performed to bound air quality impacts of the spill, subject to the uncertainties in the oil flow rate from the well.

Modeling the impact of ozone-alkene reactions on formaldehyde concentrations during the Texas Air Quality Study 2006

Beata H. Czader, Bernhard Rappengluck, Soontae Kim⁺, Daewon W. Byun^*

Department of Earth and Atmospheric Sciences, University of Houston, Houston, TX 77204, USA

⁺Present Affiliation: Ajuo University, Suwon, S. Korea

^*Present Affiliation: Air Resources Laboratory,NOAA, Silver Spring, MD20910

Formaldehyde (HCHO) is an important indicator of chemical reactivity affecting air quality. Elevated HCHO levels are found in Southeastern US rural areas with large amount of isoprene emissions. In Houston, in particular near the highly industrialized Houston Ship Channel (HSC) area, ambient formaldehyde concentrations can reach very high levels. Formaldehyde can be formed from the chemical reactions of volatile organic compounds (VOCs) and/or directly emitted from mobile and industrial sources. HCHO can not only be formed through the photochemical oxidation of VOCs by hydroxyl radical (OH), but also through the reactions of alkenes with ozone.

This work investigates the impact of ozone-alkene reactions on formaldehyde concentrations during the Second Texas Air Quality Study (TexAQS) in summer 2006 by means of utilizing the Community Multiscale Air Quality (CMAQ) model and its integrated processes and reactions analyses. Photochemical formation from abundant anthropogenic alkenes from industrial sources dominates observed formaldehyde levels during daytime. However, during nighttimes chemical reactions of alkenes with ozone add significantly to the formaldehyde concentrations.

Soil NOx model/satellite measurement intercomparisons

Heidy Plata, Dale Allen, Kenneth Pickering, Thomas Pierce, Sheryl Ehrman

Emissions of NOx from soil are one of the major contributors to ground level NOx, particularly in rural areas. These emission fluxes that arise from microbial activity are strongly influenced by soil temperature, fertilizer application, pulsing following rainfall, biomass burning, and canopy reduction. Accurate estimates of these emissions are necessary to improve air quality models used for the study of regional air pollution as well as global climate change. As part of a larger effort to improve emissions inventories for NOx, we focused on comparisons between tropospheric column satellite measurements of NO2 (using three versions of Ozone Monitoring Instrument (OMI) data: specifically OMI- standard and two different versions of OMI-DOMINO tropospheric NO2 products), estimates of NO2 emissions using BEIS3, and tropospheric columns of NO2 from the Community Multiscale Air Quality (CMAQ) model. We examined satellite observations of NO2 from March, April and May of 2005 and 2006 in the Midwest region of United States. In order to study the NO2 pulse from soil that follows precipitation, we performed spatially and temporally resolved intercomparisons for precipitation events. We focused on events that were not located near population centers, had sufficiently low cloud cover so as to allow satellite observation, and were not significantly influenced by lightning or biomass burning. After applying these screens, we had seven distinct precipitation events in which a significant response in satellite NO2 tropospheric column was observed. In all of these cases, the OMI response is greater than the CMAQ response, suggesting that soil emissions are underestimated. Since some of the cases in March showed satellite response while biogenic emissions from BEIS3 were flat, this suggests that perhaps the crop related biogenic emission enhancements from BEIS3 are not starting early enough in the spring season. Additional analysis of these episodes is underway and will be discussed.

Dynamic Model Performance Evaluation Using Weekday-Weekend and Retrospective Analyses

Jim Smith and Mark Estes

Texas Commission on Environmental Quality

Model performance evaluation (MPE) is commonly accomplished by comparing modeled concentrations for a base period with concentrations measured during the same period. While this is a necessary part of MPE, it does not address a fundamental aspect of the modeling: its ability to predict concentration changes as a function of changes in model inputs. Although critical to assessing the future state of an airshed, model responsiveness is evaluated much less frequently than base-case performance because it is often difficult and resource-intensive to accomplish.

In this paper the authors describe two approaches to dynamic MPE used in the latest State Implementation Plan revision for the Houston/Galveston/Brazoria ozone nonattainment area: Weekday/Weekend Analysis and Retrospective Analysis. In the former, the model's response to changes in emissions between weekdays and weekends (primarily traffic) are compared with observed responses, and in the latter the model is used to predict ozone design values in a prior year, then the modeled year-to-year design value changes are compared with the observed design value changes. In both these analyses, the results suggest that the photochemical model is not as responsive to emission changes as is the real airshed.

Assessing the Anthropogenic Fugitive Dust Emission Inventory and Temporal Allocation using an Updated Speciation of Particulate Matter

George Pouliot, Heather Simon, Prakash Bhave, Daniel Tong,David Mobley, Tom Pace and Thomas Pierce

Crustal materials are mainly emitted by anthropogenic and windblown fugitive dust, but also may potentially include some fly ash and industrial process emissions which are chemically similar to crustal emissions. Source apportionment studies have shown that anthropogenic fugitive dust emissions contribute on the order of 5-20% of PM2.5 (particles with an aerodynamic diameter less than 2.5 um) and 40-60% of PM10 (particles with an aerodynamic diameter less than 10 um) in urban areas that either have been or potentially may be unable to attain the National Ambient Air Quality Standards (NAAQS) for PM2.5 and/or PM10. On the other hand, air quality models suggest vastly higher contributions from current fugitive dust emission inventories, with contributions ranging from 50-80% for PM2.5 and 70-90% for PM10. These estimates are from a Desert Research Institute workshop report from May 2000 that is available from EPA's Technology Transfer Network Clearinghouse for Inventories & Emissions Factors.

This paper uses an improved speciation of the particulate matter to include, in addition to the current PM species, eight trace metals as well as separate non-carbon organic matter to assess potential improvements to the emission estimates of anthropogenic fugitive dust (unpaved and paved road dust, dust from highway, commercial and residential construction and agricultural tilling). Proposed improvements to the inventory include revisions to the temporal profiles and revisions to the estimate of the fraction of emissions that are "transported" from the source region. Revisions to the emission estimates and methodology are modeled using updates to the Community Multiscale Air Quality (CMAQ) model to track these eight trace metals and the non-carbon organic matter. We will show preliminary modeling results of these trace metals and compare them to observed values obtained from available observation networks.

Impact of Lightning-NO Emissions on Eastern United States Photochemistry During the Summer of 2006 as Determined Using the CMAQ Model

Dale Allen, Dept. of Atmospheric and Oceanic Science, UMD-College Park

Kenneth Pickering, Atmospheric Chemistry and Dynamics Branch, NASA-GSFC

Robert Pinder, Atmospheric Modeling and Analysis Division, U.S. EPA

Barron Henderson, Dept. of Environmental Science and Engineering, UNC Chapel Hill

William Koshak, Earth Science Office, NASA-MSFC

Thomas Pierce, Atmospheric Modeling and Analysis Division, U.S. EPA

Lightning-NO emissions are responsible for 15-30 ppbv enhancements in upper tropospheric ozone over the eastern United States during the summer time. Enhancements vary from year to year but were particularly large during the summer of 2006, a period during which meteorological conditions were particularly conducive to ozone formation.A lightning-NO parameterization has been developed that can be used with the CMAQ model. Lightning-NO emissions in this scheme are assumed to be proportional to convective precipitation rate and scaled so that monthly average flash rates in each grid box match National Lightning Detection Network (NLDN) observed flash rates after adjusting for climatological intracloud to cloud-to-ground (IC/CG) ratios. The contribution of lightning-NO emissions to eastern United States NO_xand ozone distributions during the summer of 2006 will be evaluated by comparing results of 12-km CMAQ simulations with and without lightning-NO emissions to measurements from the IONS field campaign and to satellite retrievals from the Tropospheric Emission Spectrometer (TES) and the Ozone Monitoring Instrument (OMI) aboard the Aura satellite. Special attention will be paid to the impact of the assumed vertical distribution of emissions on upper tropospheric NOx and ozone amounts.

Improvements in Emissions and Air Quality Modeling System applied to Rio de Janeiro Brazil

Santolim, L. C. D; Simaes; A. S.; Curbani, F.; Albuquerque, T. T. A; Morais, T. J.; Cavassani, K. N.; F. Frizzera, A. B.; Scardini; Silva, M. L. B.; Viera, M. R; Kohler, L. A; Oliveira, A. M. P; Gonsalves, T. B;

EcoSoft - Consultoria e Softwares Ambientais Ltda:Consulting, Softwares and Monitoring

It was developed an alternative tool to build a spatially (3D geo-referenced framework) and temporally resolved emissions inventory called SIA-ATMOS. All kind of sources arising from the study area are included in the software, i.e. mobile sources, industrial, commercial and residential emissions, open burning, dust resuspension and stack releases. Input data to feed the software include geo-referenced location for all the considered sources, traffic patterns, integrated emission factors for mobile sources, fleet composition, energy generation at local power plants, natural gas burning in residential and commercial places and trash and vegetation burning rates, among others information. Additional to that, an interesting aspect of this work is WRF and CMAQ models coupling into this package tool, which allows users without Linux knowledge to work with these types of models. Thus in the course of this work, it was introduced a routine into SIA-ATMOS which modifies the time independent sparse gridding matrix (GRDMAT) depending on meteorological values. This modified gridding matrix is calculated for each hour of the the year, and therefore enhances the spatial as well as the temporal resolution for certain source sectors. Additionally a vertical distribution using Plume Rise was also introduced. The newly created emissions inventory was used as input for CMAQv4.6. This paper presents a package tool to quantify and evaluate the local air quality. The main aim of the work is to implement a calculation tool to predict air pollutant concentrations and the impact of potential mitigation measures on the local air quality. As an example of the tool application, an emission inventory was created to the Metropolitan Area of Rio de Janeiro (MARJ) based on 2008 year as well as mitigation scenarios. The calculated air concentrations were then compared with measurements from the monitoring stations. A good agreement between predicted concentrations and monitoring campaigns was found. The results showed a stronger contribution in O₃ formation around the petrochemical complex influence area into the MARJ. Afterwards, it was simulated the air quality impact by reducing industrial emissions, decreasing the SO₂, VOC, NO_X, PM₁₀ emissions, according to the best available technology. Reducing the industries emissions, the hypothetical scenarios presented a maximum O₃ reduction of 30%. In some circumstances, such decrease of emissions may lead to 20% of O₃ increases. By the other hand, the NO_X, PM₁₀, VOC and SO₂ concentrations decrease 20%, 10%, 50% and 50%, respectively. The results indicated the CMAQ model can be an effective tool to reproduce the atmosphere behavior and mainly to simulate scenarios to improve the air quality in a complex Mega City like Rio de Janeiro.

Photochemical Model Assessment of PM2.5 Ammonium Nitrate in California: Emissions Sensitivity and Performance Evaluation

Kirk Baker, U.S. Environmental Protection Agency

Heather Simon, U.S. Environmental Protection Agency

Many counties in central California are in nonattainment of EPA's 24-hr average PM2.5 National Ambient Air Quality Standard. Nitrate is often the dominant constituent of PM2.5 mass during wintertime high concentration periods in this region. Nitrate estimates from the Community Multiscale Air Quality (CMAQ) modeling system are evaluated against routine 24-hour and hourly average measurements in California. Previous evaluations of this type have concentrated on short episodes. In this study, the model has been applied for the entire year of 2005 to capture a diverse set of chemical and meteorological conditions. The modeling system consistently underestimates nitrate concentrations: mean nitrate bias at Speciation Trends Network sites for 2005 is -2.1 ug/m3 (-50.0%) and more pronounced during elevated nitrate events (-12.8 ug/m3).

Nitrate concentrations are influenced by a complex set of chemical and physical processes. Possible causes of the model under-prediction are investigated including emissions of nitrogen oxides (NOx) and NH3, inaccurate partitioning behavior due to model bias in temperature and RH, total nitrate under-predictions due to model representation of fog chemistry or wind field, and dilution of modeled nitrate from overpredicted boundary layer heights. Emissions sensitivity simulations where ammonia and NOx emissions are systematically perturbed suggest the central valley of California varies synoptically and sometimes hourly between ammonia and NOx limited formation of PM2.5 ammonium nitrate. Systematic increases in either and both of these precursors rarely improved model underestimates of elevated PM2.5 nitrate. Hourly PM2.5 nitrate ion measurements at Fresno indicate the modeling system captures the larger synoptic variations in PM2.5 nitrate concentration but does not sustain elevated concentrations throughout the day. Average hourly bias at Fresno where observed PM2.5 nitrate ion is > 10 ug/m3 again indicates an underprediction tendency (-5.8 ug/m3) while relative humidity also tends to be underestimated (-9 %) and temperature is minimally biased (0.15 C).

The Contribution of Marine Organic Emissions to Coastal Air Quality

Brett Gantt (1), Nicholas Meskhidze (1), Annmarie Carlton (2)

(1) Department of Marine Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC

(2) U.S.EPA - Office of Research and Development, Research Triangle Park, NC

New emission mechanisms for marine primary organic carbon (OC) aerosols and phytoplankton-emitted isoprene and monoterpenes were implemented in Community Multiscale Air Quality (CMAQ) model Version 4.7. Model simulations were designed to examine the contribution of ocean-derived biogenic volatile organic compounds (BVOC) and primary organic aerosols, i.e., from the mechanical process of bubble bursting, to air quality over the Pacific coast of the US. The CMAQ model simulations were performed for the months of June, July, and August 2005 over a domain including the western US at a horizontal resolution of 12 x 12 km². A combination of remotely-sensed data, laboratory measurements, and model meteorology were used to calculate the marine organic emissions, with marine isoprene and monoterpenes added offline and primary OC aerosol simulated online. Model-predicted atmospheric concentrations of OC aerosols and isoprene were compared to data from coastal stations of the Interagency Monitoring of Protected Visual Environments (IMPROVE) and Photochemical Assessment Monitoring Stations (PAMS) networks. Our results indicate that the inclusion of marine organic emissions resulted in some improvement in the magnitude of model-predicted OC, especially during periods of high contribution of marine organic aerosols to total OC. Marine-source organic aerosols accounted for over 50% of the simulated surface OC aerosol mass over the open ocean. Small, but non-trivial increases in the surface concentrations of ozone (O₃) and particulate matter less than 2.5 μm (PM_2.5) were predicted by the model. For coastal cities like San Francisco, CA, surface O₃ mixing ratio increased by up to 0.2 ppb (0.5%) and PM_2.5 concentrations increased up to 0.3 μg m^-3 (5%). This study shows air quality simulations can be enhanced by the addition of marine organics to current air quality models.

Response of fine particles to reduce in precursor emissions and its implication in Yangtze River Delta (YRD), China

Juan Li, Joshua S. Fu,^* Yang Gao, Kan Huang, Guoshun Zhuang

An integrated MM5-CMAQ modeling approach was used to investigate the fine particles, which were one of the vital factors resulting in the heavy haze in Yangtze River Delta (YRD), China where is hosting 2010 World EXPO, with a focus on the response of fine particles to reduce in precursor emission. The grid domain with spatial resolutions of 3-km was designed for the target region.

Based on simulation and observation results in the whole year run in 2006, nitrate, sulfate and ammonium were main components in fine particles, accounting for 36-49% of the total mass. Therefore, the sensitivity of fine particles to reduce in precursor emissions on the mediate scale might more likely be ascribed to the change of three components.

The responses of nitrate, sulfate, and ammonium to reduction in 20% NOx emission were rather sophisticated. Generally, the simulated concentrations of nitrate, sulfate, and ammonium could slightly decrease in summer, fall, and spring under this condition, while increase with the decrease of NOx emission in cold season, especially in winter. The increase of nitrate concentration is more likely owing to the role of peroxyteldehyde nitrate (PAN) in the formation of nitrate. For it was reported that PAN is a temporal reservoir and carrier of NO₂, PAN is capable to remain for a longer time in the air if temperature is lower, while temperature rise, PAN could release NO₂ with the reversible reaction. It has been seen in this study that, PAN were well correlated with NO₃^- and NH₄⁺ and the sequence of the slopes of the regression lines in four seasons was in the order of winter>fall>spring>summer, which was coincident with the seasonal variation of temperature. These results would indicate that the lower temperature was a favorable situation to form PAN and PAN could lead to release NO₂ and then to form nitric acid while temperature slightly increase. On the other hand, PAN could highly impact on the variation of ammonium. For NH₃ concentration were positively correlated with NO and NO₂, more NO₂ would mean more NH₃. Thus, the released NO₂ from PAN would lead more NH₃ and in turn, more NH₄⁺, as NH₃ is the key precursor in the production of NH₄⁺. Besides, the meteorological condition with higher humidity and low temperature in winter could in favor of the formation of ammonium salts like NH₄NO₃, (NH₄)₂SO₄. In addition, the increase of sulfate concentration in winter was also probably due to the increase of O₃ in the case of reduction in NOx emission.

The responses of nitrate, sulfate, and ammonium to reduction in 20% VOC emission were generally the same in the whole year, i.e. the concentrations of the three species all decreased. Obviously, reduction in precursor VOC emission could directly lead to the decrease in PAN and O₃. Thus, the concentrations of nitrate, sulfate, and ammonium decrease with the decrease of PAN and O₃.

Based on the results mentioned above it seems that the reduction in VOC emission is one of the key measures to reduce the air pollution and visibility in YRD.

Integration Approach Between the MOVES and SMOKE Models

Bok H. Baek and Catherine Seppanen

Institute for the Environment

University of North Carolina at Chapel Hill

Chapel Hill, NC 27599

Marc Houyoux, Alison Eyth, and Rich Mason

Office of Air Quality Planning and Standards

U.S. EPA,Research Triangle Park, NC 27711

Allison DenBleyker, Christian E. Lindhjem and Michele Jimenez
ENVIRON International Corporation
773 San Marin Drive, Suite 2115
Novato, CA 94998

The successful use of the MOVES2010 emission factor calculations for a regional modeling requires careful planning and a clear understanding of emission rates calculation in MOVES2010. To reduce the time and effort required of the user for this process, and to help the user obtain more accurate modeling results, EPA initiated this work to integrate the MOVES2010 and SMOKE models.

The first part of the integration was to develop the meteorological data preprocessor that prepares spatially and temporally average temperatures and relative humidity to provide the meteorological conditions for both the MOVES2010 and SMOKE models. The next part was to develop the MOVES2010 driver and post-processing scripts to help users to setup and run MOVES2010 model efficiently to generate lookup tables of emission rates for all emission processes for SMOKE. The last part was to develop how to process those MOVES2010 lookup tables through SMOKE to create hourly gridded and speciated input files for a regional air quality modeling.

This paper describes how the meteorological data are prepared, so that MOVES2010 and SMOKE can read and use them for their modeling and how the MOVES2010 lookup tables are processed through SMOKE modeling system using average/real-time hourly gridded meteorology data. When user estimates on-roadway running emission processes, county-total VMT and average speed inventory are used as an input to use the 'rateperdistance' lookup table. However, SMOKE requires county-total vehicle population by vehicle type as input to use the 'ratepervehicle' and 'rateperprofile' lookup tables for off-network and parked vapor venting emissions processes, respectively.

Use of Surface Measurements and MODIS Aerosol Optical Depth for Improved Model Based PM2.5 Prediction in the United States

Sinan Sousan^(a,b), Jaemeen Baek^(b), Naresh Kumar^(c), Jacob Oleson^(d), Scott Spak^(b), Greg Carmichael^(a,b), and Charles Stanier^(a,b)

a) Chemical and Biochemical Engineering Department;

b) Center for Global and Regional Environmental Research ;

c) Geography Department;

d) Biostatistics Department.

University of Iowa

Epidemiological studies of the health effects of air pollution and particulate matter have found robust associations between PM2.5 and mortality rates. However, these studies have limited ability to investigate species- or source-specific PM health effects, and suffer from exposure misclassification resulting from the use of central site monitors. Multiple studies have shown that substantial spatial gradients exist, especially for primary pollutants emitted from low elevation sources such as motor vehicles and industrial area sources. Accordingly, PM2.5 exposure estimates for United States at high spatial and temporal resolution are highly desirable to conduct further epidemiological studies focused on identifying more and less harmful types of particulate matter. Data assimilation via optimal interpolation (OI) was used to improve the Models-3 Community Multiscale Air Quality Model (CMAQ) modeling system estimates of aerosol pollution. Model predicted concentrations over North America for 2002 with and without assimilation of MODIS satellite-based aerosol optical depth are compared to PM2.5 measurements for performance evaluation. Effect of various AOD regridding and interpolation schemes on performance statistics is discussed, as is the role of uncertainty in the conversion between PM mass and optical properties. Furthermore, the effect of quality checks and regression with AERONET on reducing bias and noise in the MODIS AOD products is discussed. Our preliminary results show that the best MODIS assimilation parameters may be region and/or season specific. OI improved the PM2.5 simulations in 78% of regions relative to IMPROVE and in 44% regions relative to STN. The OI method will be extended to include the assimilation of PM2.5 ground observations, and used to produce optimal PM2.5 estimates for the time period 2000 to 2004.

Emissions Inventory Development for Fine-Scale Air Quality Modeling

Rebecca 'Lee' Tooly, U.S. EPA, OAQPS, Research Triangle Park, NC

Steve Reid and Neil Wheeler, Sonoma Technology, Inc., 1455 N. McDowell Blvd., Suite D, Petaluma, CA 94954

In the U.S., many state and local agencies are now doing multi-pollutant fine-scale air quality modeling for State Implementation Plan (SIP) attainment demonstrations. The U.S. EPA formed a focus group of emissions inventory developers in agencies that are working to create more locally representative emissions inventories to support fine-scale air quality modeling. The group was invited to share information on the types of problems they are trying to solve and the approaches they are taking to develop the inventory. This paper will discuss the findings of the focus group and emphasize the types of data analysis identified as particularly beneficial to help plan and prioritize the inventory work. Also included will be recommendations on how the group's findings can be translated to the EPA's National Emissions Inventory (NEI). The paper will describe the types of NEI data analysis that can support state and local agencies that want to develop more locally representative emissions data for fine-scale air quality modeling.

Evaluation of the Comprehensive Air Quality Model with Extensions (CAMx) Against Three Classical Mesoscale Tracer Experiments

Bret Anderson, Kirk Baker, Chris Emery

In 2008, the US Environmental Protection Agency initiated an evaluation program of long range transport modeling systems currently used for Class I air quality impact analysis requirements under the Prevention of Significant Deterioration of Air Quality program.

In an extension of that project, the US EPA, US Forest Service, and ENVIRON are evaluatiing the Comprehensive Air Quality Model with Extensions (CAMx) against three classic mesoscale tracer experiments - the 1994 European Tracer Experiment, 1983 Cross-Appalachian Tracer Experiment, and 1980 Great Plains Tracer Experiment. These inert tracer evaluations will help determine the efficacy of the use of photochemical grid models (PGM's) for potential use in long range transport modeling for PSD Class I air quality impact analysis requirements.

Monitoring Air Quality Changes Resulting from NOx Emission Regulations over the United States Using OMI and GOME-2 Tropospheric NO2 Data
K. Pickering, A. Prados, E. Celarier, R. Pinder, S. Kondragunta

EPA's NOx Budget Trading Program has led to substantial emission reductions at power plants in the eastern and central United States. Surface measurements to track the resulting changes in NOx concentrations are not sufficient because of interferences in the measurement and inability of surface monitors to track the transport of NOx aloft. Therefore, we use the tropospheric column NO2 data from the OMI instrument on NASA's Aura satellite (afternoon overpass) and the GOME-2 instrument on the European MetOp satellite (morning overpass) to monitor these changes. In addition, CMAQ model output can be used to attribute the regions where air quality changes are detected by satellite to specific clusters of power plants. We examine changes in tropospheric column NO2 in the summers 2005 to 2009. Large NO2 decreases are found in a swath from northern Ohio and southern Michigan westward to Nebraska, Kansas and Oklahoma. A pronounced region of NO2 increase was found over western and central Pennsylvania from 2005 to 2008, but decreases were found in this region from 2008 and 2009. Nearly continuous decreases were found in the corridor from northern New Jersey to northern Virginia. We compare these ambient air quality changes with the changes seen in the Continuous Emissions Monitoring System data. The sign of the summertime air quality change as detected by satellite was the same as found in the CEMS data in nearly all regions.

Evaluation of an Advanced Reactive Puff Model using Aircraft-based Plume Measurements

Krish Vijayaraghavan¹, Prakash Karamchandani¹, Greg Yarwood¹, Sue Kemball-Cook¹, Biswanath Chowdhury² and Eladio Knipping³

¹ENVIRON International Corporation, 773 San Marin Drive, Suite 2115, Novato, CA 94998

²Sage Management, Princeton, NJ 08540

³EPRI, Washington, DC 20036

We present the application and evaluation of an advanced reactive Lagrangian puff dispersion model that comprises a state-of-the science puff model with an advanced chemistry module that accurately represents the chemistry and evolution of a plume from an elevated point source stack. The model is evaluated using aircraft-based measurements of the constituents of plumes from coal-fired power plants in northeast Texas in 2005. Meteorology and background concentrations for the model are derived from aircraft-based and radar profiler meteorological observations, 3-D modeled meteorology from MM5 and concentration fields from the CAMx air quality model. The aircraft measurements of plume concentrations used for model evaluation were collected during August and September of 2005 as part of an air quality study conducted for Northeast Texas Air Care (NETAC) and include plume traverses at multiple downwind distances from the source. Chemical species evaluated include NO, NO₂, NO_y, SO₂, CO and O₃. Implications of the results of the performance evaluation are discussed.

Assimilation of MODIS Aerosol Optical Depth for Improving CMAQ PM2.5 Simulation

Tianfeng Chai (1,2), Hyun-Cheol Kim (1,3), Daniel Tong (1,3), Pius Lee (1), Daewon W. Byun(1)

1. Air Resources Laboratory, National Oceanic and Atmospheric Administration, Silver Spring, MD 20910

2. Science and Technology Corporation, Hampton, VA 23666

3. Earth Resources Technology, Inc., Annapolis Junction, MD 20701

MODIS satellite observations provide global aerosol optical depth (AOD) data on a daily basis. They can be used to improve model prediction of air quality such as PM2.5 (particulate matter of size less than 2.5 micrometer), whose accuracy is limited by the uncertainties in the model inputs such as emissions and deposition velocities, as well as the deficiencies in physical and chemical processes in the model. In this study, the level-2 MODIS AOD observations from TEYYA platform are assimilated with the Community Multiscale Air Quality (CMAQ) model predictions. Both the optimal interpolation (OI) and variational methods are utilized in the data assimilation tests. The required background error covariance statistics of CMAQ predictions is estimated using the NMC and Hollingworth-Lannberg approaches. Inversion of the covariance matrix is done with the computationally efficient truncated singular value decomposition (TSVD) regularization. To extend the spatial coverage of the satellite data, monthly averaged AOD is used as a replacement for the daily observation field. The standard deviation among different days of the month is used as the observational error field. The effectiveness of such practices is compared with the observation data at the original measurement times. Adjustment of the speciated PM2.5 components is achieved by attributing the observation and prediction differences to the uncertainties in the emission inputs and/or deposition velocities. Implications of such attributions will be investigated and the impact of the data assimilation will be quantified utilizing the AIRNow measurements.

An Integrated Measurement-Modeling Approach to Quantify Contribution of Washington Dulles Airport Emissions to Local Air Quality
Saravanan Arunachalam, Neil Davis, B.H. Baek, Dongmei Yang, Kevin Talgo, Mohammed Omary, Uma Shankar, Adel Hanna, Brian Kim, Jawad Rachami, Roger Wayson, Steven Cliff

Significant advances have been made in estimating emissions from various sources at an airport. However, efforts to quantify the air quality impacts of airports are relatively limited. Given the projected growth of aircraft activity in the next two decades, it is critical to understand and quantify the contribution of airport emissions to local and regional-scale air quality. In this paper, we will present results from measurement and modeling based assessment of the Washington Dulles airport on local air quality. On-site measurements of various gas-phase species, (CO, NO2, O3, SO2, VOCs), PM2.5 components (total mass, sulfate, EC, OC) and various HAPs were made at multiple locations within the airport during 2 campaigns in April 2009, and January 2010, and a third is being planned during July 2010. A rotating drum impactor (RDI) to measure size-resolved PM components, and a SMPS were also employed. A near real-time application of the WRF-SMOKE-CMAQ modeling system at a resolution of 12-km and 4-km was developed and models run in near real-time using all background emissions from anthropogenic and biogenic sources except the airport emissions. Based upon aircraft activity data for Dulles, a detailed emissions inventory of the airport was developed and modeled with the EDMS modeling system, and these emissions were added to the background emissions and CMAQ modeling simulations repeated. We will present results from the 2 completed campaigns along with the air quality assessment based upon the modeling outputs as well as the measurements. We will focus on the evaluation of the PM components, by size, speciation and total mass against the field campaign data as well as against data from routine monitors such as AQS and STN within the 4-km domain.

The Remote Sensing INformation Gateway (RSIG): A Tool for Quickly obtaining Hard to Access Air Quality Data

Jim Szykman, P.I., U.S. EPA Office of Research and Development, Langley, VA

Todd Plessel, Developer, Lockheed Martin, RTP, NC

Heidi Paulsen, Project Manager, U.S. EPA, Office of Environmental Information, RTP, NC

Remote Sensing Information Gateway (RSIG) is an operational web-based tool (www.epa.gov/rsig) that enables researchers to access a variety of distributed environmental datasets in a highly efficient manner, thereby allowing them to focus more on research and less on data processing.

RSIG's objectives are: (1) to quickly provide subsetted/aggregated data, and to create composite geo-referenced visualizations of that data. As an example, downloading and processing several environmental datasets from different sources via FTP could take up to 50+ hours, whereas RSIG can accomplish this task in about 5 minutes, (2) to provide a common framework for the analysis of related datasets, such as re-gridding all data to a common grid, computing derived variables, and so on, and (3) to develop applications of the related datasets that are relevant and responsive to the overall exposure framework established within NERL.

Users can tap into a range of key measurement and model data relevant to air quality research and analysis. RSIG provides the user a direct connection to data sets include Level 1 (L1) aerosol backscatter data from the Cloud-Aerosol Lidar with Orthogonal Polarization(CALIOP) and Level 2 (L2) aerosol and cloud optical depth from the Moderate Resolution Imaging Spectroradiometer (MODIS) instruments. NOAA's Geostationary Operational Environmental Satellite (GOES) Aerosol and Smoke Product (GASP) is also available through RSIG. Later ths year we also anticipate the addition of Level 2 (L2) version 3.01 Cloud-Aerosol Lidar with Orthogonal Polarization(CALIOP) data and L2 Measure Pollution In The Troposphere (MOPITT) CO data.

Other primary non-satellite measurement data includes O3, PM2.5, SO2, NOx, RH, and T data from EPA's Air Quality System (AQS) with an current on-going effort to include Measurement of Ozone, Water Vapor, Carbon Monoxide and Nitrogen Oxides Aboard Airbus In-Service Aircraft (MOZAIC-- http://mozaic.aero.obs-mip.fr) (Restricted access based on MOZAIC data use policy) and the NOAA-EPA Ultraviolet Brewer spectrophotometer (NEUBrew) monitoring network. The observations will include O₃, H₂O, CO, and NO₂ from MOZAIC and surface UV, total column O₃, and O₃ vertical profiles from NEUBrew.

RSIG is also the only system that provides access to a limited set of full 3-dimensional Community Multi-scale Air Quality (CMAQ) model output from EPA.

RSIG also provides a user the ability to regrid data, including satellite data, onto multiple 2-D CMAQ grids and 3-D grids, with 3-D using the CMAQ vertical grid description parameters NLAYS, VGTOP, VGLVLS (sigma pressures), allowing the user to essentially create a Level 3 gridded data product on the fly.

This presentation will provide an overview of RSIG, describe key functions, show the audience how RSIG several use scenarios and provide time for questions individuals may have related to using RSIG for their own work.

Evaluation of Air Quality Models Applied to Wildland Fire Impact Simulation

Fernando Garcia-Menendez, Yongtao Hu, and M. Talat Odman
School of Civil and Environmental Engineering,
Georgia Institute of Technology, Atlanta, GA, 30332-0512, USA

During each of the past 10 years, wildland fires have burned, on average, over 9 million acres of vegetation across the Unites States. Pollutants emitted from wildland fires may be transported and react to form other pollutants, contributing to poor air quality and leading to a diverse set of problems including detrimental health effects. Numerical models can be applied to simulate the effects of wildland fires on air quality and offer a valuable decision-making tool. In this study, we have evaluated and compared the ability of several air quality models with different characteristics to simulate the impacts of wildfires on pollutant levels. We have identified the strengths and weaknesses of each modeling system by simulating past wildfire events and systematically evaluating modeled concentrations with measured smoke data. The models considered include CMAQ, CALPUFF, and DAYSMOKE. CMAQ simulations were carried out with standard, adaptive-grid, and DAYSMOKE-coupled versions of the model. The selection of models also provided a comparison of different modeling approaches: dispersion or photochemical, stochastic or full-physics based, and reactive or inert.

Model evaluation and comparison were accomplished by simulating several selected fire episodes in the Southeast. These involved both prescribed burns and wildfires affecting urban environments, as well as fires studied within field campaigns. The pollutant measurements utilized for model evaluation included observations from monitoring networks and data from field campaigns. Statistical measures such as bias and error were estimated to compare the performance of different models in replicating fine particulate matter and ozone concentrations. However, evaluations were extended to a detailed diagnostic level. The results of each modeled episode were scrutinized to identify the reasons behind each successful or failed simulation. Visualization of simulation results and detailed analysis of individual processes within the models were used extensively towards this end. The results from this study provide insight into the applicability of different smoke models and modeling approaches. Additionally, the results will be further applied to quantify the uncertainties associated with predicted pollution levels in future work.

Evaluation of Modeled Ozone Biases Using Satellite Data and Surface Measurements

Yunsoo Choi12, Daewon Byun1, Pius Lee1, Rick Saylor1, Ariel Stein12, Daniel Tong12, Hyun-Cheol Kim12, Fantine Ngan13, Tianfeng Chai14, Marina Tsidulko5, and Ivanka Stajner6

¹NOAA/OAR/ARL, 1315 East West Hwy, Room 3461, Silver Spring, MD 20910.

² Earth Resources & Technology ,Inc, Annapolis Junction, MD

³ University Corporation for Atmospheric Research, Boulder, CO.

⁴Science and Technology Corporation, Hampton, VA.

⁵ Science Applications International Corporation, Camp Springs, MD.

⁶ Office of Science and Technology, National Weather Service, Silver Spring, MD.

Simulation results from the NOAA National Air Quality Forecasting Capability (NAQFC) for August 2009 are being analyzed to investigate the causes of recurrent summer daytime ozone overpredictions. The NAQFC uses a version of CMAQ (currently based on CMAQ 4.6) adapted for daily production of next day ozone forecasts using the NOAA National Center for Environmental Prediction North American Model (NAM) as the meteorological forecast driver. Over Eastern US, peak summer ozone concentrations have generally been overpredicted, by as much as 5-10 ppbv. Because the overprediction bias was more pronounced during August 2009, we conducted model sensitivity studies to compare results for that month with surface measurements (AIRNow NO₂ and O₃, PAMS isoprene, and SEARCH NO₂, HNO₃ and NO_y) and satellite data (OMI column-integrated NO₂ and HCHO) to investigate the underlying causes. Initial results suggest that overprediction of O₃ in the nighttime boundary layer may be associated with next-day afternoon high biases. Continuing work will attempt to understand the reasons for O₃ overpredictions, including nighttime reactive nitrogen chemistry, simulation of the nocturnal boundary layer and nighttime dry deposition parameterizations. We are also investigating sensitivity to errors in estimated input emissions, which were based on pre-recession estimates of economic activity. The paper will describe the sensitivity studies performed and summarize our findings to-date.

2006 Annual Operational Evaluation of the Environment Canada Air Quality Model System

Jack Chen, Sophie Cousineau, Didier Davignon, Annie Duhamel, Samuel Gilbert, Valerie Manard, Jacinthe Racine, Mourad Sassi, Mehrez Samaali

A modelling platform for the purpose of air quality policy scenario assessments is setup and evaluated for the year 2006 at Environment Canada (EC). The core air quality transport model is the AURAMS off-line regional air quality model. The system is driven with meteorology from the forecast Global Environment Multiscale (GEM) model, and anthropogenic emissions inventory from the latest 2006 Canada National Pollutant Release Inventory (NPRI), the US EPA NEI 2005v4 inventory, and the 1999 Mexican emissions. Anthropogenic emissions are processed using the SMOKE model, and biogenic emissions are calculated online via the BEIS3 system. The annual simulation was conducted concurrently in three sections, each covering a 4-month period. The model is run first on a parent grid covering continental North America at 45-km resolution. Two nested grids covering eastern and western Canada at 22.5-km resolution are then simulated. Evaluation of the model is done on an integrated verification system developed with PostgresSQL and PostGIS database with geospatial capabilities. Model results are compared with routine, high resolution surface measurements from NAPS, and CAPMoN networks in Canada, as well as EPA AQS and IMPROVE networks in the US. The species include O₃, PM_2.5, PM₁₀ and NO₂, that are used to calculate the Canada Air Quality Health Index (AQHI), as well as PM_2.5 component species. The presentation will provide an overview of the modelling platform, the verification system, as well as evaluation results from the 2006 annual simulation.

Assessment of impact of fire emissions during the Second Texas Air Quality Study in summer 2006 with satellite fire observations

Hyun Cheol Kim ^1,2,3, Daewon W. Byun ^1,2, Daegyun Lee ^1,4, Soontae Kim ^1,5, Fong Ngan ^1,2, Warren E. Heilman ⁶, Jesoph J. Charney ⁶, Xindi Bian⁶, Bryan Lambeth ⁷, Shobha Kondragunta ⁸, and Robert Griffin⁹

1. University of Houston, Houston, TX
2. Now in NOAA Air Resources Laboratory, Silver Spring, MD
3. Earth Resources Technology, Inc
4. Now in National Institute of Environmental Research, Korea
5. Now in Ajou University, Korea
6. USDA Forest Service, North Central Research Station, East Lansing, MI
7. Texas Commission on Environmental Quality (TCEQ)
8. NOAA NESDIS, Camp Springs, MD
9. Rice University, Houston, TX

We have utilized the NOAA NESDIS Hazard Mapping System (HMS) data and a fire emission retrieval algorithm developed by NOAA and NCAR to generate wildfire emissions during the Second Texas Air Quality Study (TexAQS-II) in summer 2006. Using the MM5 and the CMAQ for meteorological and chemistry simulation models, we have investigated the impact of fire emissions on regional air quality in the 36 km resolution CONUS domain, and the 12 km resolution domain covering the Houston, TX region. Results show some components of PM simulation can be much improved by the additional fire emissions, demonstrating the need to provide proper fire emissions and accurate simulation of meteorological fields, especially wind and PBL heights. Spatial distributions of carbonaceous aerosols and their ratios to the total PM mass in the eastern Texas region are investigated. Compared to measurements from the previous study, the base case CMAQ OC/PM and EC/PM ratios are lower, implying underestimation of OC and EC, or overestimation of other PM species coming into the Texas region during the simulation period. For the 15 days period, total OC mass transported is 22.3Gg and 30.3 Gg for without and with fire emission, respectively, showing 36% increase in the fire case, and total EC mass transported is increased by 41%. Sensitivity test on the plume rise option is also done by changing vertically allocated amount of fire emissions. Assuming stronger fire plume, by doubling the amount of fire emission going over the PBL height, the OC mass budget showed around 5% decrease in the lower layers.

Evaluating NAM/CMAQ Performance over U.S. West Coast using Multiple Data Sources

Youhua Tang^1,2 (youhua.tang@noaa.gov), Jeffery T. McQueen² (jeff.mcqueen@noaa.gov), Jianping Huang^1,2 (jianping.huang@noaa.gov), Marina Tsidulko^1,2 (Marina.Tsidulko@noaa.gov), Stuart A. McKeen³ (stuart.a.mckeen@noaa.gov), Daewon Byun⁴ (Daewon.Byun@noaa.gov), Pius Lee⁴ (pius.lee@noaa.gov), Ivanka Stajner⁵ (Ivanka.Stajner@noaa.gov) and CalNex 2010 Measurement Teams

1. Scientific Applications International Corporation, Camp Springs, MD 20746, USA

2. Environmental Modeling Center, NOAA National Centers for Environmental Prediction, 5200 Auth Road

, Camp Springs, MD 20746, USA

3. NOAA Earth System Research Laboratory, Boulder, CO 80305

4. NOAA Air Resource Laboratory, Silver Spring, MD

5. Noblis Inc, Falls Church, VA

The CalNex field experiment (http://www.esrl.noaa.gov/csd/calnex) was run during the summer 2010 over California to improve understanding of regional air quality under changing climate conditions. The field data provided extensive above-ground chemical information that we use for evaluation of the operational National Air Quality Forecast Capability (NAQFC) modeling system run at NOAA/NWS/NCEP. The NAQFC includes the North American Model NAM (WRF-NMM) meteorological model coupled to the CMAQ model to produce 48-hour air quality predictions over the Continental U.S. Comparisons with CalNex data complement NAQFC verification already performed in near real-time with EPA AIRNOW surface O₃/PM2.5 data. The use of dynamic CMAQ lateral boundary conditions from the RAQMS (Realtime Air Quality Modeling System) global model was also tested. These dynamic lateral boundary conditions had a strong impact on forecasted ozone in the middle and upper troposphere. We also compared surface versus higher-altitude pollutant biases calculated from aircraft and ozonesonde data. For primary emitted species over polluted areas, this relationship depends strongly on the altitude at which the comparisons are made. Ozone biases at higher altitudes are weakly correlated to surface ozone bias, but strongly related to regional background biases in the model.

Impacts of NOx Emission Reductions in Georgia on PM2.5

Jim Boylan¹, Byeong-Uk Kim¹, Michelle Bergin¹, Jim Kelly¹

¹Georgia Department of Natural Resources (GA DNR)

A number of annual CMAQ simulations (using 2009 emissions) were performed to quantify the impacts of NOx emission reductions in Georgia on annual PM2.5 concentrations in nonattainment areas. Sensitivities to statewide NOx emission reduction (30%, 50%, 70%, and 100%) were modeled and evaluated. The maximum predicted impacts on annual PM2.5 concentrations were approximately 1.2 mg/m³ for 100% statewide NOx removal, 0.4 mg/m³ for 70% removal, 0.22 mg/m³ for 50% removal, and 0.11 mg/m³ for 30% removal. However, the chemical mechanisms embedded in current photochemical air quality models are not designed to adequately characterize the impacts resulting from large statewide NOx reductions (i.e., 100% and 70%) on nearby receptors. Also, the primary method expected for NOx to impact PM2.5 is by changes in nitrate concentrations. However, the statewide 100% NOx "zero-out" run predicts that the majority (up to 85%) of the impacts on PM2.5 concentrations are coming from secondary organic aerosol (SOA), a species with high modeling uncertainties. Comparisons of measured organic carbon (OC) with model-predicted OC show very significant over-predictions by the model at the same hours that the model predicts a large response in SOA concentrations due to reductions in NOx emissions. In addition, a technical analysis of potential very aggressive upper bound state-wide NOx emission reductions in Georgia from mobile (on-road and non-road), point, and area sources show maximum achievable NOx reductions of approximately 31% by 2014 and 47% by 2030. Modeling results for the 30% and 50% NOx reduction runs results in a maximum impact on PM2.5 of 0.22 μg/m³. This value is far less than the lowest proposed Prevention of Significant Deterioration (PSD) significant impact levels (SILs) for direct PM2.5 emissions from a single stationary source (i.e., 0.3 μg/m³). Due to the modeling uncertainties associated with very large NOx reductions on SOA and the low predicted impact of achievable NOx reduction on PM2.5, GA EPD has concluded that NOx is insignificant as a precursor to PM2.5 in the Georgia nonattainment areas.

High Resolution Air Quality Modeling for the Mexico City Metropolitan Zone using a Source-Oriented CMAQ model Emission Inventory and Base Case Model Results

Sajjad Ali¹, Iris V. Cureao², Hongliang Zhang¹, AdrianMaran², Qi Ying^1,*, Humberto A. Bravo^2,*, Rodolfo Sosa²

¹Department of Civil Engineering, Texas A&M University, College Station, Texas, USA 77845-3136

² Centro de Ciencias de la Atmosfera, Seccion de Contaminacion Ambiental, Universidad Nacional Autonoma de Mexico, Ciudad Universitaria, Mexico D.F. CP 04510

^*Co-corresponding authors. Email: qying@civil.tamu.edu(Q. Ying), hbravo@servidor.unam.mx(H. A. Bravo)

The severe air pollution problems in Mexico City Metropolitan Zone (MCMZ) significantly affect the health of approximate 20 million people living in this area and surrounding regions. Proper air pollution control strategies cannot be developed without a proper modeling tool to understand the contributions from different sources and evaluate the effects of emission control strategies.

In this study, a source-oriented Community Multiscale Air Quality Model (CMAQ) is applied to model air quality in the MCMZ and surrounding regions that cover an area of 200x200 km² during six-day air quality episode from March 2-7, 2006 with 1 km spatial resolution. The most updated 2006 anthropogenic emission inventory for the MCMZ was provided by the Mexico City's Secretary of Environment. Point sources outside of the MCMZ area are generally based on the Mexico National Emission Inventory, 1999. Emission rates of NOx and SO₂ from the Francisco Perez Rios power plant, located in the Tula industrial complex, Hidalgo State, are updated to better represent actual emissions in 2006. In addition, SO₂ emission from Popocatapetl, an active volcano 70 km southeast of Mexico City, is also included. The meteorology conditions during this episode are generated using both MM5 and WRF and are compared extensively with observations. Predicted hourly O₃, NO, NO₂, CO, SO₂, PM2.5, PM10 and 24-hour averaged VOC species using different meteorology inputs are compared with all available observations throughout the entire MCMZ region. The sensitivity of the model predictions to the boundary conditions, and emissions from the power plant in Tula and Popocatapetl volcano are also investigated. Back-trajectory analysis was performed using a HYSPLIT to identify the potential source regions of SO₂.

The high resolution emission inventory used in this study is so far the most complete and accurate air quality modeling inventory developed for the MCMZ. It is also the first time that the CMAQ model is applied and extensively evaluated in this area. The CMAQ model, along with the emission and meteorology data, will be used as a foundation for future air quality modeling exercises.

Assessment of VOCs and NO2 photochemical model predictions in U.S. urban areas: potential implications for ozone control strategies

Kirk Baker, U.S. Environmental Protection Agency

Annmarie Carlton, U.S. Environmental Protection Agency

Inadequate representation of the relative mix of VOCs and NO_x by air quality models may preclude appropriate interpretation of O₃ control scenarios. CMAQ predictions of O₃, NO₂, total gas-phase non-methane organic carbon (TNMOC) and isoprene are evaluated with urban-centric PAMS monitoring locations throughout the continental U.S. for 2005. Model estimated TNMOC:NO₂ ratios, which may be an indicator of O₃ formation regime, do not always match the magnitude or spatial and temporal variability in observations. The disparity between model and observed ratios is often related to daytime under-estimates of NO₂, though a combination of modeling system errors is likely. Isoprene is under-predicted throughout the network for most months, hours of the day, and in particular during periods of elevated O₃ concentrations. Errors in isoprene may be partly attributable to poorly characterized urban vegetative land use. Improved urban emissions of NO_X and highly reactive VOCs (e.g., isoprene) would likely improve the modeling system skill in estimating peak O₃ in urban areas. When NO_X and VOC emissions are poorly characterized and predicted O₃ production regimes are not accurate, photochemical models may not correctly describe changes in air quality in response to planned emission reductions. This evaluation indicates that each individual area should do a conceptualization of its particular O₃ problem to develop adaptive and holistic control strategies that change with time and space for a given metropolitan area when developing their air quality management plan or SIP. A photochemical model may appropriately capture the O₃ formation regime if emissions are well characterized, although the addition of measurements to support the modeled assessment is valuable.

The Influence of Short-lived Ozone Precursor Emissions on Radiative Climate Forcing by Ozone and Methane

Land-sea Breeze Effects In Pearl River Estuary

Roger Kwok, Jimmy C.H. Fung, Alexis K.H. Lau

Institute for the Environment, Hong Kong University of Science & Technology

Seasonal variations of air pollutant concentrations in Hong Kong/Pearl River Delta (HKPRD) region have been studied extensively, using meteorological model MM5 and air quality model CMAQ to simulate cases in January, April, July and October of 2004. Further analyses on the model results reveal that local land-sea breeze circulations are also important in mixing and accumulating pollutants originated from within HKPRD cities, when background winds are less dominant. In this presentation, we will look at the land-sea breeze effects with several perspectives, such as day-night contrast in pollutant level, contribution to Hong Kong pollutant level from other PRD cities based on the results of Tagged Species Source Apportionment-implemented CMAQ (TSSA), meteorological conditions prior to land-sea breeze events, etc.

Impacts of Future Climate Change on US Energy Demands and Associated Emissions

Jessica Montanez^1,2; Dan Loughlin²; Cynthia Gage², Bryan Hubbell², J. Jason West¹,
¹Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, NC 27711, USA
²U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA

Future climate change will influence the demands for energy for heating and cooling.  We aim to quantify the impact of climate change on heating and cooling demands in the United States to 2050.  We also aim to quantify the corresponding effects on the energy system and air pollutant and carbon dioxide (CO₂) emissions.  We use the MARKAL (MARKet ALlocation) linear programming energy systems economic-optimization model and U.S. EPA nine-region MARKAL database.    Future climate conditions are based on the Intergovernmental Panel on Climate Change Special Report on Emission Scenarios (IPCC SRES) A2 and B1 scenarios.  Changes in heating degree days and cooling degree days in the residential and commercial sectors are used to estimate changes in energy demands in each of the nine MARKAL regions.  For each scenario, MARKAL estimates the set of energy technologies needed to meet demands at least cost.  We will present estimates of changes in technology penetration with and without climate change, as well as fuel use by type and region, and future emissions of criteria pollutants and CO₂. We will also present estimates of how future energy demand is further affected under different assumptions about future air pollution and greenhouse gas policies.

Impact of Climate-Responsive Controls on Regional Air Quality

Alexandra P. Tsimpidi¹, Peng Liu¹, Yongtao Hu¹,Yuhang Wang¹, Athanasios Nenes¹, Proveen Amar², Armistead G. Russell¹

¹Georgia Institute of Technology, Atlanta, GA

²NESCAUM, Boston, MA

In this work WRF-SMOKE-CMAQ system will be evaluated and applied to investigate how climate-responsive emissions controls and forest management practices will impact future regional air quality. Annual simulations of air quality in current (2010) and future (2050) years will be conducted with business-as-usual base-case emissions and under different climate-responsive emission scenarios. The modeling domain will cover the continental USA and a part of Canada and Mexico using the Regional Planning Organization (RPO) grid which consists of 148x112 grids with 36-kmx36-km horizontal resolution and 15 vertical layers. Meteorological variables fields will be derived from the GISS GCM (Rind et al., 1999), which was applied at a horizontal resolution 4^o latitude and 5^o longitude to simulate current and future climate at global scale. The Weather Research and Forecasting model (WRF, version 3.1.1) will be used to downscale the GISS results to the regional scale using spectral nudging technique. Base case emissions inventory for 2010 will be projected from the 2005 national emissions inventory (EI) by applying EGAS growth factors and control factors derived form federal and local control strategies such as CAIR, MACT and Stage II etc. The similar method will be used to project 2010 EI to year 2020 and ultimately to obtain base case 2050 EI (Woo et al., 2008). Climate-responsive control approaches will be modeled using the EPA MARKAL 9R that based on economic/energy modeling. The emission changes to be caused by these climate-responsive controls will be made through SMOKE modeling and applied both temporally and spatially. The air quality impacts study will focus on the Eastern US, with specific considerations given to the Northeast and Southeast. These two regions suffer from poor air quality, but differ in their emission sources, potential policy approaches to global change (e.g., cellulosic ethanol production in the Southeast) and future climate-derived perturbations. The sensitivities and dominant uncertainties associated with the projected EI and the meteorological fields will also be investigated.

References

Rind D, Lerner J, Shah K, Suozzo R, Use of on-line tracers as a diagnostic tool in general circulation model development 2. Transport between the troposphere and stratosphere, J. Geophys. Res., 104, 9123-9139, 1999

Woo JH, He S, Tagaris E, Liao KJ, Manomaiphiboon K, Amar P, Russell AG, Development of North American Emission Inventories for Air Quality Modeling under Climate Change, J. Air Waste Manage. Assoc., 58, 1483-1494, 2008

In-situ measurement and CMAQ simulation of SO2 over central China

Hao He, Can Li, Russel Dickerson, Zhanqing Li

Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD 20742

A month-long aircraft campaign was carried out over central China in spring 2008 to measure the spatial and temporal distribution of SO₂. We present a numerical simulation for this campaign with CMAQ Version 4.6 and compare those results to aircraft and satellite observations during the campaign. To drive the CMAQ model, WRF Version 3.0.1 and the NASA 2006 INTEX-B emission inventory provide meteorological fields and emission input data, respectively. The CMAQ results show good agreement with both in-situ measurements and the NASA-GSFC OMI SO₂ product. However, the comparison of CMAQ results to vertical profiles form the airborne observations reveals that CMAQ underestimates SO₂ concentration within the planetary boundary layer and overestimates SO₂ concentration in the free troposphere. With the analysis of CMAQ outputs, the gaseous and aqueous phase chemistry of SO₂ is also investigated.

AQMOS: Air Quality Model Output Statistics from CMAQ Model Forecasts

Dianne Miller, Clinton MacDonald, Timothy Dye, Kenneth Craig, and Daniel Alrick
Sonoma Technology, Inc., 1455 N. McDowell Blvd., Suite D, Petaluma, CA 94954

Air quality forecasters routinely use numerical air quality models for forecast guidance when issuing local forecasts. Models used by forecasters include the National Oceanic and Atmospheric Administration's National Weather Service Air Quality Forecast Guidance model for ozone, as well as the BlueSky Gateway Experimental model for ozone and PM_2.5. While these Community Multiscale Air Quality (CMAQ) model forecasts provide useful regional information, like all numerical models, they are still evolving and their site-specific forecasts could be improved. Therefore, the Air Quality Model Output Statistics (AQMOS) tool was developed to provide additional value to forecasters by adjusting the available model predictions with recent observations of ozone and PM_2.5. AQMOS is a web-based software tool that computes on a daily basis regression equations between recent historical air quality model predictions and observations. Each regression equation is model- and city-specific. Each day's equations are applied to the current model predictions for over 300 forecast cities in the AIRNow Program (www.airnow.gov). In this paper, we will describe how AQMOS works and show that AQMOS improves predictions substantially.

Evaluating the Impact of the US Aviation Sector on Exceedances of the SO2, NO2, O3, and PM2.5 Primary Standards

Jeffrey Rissman, Saravanan Arunachalam, J. Jason West

The EPA has announced changes to the standards for three criteria pollutants, SO2, NO2, and O3, while the current PM2.5 standard was revised in 2006. The objective of this work was to determine what effect aviation emissions may have in causing locations around the country to experience exceedances of these four standards. The effects of aircraft on pollutant concentrations were analyzed via four year-long simulations using the CMAQ modeling system. Test cases with and without landing and takeoff cycle (LTO) aviation emissions from 99 major airports in the U.S., as estimated by the NextGen Joint Planning and Development Office, were run for the years 2005 and 2025. It was found that aircraft can be an important source of NO2 in areas which experience large-magnitude exceedances, accounting for between 1.3% and 11.2% of maximum NO2 concentrations in the top four counties by 2025. However, some urban areas with NO2 exceedances experienced a below-average aircraft contribution. Aircraft effects on O3 concentrations are mixed: aircraft often decrease ozone concentrations in urban areas while increasing concentrations near these same areas and throughout the southern and western parts of the country. A reduction in aircraft contribution to O3 may cause modest (<0.50%) reductions in ozone in many exceedance areas while increasing O3 concentrations in urban centers by a larger margin. On average, aviation is responsible for only 0.004% (2005) or 0.001% (2025) of SO2 concentrations during exceedance events, which is one to two orders of magnitude smaller than aviation's contribution to total SO2 concentrations (0.05% and 0.14% for 2005 and 2025 respectively). It is unlikely that aircraft will significantly contribute to most areas' non-attainment of the proposed SO2 standard. For all three pollutants, aircraft will play a more important role in determining concentrations in 2025 than they did in 2005. Results for PM2.5 are not yet available, but PM2.5 will also be discussed at CMAS 2010.

A conceptual framework to represent the subgrid concentration variability (SGV) to compliment CMAQ modeling for use in enhanced exposure assessments

Mohammed A Majeed⁽¹⁾ and Jason Ching ⁽²⁾

¹Delaware Dept. of Natural Resources & Environmental Control, New Castle, DE

²Atmospheric Modeling Division, NERL, ORD, USEPA RTP, NC

Air quality modeling systems (such as CMAQ) can provide an important information base for conducting human health multiple pollutant exposure assessments. For such applications, however, the computational requirements are quite demanding especially given the wide range of time intervals typically needed for the assessments. While acute exposures can take place over short time intervals (hourly to weeks); the time frame for assessing exposures for chronic conditions can extend to periods ranging from annual to lifetimes. Currently, annual runs (of CMAQ) on regional scales based on grid sizes of 12 km at hourly time intervals are now feasible, thus providing a capability to cover a wide range of assessment conditions. However, pollutant concentrations may vary widely within any 12 km grid area, especially in urban and/or highly industrialized areas. Such spatial details can, in principle, be obtained based on running air quality models at finer resolution, or local scale models, or some hybrid combinations of such tools. However, this is achieved at a high computational cost and is clearly impractical especially when considering the broad band of time intervals required for performing chronic health exposure assessments.

We propose to meet this challenge with an approach based on a paradigm of utilizing regional scale modeling and supplementing the within-grid concentration variation distributions as an adjoint to each and all such model grids and on hourly bases. For our approach, we present our conceptual framework and describe in detail a template that features: (1) a practical means to develop and describe SGVs of within-grid modeled concentrations based in terms of (2) analytic functions representing each grid's SGV concentration distribution, and (3) derived online running of regional scale CMAQ models. We illustrate and discuss details and nuances regarding an implementation resulting from applying this template showing results from a case study of benzene for several 12 km grid cells in the Wilmington Delaware area.

A hybrid approach for particulate matter source apportionment: Combining receptor modeling with chemical transport modeling

Yongtao Hu, Sivaraman Balachandran, Jorge Pachon, M. Talat Odman, James A. Mulholland and Armistead G. Russell

School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA, 30332

Receptor-oriented modeling (RM) approaches have been traditionally used for particulate matter (PM) source apportionment. RMs rely on using observed concentrations of the PM composition at receptors to solve a species balance equation that estimates the source impacts, along with however limited knowledge of emissions, e.g. emissions source profiles. Here we develop a hybrid, iterative PM source apportionment approach based on species balances that utilizes a chemical transport model (CTM) equipped with a sensitivity analysis tool to provide physically and chemically consistent relationships between sources and receptors. This hybrid approach enhances RM approaches by adding first principles knowledge through a CTM model. This approach will overcome problems that limit the ability of RMs to identify and quantify source impacts. For example, RMs are based on assumptions that are often not strictly correct as applied such as all sources being included and chemical reactions being minimal. A further issue is that some source profiles are similar, leading to colinearity problems. In contrast, source-oriented modeling (SM) approaches such as CTMs that follow a first principles approach, describing emissions, transport, transformation and loss of chemical species in the atmosphere and estimating ambient concentrations, remove most of such limitations, and provide complementary but critical knowledge to source apportionment.

We apply this hybrid approach to conduct PM source apportionment at Speciation Trend Network sites nationwide with a summer month and a winter month simulations of PM over much of North America. Ambient PM concentrations at the receptor sites are apportioned to thirty two separate sources during both summer and winter days. Source impact uncertainties of the approach are estimated at the same time. The method can be readily applied to large domains and long (such as annual) time periods to provide source impact estimates for health related studies.

Vehicular emission inventory using SMOKE in the Metropolitan Area of Sao Paulo - Brazil

Ynoue¹, R. Y. ; Albuquerque¹, T. T. A.; Nascimento², E. S.

¹Atmospheric Sciences Department/ University of Sao Paulo

²Computer Science Department/University of Espirito Santo

The SMOKE emissions model was applied to build a spatially and temporally resolved vehicular emissions inventory for the Metropolitan Area of Sao Paulo and its surroundings. Mobile spatial distribution was done using satellite nighttime lights map from the Defense Meteorological Satellite Program - Operational Linescan System (DMSP-OLS)¹ as surrogate. This satellite map was used by Martins et al. (2010)² to estimate vehicular density. Emission factors derived from tunnel experiments as described in Martins et al. (2006)³ were used for NOx, SO₂, CO, VOC's and Particulate Material emissions. For NH₃, we used the emission factor estimated by Fraser and Cass (1998)⁴. Two temporal profiles were used, one for heavy and another one for light duty vehicles. These temporal profiles were estimated by the local Environmental Protection Agency (CETESB). Organic gases chemical speciation was done according to findings of Martins et al (2006)³ and particulate material partitioning, as of Sanchez-Ccoyllo et al. (2008)⁵. Vehicular density was recalibrated assuming that, on a diurnal basis average, only 60% of the 9.2 million vehicles fleet is actually on the streets. The total number of vehicles was estimated as approximately 5,5 million in the Metropolitan Area of Sao Paulo in 2008. A comparison with the official vehicular emissions inventory (CETESB, 2009)⁶ was made. Our inventory was very similar for CO and NOx emissions, being only 5 and 17% higher, respectively. For particulate material, our emissions were 20% lower. Larger discrepancies were found for VOC's (less than half of the official) and SO₂ (more than twice of the official emissions).

¹Image and data processing by NOAA's National Geophysical Data Center. DMSP data collected by US Air Force Weather Agency; available at http://www.ngdc.noaa.gov/dmsp/dmsp.html

²J. A. Martins, C. R. Mazzoli da Rocha, M. G. L. Oliveira, R. Y. Ynoue, M. F. Andrade, E. D. de Freitas, L. Droprinchinski Martins: Desenvolvimento de inventarios de emissio de alta resoluo: Intensidade de luzes noturnas e distribuio espacial de veiculos. Submitted to the XVI CONGRESSO BRASILEIRO DE METEOROLOGIA - 13/09/2010 a 17/09/2010 - Belim - PA, Brazil.

³L. D. Martins, M. F. Andrade, E. D. Freitas, A. Pretto, L. V. Gatti, E. L. Albuquerque, E. Tomaz, M. L. Guardani, M. H. R. B. Martins, and O. M. A. Junior. Emissions Factors for Gas-Powered Vehicles Traveling through Road Tunnels in Sao Paulo, Brazil. Environmental Science & Technology, v.40, p.6722 - 6729, 2006.

⁴Fraser, M.P., and G.R. Cass. 1998. Detection of excess ammonia emissions from in-use vehicles and the implications for fine particle control. Environmental Science and Technology 32:1053-1057.

⁵O. R. Sanchez-Ccoyllo, R. Y. Ynoue, L. D. Martins, R. Astolfo, R. M. Miranda, E. D. Freitas, A. S. Borges, A. Fornaro, H. Freitas, A. Moreira and M. F. Andrade: Vehicular particulate matter emissions in road tunnels in Sao Paulo, Brazil. Environmental Monitoring and Assessment, v. 1, p. 1, 2008.

⁶CETESB (2009). Relatario Anual de Qualidade do Ar no Estado de Sao Paulo 2008. CETESB-Companhia de Tecnologia de Saneamento Ambiental, Sao Paulo, Brazil.

Inorganic Aerosols Response to SO2 emissions reductions in the Metropolitan Area of Sao Paulo - Brazil

Taciana T. de A. Albuquerque¹, J. Jason West ², Rita Yuri Ynoue¹, Maria de Fatima Andrade¹

¹University of Sao Paulo - Atmospheric Science Department

²University of North Carolina - Department of Environmental Sciences & Engineering

The Models-3 Community Multiscale Air Quality Modeling System (CMAQ) was used to investigate the spatial and temporal variability of the efficacy of emissions control strategies in the Metropolitan Area of Sao Paulo (MASP), Brazil. In particular, we investigate the response of inorganic aerosols to changes in precursor (SO₂) concentrations. An aerosol sampling campaign was performed during 10 days of the winter 2008 (Aug. 12 - Aug. 22) in MASP to compare with modeling results. Meteorological fields were modeled using the Weather Research and Forecasting model WRFv3.1, for the 10-day period, using three nested domains with 27-km grid resolution (34 x 34 cells), 9-km (52 x 52 cells), and a high resolution domain of 3-km (109 x 76 cells). Only the 3-km domain was aligned with the CMAQ domain, which covers the most polluted cities (Campinas, Sorocaba, Sao Jose dos Campos and Cubatao) surrounding the metropolitan area. CMAQ modeling simulations were conducted over 306 hours from 11 August to 23 August, 2008. The SMOKE model was used to provide input files to CMAQ, based on vehicular emission factors and information relating the number of vehicles and urbanized area from satellite images. The urbanization mapping was essential to simulate emission inventories with SMOKE model. The CMAQ and SMOKE domains consist of 102 x 69 grid cells with 3 km horizontal spacing and 20 vertical layers. The air quality simulations use measured concentrations as initial and boundary conditions. Aerosol processes and aqueous chemistry in CMAQ (AERO4) were used, as well as the Carbon Bond V gas phase mechanism. Three different scenarios were simulated considering the current emission inventory, a reduction of 50% of SO₂ emissions, and a scenario considering no SO₂ emissions. Preliminary results show that between the different scenarios at measurement stations, SO₂ concentration was seen to vary substantially as SO₂ emissions changed, but PM_2.5 showed much less variation due to the slow conversion of SO₂ to sulfate and the contribution of other PM2.5 species. Focusing on the system of sulfate-nitrate-ammonium, the smaller Aitken mode, which represents fresh particles either from nucleation or from direct emission, showed a mass concentration response after SO₂ emissions reductions, while the larger accumulation mode, which represents aged particles, showed very little change in concentration. Evaluating the whole domain there are many places where the inorganic aerosols and PM_2.5 concentrations changed more than at the measurement sites inside the MASP. We aim to fully quantify the responses of PM2.5 to changes in SO2 and explain these changes by chemical responses in the sulfate-nitrate-ammonium system. We also aim to use these simulations to evaluate the efficacy of reducing SO₄, NH₃ and NO_xemissions on PM2.5.These results indicate that a combination of controls on SO₂ with the precursors gases (NH3 and NOx) of the inorganic aerosols, may be more effective at reducing PM2.5 concentrations.

Development of an open biomass burning emissions modeling system for Asia using BlueSky Framework

Ki-Chul Choi, Jung-Hun Woo, Bok H. Baek, Meongdo Jang, Seung Heon Yoo

Open biomass burning is an important contributor to air pollution in Asia region. Estimation of fire emissions, however, have been problematic primarily because of uncertainty in the size and location of sources, and of their temporal and spatial variability. Having more comprehensive tools to estimate wildfire emissions, which can characterize their temporal and spatial variability, are needed. An emissions processing system that can generate speciated, gridded, and temporally allocated emissions is needed in support of Models-3/CMAQ air quality modeling over East-Asia. For these reasons, we have developed a biomass burning emissions modeling system based on satellite imagery in order to better take into account spatial and temporal emissions distributions. The BlueSky Framework, which was developed by USDA Forest Service and US EPA, was used for our Asian biomass burning emissions modeling system. We focus on not only Asia but Siberia to assess the Siberian fire impacts. Fire information and vegetation map are derived from MODIS satellite imagery. The models used for this study were the Fuel characteristic classification System(FCCS), CONSUME, and Emissions Production Model(EPM). We integrate BlueSky into SMOKE modeling system to support air quality modeling and analysis in Asia region. The results will be presented at the conference.

An anthropogenic emissions processing system for Asia using SMOKE

Jung-Hun Woo, Ki-Chul Choi, Bok H. Baek, Meongdo Jang, Young-Il Ma

Air quality modeling is a useful methodology to investigate air quality degradation in various locations and to analyze effectiveness of emission reduction plans. A comprehensive air quality model usually requires a coordinated set of emissions input of all necessary species. It is important, therefore, that these emission fields correctly reflect spatial and temporal characteristics of emission sources. A comprehensive emissions modeling has not been conducted for Asia which is a major contributor to global anthropogenic emissions. For these reasons, we have developed an anthropogenic emissions processing system for Asia in support of air quality modeling and analysis over Asia. The SMOKE(Sparse Matrix Operator kernel Emissions) system, which was developed by U.S. EPA and has been maintained by the Carolina Environmental Program(CEP) of the University of North Carolina, was used to develop our emissions processing system. A merged version of INTEX 2006(Zhang et al., 2009) and TRACE-P 2000(Streets et al., 2003) inventories was used as a starting point Asian emissions inventory. Inventory reclassification for both geographical regions and energy use activities was conducted to extend the original inventory. The IDA(Inventory Data Analyer) format, a major SMOKE-ready emissions format, was used to create SMOKE-ready emissions. Source Classification Codes(SCCs) and country/state/county (FIPS) code, which are the two key data fields of SMOKE IDA data structure, were created for Asia. US EPA's MIMS(Multimedia Integrated Modeling System) Spatial Allocator software, along with many global and regional GIS shapes were used to create spatial allocation profiles for Asia. Temporal allocation profiles and chemical speciation profiles were partly regionalized using Asia-based studies. We expect the result of this study can provide better air quality modeling inputs, which can be act as a major step to improve our understanding of Asian air quality.

A Fertilizer Emission Scenario Tool for CMAQ (FEST-C)

Ellen Cooter

USEPA/ORD/NERL, Research Triangle Park NC

LiMei Ran, Verel Benson and Qun He

UNC Institute for the Environment, Chapel Hill, NC

The 2011 CMAQ release will contain a research option for modeling regional scale bi-directional ammonia flux in which current emission estimates for inorganic commercial N fertilizer are replaced by a dynamic, time-varying estimate calculated within CMAQ. Results of a bi-directional ammonia flux pilot for the Eastern U.S. CMAQ domain suggest that agreement between simulated and observed estimates of atmospheric nitrogen deposition may improve with the implementation of the bi-directional flux approach. CMAQ estimation of ammonia emissions associated with the use of inorganic N fertilizer requires information regarding amount, timing and depth of fertilizer application. Preliminary estimates of regional inorganic N fertilizer application was provided to the pilot study, but full CMAQ implementation for the CONUS requires that more detailed regional agricultural land use and land management information be provided, and that the overall input estimation process be further refined and automated. An added capability to simulate plant demand-based fertilizer input response to scenarios of near-term interannual weather variability and land management change as well as future land use, land cover, land management and climate changes will allow quick screening of such scenarios so that those that best characterize client-driven regional-scale science questions can be identified for full CMAQ simulation. Simulations using this coupled regional agricultural land management/CMAQ system will support assessments of the implications of such scenarios for air quality and the provision of ecosystem services related to clean air, clean water, food, fiber and bio-fuel feedstocks.

The Fertilizer Emission Scenario Tool for CMAQ (FEST-C) has been developed to estimate regional inorganic N fertilizer application rates and application timing and to output these estimates as a CMAQ-ready input file. Options currently available in FEST-C will be described and examples of CMAQ input variables as well as a variety of additional edge-of-field biogeochemical FEST-C output variables will be displayed. Plans for evaluation and future enhancements will also be described. The tool will be available for download from CMAS at the time of the 2011 CMAQ release.

Modeling chemical composition of wind-blown dust particles and comparison with speciated PM measurements in the United States

 

Mo Dan, Daniel Tong, Daewon Byun

NOAA Air Resources Laboratory (ARL), Silver Spring, MD

Wind-blown dust particles, an integral constituent of the modern atmosphere, exert variable effects on air quality, human health and global climate. Dust particles are typically modeled as an unbreakable model species in atmospheric models, making it difficult to use such model results in specific applications (e.g., health impact assessment, ocean fertilization, or climate forcing), or to compare the results directly with speciated particulate matter (PM) observations. By combining seven year (2001-2007) PM chemical component data observed from the IMPROVE and CSN networks, together with satellite data and fingerprint element ratios, we identify cases influenced by local dust sources, as well as sources outside the United States. We then compose region-dependent chemical profiles for wind-blown dust that can be used to better represent dust aerosols in CMAQ or other models simulating dust storms in the United States. By modifying the dust emission model, FENGSHA, we split the bulk dust emissions into nutrients, crustal elements, trace metals, and others. Using a modified version of CMAQ4.6, we conduct several air quality simulations to reproduce the observed dust events as recorded by the monitoring networks. The model output is then compared with monitoring species directly. This effort, together with the further speciation of “PM Other” from anthropogenic sources, provides a unique opportunity to reduce model uncertainties associated with dust source emissions.

Application of the SMOKE-CMAQ system to understand air quality problems over Seoul Metropolitan Area, South Korea

Soontae Kim, Nankouyng Moon¹, Jung-Hun Woo², Kweon Jung³, Jun-Bok Lee³, Cheol Yoo⁴, Daegyun Lee⁴

 

Division of Environmental, Civil and Transportation Engineering Ajou University, Suwon, South Korea

¹Division of Strategic Assessment, Korea Environment Institute, Seoul, South Korea

²Department of Advanced Technology Fusion, Konkuk University, Seoul, South Korea

³Seoul Metropolitan Government Institute of Health & Environment, Seoul, South Korea

⁴National Institute of Environmental Research, Incheon, South Korea

Emission control measures to improve air quality over Seoul Metropolitan Area (SMA) in South Korea have been implemented recently.  Air quality trends over the region show that ambient PM₁₀ concentrations has been decreased but still higher than the national ambient air quality standard (50 μg/m³). Ozone and oxidant concentrations keep increasing even though the SMA is considered as a NOx-rich area due to heavy traffics and no large VOC-emitting facilities in the area.  To elucidate the air quality problems and to help prepare efficient control measures, air quality simulations using the SMOKE-CMAQ system were introduced.  To prepare CMAQ-ready emissions, Korean national emissions inventory from the National Institute for Environmental Research and INTEX-B emissions data were processed with SMOKE according to the coverage of target domains while biogenic emissions were estimated through BEIS3 and MEGAN depending on vegetation data availability.  In this paper, we discuss QA/QC procedures to check the uncertainties in Korean emissions inventory after comparing CMAQ results to the observations, and then present emission sensitivity test results to explain source contributions over the region.

REALITY: Road Emission Activity-Link based InvenTorY: A dynamic on-road pollutant emissions model

M. Lebacque

The objective of this paper is to introduce a new emission processing model called REALITY, (Lebacque, 2010). REALITY stands for Road Emission Activity-Link based InvenTorY. The function of this program is to create customized mobile source emission inventory ready to be used in AQMs (Air Quality Models) such as the model developed by the research team of CEREA called POLYPHEMUS, (Mallet et al, 2007), or it can be used as an stand alone model. The model can calculate pollutant emissions on a network as a function of traffic volumes that vary due to variations in velocities on the roads by time of day.

REALITY assimilates the output data from any dynamic assignment model such as the model of dynamic assignment developed by the research team of LVMT, called LTK, (Aguilera and Leurent, 2009). The output of the dynamic assignment should at least contain the latitude and the longitude of the entry and the exit points of a link, average traffic flows, and average speeds on links for specified time steps. In addition, the model needs some supplementary data such as fleet composition, weather information (temperature, wind intensity and direction, and humidity level), and speed profiles on each link per unit of time. The problem of multi-dimensionality of data is addressed and a new solution is proposed using the technique of Random Matrix Theory, (Bouchaud, et al. 2005). The problem of applying correction factors is addressed and a solution is proposed using statistical method of bootstrapping and confidence interval calculation, (Frey et Zheng, 2002).

Two new algorithms will be integrated into the model REALITY in the near future. DYNABURBS stands for (Dynamic Assignment for Suburbs). This algorithm calculates dynamic assignment for the case of trip changing, and includes parking option. Trip chaining refers to the number of stops between an origin and a destination due to non-work activities such as dropping kids off at school on the way to work, or grocery shopping, or a visit to doctor, or a dash to drugstore, or some recreational activity. Trip chaining includes parking either on side streets or in indoor, or outdoor parking garages. The outcome of this model is used in calculating cold start emissions. DYNABURBS is specially developed for urban networks, (excluding all highways and expressways). The model finds a dynamic assignment for a given number of origins and destinations.

APOLARIS stands for (Atmospheric Pollution Activity-Road Initiated Source). This model is developed to calculate pollutant concentration and speciation on each link of an urban network for each time step. Linking the output of this model which is at high spatial and temporal resolution with a low resolution AQM models will give better estimates of pollutant concentration and speciation.

A Novel Parameterization of the Marine Primary Organic Aerosol Emission for Regional and Global Models

Nicholas Meskhidze, Brett Gantt

Department of Marine Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC

In this study new formulation for marine primary organic aerosol emission function was developed using aerosol chemical composition observations from coastal stations in the Interagency Monitoring of Protected Visual Environments (IMPROVE) network and satellite-derived ocean and meteorological parameters. Natural aerosol emission sources over the remote oceans are some one of the largest uncertainties for air quality and climate change. Although sea salt is recognized as a major component of marine aerosol mass flux over regions with high wind speeds and/or when other aerosol sources are weak, recent studies highlight important role of organic carbon (OC) aerosols in marine environment. Many organic compounds accumulate at the seawater/air interface, either due to their low solubility or active transport by sub-millimeter size bubbles to the surface. Bubbles that burst in the presence of the microlayer can become considerably enriched with organic compounds in the aerosol relative to bulk seawater concentrations which is often observed over the productive waters of the ocean.

Studies suggest that the OC mass fraction of sea spray could be a function of diverse oceanic parameters such as the concentration of dissolved organic carbon ([DOC]), particulate organic carbon ([POC]), chlorophyll-a ([Chl-a]), the type of OC, the concentration of surfactants in the ocean microlayer, and the path length, size and "age" of bubbles in sea water. Multi-variable regression analysis conducted in this study (for 2000-2007) suggested that DOC and the surface wind speed (U₁₀) were likely the key variables in determining the OC fraction of sub-micrometer sized aerosols arriving at the coastal sites via marine trajectories. New ([DOC]-based) parameterizationpredicts that primary marine organic aerosol emissions may not be limited to high productivity oceanic regions only (as suggested by the existing ([Chl-a]-based) formulations)and could comprise large portions of the coastal and open oceans. New [DOC]-based emissions parameterization is designed to run online in existing regional and global models. Offline versions with different special resolutions have also been created.

Development of a High-Resolution Mercury Emission Inventory in Canada

Xin Qiu(1), Sophie Cousineau (2), Daniel Figueras(2) and Scott Penton(1)

(1) Novus Enviornmental Inc.

(2) Environment Canada

To assess the impact of changing mercury emissions on Canadian ecosystems, Environment Canada requires accurate emissions of Canadian anthropogenic sources of mercury for 2006 at high spatial resolution. To accomplish this goal, several works have been accomplished in early 2010: review of the Canadian mercury inventory for 2006 and establishment of a detailed work plan to achieve temporal and spatial allocation of mercury emissions in both the horizontal and vertical dimensions (based on the latest Canadian landuse data), 15-km Canada-wide spatial allocation of non-point sources of mercury emissions using the UNC spatial allocator tool, temporal profile analysis to temporally allocate mercury emission on a monthly basis, gridding of emission files for all point and non-point sources in the inventory using the SMOKE emission processing system, conversion of the SMOKE emission files to GRAHM model ready input files in RPN standard file format, andtesting of the mercury emissions for consistency and mass balance. As a result of the work, the Canadian Mercury emission inventory has been considerably improved in its quality, spatial and temporal resolutions.

The Development and Uses of EPAs SPECIATE Database

Heather Simon⁺, Lee Beck^*, Prakash Bhave⁺, Frank Divita^#, Ying Hsu^#, Deborah Luecken⁺, David Mobley⁺, George Pouliot⁺, Adam Reff^&, Golam Sarwar⁺, and Madeleine Strum^&

⁺Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, US Environmental Protection Agency, Research Triangle Park, NC 27711

^*US EPA, National Risk Management Research Laboratory, Research Triangle Park, NC 27711

^#E.H. Pechan and Associates, 3622 Lyckan Parkway, Durham, NC 27707

^&Office of Air Quality Planning and Standards, Environmental Protection Agency, Research Triangle Park, NC 27711

SPECIATE is the U.S. Environmental Protection Agency's (EPA) repository of volatile organic compounds (VOC) and particulate matter (PM) speciation profiles of air pollution sources. These source profiles can be used to (1) provide input to chemical mass balance (CMB) receptor models; (2) associate the factors derived from ambient measurements by multivariate receptor models (e.g., factor analysis and positive matrix factorization) with emission source signatures; (3) interpret ambient measurement data; and (4) create speciated emission inventories for regional haze, climate, and photochemical air quality models. Here we combine information from the SPECIATE v4.2 database with the National Emissions Inventory (NEI) to provide estimates of source-specific emissions for individual VOC compounds. We present the major sources of compounds with high ozone-formation potential and major sources of toxic VOCs. In addition we provide several examples of applications of this database.

Inverse Modeling of NH3 Precursor Emissions Using the Adjoint of CMAQ

Matthew Turner and Daven Henze

Air quality models are utilized by the US EPA to develop emission control regulations for air quality improvements, but air quality models contain large amounts of uncertainty. Developing a constraint on emissions using the 4D-Variational data assimilation technique is one method of reducing the uncertainty in sources of aerosols. In an effort to provide more accurate predictive capabilities, the current CMAQ adjoint model will be updated to include an adjoint of aerosol dynamics. Inverse modeling will be performed on the model domain to constrain spatial variability and seasonal cycles of NH3 emissions. Since the initial conditions provide a small influence when compared to emissions over the course of several months, inversions for each season and year can be performed separately.
A combination of speciated aerosol concentrations taken from a network of air quality surface monitoring stations, and satellite measurements taken from the TES and OMI instruments aboard the Aura satellite will be used. Ammonium measurements will be taken from the CASTNet of monitoring stations, which provides weekly average measurements of sulfate, ammonium, and nitrate. Tropospheric NH3 vertical columns from the TES instrument aboard the Aura satellite will also be used.

Emissions Uncertainty: Focusing on NOx Emissions From Electric Generating Units

Emily Wisner
Statistics Department, North Carolina State University, Raleigh NC

David Mobley and George Pouliot
Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, Environmental Protection Agency, Research Triangle Park, NC 27711

Emission factors are important for estimating and characterizing emissoins from sources of air pollution. An emission factor is a representative value that attempts to relate the quantity of a pollutant released into the atmosphere with an activity associated with the release of that pollutant. These factors are usually expressed as the weight of pollutant divided by a unit weight, volume, distance, or duration of the activity emitting the pollutant (e. g., kilograms of particulate emitted per megagram of coal burned). Such factors facilitate estimation of emissions from various sources of air pollution based on: pollutant class, type of combustion, and fuel source. In most cases, these factors are simply averages of all available data of acceptable quality, and are generally assumed to be representative of long-term averages for all facilities in the source category (i. e., an estimated population average).

The objectives of this presentation are to: (1) Verify the AP-42 NOx emission factors from combustion sources from Electric Generating Units (EGUs) with currently available continuous emission monitoring data; (2) Develop quantitative uncertainty indicators for the EPA's A through E rated emission factors on NOx emissions from combustion sources. We found that the AP-42 emission factor values were accurate for some SCCs (Source Classification Codes), which accounted for over two thirds of our data. However, in general, the AP-42 values were not accurate for over half of our SCCs. We were also able to quantify the uncertainty of the AP-42 letter grades.

Assessing Internal Variability and Internal Grid Nudging in the WRF Model for Regional Climate Modeling Applications

Jared H. Bowden, Tanya L. Otte, Christopher G. Nolte, Benjamin T. Wells, Jonathan E. Pleim, Jerold A. Herwehe, Russell Bullock, Martin J. Otte

Determining a model configuration for regional climate modeling requires testing multiple model options which can become an overwhelming task when considering that climate statistics are needed for proper evaluation. A potential option is to use a smaller integration period, but the model internal variability, i.e. model results sensitive to initialization or domain size, may influence the robustness of the simulation when testing multiple model configurations. This study is a first attempt to demonstrate the WRF model sensitivity to internal variability by conducting an ensemble of annual simulations initialized at different times driven at the lateral boundaries with NCEP/DOE AMIP II data.

Understanding the internal variability can provide robust estimates in model configuration. This study examines the options available for interior gird nudging to keep the large scales in the model interior consistent with the driving data. The sensitivity to internal grid nudging is examined using annual simulations initialized at different times in combination with a base (no nudging) case, analysis (grid point) nudging and spectral nudging simulations. For each internal grid nudging technique, we explore the strengths and limitations of each approach including the mean error and the potential to suppress the model variability at certain length scales.

Improving Cloud Impacts on Photolysis Using anOn-Line Radiation Model in CAMx

Chris Emery, Jaegun Jung, Jeremiah Johnson, Greg Yarwood, Sasha Madronich, Georg Grell, and Doug Boyer

Clouds strongly influence tropospheric oxidant chemistry by altering photolysis reaction rates. The CAMx photochemical grid model has treated cloud effects on photolysis reaction rates using a parametric approach developed for the Regional Acid Deposition Model (RADM), based on radiative transfer model (RTM) calculations for idealized scenarios. Improvements in RTM efficiency and computer speed make it feasible to embed an RTM within a photochemical grid model to provide "on-line" cloud adjustments for photolysis reaction rates. In this paper, we describe improvements on two related fronts: (1) modifying the Weather Research and Forecasting (WRF) mesoscale model to specifically output sub-grid cloud parameters from the Grell cumulus model

Assessing the impact of bi-directional ammonia transport on nitrogen fertilizer emissions and fate in the Eastern U.S.

Megan L. Gore, Ellen J. Cooter, Robin Dennis, Jon Pleim, Jesse O. Bash, Viney P. Aneja

Atmospheric ammonia (NH₃) plays a role in the formation of particulate matter and can have adverse effects at elevated concentrations on terrestrial and aquatic ecosystems via wet and dry deposition. Large uncertainties exist in quantifying NH₃ emissions, particularly in the agricultural sector, and modeling subsequent environmental processes. A pilot study assessing bi-directional NH₃ transport using the Community Multi-scale Air Quality (CMAQ) Model was completed to develop and test bi-directional flux algorithms, explore methods of providing agricultural fertilizer information into CMAQ using a dynamic soil emission potential component, and clarify possible NH₃ and overall one-atmosphere chemical budget changes. Two 2002 annual simulations, a bi-directional and a base CMAQ v4.7.1 case (i.e., uni-directional), were run over the eastern continental United States.

Preliminary analysis indicates that the soil and canopy flux in the bi-directional simulation have a spatial pattern similar to that of the base fertilizer emissions. Soil flux values were several times larger than the base fertilizer emissions, which were reduced by canopy uptake to give a canopy flux that is higher but comparable to the base emission values. The largest difference between the two simulations was considerably higher values in the Upper Midwest in the bi-directional case. Additionally, dry deposition was less overall and wet deposition was greater in the warmer months in the bi-directional run, with the largest differences for both also occurring in the Upper Midwest. The results indicate the sensitivity of deposition to fertilizer emissions, and the need for accurate spatial and temporal representation. Full implementation of the bi-directional flux option is planned for the 2011 CMAQ release.

Recent Updates to the Visualization Environment for Rich Data Interpretation (VERDI)
Qun He1, Donna Schwede2, Kirk Baker3,Todd Plessel4, Tommy Cathey4, Liz Adams1, Mary Ann Bitz5, Nicholson Collier5

1 Center for Environmental Modeling for Policy Development, Institute for the Environment, University of North Carolina at Chapel Hill, Chapel Hill, NC
2 Atmospheric Modeling and Analysis Division, U.S. Environmental Protection Agency, Research Triangle Park, NC
3 Air Quality Analysis Division, U.S. Environmental Protection Agency, Research Triangle Park, NC
4 Lockheed Martin Information Technology, Research Triangle Park, NC
5 Argonne National Laboratory, Decision and Information Sciences Division, Argonne, IL

An initial version of the Visualization Environment for Rich Data Interpretation (VERDI), an open-source Java tool for visualizing the results from the Community Multiscale Air Quality Model (CMAQ) and associated programs, was released in 2007. Development of VERDI has continued and an overview of the new features is presented. Batch scripting capability from both the VERDI GUI window and the command line has been included in the current release (version 1.3). Multiple plots with various predefined configurations can be drawn using a single batch script file without invoking the VERDI GUI. VERDI's new batch scripting will automate the production of graphics used for model evaluation. Multiple observational datasets can be overlaid on the Fast Tile Plot to facilitate comparisons between CMAQ and observational data from different networks. VERDI's new remote file access capability allows users to run VERDI locally on their PC and to access NetCDF data from remote computer servers through ssh connections. Users can use the Areal Interpolation Plot to map gridded data to polygon areas such as watersheds and census tracts. VERDI is supported via the CMAS center and has a website (http://www.verdi-tool.org/) which provides information about the Java tool as well as software and documentation downloads. A SourceForge repository was created for source code version control and distribution. Contributions to the development of VERDI by the user community are encouraged.

The Implications of Acetone Condensation on Remote Conditions
Henderson, B.H.; Pinder, R.W.; Mathur, R.

The Carbon Bond chemical mechanisms, as implemented in CAMx and CMAQ, have condensing assumptions that lead to high-biased organic nitrates and enhanced ozone. Carbon Bond lumps alkanes, acetone, and higher ketones together into the model species PAR. PAR's OH reactivity and organic nitrate yield is calculated as the weighted average of 23 alkanes measured in Los Angeles in 1988. In urban conditions, the reactivity profile of organics support this assumption. In the upper troposphere and rural lower troposphere, acetone concentrations are high and are a substantial fraction of VOC reactivity. Compared to alkanes, acetone has a low yield of organic nitrates. When used in hemispheric simulations, organic nitrates are over-produced and accumulate. The organic nitrates have a low reactivity and act as a long-range NOx reservoir (local sink and remote source). Correcting the organic nitrate yield for acetone, increases tropospheric NOx lifetimes, but decreases remote ozone concentrations. The Carbon Bond condensation assumptions may have little influence on urban simulations, but should be kept in mind when researching continental, hemispheric, global, or preindustrial conditions.

Restructuring of the CMAQ Aerosol Module

Steve Howard, Prakash Bhave, Jeff Young, Sergey Napelenok, and Shawn Roselle

As the CMAQ model increases in complexity and functionality, certain components of its modular design are being challenged.  In particular, the CMAQ aerosol module has become cumbersome to modify because all of the aerosol species are hard-coded across various sections of the FORTRAN code.  Jiang and Roth (Models-3 Users’ Workshop, “Development of a Modularised Aerosol Module in CMAQ,” 2002) described the development of a new aerosol module with process-level modularity and implemented it as an alternative to the AERO2 module in CMAQ v4.1.  Since that implementation, the scientific algorithms within the CMAQ aerosol module have been enhanced and the overall demands on the entire modeling system have grown.  With the aerosol algorithms changing rapidly, there have been few opportunities to pause scientific development and focus on code structure.

In this poster, we will describe a major effort to restructure the aerosol module code in preparation for the next public release of CMAQ.  This effort borrows ideas from the earlier work by Jiang and Roth (2002) and uses Fortran 90 constructs to encapsulate all aerosol-related parameters in a manner that enhances modularity.  The primary objective of this restructuring effort is to facilitate the addition (or removal) of aerosol species from the CMAQ model in a manner that requires minimal changes to the source code.  Our code changes have negligible effects on the model results and will be portable to several variants of the CMAQ model that are in the public domain (e.g., standard configuration, multi-pollutant model, sulfate tracking, primary carbon apportionment, DDM).

Development of National Air Quality Forecasting Capability at 4 km horizontal resolution for the CONUS

Pius Lee^1*,Fantine Ngan², Hyuncheol Kim³,David Wong⁴, Daniel Tong³,Tianfeng Chai³, Yunsoo Choi³, Daewon Byun¹, Rick Saylor¹, Ariel Stein³,

Youhua Tang⁵, Jeff McQueen⁶,Marina Tsudlko⁵,Jeff Young⁴,Ho-Chun Huang⁵,Sarah Lu⁵,

Catarina Tassone⁵, Ken Carey⁷, and Ivanka Stajner⁷

^{Corresponding Author Address: Pius Lee,}

^{1 NOAA/OAR/ARL, 1315 East West Hwy, Room 3461, Silver Spring, MD 20910.}

^{2 University Corporation for Atmospheric Research, Boulder, CO.}

^{3 Earth Resources Technology Corporation, Annapolis, MD.}

^{4 EPA, Research Triangle Park, NC.}

^{5I.M. System Group, Inc., Rockville, MD.}

^{6 NOAA/NWS/National Centers for Environmental Prediction, Camp Springs, MD.}

^{7 Noblis, Inc., Falls Church, VA.}

The National Air Quality Forecasting Capability (NAQFC) is providing 48 hour forecasts of surface ozone concentration and its corresponding daily one-hour and 8-hour maxima for the contiguous 48 states (CONUS). The NAQFC numerical modeling system operationally links the National Centers for Environmental Prediction (NCEP) North American Mesoscale Model (NAM) with CMAQ. Both NAM and CMAQ are running at 12 km horizontal resolution. Currently, the NAM consists Weather Research and Forecasting) Non-hydrostatic Multiscale Model (WRF-NMM). However the Earth Systems Modeling Framework (ESMF) based National Environmental Modeling System (NEMS-NMMB) will replace the current NAM with multiple nested grids in the near future, one of which will be 4km horizontal resolution grid covering CONUS. In anticipation of this planned upgrade of the meteorological model, studies are being made to evaluate requirements and performance of CMAQ running at 4km horizontal resolution over CONUS. A prototype version of NAQFC on a finer horizontal grid is used to investigate advantages of the improved spatial resolution of emission inputs. Spatially refined representation of emission inputs for CMAQ can be generated for all the relevant pollutants using geo-spatially based tools such as the Spatial-Locator, but require a considerable data processing effort. Therefore, this study focuses on the impact of spatial refinements in the representation of point and mobile source emissions. The impact of finer resolution biogenic emissions are not included in this initial first step since these are expected to have less variation on fine spatial scales. Finer resolution representation of mobile and point sources allows CMAQ to capture different NO_x/VOC ratios across various chemical regimes and degrees of urbanization. This study reinforces the importance of accurate emission modeling. Comparison of 4km and 12 km CMAQ forecasts for a late Spring and a Summer case were performed to understand the impacts of improved spatial resolution of emissions on CMAQ simulations. Finer spatial resolution of mobile emissions is shown to be essential for capturing the correct chemical make-up of air pollutants in urban and suburban air-sheds.

New Tools and Updates in the Spatial Allocator for Meteorology and Air Quality Modeling
Limei Ran, Uma Shankar, Ellen Cooter, Aijun Xiu, Neil Davis

Center for Environmental Modeling for Policy Development
Institute for the Environment, University of North Carolina at Chapel Hill
137 E. Franklin St. CB #6116
Chapel Hill, NC 27599-6116

The Spatial Allocator (SA) is a set of geospatial tools we have developed over the past few years to help users manipulate and generate geospatial data files related to emissions and air quality modeling. The tools are designed to support some of the unique aspects of the file formats used for Sparse Matrix Operator Kernel Emissions (SMOKE), Weather Research and Forecast (WRF), and Community Multiscale Air Quality (CMAQ) modeling. The SA tools use GIS shapefile files in vector format, IOAPI and WRF NetCDF files, image data in different raster formats, and plain text data files as input and output data. The current release of the SA contains three parts of tool sets: Vector, Raster and Surrogate Tools. Originally, SA contains only the Vector Tools developed in C programs for emission surrogate computation, Biogenic Emissions Landuse Database, version 3 (BELD3) re-gridding, surf zone computation for sea salt emission, and some other spatial overlay and attributes allocation functions. In order to reduce the tedious work to create a script file for each surrogate computation and to make surrogate computation and surrogate quality control easy and consistent, we developed the Java-based Surrogate Tools to generate and quality control all surrogates at one program run using text format CSV input tables. The Java tools will create each script file to call the Vector Tools for surrogate computation.

In recent years, with the increasing interests in using satellite data we added the Raster Tools to the SA even though it is completely independent software tools from the SA Vector and Surrogate Tools systems. The tools were first developed to process 2001 National Land Cover Database (NLCD) and Moderate Resolution Imaging Spectroradiometer (MODIS) land cover data for WRF modeling. Since then, with supports from NASA ROSES and US EPA projects we have enhanced it to process MODIS cloud and Aerosol optical depth (AOD) products and OMI AOD and NO2 products at both swath (L2 ) and gridded (L2G and L3) level as well as Geostationary Operational Environmental Satellite (GOES) data. We also developed a tool in the SA Raster Tools system to convert mesoscale model meteorology data for the Environmental Policy Integrated Climate (EPIC) fertilizer modeling and we are developing a tool to convert EPIC fertilizer output for CMAQ ammonia bi-directional flux modeling. In addition, we have developed tools in the Atmospheric Model Evaluation Tool (AMET) to compare gridded satellite data with CMAQ output. We will demonstrate gridded satellite images, some satellite data comparisons with CMAQ output, and output from EPIC fertilizer modeling. The goal of this presentation is to show the new tools developed in the new release of the SA Raster Tools system and important updates in the Vector Tools.

Fine resolution modeling with CMAQ-adjoint

Jaroslav Resler, Krystof Eben, Pavel Jurus

Effects of uncertainties in vertical mixing schemes on ozone simulations

Wei Tang^a,*, Daniel S. Cohan^a, Gary A. Morris^b, Daewon W. Byun^c

^aDepartment of Civil and Environmental Engineering, Rice University, 6100 Main Street MS 317, Houston, TX 77005, USA; Email: Wei.Tang@rice.edu

^bDepartment of Physics and Astronomy, Valparaiso University, Valparaiso, IN 46383, USA

^cAir Resources Laboratory, NOAA, Silver Spring, MD 20910,USA

Vertical mixing is an important physical process in the numerical modeling of air pollutants. Uncertainties resulting from different formulations of vertical diffusion schemes and from associated parameterization errors in eddy diffusivity and dry deposition velocities may have substantial impacts on modeling results. In this study, new capabilities of conducting sensitivity analysis to dry deposition velocities and eddy diffusivity were incorporated into the CMAQ (Community Multiscale Air Quality)-HDDM (High-order decoupled direct method) model to quantify the influence of parametric uncertainty in these parameters. Evaluations of simulated ozone concentrations under two vertical diffusion schemes, EDDY and ACM2, were performed using ground, aircraft, and ozonesonde measurement data from the Texas Air Quality Study II to quantify the influence of structural uncertainty in vertical mixing schemes. The results indicate that the structural uncertainty generated from two vertical diffusion schemes and parametric uncertainties in dry deposition velocities and eddy diffusivity are not leading contributors to discrepancies between modeled and observed ozone data. However, the HDDM modeling shows that these structural and parametric uncertainties in vertical mixing may have significant effects on ozone sensitivities to nitrogen oxides (NO_x) and volatile organic compound (VOC) emissions.

CMAQ Redesign

Jeffrey Young

CMAQ has undergone a redesign of the method by which the model utilizes its species specifications, which includes the gas-phase, aerosols, non-reactive and possibly, inert tracers. The model species specification information in the current approach, which has been used since the beginning, is provided through compiling the model with Fortran include files. These include files are stored separately along with the other code in the CVS archive. With the many mechanism versions of CMAQ, the management of these include files is burdensome. The CMAQ redesign implements the species specifications by means of Fortran Namelist files, which replace the include files but are generated in a somewhat similar fashion. Each Namelist replaces as many as a dozen include files. The model reads the Fortran Namelist files at run time.

This approach eliminates the need to maintain a wide variety of include files with the CMAQ code and enables more flexibility. If desired, some simple, careful changes to the model species dispositions can be made directly to the Namelist files without recompiling the model. Testing the redesign against a base version shows some differences, which are generally small. The run times are about 2% longer with the Namelist approach for a one-day, 12-Km Eastern US run on the SGI Altix machines.

Evaluation of a 2006 Annual CMAQ v4.7.1 Simulation for the Continental United States using 12-km Grid Spacing, WRF Meteorology and GEMS Boundary Conditions

K. Wyat Appel, Shawn Roselle, Brian Eder, Daiwen Kang, Thomas Pierce, Kenneth Schere, S.T. Rao and Stefano Galmarini

The Air Quality Model Evaluation International Initiative (AQMEII) is a model evaluation project involving numerous research groups from North American and Europe with the goal of advancing the way regional scale air quality modeling systems are evaluated.  As part of the AQMEII project, the Atmospheric Modeling and Analysis Division (AMAD) of the United States Environmental Protection Agency (USEPA) has performed an annual 2006 CMAQ simulation for the continental United States.  This CMAQ simulation is unique from previous CMAQ simulations performed by AMAD in the past for several reasons.  First, the simulation was performed over a single domain that covers the entire continental United States (CONUS) and a large portion of Canada using 12-km by 12-km horizontal grid spacing.  In the past, two separate simulations covering the eastern and western United States have been used instead of single, continuous domain.  Second, the simulation utilizes meteorology provided by the latest version of the Weather Research and Forecasting (WRF) model.  Previous CMAQ simulations have typically utilized meteorology provided by the 5^th Generation Mesoscale Model (MM5) or performed CMAQ simulations using WRF meteorology for only a few months of the year, so this simulation represents a significant increase in the use of WRF meteorology.  Finally, the CMAQ simulation utilizes boundary conditions provided by the Global and regional Earth-system Monitoring using Satellite and in-situ data (GEMS) product.  As the name suggests, the GEMS product is a global reanalysis product that combines both modeled data and observations (both surface and satellite) to provide data for meteorology and atmospheric gases including greenhouse gases, global reactive gases and global aerosols.  The performance of the CONUS CMAQ simulation will be compared to available surface and upper-air (e.g. ozonesondes) air quality observations and to previous CMAQ simulations for the same time period in order to 1) assess the overall operational performance of the simulation and 2) determine how and possibly why the results of the simulation differ from previous CMAQ simulations.  The results of these comparisons will be presented.

Establishing Boundary Conditions for Bogota's Air Quality Model Using CHIMERE

Juan Pablo Aristizabal, Nestor Y. RojasRojas(1), Alain Clappier, NadageBlond(2),

(1) Air quality group. Chemical and Environmental Enginnering Departament, Universidad Nacional de Colombia, sede Bogota. Carrera 45 No 26-85, Ed. 412 Of. 206. Bogota D.C.

Bogota's air quality model (AQM), based on a meteorological Finite Volume Model (FVM) and a Transport and Air Pollution Model (TAPOM), was developed in 2004 for the District's Secretary of Environment. Unfortunately, the model has had little use in assessing the effectiveness of air quality improvement policy measures.

As an effort to have better results and a continuous use of the model we focused this job in obtaining boundary conditions. The WRF (Weather Research and Forcasting) model was used to obtain the meteorological boundary conditions and the European model CHIMERE to simulate the atmospheric photochemical processes. All simulations were run over a mesoscale domain. Results of these simulations were compared with measurements of several weather and air quality monitoring stations.

Simulations made with WRF and CHIMERE showed that both models are able to set the boundary conditions for the TAPOM model. We compared qualitatively the results of WRF simulations as the first approach and the results at mesoscale were accurate. CHIMERE simulations compared against records from weather and air quality monitoring stations showed that, in all cases, measurements were higher than the values obtained trough simulation for primary pollutants and lower for secondary pollutants (e.g. ozone). This was mainly due to the location and type of the air quality monitoring stations, which had a strong influence of nearby sources.

Evaluation of EDMS / AERMOD at Los Angeles International Airport
Neil Davis, Kevin Talgo and Saravanan Arunachalam

EDMS was developed by the FAA for regulatory modeling of airport impacts on air quality, and makes use of the AERMOD dispersion model. This research was undertaken to perform an in-depth analysis of these tools with a specific focus on the plume parameters, as recent research has been undertaken to better diagnose these plume parameters with use of LIDAR data. We used the Los Angeles International Airport Air Quality and Source Apportionment Study recently performed for Los Angeles World Airports (LAWA). As part of this study, a detailed emissions inventory was developed for the LAWA airport using EDMS and a measurement campaign for the summer of 2008 was undertaken, with a focus on PM2.5 and several gas-phase species in and around the LAX airport. This paper will focus on modeling the LAX airport emissions using EDMS/AERMOD and evaluating the model against observations from the 2008 field campaign.

The first part of our evaluation focused on AERMOD's ability to replicate a LIDAR scan which was taken at LAX in May 2001. To facilitate this comparison, we set up a receptor grid perpendicular to the runway, which was in line with the LIDAR scan. We then examined vertical cross-sections (from AERMOD using flag-pole receptors) along this line to determine how the plume parameters affected the shape of the plume in this cross-section, and which settings best replicated the LIDAR scan provided.

The second part of our evaluation looked at AERMOD's performance compared to surface measurements recorded in and around LAX during the 2008 measurement campaign, for both hourly and daily average concentrations. Using these results, we will comment on EDMS/AERMOD's ability to model aircraft sources and to match the field campaign data.

Impact of boundary layer processes on Taiwan's air quality simulation

Fang-Yi Cheng

Taiwan is located in Eastern Asia and surrounded by the oceans. The central mountain range is the principal range of mountains in Taiwan, and runs from north of the island to the south. The peak of the mountain is about 4000 m. The complicated topography makes the difficulty for meteorological model to capture the local circulations in the boundary layer. To understand how the boundary layer processes affect the mixing and transport processes of air pollutants, the Community Multiscale Air Quality (CMAQ) modeling system and Weather Research Forecasting (WRF) model were used to study effect of boundary layer processes on air quality simulations.

The WRF model is simulated with various PBL parameterizations and land surface models (LSM) to generate sets of meteorological fields. CMAQ sensitivities are performed with different meteorological fields. The emission input is fixed for all the CMAQ sensitivities.

The sensitivities of WRF to various PBL and LSM schemes were compared with meteorological observations to assess the model predictive capability in terms of the mixing and transport fields. The preliminary results show apparent difference in the wind fields and PBL height. The CMAQ sensitivities are compared with air quality monitoring networks and simulation results also show substantial difference of the NOx and VOCs concentrations.

The goal of this study is to explore the capability of the mesoscale meteorological modeling and air quality modeling on simulating the vertical mixing, local transport; capturing the meteorological factors such as the wind, temperature, boundary layer height, momentum and heat flux components as well as enhancing the understanding of boundary layer processes on Taiwan's air quality simulation.

Developing CMAQ for Many-Core and GPGPU Processors
George Delic,
P.O. Box 569, Chapel Hill, NC 27514-0569

The steps in porting CMAQ to current generation many-core or multi-thread commodity computer architectures are analyzed and described. The systems considered include both many-core commodity processors (MCCP), and general-purpose computing on graphics processing units (GPGPU). The behavior of vendor-supported compilers is evaluated with a view to discovering the opportunities they provide in the porting steps. One purpose of this analysis is to uncover the prospects for workload throughput improvement for CMAQ when ported to such systems. Another purpose is to itemize the changes required in CMAQ to reap the benefits that these recent parallel programming paradigms offer. In this analysis the three gas chemistry solvers (EBI, Rosenbrock, and Gear) are compared to explore potential benefits when utilizing such systems. A case study uses the thread-safe version of the CMAQ Rosenbrock solver reported on in the previous year's CMAS conference. It is anticipated that preliminary results will be available with both MCCP and GPGPU systems for several full 24 hour episodes on a 279 X 240 Eastern US domain at 12 Km grid spacing with 34 vertical layers. An interpretation of CMAQ behavior on such many-core and multithread systems will be discussed and prospects for enhancing CMAQ workload throughput summarized. This work is performed by HiPERiSM Consulting, LLC as subcontractor to Computer Sciences Corporation, under U.S. EPA SES3 Contract GS-35F-4381G BPA 0775, Task Order 1522.

A comparison of CMAQ-based and observation-based statistical models relating ozone to meteorological parameters

Jerry Davis, USEPA Office of Air Quality Planning and Standards; Bill Cox, USEPA Office of Air Quality Planning and Standards; Adam Reff, USEPA Office of Air Quality Planning and Standards; and Pat Dolwick, USEPA Office of Air Quality Planning and Standards

The statistical relationships between ambient ozone and ambient meteorology was compared with the statistical relationships between CMAQ simulated ozone and input meteorology from the MM5 meteorological model.  Comparisons were made for 74 cities over the Eastern U.S.for the ozone season (April 1 to September 30) during a five-year period (2002-2006).  The ozone data consisted of the daily maximum 8-hr average; the main meteorology covariates were the daily maximum temperature, the daily average relative humidity, the average morning wind speed (7 to 10 AM) and the average afternoon wind speed (1 to 4 PM).  The natural log of the ozone values was used as the response variable.  The same covariates were used for both regression models.  The approximate linear effect on the ozone response variable was defined as the difference between the model predicted ozone at the 75% quartile and the 25% quartile of the meteorology covariate divided by the difference between the two percentiles. The standard errors of these statistical- model generated effects are obtained by bootstrap re-sampling of the original data 200 times followed by a complete refitting of the data.  R-squared values above 60% were obtained for the majority of the locations in the analysis for both the ambient and CMAQ statistical models.  Covariate-by-covariate comparisons were made based on the natural cubic spline output from the two statistical models.  In general, there was good agreement for like covariates from each statistical model.  The effects values from both statistical models were compared using a t test, which indicated that for most locations and covariates one could not reject the null hypothesis that there was no statistically significant difference between the comparable effects values.  This comparison between the two statistical models indicates that the relationship between ozone and meteorology is similar but not equivalent.  In particular, there appears to be a tendency for the air quality model to underestimate how ozone increases with temperature.  This finding may have impacts on projections of future climate effects on ozone air quality via global and regional climate modeling.

Use of ensemble WRF meteorological fields in July 2005 CMAQ simulations over the eastern U.S.

Brian Etherton, Renaissance Computing Institute, Kirk Baker, USEPA Office of Air Quality Planning and Standards, Pat Dolwick, USEPA Office of Air Quality Planning and Standards, Saravanan Arunachalam, UNC Institute for the Environment

Accurate prediction of past air quality concentrations within a chemical-transport model requires an accurate characterization of the meteorological conditions that occurred during the evaluation base case.  Numerical weather prediction models, such as the Weather Research and Forecasting (WRF) model, are typically used to develop the requisite gridded meteorological data that are input to the air quality model.  Due to the complexity of the atmosphere, meteorological models like WRF often parameterize key atmospheric processes that occur at scales smaller than the model grid resolution.  The purpose of this analysis was to model an ensemble of potential WRF configurations looking at the impacts of the differing planetary boundary layer (PBL) parameterizations and differing land surface models.

A set of 16 WRF simulations was completed over an eastern U.S. domain for the month of July 2005, a period in which air quality was relatively poor.  Each of the WRF ensemble members were evaluated against existing ambient meteorological data to assess which configurations best matched the historical meteorological conditions.  Additionally, each of the meteorological simulations were input into the Community Multiscale Air Quality (CMAQ) model to assess how differing meteorological model performance affected the accuracy of the air quality model, specifically on predictions of O3 and PM2.5.  The three specific goals of this analysis were: 1) to better understand how meteorological model performance affects downstream air quality model performance (i.e., how correlated are WRF bias/errors with eventual CMAQ bias/errors), 2) to better understand which meteorological parameters are most important to replicate in order to maximize AQ model performance, and 3) to better understand the potential variability in meteorological model parameterizations that can affect the resultant met inputs and AQ model outputs.

Incorporating Principal Component Analysis into the evaluation of CMAQ

Brian Eder, K. Wyatt Appel and Thomas Pierce

Emissions and Model Evaluation Branch

Atmospheric Modeling and Analysis Division

 National Exposure Research Laboratory

Environmental Protection Agency

 A five year (2002-2006) simulation of CMAQ (Version 4.7) covering the eastern United States will be evaluated using Principle Component Analysis (PCA), in order to identify and characterize statistically significant patterns of model bias and error.  This evaluation will utilize approximately 50 rural monitors from the Clean Air Status and Trends Network (CASTNet).  One of the purposes of CASTNet, which was implemented by the U. S. Environmental Protection Agency, is to identify and characterize broad-scale spatial and temporal trends of various air pollutants, including gaseous SO₂, HNO₃, O₃ and  particulate SO₂, NO₃, and NH₄, in order to facilitate model evaluation.

The main objective of using PCA (which has been extensively in examination of air quality data, but, as far as the author’s are aware, never in the evaluation of air quality models) will be to identify characteristic and recurring modes of CMAQ bias and error across a myriad of spatial and temporal scales.  PCA is especially useful in that it can:  1) identify areas of poor model performance across space and time; 2) facilitate understanding of the probable mechanisms (emissions, meteorological, and/or chemical) responsible for said poor performance; and 3) designate specific CASTNet stations that can be used in providing in depth diagnostic analysis.  Examination of the time series of the principal component scores will then be performed, using spectral density analysis.  Such analysis can identify periods of especially good or bad performance as well as any cycles or periodicities (i.e. annual or seasonal) that may be associated the principal components.

A demonstration of this approach will focus on CMAQ’s simulation of ambient air concentrations of SO₄ over the five year period.  Select results for the simulation of the remaining species will be provided as time and space permit. 

Evaluation of Retrospective, Multi-year, Continental Scale WRF (Version 3.1) simulations that used the PX LSM and ACM2 PBL

Robert Gilliam and Jonathan Pleim

The Pleim-Xiu Land Surface Model, Pleim surface-layer and Asymmetric Convective planetary boundary layer Model version 2 has been adapted to the Advanced Research WRF model since version 3.0. These physics have been used extensively in MM5 for retrospective air quality applications. Recently three annual retrospective simulations (2002, 2006 and 2007) using WRF-ARW with these physics have been completed and represent the first large scale use of WRF to drive the Community Multi-scale Air Quality model. Since meteorology is a key driver of air quality modeling systems it is important to fully evaluate using surface and above surface observations. This research provides an assessment of the 2-m temperature, 2-m mixing ratio, 10-m wind, seasonal precipitation, boundary layer wind and boundary layer temperature.

Results of Recent Dynamic Evaluation Analyses with the CMAQ Model

J. Godowitch, T. Pierce, S. Napelenok, R. Pinder, S.T. Rao, and A. Gilliland

Since photochemical air quality models are being applied in regulatory arenas to determine how particular emission control strategies impact air quality for criteria pollutants, it is critical that a model's response to emission changes is investigated and better understood. The dynamic evaluation approach, an emerging technique being employed as an integral component of a comprehensive model evaluation program, is specifically designed to assess a model's ability to reproduce observed changes in pollutant concentrations that could be attributed to changes in emissions and/or variability in meteorological conditions spanning multiple years. Consequently, the dynamic evaluation approach focuses on examining the changes in modeled and observed concentrations in relative (percentage) and absolute terms rather than computing statistical metrics with concentration pairs.

An overview is provided of recent results generated from various dynamic model evaluation efforts using the Community Multi-scale Air Quality (CMAQ) modeling system. Based on CMAQ annual simulations spanning 2002 through 2006, which included the 2003-2004 implementation of NO_X emission reductions in the electrical utility sector from the U.S. Environmental Agency's NO_X SIP (State Implementation Plan) Call program, results are presented of observed and modeled changes in the diurnal variation of hourly ozone concentrations, as well as in maximum 8-hour ozone from summer seasons over this 5-year period. To assess the impact of the decline in the mobile NO_X emission sector from 2002-2006, changes in urban weekday morning (6-9 AM) modeled and observed NO_X concentrations are analyzed and compared. An investigation of long-term changes in the weekend / weekday effect was performed which covered the 18-year period from 1988-2005 which involved analysis of observed and modeled concentrations of maximum 8-h ozone, NO_X and CO from a weekday and a weekend day spanning this extended period. Additionally, the CMAQ/DDM (Direct Decoupled Method) model was applied in a sensitivity analysis of NO_X emissions uncertainty to assess variability in modeled ozone change compared to observational change between the summer seasons of 2002 and 2006. Selected results from each of these efforts will be summarized.

Simulating Ozone: A Comparative Analysis of CMAQ and WRF/Chem

Jerold A. Herwehe, Tanya L. Otte, Rohit Mathur, S. T. Rao

U.S. EPA/ORD/NERL/Atmospheric Modeling and Analysis Division

The interaction of meteorology and chemistry is a fundamental part of any air quality modeling system. The Community Multiscale Air Quality modeling system (CMAQ) is an offline coupled model which ingests stored hourly values of meteorological variables produced by regional scale numerical weather prediction models (such as WRF or MM5) in order to drive its chemical processes and transport. In contrast, the Weather Research and Forecasting with Chemistry model (WRF/Chem) is an online coupled model which solves the meteorology and chemistry together at each time step, thereby permitting bidirectional feedbacks between the chemistry, aerosols, physics, and meteorological dynamics during the simulation. The degree of coupling between meteorology and chemistry may have a significant effect on the simulated air quality results, depending on the resolved temporal and spatial scales of the simulated scenario.

The present study examines the ability of WRF-driven CMAQ and WRF/Chem to simulate ozone (O₃) for August 2006, a summer retrospective study period that coincides with a portion of the TexAQS 2006 field campaign. This model intercomparison also includes model evaluation against available surface network observations (such as the Air Quality System (AQS) and the SouthEastern Aerosol Research and Characterization (SEARCH) Study) and vertical sounding observations (such as from the INTEX-B Ozonesonde Network Study 2006 (IONS06)). The modeled domain covers the eastern two-thirds of the United States using 12 km horizontal grid spacing and 34 vertical layers. To improve model compatibility for comparison purposes, the 2005 Carbon Bond chemical mechanism (CB05) was implemented into WRF/Chem and then linked to the Modal Aerosol Dynamics Model for Europe/Secondary Organic Aerosol Model (MADE/SORGAM) scheme that was already part of WRF/Chem. In addition to the same gas-phase chemical mechanism, CMAQ and WRF/Chem used the same emissions and the same initial and boundary conditions for their August 2006 simulations. Both model simulations will be compared by examining differences in such processes and quantities as averaged surface concentrations, boundary layer ventilation, ozone production efficiency, air mass photochemical age, the nocturnal transport of ozone precursors, oxidizing capacity, and coefficient of variation, as well as other statistical analyses.

Atmospheric Chemistry Model Data for summer 2006 utilizing the RAQMS-CMAQ linkage

Daegyun Lee, Daewon Byun, Hyuncheol Kim, Fong Ngan, Brad Pierce, Jassim Al-Saadi, Soontae Kim

 

 The Community Multiscale Air Quality (CMAQ) model is a state-of-the-art science atmospheric chemistry model which has been widely used to study and simulate multi-scale air quality issues. The CMAQ model is capable of providing high quality atmospheric chemistry profiles through the utilization of high resolution inputs relating meteorology and emissions with chemical reactions. However, it cannot simulate air quality accurately if input data are not appropriate and reliable. One of the most important inputs required by CMAQ is lateral boundary conditions (LBCs), which continue to affect model predictions throughout the simulations. Although a nesting technique may be used to reduce uncertainties of boundary conditions in the urban-scale domain, this technique is not applicable for the regional-scale domain or in the case when the in-flux mass of pollutants is not negligible. Since the current CMAQ model uses a set of constant lateral background condition profiles of the pollutant species, without reflecting temporal and spatial variations at the boundaries, it is critical to generate proper model-ready boundary data for model inputs.

The key hypothesis of this study is that such limitations can be improved by the utilization of the NASA’s substantial archives of earth science remote sensing and modeling data products. The NASA LaRC-University of Wisconsin Realtime Air Quality Modeling System (RAQMS) model with satellite observations assimilated using a statistical digital filter can provide such dynamic lateral boundary conditions for CMAQ.

The objective of this study is to improve predictability of CMAQ modeling by means of the lateral boundary conditions generated from RAQMS results. Based on the previous research, we updated the RAQMS-CMAQ linking tool by adding the CMAQ aerosol modules (AERO3 and AERO4) as well as additional gas phase species available from RAQMS. We investigated the boundary condition impacts on CMAQ simulation, and verified and characterized the model results by comparing with various measurement data available in this study.

Simulation of Wintertime High Ozone Concentrations in Southwestern Wyoming Using the CALMET/CALGRID Modeling System

Jason F. Reed (TRC), Ken Rairigh (WYDEQ), Michael B. Newman (TRC), David Strimaitis (TRC) and Gale F. Hoffnagle (TRC)

Recent, winter-season high ozone concentrations in the Upper Green River Basin (UGRB) of southwest Wyoming led to the Upper Green River Winter Ozone Study (UGWOS) during late-winter 2008, sponsored by the Wyoming Department of Environmental Quality (WYDEQ). As discussed in the State's ozone nonattainment support document, the UGRB's unique natural and anthropogenic factors make rapid winter-time ozone formation possible under the right seasonal conditions. Recent development of oil and gas fields Basin-wide has increased the amount of ozone precursors (VOC and NOx) within the airshed. In addition, the Basin's terrain can block airflow from entering or exiting during certain meteorological conditions, such as anomalously low wind speeds associated with strong high pressure systems. Finally, snow cover prevalent during winter increases the available radiation for photochemistry as well as the atmospheric stability of the Basin.

In order to better understand these phenomena, WYDEQ has engaged in meteorological and photochemical grid modeling using the CALMET/CALGRID modeling system. Using the data collected from the 2008 UGWOS studies, the CALMET model was run using numerical prognostic model output, refined land cover and terrain data and meteorological observations to produce a model-ready meteorological database for the area. The CALGRID model was run with the above meteorology and a refined emission inventory processed to meet CB-IV chemical mechanism requirements. The inventory is based largely on actual, speciated emissions during the 2008 winter season for oil and gas production in the UGRB. Recent CALGRID model upgrades to its advection scheme have shown the potential for improved model performance compared to the initial modeling, which did not consistently replicate high hourly ozone across the domain during a 1 week period in February 2008.

This paper will summarize the work completed to-date, including pertinent sensitivity analyses and findings intended to understand CALGRID model performance. Several aspects of the CALGRID model will be explored including the ability of the CB-IV mechanism to replicate measured ozone during these events, sensitivity to emission inventory and speciation, meteorological inputs such as vertical velocities and upgrades to its advection schemes.

Model performance assessment of ozone, speciated PM2.5, wet deposition and other gases for multiple photochemical models: CMAQ and CAMx

Sharon Phillips, Kirk Baker, Norman Possiel

The Community Multi-scale Air Quality model (CMAQ) and the Comprehensive Air Quality Model with Extensions (CAMx) are state-of-the-science tools in use by the regulatory community for estimating the impacts of sources and control strategies on ozone and fine particle concentrations, deposition, and visibility. This evaluation consists of model simulations for the entire year of 2005 for 12-km resolution domains covering the eastern and western United States. Both models were run with a consistent 14 layer vertical structure. Common inputs include: boundary/initial concentrations from a global chemistry model (GEOS-CHEM); emissions inventories for biogenics and anthropogenic sources; and meteorological data (MM5). All emissions were processed through SMOKE to generate CMAQ appropriate emissions species. These emissions were converted to CAMx using simple processing tools to change the names of certain species and change emissions units. The model evaluation includes graphical and statistical comparisons of model-predicted ozone, PM2.5 species, nitrogen and sulfur deposition, and inorganic PM2.5 precursors to the corresponding observed data as measured at sites in the following networks: the Interagency Monitoring of Protected Visual Environments (IMPROVE), the Speciation Trend Network (STN), National Atmospheric Deposition Program (NADP), Midwest Ammonia Network, and the Aerometric Information Retrieval System (AIRS).  Differences and similarities between CMAQ and CAMx in terms of model performance will be highlighted in the presentation of results.  We will also identify possible avenues for further exploration that could provide an understanding of the reasons for differences in performance by these models.

Utilization of Geostationary Satellite Observations to Evaluate Cloud Prediction by the Weather Research and Forecasting (WRF) Model

Yun-HeePark, Richard T. McNider, Arastoo Pour Biazar, Kevin Doty

Universityof Alabama in Huntsville

 

Maudood Khan

The Universities Space Research Association

 

Bright Dornblaser

TexasCommission on Environmental Quality (TCEQ)

Clouds have a profound role in photolysis activity, boundary-layer development and deep vertical mixing of pollutants and precursors.  Yet, one of the main deficiencies of atmospheric models is their inaccurate prediction of clouds.  In this study we use Geostationary Operating Environmental Satellite (GOES) observations of clouds to evaluate the performance of the Weather Research and Forecasting (WRF) model with respect to cloud prediction.  Both observed and model clouds will be classified based on their radiative properties as well as their elevation and thicknesses to create a comparable setting for evaluation.  The observations will be compared to model simulations (with different configurations) for the month of August 2006.  Also, cloud statistics for the model will be compared to that of the observations to investigate the existence of any systematic bias in the model.  A detailed discussion of the evaluation for the month of August, 2006 will be presented.

Microscale Energy Simulations by using WRF-UCM and EULAG Models: MADRID Case Study
R. San Jose1, J.L. Perez1 & R.M. Gonzalez2

1 Environmental Software and Modelling Group, Computer Science School, Technical University of Madrid (UPM), Campus de Montegancedo, Boadilla del Monte, 28660 Madrid (Spain).
2 Department of Geophysics and Meteorology, Faculty of Physics, Complutense University of Madrid (UCM), Ciudad Universitaria, 28040 Madrid (Spain)

Principal Contact: R. San Jose, Professor, Computer Science School, Technical University of Madrid (UPM), 28660, Madrid (Spain), +34-91-3367465, Fax: +34-91-3367412, Roberto@fi.upm.es.

The coupling between high spatial resolution mesoscale models such as the last version of WRF model (NCAR) including the Urban Canopy Model (UCM) and microscale (CFD) models such as EULAG model is an area with intense work during the last years. Increase in computer power of nowadays models based on high scalability of the use of different processors is a fact during the last years. In this contribution we have applied the WRF-UCM model with 200 m spatial resolution over Madrid domain. This domain is nested with larger domains covering all Iberian Peninsula and Europe starting with 50 km, 16.2 km, 1.8 km and 0.2 km. We have applied the EULAG (UCAR) microscale model with 4 m spatial resolution in dynamical and diagnostic modes. We have implemented an energy balance equation into EULAG code to obtain the sensible and latent heat fluxes with 4 m resolution. Information of 1 m resolution surface reflectance for June, 25, 28, and July, 1, 4, 2008 obtained from flights over Madrid, are also assimilated into the microscale model simulation. The results show a detailed description of the sensible and latent heat flux in high spatial resolution over a detailed urban domain which can be used for better urban planning of the energy efficiency.

Using Smog Chamber Data to Improve the Understanding of SOA Formation

Manuel Santiago¹, Ariel F. Stein², Marta G. Vivanco¹, and Rick Saylor³

1 CIEMAT (Research Center for Energy, Environment and Technology). 28040 Madrid. SPAIN

2 Earth Resources & Technology on assignment to NOAA's Air Resources Laboratory, Silver Spring ,MD.

3 NOAA's Air Resources Laboratory, Silver Spring ,MD.

The organic fraction of secondary particles, commonly known as secondary organic aerosols (SOA), constitutes a significant part of fine aerosols. The formation of SOA is subject to complicated coupling among gas-phase chemical reactions, aqueous-phase, aerosol-phase, and meteorological processes. To correctly predict total particle concentrations in the atmosphere it is necessary to understand the physical and chemical processes producing organic aerosols. Considering that the whole complexity of the processes and factors involved in SOA formation has not been completely understood, there is a need to isolate the chemical contribution in three-dimensional photochemical models from other SOA formation origins. Measurements made under controlled environmental conditions, such as those performed in a smog chamber, offer a unique opportunity to study the chemical processes leading to SOA production. Therefore, the comparison with chamber data allows the evaluation of the chemical processes of SOA formation simulated by the chemical and aerosol modules used in CMAQ.

CMAQ is typically configured as an Eulerian, three-dimensional, long-range transport, transformation and deposition ozone and particulate matter air quality model, but also can be configured as a multiphase lower-dimensional (0-D box or 1-D column) model to investigate small-scale gas- and aerosol-chemical and physical processes.

In order to investigate the formation processes of SOA in the smog chamber we set the box model version of CMAQ to the initial conditions of temperature, humidity, solar radiation and chemical concentrations measured at the EUropean PHOtochemical REactor (EUPHORE) chamber. The evolution of the concentration of each measured chemical have been modeled and compared for each experiment using the current chemical configuration of the developmental National Air Quality Forecast Capability (CB05-AERO4). The comparison of the SOA formation along with the consumption of precursors with different initial NO_x and VOC concentrations as measured in the smog chamber determines if the chemical mechanism used in the CMAQ model can reproduce not only the maximum SOAs but also their formation rate.

Impact of Model Grid Spacing on Regional Air Quality Predictions of Organic Aerosol

Craig Stroud, Paul Makar, Michael Moran, Wanmin Gong, Sunling Gong, Junhua Zhang, Kathy Hayden, Cris Mihele, Jeff Brook, Jonathan Abbatt and Jay Slowick

Regional scale chemical transport model predictions of urban organic aerosol in the literature tend to be biased low relative to observations which have important implications for applying models to human exposure health studies. We performed air quality predictions of organic aerosol with a nested version of Environment Canada’s AURAMS model (42-15-2.5km nested grid spacing) for a temporal and spatial domain corresponding to the Border Air Quality and Meteorology Study (BAQS-Met) field study which took place in southwestern Ontario in the summer of 2007. A domain wide average for the 2.5km domain and a windowed 15km domain yielded very similar organic aerosol averages (4.8 vs. 4.3 mg m^-3, respectively). On regional scales, secondary organic aerosol dominated the organic aerosol composition and was adequately resolved with the 15km model grid spacing. The shape of the 2.5km resolution model’s organic aerosol histogram for Windsor improved relative to that from the 15km resolution model. The model histograms for Bear Creek and Harrow were also improved in the high concentration “tail” region. The highest resolution model results captured nicely the midday July 4 urban plume at Bear Creek with very good temporal correlation. The results suggest that accurate simulation of urban scale and industrial plumes require the highest resolution model, in order to capture both the urban primary organic aerosol emissions and secondary organic aerosol production rates. The positive feedback between secondary organic aerosol production rate and existing solvent organic mass concentration is represented more accurately with the highest resolution. This non-linearity in secondary organic aerosol production may partly explain the consistent negative bias in the literature, when urban-scale organic aerosol or evaluations are made using coarser-scale chemical transport models.

Numerical analysis of air pollutants at the west part of Japan during an intensive observational campaign in 2009

Kazuyo Yamaji (Japan Agency for Marine-Earth Science and Technology), Li Jie (Institute of Atmospheric Physic),Itsushi Uno (Kyushu University), Yugo Kanaya (Japan Agency for Marine-Earth Science and Technology), Fumikazu Taketani (Japan Agency for Marine-Earth Science and Technology), Hiroshi Tanimoto (National Institute for Environmental Studies), Satoshi Inomata (National Institute for Environmental Studies)

An intensive observational campaign were performed to understand regional atmospheric pollution of north east Asia in 2009 at Fukuke island which is a remote island bordering East China Sea placed on the west part. Considerable high atmospheric pollutants concentrations including gasses and aerosol species were observed from April to June. Firstly, reproducibility of simulated gasses and aerosol concentrations using WRF/CMAQv4.7 was evaluated by comparison with these observational data.  This model captured well these concentration levels, the day-to-day variations, and the high polluted peaks, though simulated some species (e.g. CO) were lower than observed ones.  Additionally, analysis of air pollutants were performed by the model associating with observational results, and that showed the mechanism of regional atmospheric pollution of north east Asia in spring.

GLIMPSE II: Future emission scenario development for short lived climate forcers using MARKAL

Farhan Akhtar, US EPA

Rob Pinder, US EPA

Dan Loughlin, US EPA

Modeling the climatic effects of proposed future environmental policies and energy programs is a costly and time-intensive process. All of the possible emissions outcomes from future changes in technology, economics, and politics cannot be discretely modeled by climate simulations. Screening tools are needed to develop scenarios that are plausible estimates of future emissions while demonstrating that their potential climatic effects are compelling enough to be modeled further. In previous decades, many possible future emission scenarios have been developed and modeled for long-lived greenhouse gases. Increasingly however, emissions of short-lived climate forcers such as black carbon, organic carbon, and sulfate are coming into scrutiny for their effects on the global radiative budget in addition to prior concerns over their impacts on human health and air quality.

The GLIMPSE project focuses on screening potential emission scenarios for short-lived climate forcers by linking the atmospheric radiative forcing sensitivity model, GEOS-Chem/LIDORT adjoint, with an energy system market model, MARKAL. In this presentation, we describe the expansion of the EPA 9-region MARKAL database to include emissions of black carbon and organic carbon. With the radiative forcing impacts from these pollutants and sulfate calculated using the GEOS-Chem/LIDORT adjoint model, future emissions scenarios from the US energy production and transportation can be constrained. We describe how the MARKAL system is modified to simultaneously limit both radiative forcing and air quality impacts and present several possible future energy production and transportation scenarios which later may be investigated further using a coupled atmosphere-ocean general circulation model.

Impacts of Global, Regional, and Sectoral Black Carbon Emission Reductions on Human Mortality

Susan Casper Anenberg, Kevin Talgo, Saravanan Arunachalam, J. Jason West

Black carbon (BC) is a component of fine particulate matter (PM_2.5) released during incomplete combustion of fuel. BC is associated with atmospheric warming and deleterious health impacts, including premature mortality due to cardiopulmonary disease and lung cancer. Controlling emissions may therefore have dual benefits for climate and health. Recent studies have focused on quantifying the potential impacts of reducing BC emissions on radiative forcing, but the impacts on human health have been less well studied. Here, we use a global chemical transport model and a health impact function to quantify the surface air quality and human health benefits of controlling BC emissions. We use the MOZART-4 global chemical transport model to simulate a base case and several emissions control scenarios, where anthropogenic BC emissions are reduced by half globally, individually in each of several world regions, and from each major economic sector. Meteorology and biomass burning emissions are for the year 2002 with anthropogenic BC and organic carbon emissions for 2000 from the IPCC AR5 inventory. Model performance is evaluated by comparing to global and US surface measurements of PM_2.5 components. Avoided premature mortalities are calculated with a health impact function using the change in PM_2.5 concentration between the base case and emissions control scenarios, and a concentration-response factor for chronic mortality from the epidemiology literature.   

Improved CMAQ Wet Deposition Fields Using a Precipitation Based Bias Correction

Kristen M. Foley, Robin L. Dennis, K. Wyat Appel

Spatial interpolation of observed wet deposition values from the National Atmospheric Deposition Program/National Trends Network (NADP/NTN) is often used to estimate past and current loads of acidic (S+N) and nutrient (N) deposition on sensitive ecosystems for critical loads studies. Due to siting criteria, such approaches can miss important emission sources and geographic features that impact deposition, e.g. orographic effects on precipitation amounts. The Community Multiscale Air Quality (CMAQ) model provides spatial fields of wet and dry deposition that explicitly account for emission sources across the United States as well as geographic features of the domain. However, errors in modeled precipitation (such as from MM5 or WRF) and in emission inputs can lead to significant bias and error in the wet deposition predictions compared to observed values. We present an approach to post-process the CMAQ model output to adjust for errors in precipitation using observation-based gridded precipitation data. We further correct the model output by applying a bias adjustment based on observed wet deposition levels at the NADP/NTN sites. The final adjusted spatial fields of annual total wet deposition values (specifically SO₄^2- , NO₃^- , and NH₄⁺ ) have less bias and are more highly correlated with observed wet deposition values compared to the base model output. Results are presented for 2002-2006 model simulations over the Eastern half of the United States.

Impacts of photo-excited NO2 chemistry on ozone SIPs (State Implementation Plans) in complex terrain

Yunhee Kim¹, Joshua S. Fu¹, and Golam Sarwar²

¹ Department of Civil & Environmental Engineering, University of Tennessee, Knoxville, TN

²National Exposure Research Laboratory, US. Environmental Protection Agency (USEPA), RTP, NC 27711, USA

This study investigates the effects of the excited the nitrogen dioxide (NO₂) chemistry on daily maximum 8-hr ozone concentrations in the eastern Tennessee. CMAQ4.7 with excited NO₂ chemistry was used in photochemical simulations on a matrix of modeling scenarios permitting an examination of the water vapor effects of meteorological influences on maximum ozone levels over a 4-month period during the summers of 2002 in complex terrain. We conducted with and without excited NO₂chemistry as well as decreased water vapor (80%) and increased water vapor (120%) on meteorological variables at 4-km grid resolution. Our results show that it can potentially play a more important role in dry and urban areas and mountain areas as well as it gave better model performance for daily maximum 8-hr ozone concentration with 60 ppb cutoff at all sites. In addition, the largest daily maximum 8-hr ozone increased by 4.2 ppb in the mountain areas was occurred while 8-hr O₃increased by 2.4 ppb in the valley areas was noted. Overall, the NO₂chemistry can enhance more O₃ in mountain sites than valley sites at dry atmospheric chemical conditions.

Air Quality Impacts of Increased Use of Ethanol Under the Energy Independence and Security Act

Sharon Phillips, Rich Cook; Pat Dolwick, Marc Houyoux, Rich Mason, Catherine Yanca, Margaret Zawacki, Ken Davidson, Harvey Michaels, Craig Harvey, Joseph Somers

In recent years, use of ethanol as a component of vehicle fuel has been heavily promoted for a variety of reasons, including reduction of reliance on fossil fuel, reduction in global warming, and reduction in ambient concentrations of a variety of air pollutants. Whether increased use of ethanol made from various feedstocks and production pathways results in net increases or decreases in greenhouse gases is currently the subject of considerable scientific debate. In addition, increased use of ethanol impacts other pollutants, such as "criteria" pollutants for which the U. S. Environmental Protection Agency (U. S. EPA) has set National Ambient Air Quality Standards (NAAQS) and those pollutants referred to as air toxics, based on potential for adverse cancer and non-cancer health effects. This paper focuses on these pollutants and how increased use of ethanol, as mandated by the U. S. Energy Independence and Security Act of 2007 (EISA), is likely to impact their emissions and ambient levels in the United States. The assessment of impacts was done for calendar year 2022, when renewable fuel goals of EISA are likely to be achieved.

This assessment addresses both impacts of increased ethanol use on vehicle and other engine emissions, referred to as "downstream" emissions, and "upstream" impacts, i.e., those connected with fuel production and distribution. These impacts come from changes in agricultural processes, feedstock transportation, and the production and distribution of biofuel. These processes occur domestically and internationally. The assessment takes into account how EISA impacts direct emissions, as well as precursor emissions which can photochemically react to form these pollutants.

Air quality modeling was performed using Community Multi-scale Air Quality Model (CMAQ), version 4.7. CMAQ was modeled using 12 kilometer square grids in the Eastern U. S. and Western U. S. Pollutants included in the assessment are ozone, particulate matter, acetaldehyde, ethanol, formaldehyde, acrolein, benzene and 1,3-butadiene. Implications of modeling results for the ability of States and local governments to achieve attainment for particulate matter and ozone, and the potential for adverse health effects due to air toxics exposures, will be discussed as well.

U.S. Air Quality Impacts of Particulate Filter Retrofits on Non-Road Diesel Engines

Christopher Werner, Saravanan Arunachalam, J. Jason West

Black carbon (BC) is a component of PM2.5 emitted from many types of sources, with diesel engines being the largest contributor in the U.S. and non-road engines making up a large portion thereof. Newly available emissions controls nearly eliminate PM or BC emissions, through the use of particulate filters. As a nationwide program mandating retrofits does not exist, retrofitting particulate filters represents an opportunity for reducing emissions of BC that would otherwise continue through the long service lifetime of diesel equipment, with co-benefits of reducing emissions of SO2, NOx, CO, and hydrocarbons.

Our research intends to investigate the air quality benefits of a hypothetical national program that retrofits half of existing nonroad diesel equipment with particulate filters. Using a 2005 base case with the EPA National Emissions Inventory (NEI), SMOKE is employed to simulate emission reductions of multiple precursor species, and CMAQ is run to demonstrate the effects of these reductions on air quality parameters (including ozone, PM, CO, NOx, and SO2).  We will use an annual simulation across the continental United States at 36 km resolution. These results can then support an analysis comparing the costs of such a retrofit program against benefits to air quality and human health.

Analysis of air pollution reduction effects by regional implementation plan of Seoul metropolitan area in Korea

Chul Yoo, Yong-mi Lee, Dae-gyun Lee, Jong-chun Kim, Suk-jo Lee 

  The Government had devised legislation of Special Act and drew up guidelines for improving air quality of Seoul Metropolitan area in Korea. Local governments of Seoul, Incheon and Gyeonggi were made detailed plans so that this Special Act would perform effectively. In 2007 each local government in Seoul metropolitan area conducted performance evaluations of application policy by reduced air pollutants emission for the first time. Although there was reduction of air pollutant emission in each local government, it didn′t work as expected for improving air quality using air pollution monitoring database. Therefore we worked out a way to prepare modeling input data using the performance of enforcement plan. And we simulated surface NO2 and PM10 before and after decrease in air pollutants emission and examine reduction effects of air pollution according to enforcement regulation except other influence, by using MM5-SMOKE-CMAQ system. Under the enforcement plan, each local government calculated the amount of emissions which was reduced before and after application policy. In this study the ratios of emission reduction were classified into detailed source and fuel codes using code mapping method in order to allocate the decreased emission.

  As a result of prediction using the reduction of NOx emission, NO2 concentration was decreased from 19.1ppb to 18.0ppb in Seoul. In Gyeonggi and Incheon NO2 concentrations were down to 0.65ppb and 0.68ppb after application of enforcement plan. PM10 concentration was reduced from 18.2ug/m3 to 17.5ug/m3 in Seoul. In Gyeonggi PM10 concentration was down to 0.51ug/m3 and in Incheon PM10 concentration was decreased about 0.47ug/m3 which was the lower concentration than any other cities. We determined the change of spatial distribution of air pollution according to the reduction of air pollutant emission using GIS program. The result indicated that the spatial distribution of air quality as reduction effect depended on distribution of air pollutant emission. Because in this study the code mapping method was not considered spatial distribution which was one of the most important elements, the spatial distribution using air monitoring data did not match those of emission and air pollution reduction. Therefore it indicated that considering spatial distribution of enforcing reduction for air pollutant emission would produce the improved result for predicting the reduction effect of air quality and analyzing the policy impacts.

High ozone events and attainment demonstrations in Houston, Texas

Evan Couzo, Harvey Jeffries, Jason West, William Vizuete

The Houston-Galveston-Brazoria (HGB) area has had multiple decades of persistent high ozone (O₃) values, but has attained the current federal 0.08 ppm 8-h standard within the last three years according to recent monitoring data. We have analyzed ten years of ground-level measurements at 25 monitors in Houston and found that peak 1-h O₃ concentrations were often associated with large hourly O₃ increases.  A non-typical O₃ change (NTOC) – defined here as an increase of at least 40 ppb/hr or 60 ppb/2hrs – was measured 25% of the time when concentrations recorded at a monitor exceeded the 8-h O₃ standard.  CAMx model simulations (120 total days in 2005 and 2006) used to support the 2010 State Implementation Plan for the HGB non-attainment area were found to be limited in their ability to simulate NTOCs and under predicted the maximum observed rate of change by more than 50 ppb/hr.  We show that the regulatory model, using "average" emissions in accordance with current EPA methodology, has difficulty simulating spatially isolated, high O₃ events measured at monitors that routinely violate the 8-h O₃ standard.  When a day-specific emissions inventory is used, 1-h O₃ predictions are nearly identical to simulations using the "average" emissions inventory.  Peak 1-h predictions and hourly O₃ changes were only 8 ppb and 3 ppb/hr greater when using day-specific emissions, which more accurately represents the emissions profile of the HGB region.  Our results suggest that this modeling system will be unable to guide selection of effective control strategies as the regulators shift their attention to meeting a more stringent federal 8-h O₃ standard.

The Houston Regulatory Model and Simulating Non-typical Ozone Changes

Adeola Olatosi, Evan Couzo and William Vizuete

 

The Houston-Galveston-Brazoria (HGB) areais currently classified as a severe non-attainment area under the 1997 ozone eight-hour standard, and the Texas Commission on Environmental Quality is required to produce a State Implementation Plan (SIP) to show future attainment. Our analysis of ten years of surface measurements revealed two types of ozone changes on days that violated the federal ozone standard. On some days measured ozone concentrations at violating monitors showed gradual increases of 10-30ppb/hr. On other days, however, there were rapid increases in ozone concentration equal to or greater than 40ppb/hr, which we have labeled as non-typical ozone changes (NTOC). We found that including days with NTOC characteristics can increase a monitor’s baseline design value by up to 10 ppb. The regulatory model used in the HGB SIP does not simulate these rapid hourly ozone concentration changes. The goal of this study is to evaluate the processes that limit the regulatory model's ability to reproduce measured NTOC behavior. We have assessed the model's response to episodic releases of VOCs. In previous research, rapidly increasing ozone concentrations have been linked to upwind emissions of VOCs. Here we quantify the impact of episodic emissions on simulated ozone concentrations. We have also used process analysis tools to quantify changes in model processes and identified those processes that are most influential in predicting rapid increases of ozone. These data will help assess the regulatory model's ability to predict NTOC behavior and ultimately its value in generating effective control policy. 

Effects of Varying Horizontal Resolution to Model Responsiveness for PM2.5

Karen Wesson, Brian Timin, Darryl Weatherhead, and Larry Sorrels

With our ever increasing technical and scientific ability to apply photochemical models at finer horizontal resolutions, we have witnessed an increase in the number of these applications at horizontal scales below 12km to assess regulatory impacts. In defining the appropriate horizontal resolution for regulatory modeling, it is important to understand how varying horizontal resolutions may impact model responsiveness. To do this, we examine how CMAQ responds to emissions reductions between a current year base case and a future year base case and emissions scenario. CMAQ v4.7 was run for several example cities at horizontal scales of 12km and 4km using a base case of 2005, a future emissions scenario of 2020, and some example emissions reductions of local PM_2.5 and SO₂ sources. We then examine the impact of horizontal scale on the calculated changes in PM_2.5 Design Values (DVs) and Relative Response Factors (YYFs) for both the annual and daily PM_2.5 standards. We further explore how the magnitude and location of the emissions reductions impact the concentration changes for the different horizontal scales. How these results may impact regulatory decisions, in particular, SIP attainment demonstration modeling, is then summarized and discussed.