Jonathan E. Pleim, Shawn J. Roselle, Rohit Mathur, Jeffery Young, Prakash Bhave, Heather Simon, Deborah Luecken, Bill Hutzell, Golam Sarwar, John Streicher, David Wong, Havala Pye, Kathleen Fahey, Rob Gilliam, George Pouliot

Atmospheric Modeling and Analysis Division, NERL, USEPA, Research Triangle Park, NC

A new version of the Community Mutiscale Air Quality model (CMAQv5.0) is scheduled for release to the public in October 2011. The new version includes science updates and code restructuring. In addition, an option to run the WRF-CMAQ as a 2-way coupled model is included. This presentation will outline the new features of CMAQv5.0 and introduce the other presentations describing these model upgrades in more detail.

The aerosol module was redesigned to eliminate dependencies and duplication across modules. The speciation of PMOther was improved by adding 9 new PM2.5 species, including primary non-carbon organic matter (NCOM), and particulate Fe, Al, Si, Ti, Ca, Mg, K and Mn. An oxidative aging reaction was added for primary organic aerosol. There were some minor updates to the SOA yield parameterization. A new aerosol thermodynamics module was included (ISOYYOPIAv2). An optional algorithm for windblown dust was added.

The Carbon Bond (version 05) chemical mechanism (CB05) now includes updated the toluene chemistry, revised rate constants for N2O5 hydrolysis based on the latest recommendation of IUPAC, and additional reactions of toluene and xylene with chlorine radical. A new version of the SAPRC mechanism (SAPRC07TB & TC) was integrated into CMAQ, and includes fully updated organic and inorganic reactions, updated photolysis rates, operator species to better represent chemical reactions in low-NOx conditions, explicit treatment of additional species with high emission, high toxicity, or high SOA formation, and an updated isoprene mechanism.

The inline photolysis rate module was improved to increase flexibility by reading input data from an ASCII file (created by a pre-processor) for easy modification/introduction of associated data (e.g., quantum yield, cross section) for new reactions, and an algorithm is now included that calculates the surface albedo based on land use categories, zenith angle, seasonal vegetation, snow and sea ice coverage.

A new option was introduced to add production of nitrogen oxide (NO) from lightning. This algorithm allows users to incorporate lightning NO by either reading an offline 4-D file, computing inline informed by lightning detection network data, and a fully inline option.

A new technique for computing vertical velocity and vertical advection that eliminates excessive diffusion in the upper model layers has been developed. Also, a new stable boundary layer scheme has been applied to both WRF and CMAQ.

Other structural updates were also included to improve the flexibility of the CMAQ code. A namelist option was included to eliminate the numerous include files, which allows users to make some changes to the model species treatment at runtime.

Rohit Mathur, Jonathan Pleim, David Wong, Francis Binkowski, Tanya Otte, Rob Gilliam, Aijun Xiu, Shawn Roselle, Jeffrey Young

¹ Atmospheric Modeling and Division, NERL, U.S. EPA, RTP, NC 27711
² Institute for the Environment, University of North Carolina, Chapel Hill, NC 27599

CMAQv5.0 will include the capability to run the model either in the traditional off-line manner (in which output from meteorology model is used to drive the chemistry-transport calculations) or with 2-way coupling with the WRF model. A modeling framework that facilitates coupled on-line calculations is desirable since it (1) provides consistent treatment of dynamical processes and reduces redundant calculations, (2) provides ability to couple dynamical and chemical calculations at finer time-steps and thus facilitates consistent use of data, (3) reduces the disk-storage requirements typically associated with off-line applications, and (4) provides opportunities to represent and assess the potentially important radiative effects of pollutant loading on simulated dynamical features.

The design of the 2-way coupled WRF-CMAQ system facilitates coupling of the dynamical and chemistry/transport calculations at flexible user defined intervals as well as the investigation of the feedback effects of pollution loading on radiation and subsequent simulation of the dynamical state of the atmosphere, thereby enabling "2-way" interactions. Applications with this modeling system to study and examine the various issues related to atmospheric pollution as well as air quality-climate interactions will be discussed.

Jesse O. Bash, Jon Pleim, Ellen Cooter, John T. Walker, Robin Dennis

Ammonia (NH₃) is an important precursor for particulate matter, yet NH₃ emissions are challenging to estimate and concentrations are difficult to measure. It is critical to understand the factors that lead to episodes of poor air quality for effective mitigation strategies. As climate change leads to increased variability in meteorology, relying on seasonal averages as the drivers for emissions estimates adds additional uncertainty to model simulations. It is necessary to capture the dynamic and episodic nature of ammonia emissions, including the influences of meteorology, air-surface exchange, and human activity to reduce uncertainty in model scenarios of NH₃ emissions mitigation strategies, agricultural food production and climate change. This presentation will cover the implementation of the bi-directional NH₃ exchange model in CMAQ 5.0 and evaluate an annual simulation against observations and an annual CMAQ v4.7.1 simulation. Bidirectional exchange algorithms will be evaluated using observations from a flux measurement campaign and regional scale model results will be evaluated using monitoring network ambient ammonium aerosol concentrations and NHx wet deposition observations. Preliminary results show that the bidirectional NH₃ exchange model reduces the model bias in seasonal and annual NH_x wet deposition estimates and regionally reduces the bias in NO₃^- and NH₄⁺ aerosol estimates. The bidirectional NH3 exchange model in CMAQ 5.0 moderately improved the model performance while more directly linking agricultural management practices and land use to air-quality.

S. Trivikrama Rao¹, Rohit Mathur¹, Stefano Galmarini²

¹U.S. Environmental Protection Agency
² European Commission Joint Research Centre

With the endorsement and strong support by the U.S. Environmental Protection Agency, European Commission, and Environment Canada, , the project titled Air Quality Model Evaluation International Initiative (AQMEII) was launched in 2009 by bringing together scientists from Europe and North America (Rao et al., 2011). Several regional-scale numerical photochemical models were exercised over the NA and EU domains with a common emissions inventory for the full year of 2006. Several papers resulting from this collaborative effort have been submitted to the AQMEII special issue of Atmospheric Environment. A large 4-D database that has been assembled by EU JRC for the AQMEII project is now available to the scientists interested in developing innovative model evaluation techniques (Galmarini and Rao, 2011).

Further work on model evaluation is being carried out by the AQMEII investigators. Having successfully completed the first phase of AQMEII, Phase 2 activity of AQMEII will be launched at the 2011 AQMEII workshop in Chapel Hill, NC to focus on interactions of air quality and climate change. In Phase 2, coupled meteorology-atmospheric chemistry models will be exercised over the two continents with a common input database to assess how well the current generation of coupled regional air quality models can simulate the spatio-temporal variability in the optical and radiative characteristics of atmospheric aerosols and associated feedbacks among aerosols, radiations, clouds, and precipitation. The results of AQMEII Phase 2 would be useful to policy makers for developing effective policies to deal with air pollution and climate change.

References:

S.T. Rao, S. Galmarini, and K. Puckett, "Air Quality Model Evaluation International Initiative (AQMEII): Advancing the state of the science in regional-scale photochemical modeling systems", Bull. of the Amer. Meteor. Soc., January 2011.

S. Galmarini and S.T. Rao, "The AQMEII Two-Continent Regional Air Quality Model Evaluation Study", Atm. Env., March 2011.

Andrea Fraser, Geoff Dollard, Paul Willis, Trevor Davies, Justin Lingard
AEA, Fermi Avenue, Harwell IBC, Oxon, OX11 0QR, 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. It is used to produce a two day forecast for O₃, NO₂, SO₂, CO, PM₁₀, PM_2.5 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 for Environment, Food and Rural Affairs (Defra) and the Devolved Authorities (DA). This model forms one of the tools used to generate the forecast and is part of the continued improvements to the service.

During March and April 2011, the UK's Automated Urban and Rural Monitoring Network (AURN) recorded two widespread episodes of elevated PM₁₀ air pollution. Over a 12 month period our forecast generally under predicts PM₁₀ however the events in March and April 2011 were well predicted at the urban background monitoring sites. This study investigates the prevailing weather conditions and the contribution of PM components during this period to characterize the conditions under which the model performed well.

M.D. Moran, S. Menard, R. Pavlovic, H. Landry, P.-A. Beaulieu, S. Gilbert, J. Chen, P.A. Makar, W. Gong, C. Stroud , A. Kallaur, and S. Gong

GEM-MACH15 has been Environment Canada's operational regional air quality forecast model since it replaced the CHRONOS AQ model in November 2009. GEM-MACH15 is a limited-area configuration of GEM-MACH, an on-line chemical transport model that is embedded within GEM, Environment Canada's multi-scale operational weather forecast model. It is run twice daily to produce 48-hour forecasts of hourly O₃, PM_2.5, and NO₂ fields on a North American grid with 15-km horizontal grid spacing and 58 vertical levels from the surface to 0.1 hPa.

The GEM-MACH15 operational AQ forecasting system has undergone several upgrades since 2009. These have focused on but not been restricted to improvements to the emissions fields used by the model, including a switch from a 2005 U.S. emissions inventory to a 2012 projected inventory. The impacts of these changes on model performance will be described for a number of evaluation metrics and analysis techniques. The roles of AQ model output statistics and objective analysis, current challenges, and plans for further improvements to GEM-MACH15 will also be discussed.

Seiji SUGATA, Toshimasa OHARA (National Institute for Environmental Studies),

Jun-Ichi KUROKAWA (Japan Environmental Sanitation Center),

Masamitsu HAYASAKI (Toyama University)

We have developed the air pollution forecast system VENUS (Visual atmospheric ENvironment Utility System). VENUS calculates automatically concentrations on the day and the next day of nitrogen dioxide and photochemical oxidants every night, and exhibits maps of the concentrations every morning on a web. The system consists of integration of the meteorological model RAMS and the air quality model CMAQ.
The photochemical-oxidants reproducibility of VENUS was verified. Distribution of the concentration and time change in a several prefectures were investigated for the wide-range high oxidant concentration event in Japan in May, 2009. Distribution and its movement of high concentration area were well reproduced though reproducibility of the peak values was quantitatively restrictive. Moreover, skill scores, such as the correlation coefficient, between simulation and observation was investigated in May and August of the same year about hourly averaged concentration of photochemical oxidants and its daily top 5% concentration in each prefecture, As a result, the top 5% concentration in particular shows high correlation coefficient, and bias and error were small.
According to these validation results, it is demonstrated that the VENUS has the accuracy helpful to predict daily high photochemical-oxidants concentration in each prefecture in Japan.

Jeff McQueen(1), Eric Rogers(1), Hui-Ya Chuang(1), Ivanka Stajner (2,3), Pius Lee(4), Roland Draxler (4), Rohit Mathur(5) and Jon Pleim(5)

1. NOAA NWS/NCEP
2. NOAA NWS/OST
3. Noblis, Inc.
4. NOAA/OAR/ ARL
5. EPA/ASMD

The NOAA National Air Quality Forecasting Capability (NAQFC) has been in existence and providing predictions of ozone and particulate matter to air quality forecasters since 2003. NOAA in partnership with the U.S. EPA followed phased development, testing, and implementation for improvements in the NOAA/NWS National Centers for Environmental Prediction (NCEP) North American Model (NAM) and the Community Model for Air Quality (CMAQ) system. This paper will highlight some of the improvements made to improve both physical and numerical coupling that were inspired by the work of Dr. Byun and others.

With the increase in computer power, the NAQFC grew from predicting ozone over a North East U.S. domain with only gas-phase chemistry to a national domain (CONUS, Alaska and Hawaii) with both gas-phase and aerosol chemistry. NAQFC initial direction was set to a large part from the work of Byun (1999, 2010) through emphasizing and developing improved coupling between the North American Model (NAM) and CMAQ land surface, deposition, boundary layer, radiative and moist atmospheric processes. This focus also helped to also identify areas of improvement for the meteorological model driver especially for predictions during quiescent atmospheric conditions. Byun, et al. (1999) also highlighted the importance of minimizing interpolation error and preserving mass conservation in both the meteorological and air quality models by utilizing common horizontal and vertical coordinates and mass conserving dynamics as much as is possible. His work also identified the need for more realistic and real-time lateral boundary conditions from global models and the NAQFC program performed several studies on these sensitivities.

Finally, future improvements in the NAQFC as recommended by his work will be discussed and help serve as a pathway for future air quality forecast developments. Recommendations included the use of common dynamics, more up to date emissions, on-line chemistry, a cloud analyses for initializing CMAQ and reliance on additional verification metrics. Key issues for producing air quality predictions using running high resolution chemistry models are also highlighted.

Ivanka Stajner (1,2), Jeff McQueen(3), Pius Lee(4), Roland Draxler(3), Phil Dickerson(5), Kyle Wedmark(1,2), Tim McClung (1)

(1) NOAA NWS/OST

(2) Noblis, Inc.

(3) NOAA NWS/NCEP

(4) NOAA ARL

(5) EPA

The National Air Quality Forecasting Capability (NAQFC) is being built by NOAA in partnership with the U.S. EPA through phased development, testing, and implementation. NAQFC provides nationwide operational predictions of ozone and wildfire smoke available at www.weather.gov/aq. Predictions are produced beyond midnight of the following day at 12km resolution and 1 hour time intervals and they are distributed in numerical and graphical formats. Ozone prediction and developmental testing of aerosol predictions combine the NOAA National Centers for Environmental Prediction (NCEP) operational North American Mesoscale (NAM) weather predictions with inventory based emissions estimates from the EPA and chemical processes within the Community Multiscale Air Quality (CMAQ) model. Predictions of smoke and testing of dust storm predictions, both of which have highly variable intermittent sources, use the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model. Routine verification of ozone and developmental aerosol predictions relies on AIRNow compilation of observations from surface monitors, whereas verification of smoke and dust predictions relies on satellite retrievals of smoke and dust column integrals.

Air quality predictions will begin using NCEP's new version of the Nonhydrostatic Mesoscale Model on an Arakawa B-grid (NMMB), which is targeted for implementation in the summer of 2011. Impacts of this change on air quality predictions are being evaluated. Updates to dust predictions currently being tested include soil moisture as one of the factors modulating dust emissions and development of a satellite product for verification of dust predictions. Recent development is focusing on improvements in modeling of fine particulate matter (PM2.5) and the contributions from intermittent sources: dust from dust storms and smoke from forest fires. Contributions from sources external to the CONUS domain are being examined, including long range transport of dust and smoke. Assimilation of PM2.5 AIRNow observations is also being tested in order to improve developmental aerosol predictions.

Christian Hogrefe^1,2, Prakash Doraiswamy², Brian Colle³, Ken Demerjian², Winston Hao¹, Mark Beauharnois², M. Erickson³, M. Souders³, and Jia-Yeong Ku¹

¹New York State Department of Environmental Conservation, 625 Broadway, Albany, NY

²Atmospheric Sciences Research Center, University at Albany, 251 Fuller Road, Albany, NY

³School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY

This study presents an analysis of the effects of meteorological and emissions uncertainty and horizontal grid resolution on air quality forecasts over the Northeastern U.S. during a 1 month warm season period and a 1.5 month cold season period. The meteorological uncertainty is prescribed through the use of twelve different MM5 and WRF weather forecasts from a short-range ensemble forecasting system differing in their initial conditions and physics options. CMAQ air quality simulations were performed at a horizontal grid spacing of 12 km using all of these weather forecast members and a common set of emissions. Results indicate that these variations in the meteorological fields cause a typical daily maximum 8-hr ozone variability of 10% and a typical daily average PM_2.5 variability of 20%. To quantify the effects of emission uncertainties, we present results from CMAQ simulations utilizing the Decoupled Direct Method (DDM) to compute ozone and PM_2.5 sensitivities towards emission variations. Preliminary results indicate that emission uncertainties have a relatively bigger impact on PM_2.5 forecasts than ozone forecasts but exhibit significant spatial variability. Therefore, we present a detailed analysis of the spatial and temporal variability of these sensitivities and discuss the implication to the resulting air quality forecasts. Finally, we present simulations performed with nested grids of 12km, 4km, and 1.33 km horizontal spacing centered over the greater New York City region. These simulations are analyzed to quantify the effect of spatial variability within typical 12 km forecast grids and compare this variability with the variability caused by the meteorological and emission uncertainties in the simulations described above.

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 six years. We evaluate Hi-Res's air quality forecasting ability by examining O₃ and PM_2.5 as well as PM_2.5 components performance during 2006-2011. Observational datasets from SLAMS air quality monitors as well as from SEARCH and ASACA networks are utilized. The forecasting performances for meteorological variables are also examined. Note that in 2011, dramatically increased frequencies of O₃ and PM_2.5 exceedances in Georgia have been observed.

In addition to the standard O₃ and PM_2.5 concentration forecasts, we propose a "source contribution" forecast that would provide important information about how O₃ and PM_2.5 events can be avoided. We will add DDM-3D simulation to calculate the forward sensitivity coefficients of O₃ and PM_2.5 to certain emission sources on top of the current standard CMAQ simulation in operation. The sensitivity forecasts will provide source contribution information, for example emission contributions (as fractions) to the forecasted ambient concentration of O₃ and PM_2.5 from sources like EGU (power plants) and vehicular emissions and from emissions of potential prescribed burns. The forecasts of fractional contributions from major sources will be disseminated along with the Hi-Res standard O₃ and PM_2.5 concentration forecasts. The air quality impacts from burnable forests in selected areas can be forecasted using DDM-3D sensitivity simulations. Then the allowable acreage for prescribed burns in each area on the forecasting day can be derived and provided ahead of the permit issuing time. These forecast products will provide information ahead of time about how O₃ and PM_2.5 events to be impacted by type of emissions sources, which can help EPD issuing warnings aiming for specific public and GFC issuing prescribed burning permit that might results in avoiding such events.

Pius Lee^1*,Fantine Ngan², Hyuncheol Kim³, Daniel Tong³,Yunhee Kim³, Tianfeng Chai³, Yunsoo Choi³, Rick Saylor⁸, Ariel Stein³,

Youhua Tang⁴, Jianping Huang⁴, Jeff McQueen⁵,Marina Tsidulko⁴,Hochun Huang⁴,Sarah Lu⁴,

Ken Carey⁶, Ivanka Stajner⁶,and Paula Davidson⁷

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.

³Earth Resources & Technology ,Inc, Annapolis Junction, MD

⁴ I.M. Systems Group, Inc. Rockville, MD 20852.

⁵ NOAA/NWS/National Centers for Environmental Prediction, Camp Springs, MD.

⁶ Noblis, Inc., Falls Church, VA.

⁷ Office of Science and Technology, National Weather Service, Silver Spring, MD.

⁸NOAA/OAR/ARL, 456 S. Illinois Ave, Oak Ridge, TN 37830

The National Air Quality Forecasting Capability (NAQFC) currently provides 48-hour surface ozone concentration forecasts for the conterminous (CONUS) 48 states. One of the considerable challenges for the NAQFC is future quantitative forecasts of the concentration of particulate matter (PM) smaller or equal to 2.5 μm aerodynamic diameter (PM_2.5). The highly variable spatial and temporal distribution of PM_2.5 together with its complex combination of chemical composition, mixing states, coatings of organics, and characteristics of origination pose challenges to predict its mass, speciation, and distribution over space and time. An equally considerable challenge is the potential lowering of the US EPA ozone standards. There is evidence that long-range transport has caused the background concentration of ozone to the CONUS domain to increase in recent decades. Capturing real-time concentrations of the various constituents in the inflowing air mass to the domain will be even more critical to accurately forecast ozone concentrations under the lower standards. To address both of these challenges we focus on wildfires in this study. Wildfires are a significant source of primary PM_2.5 and ozone precursors. It has been reported that PM_2.5 emitted during large fires can remain in the atmosphere for multiple days and be transported long distances. Therefore, PM_2.5 forecasting should include the best available real-time emission information on intra- and- exo- domain wildfires. By the same token, precursor emissions from wildfires would help ozone forecasting. This study is an attempt to quantify possible improvement of forecasts for both ozone and PM_2.5 by including the exo-domain fires. A large North American CMAQ parent domain was created to facilitate the generation of dynamic boundary conditions for the CONUS domain. The parent domain ingests near real-time fire information within its geographical coverage but outside of the CONUS domain. Results of this nested run will be compared to that without nesting to delineate its impact. Both ground-based and airborne measurements will be used to quantify the fidelity of the forecast for early summer 2010, when numerous fires occurred in North and Central America but outside CONUS.

Fong Ngan, HyunCheol Kim, Daewon Byun, Pius Lee, Soontae Kim, Fang-Yi Cheng, DaeGyun Lee and Bernhard Rappenglueck

The East Texas Air Quality Forecasting System (ETAQ) utilizing MM5/SMOKE/CMAQ-based modeling system has been implemented to forecast air quality since June 2006 by the Institute for Multidimensional Air Quality Studies (IMAQS) at the University of Houston (UH). The ETAQS system generates 54-hour air quality prediction daily for the outer-most nest covering continental US (CONUS) in 36-km, the intermediate nest covering Texas and neighboring the Gulf of Mexico areas in 12-km, and the innermost nest for eastern Texas in 4-km horizontal grid-spacing resolution. It has been used successfully for planning and implementation of various measurement campaigns, including the Second Texas Air Quality Study (TexAQS-II) intensive field measurements. The evaluation of the forecasting results over the 5-year (2006 - 2010) period has been performed to study the spatial and temporal distribution of model errors. To investigate the frequently occurring high ozone bias by the model-predicted values during the ozone season over the southeastern US, we correlated meteorological factors such as temperature and wind speed with the ozone biases using surface meteorological measurements collected in MADIS dataset (Meteorological Assimilation Data Ingest System) and by ozone observations in AQS dataset (Air Quality System ambient observations operated by EPA). In association with the movement of frontal systems, enhancements of model ozone bias were often observed. Analyses to investigate the behavior of the model bias and identify the causes were also performed. . Conductively a few generalized conclusion on the characteristics of the ozone season of the Southeastern states has been attempted.

Eun-Su Yang1 and Sundar A. Christopher1,2

1Earth System Science Center, UAHuntsville, Huntsville, AL

2Department of Atmospheric Science, UAHuntsville, Huntsville, AL

We have developed an automated modeling system for air quality forecasts. The system uses the existing numerical models such as WRF, SMOKE, CMAQ, and SBDART, along with EPA's NEI and satellite-derived fire emissions. This work focuses on the wildfire's impact on air quality in the southeastern United States.One primary goal of this product is to forecast air quality fast enough to meet nearly-real-time decision making and services. We will show transport of fire smoke in the real-time forecast environment. We will also simulate radiance that a satellite sensor sees from the surface through the aerosol layers. The products are evaluated with satellite and ground-based observations.

Fang-Yi Cheng and Shan-Chieh Chin

In order to accurately simulate air quality modeling, the meteorological inputs need to be able to correctly represent the turbulent fluxes, wind transport and vertical mixing which strongly depends on the PBL processes. In this study, meteorological simulations are performed using Weather Research Forecasting (WRF) model by applying two different PBL schemes (YSU and MYJ). Community Multiscale Air Quality (CMAQ) modeling system is performed subsequently to study the effect of the PBL physical processes on the meteorological and air quality simulations. The comparison focused on two different atmospheric conditions. The case 1 is under the influence of the Asia continental outflow and the air pollutants is long-range transported (LRT) to northern Taiwan. The ozone (O₃) variation between two sensitivity runs is mostly caused by the PBL height difference. Case 2 is associated with the land-sea breeze flow and the locally generated O₃concentration can reach up to 120 ppb. The simulation using YSU scheme predicts higher temperature during the night and inducing weaker land breeze flow which would accumulate the nitrogen oxide (NO) in the source region. The O₃titration is stronger near the source region and lower in the outer sea area. During the daytime, the simulation using YSU scheme predicts higher temperature thus inducing stronger sea breeze flow which would carry the aged O₃back into the Taiwan Island. The stronger sea breeze from simulation using YSU predicts higher O₃toward inland and lower O₃in thee outer sea area than the one using MYJ. Comparison with the observation datasets identifies less bias of the wind speed, temperature and O₃concentration with the simulation using YSU scheme.

 Timothy Glotfelty¹, Jason Ching² , Fei Chen³, Mukul Tewari³

¹MEAS, NCSU, Raleigh NC

²CEMPD, UNC, Chapel Hill,  NC

³RAL, NCAR

               New formulations for improving numerical modeling of the urban boundary layer were developed in the community  Weather Research and Forecasting Model (WRF)-urban modeling system, which enable improved modeling of urban boundary layer frictional drag, turbulence, thermal and radiative heating processes. The performance of the WRF-urban model heavily relies on accurate description of morphological features (buildings, vegetation, impervious land use structures) in urban areas.  A unique set of Urban Canopy Parameters (UCPs) have been  generated at both 1km and 250m grid meshes derived from high- resolution airborne LIDAR data for 44 of the largest cities in the USA, which is part of the National Urban Database and Access Portal Tool/ National Building Statistics Database (NUDAPT/NBSD).  This database was implemented this summer at the National Center for Atmospheric Research (NCAR) into the WRF-urban modeling system to support the current and future single to multi-layer urban canopy modeling options.  In an initial test application, the Building Effect Parameterization (BEP) and accompanying Building Energy Model (BEM) model options in WRF were run for several cities using both this comprehensive NUDAPT/NBSD data as well as data from standard look-up tables for urban land use classes.  This study will both describe this new database capability in WRF and discuss results of this initial urban test application.  The addition of this database will provide the WRF modeling community with the necessary urban input data to be used to support future urban air quality, heat island mitigation and climate change assessment and impact studies.

Neil Wheeler, Kenneth Craig , Stacy Drury, Eric Gray, and Garnet Erdakos

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

Miriam Rorig

AirFire Team, USDA Forest Service, Pacific Northwest Research Station, Seattle, WA 98103

Currently fire weather forecasters, fire planners, and decision makers do not have easy access to information needed to verify the accuracy of, or to communicate the level of confidence in, fire weather forecasts and the fire prediction products that depend on fire weather forecasts. The Joint Fire Science Program has funded us to develop a system that produces intuitive, easily understandable meteorological model performance assessments to provide end-users with real-time information about meteorological model bias, model reliability, and overall performance of predictions of fire weather variables used in predictions of ignition risk potential. The extension of this system to parameters that are important for emissions, dispersion and transport, dry and wet deposition, photochemical and chemical reactions, and pollutant concentrations air pollution meteorology is discussed.

Farhan Akhtar, Rob Pinder, Dan Loughlin
US EPA Office of Research and Development

Daven Henze
University of Colorado, Boulder

The GLIMPSE decision support tool has been designed to generate and assess future, long-term emissions scenarios for the U.S. energy production system. Within this tool, the GEOS-Chem global atmospheric model is combined with the radiative transfer model, LIDORT. This modeling system allows the user to evaluate the global and regional radiative forcing impacts of short lived climate forcers (SLCFs) such as sulfate and elemental and organic carbonaceous aerosols. The adjoint implementation of this model further allows for the evaluation of global and regional radiative forcing from specific regional and temporal trends in emissions. This degree of high-resolution assessment information is central to the efforts of the U.S. EPA in adapting environmental policies to maximize environmental benefits while reducing possible costs to the public and industry.

To better understand the implications of possible energy policies, technologies, and costs, the GLIMPSE project integrates the assessed impacts of SLCFs gained from the GEOS-Chem/LIDORT adjoint model into the energy system optimization model, MARKAL. The U.S. EPA has developed a regional database covering energy resource supplies, electrical production within the industrial, residential, and commercial, and transportation sectors. Possible future technologies, including both alternative production sources and possible source controls, are also contained in the database. Subject to economic and environmental constraints, the MARKAL model attempts to optimize these options to minimize overall energy system cost. This tool allows users to generate long-term (through the year 2055) emission projections from the sectors contained in the database.

The emission scenarios from MARKAL give policymakers an understanding of the possible emission impacts of changing energy production technologies and controls. The scenarios generated can be used both by regional atmospheric models to model the effects on human health and visibility and by global climate models to explore relatively short-term climate mitigation options. This link between global climate models and regional air quality models will aid policy makers in adapting energy and environmental policies to reduce the overall future impacts of anthropogenic air pollution.

In this presentation, we will highlight the ability of the GEOS-Chem/LIDORT adjoint model to resolve the impact of US regional emissions of SLCFs on global and arctic radiative forcing. We will also contextualize the relative impact of US emissions against global emissions. These impacts will drive our analysis of mid-century climate outcomes of energy sector emissions projections from MARKAL. We will discuss how this tool enables policymakers to evaluate decisions regarding energy production fuel choice and emission control strategy. 

Meridith M. Fry¹, J. Jason West¹, Zachariah Adelman¹, Pat Dolwick², and Carey Jang²

¹The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599

²Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle Park, North Carolina 27709

Ozone (O₃) precursor emissions (NO_x, CO, and NMVOC) influence regional and global climate and air quality through changes in the tropospheric concentrations of ozone (O₃), methane (CH₄), and aerosols.  Here we examine the influence of regional reductions in CO and NMVOC emissions on global air quality concentrations. We simulate anthropogenic CO and NMVOC emissions reductions from nine world regions (Australia, Southeast Asia, East Asia, India, Africa and Middle East, Former Soviet Union, Europe, South America, and North America), using the global chemical transport model MOZART-4.  We use the IPCC AR5 Representative Concentration Pathway 8.5 (RCP 8.5) emissions inventory for 2005 and global meteorology from the Goddard Earth Observing System Model, version 5 (GEOS-5) for the years 2004-2005 as inputs to MOZART-4.  We first simulate base case global air quality for the year 2005, and compare the resulting distributions of O₃ and related species with observations.  We then simulate CO emissions reductions from each of the nine regions individually and NMVOC emission reductions from several regions.  The results from these simulations will be used to quantify regional and global changes in O₃ and PM_2.5 at the surface and within the troposphere, including the influence of each regional reduction on changes in long-term O₃ concentrations via CH₄ and changes in the long-range transport of O₃ and CO.  Beyond this study, we aim to use these simulations to estimate global and regional net radiative forcing.

Yong Zhang, Panos G. Georgopoulos, Sastry Isukapalli

A modeling system is presented for studying the emissions and transport of aeroallergens under different climate change scenarios. The emissions module considers major physical processes such as direct emissions and re-suspension of pollen particles, and accounts for meteorological parameters such as surface temperature trends, friction velocity, humidity, precipitation, etc. This module also incorporates Bayesian models based on analysis of historical data for estimating effects of climate change on annual flux and start date of pollen emissions. Spatiotemporal profiles of pollen emissions are estimated through the combined application of (a) Bayesian models based on analysis of historical data, (b) future meteorological data for representative climate change scenarios, and (c)future land use and land cover data. Case studies are presented focusing on the prediction of regional-scale birch pollen levels for years ranging from 2040 to 2065. Future year meteorology profiles were obtained from the Weather Research and Forecasting (WRF) model simulations driven by the General Circulation Model (GCM) for representative scenarios from the Intergovernmental Panel on Climate Change (IPCC). The generated emission profiles, future year meteorological data, and future year land use/land cover profiles were used as inputs to the Community Multiscale Air Quality modeling system for pollen (CMAQ-pollen) to estimate the spatiotemporal concentration profiles of birch pollen. Results suggest that the start date of birch pollen season is likely to be about one week earlier in these future years than in 2005. Furthermore, the annual production of birch pollen is likely to be about three times higher than in 2005, along with corresponding increases in concentration levels. These concentration profiles are being linked with modules within the Modeling Environment for Total Risk (MENTOR) framework for estimating human exposures to pollen and for characterizing the impact of climate change on allergic airway disease.

Otin M. Grimolizzi^{1, 2}, Benitez, L. M.¹, Frenzel de Llomparte, A.M.^{1, 2}

¹Laboratorio de Estudios de Baja Atmosfera (LEBA) - Instituto de Riesgo Geologico y Sistematizacion Territorial (IRGIST) - Facultad de Ciencias Naturales e Instituto Miguel Lillo - Universidad Nacional de Tucuman, Miguel Lillo 205, Tucuman, Argentina

² Consejo Nacional de Investigaciones Cientificas y Tecnicas (CONICET), Argentina

Mail to: grimolizzi@aol.com

Two solar UV radiometers were built and simultaneously operated during the past dry season in Tucuman province, NW Argentina. Ampimpa, an astronomic observatory in the mountains was adopted as a reference, non polluted monitoring site, while, the INTA (National Institute for Agricultural Technology) meteorological facilities at Famaille, in the plains was chosen as an air pollution test site, its sets of data to compare to those from Ampimpa.

The quotient of simultaneous radiometry data between the reference site, and the polluted site, was very close to unity in clear days. In cloudy, misty, smoked or rainy days this quotient does not hold, a fact that is reflected by loss of statistical significance in statistical tests performed on corresponding data sets.

By making use of the special geographic characteristics of the region, it was possible to estimate particulate matter (PM) content by taking air samples and relating them to radiometry data. Also some results of a 2006 short term campaign using sun photometers are displayed together with particulate matter content determination.

Keywords: radiometer, ANOVA, transmittance, particulate matter.

Hyun Cheol Kim¹, Ariel Stein¹, Yunhee Kim¹, Pius Lee², Fong Ngan³, Tianfeng Chai¹, Rick Saylor², Youhua Tang⁴, Jeff McQueen⁵, Chuanyu Xu⁶, Shoba Kondragunta⁶, Roland Draxler², Tara Strand⁷, Sim Larkin⁷, and Ivanka Stajner⁸

¹ ERT on assignment to NOAA/ARL, Silver Spring, MD

²NOAA/ARL, Silver Spring, MD

³UCAR on assignment to NOAA/ARL, Silver Spring, MD

⁴SAIC on assignment to NOAA/NCEP, Camp Spring, MD

⁵NOAA/NCEP, Camp Springs, MD

⁶NOAA/NESDIS, Camp Springs MD

⁷Pacific Northwest Research Station, US Forest Service

⁸Noblis, Inc., Falls Church, VA

Emissions from forest fires and prescribed burns play a significant role in regional air quality significantly contributing to the increase in particulate matter (PM) concentrations over the US. Simulating the impact from fire emissions, however, is challenging due to high uncertainties in many processes, including the detection of the fire location and time duration and the determination of the emission injection heights. In this study, we have compared outputs from two fire forecasting systems with various observations to evaluate the models’ ability to simulate physical extent of the fire smoke plume along with a quantitative verification of its chemical composition. The first one is the National Oceanic and Atmospheric Administration’s (NOAA) Smoke Forecasting System (SFS) system. This system utilizes the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model for calculating the transport, dispersion and deposition of smoke driven by the National Centers for Environmental Prediction (NCEP) North American Model (NAM) meteorological model, the Hazard Mapping System (HMS) for fire detection, and the US Department of Agriculture (USDA) Forest Service (FS) BlueSky framework for estimating the emissions.The second modeling system analyzed uses the Community Multiscale Air Quality (CMAQ) model. Observations from various sources are selected to separate the signal originating from fire emissions. From surface observations, we have utilized total PM and speciation data (e.g. Organic carbon (OC) and Elemental carbon (EC)) from AirNOW and Environmental Protection Agency (EPA) Air Quality System (AQS). Visual imagery of fire smoke analysis from HMS have been used to evaluate the horizontal transport of the simulated fires. Aerosol Optical Depth (AOD) measurements from Moderate Resolution Imaging Spectroradiometer (MODIS) and Geostationary Operational Environmental Satellite (GOES) have also been compared.  

M. Makarova, D. Ionov, A. Orlov, A. Rakitin,

Saint-Petersburg State University, Physics Department, Russia

Accurate assessment of air quality and understanding contributions of various sources is critical for developing mitigation strategies to improve air quality and protect human health. However, air quality monitoring data in Russia are very limited. Thus, chemical transport models are the only tools to assess the impact of anthropogenic emissions on air quality in the North-Western region of Russia. In this study, we demonstrate how air quality modeling was used to characterize the impact of urban sources in Russia on air pollution in Northern Europe. We used CMAQ to simulate concentrations of various air pollutants (CO, NO2, O3 and PM) in the North-Western region of Russia. We used WRF to create meteorological inputs and the official emission inventory for Russian Federation processed with SMOKE to prepare emissions inputs. CMAQ model results were compared against existing monitor data from Russian observational network and also with satellite data. We found that model results compared favorably with monitor data. Our results suggest that the impact of anthropogenic emissions from Saint-Petersburg region could be noticeable at distances more than 300 km away from the megacity (e.g. in border countries Finland and Estonia).

Tianfeng Chai ^1,2, Hyun-Cheol Kim^1,2, Daniel Tong^1,2, Pius Lee¹

1. NOAA Air Resource Laboratory, Silver Spring, MD

2. Earth Resources & Technology (ERT)

During the past several years, PM_2.5 forecasts from the National Air Quality Forecast Capability (NAQFC) developmental runs have exhibited large biases and errors when compared against AIRNow ground observations. Constraining model predictions with MODIS (MOderate Resolution Imaging Spectroradiometer) aerosol optical depth (AOD) and AIRNow PM_2.5 in-situ observations through chemical data assimilation will be tested in this study. Using the NMC (National Meteorological Center, now. National Centers for Environmental Prediction) approach, i.e. using the differences between two NAQFC forecasts as surrogates of model errors, we will initially investigate the CMAQ model error statistics in detail. How the model errors are correlated in space, in time, and between species will be investigated. The results are compared with our previous study using the Hollingworth-Lannberg approach. The statistics are then utilized in the subsequent chemical data assimilation tests. Different from our previous MODIS AOD assimilation study where the observations were inserted only once a day, the measurement data will be assimilated on an hourly basis. First, the MODIS AOD and AIRNow PM_2.5 observations will be assimilated separately to study their effect in improving the PM_2.5 predictions. Then, the observations from both sources will be assimilated simultaneously. The effect of the chemical data assimilation on surface PM_2.5 predictions will be investigated using both optical interpolation (OI) and three-dimensional variational (3D-Var) methods in this study.

Yunsoo Choi^1,2, Rick Saylor³, Ariel Stein^1,2, Pius Lee¹, Hyuncheol Kim^1,2, Yunhee Kim^1,2, Fantine Ngan⁴, Youhua Tang⁵, Jeff McQueen⁶, Ivanka Stajner⁷

¹NOAA Air Resources Laboratory, Silver Spring, MD

²Earth Resources Laboratory, Inc

³NOAA Air Resources Laboratory, Atmospheric Turbulence and Diffusion Division, Oak Ridge, TN

⁴University Corporation for Atmospheric Research, Boulder, CO

⁵Scientific Applications International Corporation, Camp Springs, MD

⁶NOAA/NWS/National Centers for Environmental Prediction, Camp Spring, MD

⁷Noblis, Inc., Falls Church, VA

Simulation results from the Community Multiscale Air Quality (CMAQ) model version 4.7 over the Conterminous United States (CONUS) for August 2009 are analyzed to evaluate the variation of the hourly-averaged ground-level ozone (O₃) at stations classified according to their O₃-NO_x-VOC chemical sensitivity regimes. The O₃ sensitivity regimes (NO_x-saturated/mixed/NO_x-sensitive) are inferred from the values of photochemical indicators based on the HCHO to NO₂ ratio of column data obtained from the Global Ozone Monitoring Experiment 2 (GOME-2) and model. When compared to the U. S. EPA's Air Quality System (AQS) observations, the CMAQ model overpredicts hourly-averaged O₃ concentrations over the photochemical indicator based regimes (NO_x-saturated: +5.4%; mixed: +22.1%, and NO_x-sensitive: +29.9%) with high correlation coefficients of R>0.92. Compared with the base CMAQ run, a 30% NO_x emissions reduction reduces the high O₃ biases over the NO_x-sensitive regime by 8.7% and a 30% VOC reduction in the model reduces the O₃ biases over the NO_x-saturated regime by 7.2%, implying that possible chemical or physical mechanisms to reduce NO_x over the NO_x-sensitive regime or to reduce VOC over the NO_x-saturated regime are not well represented in the model. Results from sensitivity studies to further explore the causes of these O₃ biases will be presented.

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

The ability of air quality modeling systems to accurately predict ozone and PM_2.5 concentrations is dependent in part on the quality of the emissions used and the associated uncertainties. In an air quality forecasting context, anthropogenic emission inventories are typically provided as annual totals that are then allocated to each hour based on typical temporal profiles for each source category. Previous work (Doraiswamy et al., 2010) has shown that different approaches for temporally allocating emissions from Electric Generating Units (EGUs) resulted in non-negligible effects on predicted ozone concentrations. Further work showed that emissions from units operated on days of high energy demand ("peaking units") may be significantly underestimated on those days when using the typical approach described above for temporally allocating emissions in air quality forecasting. Our group is currently exploring approaches to refine the characterization of EGU emissions in near real-time for forecasting applications. The approach that is being examined consists of quantifying possible relationships between energy demand forecasts and the EGU emissions as measured by the continuous emission monitors. Energy demand forecasts are performed by the regional Independent System of Operators (ISOs). For NY State, it is provided, and made readily available, by the NY ISO. The objective is to be able to estimate real-time emissions that would take into account forecasted fluctuations in energy demand. This study will use archived energy demand (load) forecasts in NY State and actual EGU emissions/heat input in NY and surrounding states from 2001 to 2010 to develop relationships. The developed relationships would be applied to estimate/refine emissions of select EGUs for a test-case during May-Sep of 2007 and compare against actual emissions. This work will present a summary of the approach followed, estimates of correlations obtained, and the limitations.

Jianping, Huang^1,2*, Jeff McQueen², Brad Ferrier^1,2, Youhua Tang^1,2, Marina Tsidulko^1,2, Ho-chun Huang^1,2, Sarah Lu^1,2, Bill Lapenta², Geoff DiMego², Michael Ek², Pius Lee³, and Ivanka Stajner^4,5

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

¹I.M. Systems Group Inc., Rockville, MD

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

3NOAA Air Resources Laboratory, Silver Spring, MD

⁴Noblis Inc., Falls Church, VA

⁵Office of Science and Technology, NOAA/National Weather Service, Silver Spring, MD

Land use and land cover (LULC) characteristics impact air quality forecasting through their effects on surface-atmosphere interactions, emissions, and dry deposition velocities. NOAA's operational air quality prediction is planned to start using the National Environmental Modeling System (NEMS) Non-hydrostatic Multi-scale Model on the Arakawa staggered B-grid (NMM-B) coupled with the Community Multi-scale Air Quality (CMAQ) modeling system in the summer of 2011. One of the notable changes in the system is that the NMM-B uses the International Geosphere-Biosphere Programme (IGBP) instead of the U.S. Geological Survey (USGS) LULC data that were used by the previous version of operational air quality predictions. The IGBP LULC utilized by NMM-B contains a more recent characterization of land use, which is based on the MODerate resolution Imaging Spectroradiometer (MODIS) land cover science data product. In this study, several sensitivity experiments are conducted with NMM-B/CMAQ to investigate the impact of IGBP LULC on prediction of meteorological variables and chemical species for the continental US (CONUS) during August 2011. We examine the differences in several meteorological fields important for air quality prediction including surface radiation, sensible and latent heat fluxes, and the planetary boundary layer height (PBLH). The impact of LULC data on biogenic emissions that are used within CMAQ is evaluated. CMAQ predictions of several key chemical species, including surface ozone and PM_2.5, are compared when different LULC data are used. The predictions are compared with AIRNow observational data and the results are verified using the NCEP Forecasting Verification System (FVS). Finally, we discuss possible modifications to the coupling of NMM-B with CMAQ to improve the use of IGBP LULC.

Hyun Cheol Kim ^1,2, Daewon W. Byun¹, Clint Harper³, Fong Ngan^1,4, Daniel Tong^1,2, Pius Lee¹, and Ivanka Stajner⁵

¹NOAA/ARL, Silver Spring, MD

²ERT, Laurel, MD

³UCAR, Boulder CO

⁴TCEQ, Austin TX

⁵Noblis, Inc., Falls Church, VA

A new tool for fine resolution geospatial data processing has been developed. Fast and accurate Geographic Information System (GIS) data processing tools are essential in air quality studies; especially in preparing model emission inputs as finer resolution air quality simulations become more commonplace. An IDL-based Geospatial Data Processor (IGDP), written in Interactive Data Language (IDL by ITT Visual Information Solutions), has been created and can process GIS data both in vector format (e.g. ESRI shapefiles) and raster format (e.g. GEOTIFF and IMG) for any given domain. Processing speeds have been improved through the use of polygon-clipping routines and other algorithms optimized for particular applications. The raster tool utilizes a histogram reverse-indexing method, which enables easy access of grouped pixels, so it can generate statistics of pixel values within each grid cells, with improved speed and enhanced control of memory usage. IGDP supports map projection conversions between Lambert Conformal Conic (LCC), Rotated latitude-longitude (RLL, used in National Centers for Environmental Protection (NCEP) North American Model (NAM)), geographic latitude-longitude coordinates, and also is capable of easy expansion to all 40 projections that IDL supports.

The basic role of IGDP is to generate emission surrogate files, but numerous other applications are possible, including re-gridding of emission data, satellite data processing, ocean/land file processing, and Land Use Land Cover (LULC) data processing. Examples of applications will be presented, including typically encountered errors when using inappropriate map projection information. We have tested the tool with high-resolution grid settings and high-resolution data, such as 1 km CONUS domain (columns = 5200 and rows = 3200) and 30-m National Land Cover Data (NLCD) data (161190 x 104424 pixels).

Yunhee Kim¹, Rick Saylor², Yunsoo Choi¹, Tianfeng Chai¹, Hyun-Cheol Kim¹,

Daniel Tong¹, Pius Lee³ , Marina Tsidulko⁴ ,and Jeff McQueen⁵

¹Earth Resources & Technology, on assignment to NOAA’s Air Resources Laboratory, Silver Spring, MD

²NOAA Air Resources Laboratory, Atmospheric Turbulence and Diffusion Division, Oak Ridge, TN

³NOAA Air Resources Laboratory, Silver Spring, MD

⁴I.M. Systems Group, Inc. Rockville, MD 20852

⁵NOAA/NWS/National Centers for Environmental Prediction, Camp Springs, MD

The National Oceanic and Atmospheric Administration’s  National Air Quality Forecasting Capability (NAQFC) developmental version utilizes the Community Multiscale Air Quality (CMAQ) model to forecast ground-level PM_2.5 total mass concentrations over the conterminous U. S. (CONUS) domain.  Currently, the PM_2.5 concentrations generated by CMAQ are overpredicted from September through April and underpredicted from June through August.  In an effort to better understand the causes of these seasonal biases, a detailed comparison of total mass and speciated PM_2.5 measurements from several networks with modeled concentrations has been conducted for two periods in CY2009.  The goal of these comparisons is to gain insight into specific CMAQ model processes or inputs that are deficient and in need of improvement to reduce the observed seasonal biases.

PM_2.5 mass and speciation data for January and August 2009 from the Southeastern Aerosol Research and Characterization (SEARCH) study network and the Interagency Monitoring of Protected Visual Environments (IMPROVE) network as well as PM_2.5 mass concentrations from the U. S. EPA AIRNow network have been compared against both CMAQ 4.6 and 4.7.1 model predictions at the network locations.  Results of these analyses clearly indicate particular emission inputs and processes that need to be improved to lessen the observed seasonal biases.  The overall approach and results of these analyses will be presented and discussed.  Additionally, results from preliminary retrospective CMAQ simulations for January and August 2009 will be presented to demonstrate proposed modifications to the modeling system to improve its PM_2.5 forecasting capabilities.

Chris Misenis, Nick Witcraft, Elliot Tardif, Cassie Mentha

The wide-ranging health impacts due to smoke from wildfires can have significant impacts on the general population. Throat and eye irritation, respiratory distress and even premature death can occur with elevated levels of fine particles (PM2.5). In a particularly dry and hot summer, North Carolina (NC) has been impacted by several wildfires in the eastern half of the state. At times, smoke inundation has occurred as far west as the Triangle and Triad regions, leading air quality officials to issue health notices and collaborate with National Weather Service offices to relay these alerts. Additionally, procedures were implemented regarding a smoke forecast in addition to the NC Division of Air Quality's (DAQ) forecasts for ozone and PM2.5. As part of this additional forecast, DAQ relayed a visibility guide to the public that correlates visibilities to current air quality conditions.

Youhua Tang^1,2 (youhua.tang@noaa.gov), Jeffery T. McQueen² (jeff.mcqueen@noaa.gov), Sarah Lu^1,2 (sarah.lu@noaa.gov), Ho-Chun Huang^1,2 (ho-chun.huang@noaa.gov), Marina Tsidulko^1,2 (Marina.Tsidulko@noaa.gov), Jianping Huang^1,2 (jianping.huang@noaa.gov), Pius Lee³ (pius.lee@noaa.gov), Ivanka Stajner⁴ (Ivanka.Stajner@noaa.gov)

1. I.M. Systems Group Inc., 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 Air Resource Laboratory, Silver Spring, MD

4. Noblis Inc, Falls Church, VA

The Global Forecast System-Goddard Chemistry Aerosol Radiation and Transport (GFS-GOCART) global aerosol model is running in inline and offline modes to predict mineral dust. The inline GFS-GOCART has the dust emission, advection, diffusion and removal processes on GFS's native grid (T126) at the time step of 450 seconds while the offline GFS-GOCART used interpolated GFS meteorology to drive a stand-alone GOCART model in 1 degree resolution. These two systems use different advection schemes and parameterizations of moist convective processes. The National Air Quality Forecast Capability (NAQFC) modeling system is running at NOAA/NWS/NCEP. Developmental testing of NAQFC predictions of fine aerosols relies on the North American Model NAM (WRF-NMM) meteorological model coupled to the CMAQ V4.6 model (CB5 gas phase and AERO-IV aerosol chemistry with fugitive dust emissions only) to produce 48-hour air quality predictions over the Continental U.S. We use the NAQFC system and GFS-GOCART provided lateral boundary conditions (LBCs) to study long-range transport of dust over the continental USA. In the summer of 2010, a dust layer that originated in the Sahara desert was transported across the Atlantic to reach the Southeastern USA in the lower troposphere. Surface AIRNOW PM data are used to evaluate the performance of developmental NAQFC aerosol predictions and compare the impacts of LBCs from the offline and inline GFS-GOCART models.

M. Tsidulko^1,2, J. McQueen², G. DiMego², M. Ek², B. Ferrier^1,2, Y. Tang^1,2, J. Huang^1,2, P. Lee³, I. Stajner^4,5

¹ I.M. Systems Group Inc., Rockville, MD

² NOAA/NWS/NCEP/Environmental Modeling Center, Camp Springs, MD

3 NOAA Air Resources Laboratory, Silver Spring, MD

⁴ Noblis Inc., Falls Church, VA

⁵ Office of Science and Technology, NOAA/National Weather Service, Silver Spring, MD

The configuration of the NOAA Air Quality Forecast System targeted for implementation in the summer of 2011 consists of newly developed NOAA Environmental Modeling System (NEMS)/NMMB (B-grid version of the Nonhydrostatic Multiscale Model) and Community Model for Air Quality (CMAQ). Several significant episodes of AQ forecasts are analyzed to determine impacts of different meteorological parameters predicted by NMMB on ozone forecast. Temperature, wind, humidity, cloud cover and other parameters are analyzed and verified with observational data. Boundary layer depths as a key parameter in air quality modeling determining extent of turbulence and dispersion for pollutants are examined and related to ozone forecasts provided by CMAQ. Verification of PBL depth with aircraft observations (ACARS) is provided. PBLs both from NMMB and CMAQ models are considered. The study emphasizes comparison between meteorological parameters from 12 km resolution full-domain NMMB forecast and 4 km resolution Continental US (CONUS) nest forecast - which is proposed as meteorological input for future higher resolution AQ predictions.

Nan Zhang, Yaosheng Chen, and Yang Zhang

Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695

In this study, the Weather Research and Forecasting model with Chemistry (WRF/Chem) with the Model of Aerosol Dynamics, Reaction, Ionization, and Dissolution is employed for RT-AQF over the southeastern U.S. during three ozone (O3) seasons (May to September of 2009 - 2011) and two winters (December 2009 to February 2010 and December 2010 to February 2011). The objectives of this study are to evaluate the model's forecasting skill and to identify possible causes of biases for model improvement.

Zachariah Adelman¹, James Boylan², Mohammad Omary¹, Qun He¹, Jason Zhao¹, Dongmei Yang¹

¹Institute for the Environment, University of North Carolina, Chapel Hill, NC

²Environmental Protection Division – Air Protection Branch, Georgia Department of Natural Resources, Atlanta, GA

Under Part 75 of Volume 40 in the Code of Federal Regulations, continuous emissions monitoring (CEM) and reporting is required for large electricity generating units (EGUs) and industrial facilities. The U.S. EPA Clean Air Markets Division collects and distributes hourly CEM data for NOx and SO2 emissions (lbs/hr), heat input (mmBTU), gross load (MW), and steam load (1000 lbs/hr) for thousands of U.S. sources from the year 1995 to the present.  Some units are required to report hourly emissions year-round (annual reporters), while other units are only required to report hourly emissions for part of the year (partial year reporters).  To satisfy the Part 75 requirement that CEM data are reported for every operating hour at units that are required to report emissions, a complex process for reporting and filling in missing data has been defined.  Many times, missing emissions are substituted with values that are much larger than the actual emissions that were emitted.  In order to properly deal with the issues described above, three steps must be followed to correctly simulate the emissions from these sources.

1. Anomalous data points in the CEM database that resulted from the Part 75 substitution methodology must be identified and corrected.
2. Differences between the reported CEM emissions and annual emissions estimates reported by the states for the CEM units must be reconciled and simulated accordingly
3. Hourly emissions for the non-reporting periods need to be generated.

This presentation describes a methodology developed under the Southeastern Modeling, Analysis and Planning (SEMAP) project to complete these three steps and improve the CEM database for conducting regulatory air quality modeling.  Analysis and data augmentation utilities were developed to implement these steps in a systematic and reproducible approach.  Details of these utilities, the algorithms and equations used to improve the CEM data, and results for several CEM units in the Southeastern U.S. are presented.

Taciana T. de A. Albuquerque^a , J. Jason West^b, Rita Yuri Ynoue^c, Maria de Fatima Andrade^c

^a Environmental Engineering Department/ Federal University of Esperito Santo. albuquerque.taciana@gmail.com

^b Environmental Sciences & Engineering / University of North Carolina at Chapel Hill.

^c Atmospheric Sciences Department/ University of Sao Paulo.

The objective of this study was to evaluate the response of PM_2.5 concentrations to changes in precursor gases and primary particles emissions. 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 Sao Paulo State, Brazil. Meteorological fields were modeled using the Weather Research and Forecasting model WRFv3.1, for the 10-day period (10 - 21 Aug, 2008) and after the SMOKE emissions model was applied to build a spatially and temporally resolved vehicular emissions inventory for a high resolution domain of 3-km (109 x 76 cells). Seven different scenarios were simulated considering the current emission inventory, called base case, a future scenario considering a reduction of 50% of SO₂ emissions (Case 2), a scenario considering no SO₂ emissions (Case 3), a reduction of 50% of SO₂, NOx and NH₃ emissions (Case 4), a scenario considering no sulfate (PSO4) and nitrate (PNO3) primary particles emissions (Case 5), another considering only excluding the PSO4 emissions (Case 6) and the last one considering no PNO3 emissions (Case 7). To compare all the numerical results over the whole domain, it was considered the average of PM_2.5 for each case during the entire period of study, after that, it was made the difference between each hypothetical case and the base case. It was also observed the temporal evolution of PM_2.5 concentrations at nine grid points (Surface Local Stations). The main results showed that reductions only in SO₂ emissions are likely to be less effective than expected at reducing PM_2.5 concentrations at many locations of Sao Paulo State. Case 2 presented an average, a decrease of 3 mg/m³ on PM_2.5 concentrations, but in some areas there were an increase of 1.2 mg/m³. Evaluating the ammonia gas availability between the base case and case 2, it was verified an increase of its concentrations in the south area of the grid, and Nitric Acid showed a decrease of its concentrations. This result could indicate that nitric acid may was transferred to the aerosol phase through the reaction with ammonia gas, originating nitrate aerosol. Case 3 was irrelevant, showing only a decrease of 0.3 mg/m³ in whole area. Case 4 showed the largest PM_2.5 reduction for entire domain, not showing an increase of the PM_2.5 concentration, in average. In case 5, at all stations, was verified a decrease of PM_2.5 average concentrations. However, there are some places of the grid showing an increase of PM_2.5 concentrations, which varies from 0.3 to 1.2 mg/m³, as also observed in case 2. Case 6 showed the same results that were observed on case 5. Case 7 did not show a significant result, presenting a small increase for the entire domain (0.3 mg/m³). This result may indicate that reductions in sulfate concentration may cause inorganic fine particle matter (PM_2.5) to respond nonlinearly, as nitric acid gas may transfer to the aerosol phase. The spatial and temporal distribution of concentration varies in the whole domain. In conclusion, the largest reduction in PM_2.5 was obtained when occurred a reduction of 50% of SO₂, NOx and NH₃ emissions, considering the average at one point (surface stations) or the average over the whole domain. We suggest that the role of secondary organic aerosols and of Black Carbon particles need to be considered when making policy decisions to control the PM_2.5 concentrations because together they represent around 70% of the PM_2.5 mass concentration in Sao Paulo, Brazil.

Carlie J. Coats, Jr., John N. McHenry, and Jeffrey Vukovich

We evaluate the effects of introducing 30-meter-resolution land cover data from the National Land Cover Database (NLCD), sub-gridscale meteorology downscaling, and a simple sub-gridscale parameterization of terrain effects on vertical emissions allocation.  We take the VISTAS 2002 36-KM study as a base case, and consider computations on both the 36-KM VISTAS grid and the 3-KM BELD3 grid (with emission aggregation back to 36 KM), which happens to be a 12:1 refinement of the 36-KM grid.  We give a multi-way comparison for both biogenic emissions and for resulting CMAQ air quality involving the following

    (1) Effects of native 36-KM  vs. downscaled 3-KM meteorology inputs to (temperature-sensitive)  biogenic emissions;

    (2) BELD3 vs. BELD3-modulated-NLCD land cover;

    (3) surface-layer-only vs. sub-gridscale terrain-parameterized 3-D emissions allocation.

This poster accompanies the "Part I" oral presentation of methods, algorithms, and "gotchas" for these sub-gridscale runs.

David Mobley, Lee Beck, Marc Houyoux, George Pouliot, Heather Simon, Prakash Bhave
US EPA

The objective of this poster is to describe EPA's SPECIATE database and to announce the recent update of the database to version 4.3. SPECIATE is the U.S. EPA repository of volatile organic gas and particulate matter (PM) speciation profiles of air pollution sources. Among the many uses of speciation data, these emission profiles may be used to: (1) create speciated emissions inventories for regional haze, PM, greenhouse gas (GHG), and photochemical air quality modeling; (2) estimate hazardous air pollutant (HAP) and toxic air pollutant (TAP) emissions from PM and organic gas primary emissions; (3) provide input to Chemical Mass Balance (CMB) receptor model; and, (4) verify profiles derived from ambient measurements by multivariate receptor models (e.g., factor analysis and positive matrix factorization). The first electronic copy of SPECIATE was released in 2002 (version 3.2). Subsequent versions have updated and added profiles. Useful tools have also been added to accompany the database (e.g., online data browser and SCC linkages). The newly released version 4.3 contains a total of 5,592 PM, volatile organic compounds (VOC), total organic gases (TOG), and Other Gases profiles.

Alexis Zubrow, Alison Eyth, Rich Mason

EPA's Office of Transportation and Air Quality (OTAQ) and Office of Air Quality Planning and Standards (OAQPS) have developed an integrated system for mobile emissions, SMOKE-MOVES. The major motivations for this new system are: to closely integrate MOVES (Motor Vehicle Emission Simulator) into the emissions process, the sensitivity of many pollutants to temperature and humidity, and the computational demands for running MOVES. We will present a brief overview of the system and will highlight recent improvements to the integrated system. In addition, we will present a comparison of SMOKE-MOVES with previous modeling results using MOVES and MOBILE6 in inventory mode. The analysis will focus on the impact of temperature on the emissions, both the annual emissions and the temporally resolved emissions.

Jared H. Bowden

Tanya L. Otte

Christopher G. Nolte

20 year 108-36km two-way nested WRF integrations driven by NCEP-DOE AMIP II Reanalysis have been completed to evaluate grid point and spectral nudging for regional climate modeling. Initial statistical evaluation of near surface fields, including 2m temperature and precipitation, revealed that interior nudging improved both the mean climate and variability on time scales ranging from interseasonal to interannual. This study is an initial attempt to provide answers as why nudging improves the representation of the climate. Specifically, we examine the synoptic scale atmospheric dynamics associated with the Bermuda High during the summer season because of its potential impact on the regional climatology, air quality, and climate change. This study demonstrates that the Bermuda High is poorly represented in the absence of interior nudging and has a signficant impact on the regional climatology and variability over the southeast US. These results support the philosophy that interior grid nudging is to improve the representation of the large-scale circulation that arise in part because of numerical limitation of only using lateral boundary forcing.

Robert W. Pinder, Dale J. Allen, Kenneth E. Pickering, K. Wyat Appel, Kristen M. Foley, Jeffrey O. Young

This poster describes the implementation and guidance for users for nitrogen oxide production of NO from lightning in CMAQv5. There are four possible settings: (1) no lightning NO production, (2) lightning NO production from a user-specified 3-D, time-varying file, (3) online lightning NO production that is scaled from convective precipitation, and (4) the recommended approach using National Lightning Detection Network data integrated with CMAQ convective precipitation to reconcile the observations of lightning with the precipitation simulation. Each of these options will be described in detail. To demonstrate the value of each option, CMAQ simulations using each of these options will be compared with observations of ozone, NOx in the free troposphere, and nitrate wet deposition. 

Havala O. T. Pye, George A. Pouliot, Michael Lewandowski, John H.
Offenberg, Tadeusz E. Kleindienst, Deborah Luecken

National Exposure Research Laboratory, US Environmental Protection
Agency, Research Triangle Park, North Carolina

Secondary organic aerosol (SOA) from the oxidation of long-chain (C8 through C19) alkanes and polycyclic aromatics hydrocarbons (PAHs) is predicted using the Community Multiscale Air Quality Model (CMAQ). Emissions of alkanes by length and structure (cyclic, linear, or branched) are obtained by speciation of the National Emission Inventory (NEI). SOA formation is then parameterized based on the yield of SOA from n-dodecane with adjustments for different length and structure alkanes. Emissions of PAHs are obtained in a similar manner with SOA yields parameterized based on naphthalene. Although SOA from anthropogenic hydrocarbons is predicted to be relatively minor, alkanes and PAHs generally contribute 5-10% and as much as 75% of the SOA from anthropogenic hydrocarbons (single-ring aromatics, alkanes, and PAHs) in the eastern U.S. and California. Higher contributions to ambient OA might be expected if emissions of long chain alkanes >C12 and PAHs were missing from current emission inventories.

Golam Sarwar, Sergey Napelenok, Shawn Roselle, Kathleen Fahey, Rohit Mathur

We examine differences in aerosol sulfate predictions in the United States between two versions of the Community Multiscale Air Quality (CMAQ) models (CMAQv4.7.1 and CMAQv5.0). The recently released CMAQv5.0 includes several updates to the representation of sulfur chemistry. These include updating the ionic strength and hydrogen ion calculations to account for the newly added crustal materials in the model and using predicted iron and manganese concentrations for sulfur oxidation via metal catalysis instead of specified background concentrations. They also include updated gas-phase reaction rate for the oxidation of sulfur dioxide by the hydroxyl radical as well as the aqueous-phase reaction rates of the five sulfur oxidation pathways represented in the model. To maintain consistency between the two models, the off-line photolysis calculation is used. The study uses anthropogenic emissions estimates from the National Emissions Inventory and biogenic emissions estimates prepared using the Biogenic Emissions Inventory System (BEIS). The meteorological driver for this CMAQ model application is the PSU/NCAR MM5 system (version 3.5). Model results obtained with CMAQv5.0 are compared to those obtained with CMAQv4.7.1. Predictions from both models are compared to the observed data from the available surface monitoring networks. Finally, model sensitivity runs are conducted to better understand the importance of selected key parameters on aerosol sulfate predictions.

Heather Simon, Prakash V. Bhave

            Organic mass (OM) is a major component of fine particulate matter (PM_2.5), but it is the most difficult to represent in numerical air quality models.  To accurately simulate ambient OM concentrations, one must represent both organic carbon (OC) and non-carbon organic mass (NCOM).  To date, these two aerosol components have been distinguished in very few models.  In this study, we add primary emissions of NCOM to the Community Multiscale Air Quality (CMAQ) model by assigning source-specific OM/OC ratios to the primary OC estimates from the National Emissions Inventory: 1.7 for biomass combustion, 1.25 for automobile exhaust, and 1.4 for other sources.  We estimate 600,000 tons/year of NCOM emissions from the U.S. alone, which is 12% of the national PM_2.5 inventory (Reff et al., 2009).  After OC and NCOM emissions are injected into the photochemical model, we simulate oxidative aging of the primary OM which produces secondary NCOM.  This oxidative aging is simulated as a second order reaction between primary organic carbon (OC_pri) and OH radicals. 

            Recently, we devised a method for estimating the spatial and seasonal distributions of NCOM from Interagency Monitoring of Protected Visual Environments (IMPROVE) measurements across the U.S. (Simon et al., 2011).  When compared against these field data, the standard CMAQ model systematically underestimates summertime OM/OC ratios and grossly underestimates the spatial and temporal variability of NCOM.  In this poster, we evaluate OM/OC ratios from our new model formulation against those derived from ambient measurements. The effects of (i) using source-specific OM/OC ratios during emission processing, and (ii) treating the oxidative aging of primary OM in the atmosphere, are evaluated separately and in combination for two month-long simulations (January 2002 and July 2002).  The new model treatment allows for better estimates of seasonal and regional trends in OM/OC ratios. 

Reff, A., Bhave, P., Simon, H., Pace, T., Pouliot, G, Mobley, D., Houyoux, M.  Emissions Inventory of PM_2.5 Trace Elements across the United States.  Environmental Science and Technology, 2009; 43(15): 5790-5796

Simon, H., Bhave, P.V., Swall, J.L., Frank, N.H.  Determining the Seasonal Variability in OM/OC Ratios across the United States Using Multiple Regression.  2011,  Atmospheric Chemistry and Physics, 11, 2933-2949

Che-Kai Yeh, Tu-Fu Chen, Ken-Hui Chang

The result of air quality modeling can be used for control strategy planning and assessment. Thus, the accuracy of air quality model's simulation is one of key issues to model developer and user. In air quality model,gas-phase chemical mechanism is the major part of atmospheric chemistry mechanism (Stockwell et al., 1997). Chemical reactions are extremely complicated in real atmosphere. Modeling result will be more accurate and closer to the real atmospheric environment if model contains more chemical reaction or uses more realistic reaction parameter like photolysis rate constant.

This research is subject to modify the gas-phase chemical mechanism in Models-3/CMAQ that includes development of RACM code on the basis of RADM2 code and updating of photolysis rate constants according to the database of Jet Propulsion Laboratory of NASA and MPI (Max Planck Institute). In order to realize the difference for changing chemical mechanism and updating rate constants, one original version (RADM2 + original photolysis rate constant) and three new versions of air quality models are simulated and compared each other using MM5 modeling results as meteorological input.

An ozone event from May 22 to29, 2003 is selected to simulate meteorological condition and air quality in this study. The results show that the averaged daily maximum ozone concentration, compared to the original version, increase 12 ppb (10 %) in Pingtung area for updating photolysis rate constant only and remaining RADM2 mechanism, while it increase 20 ppb (16 %) for using RACM mechanism and remaining original photolysis rate constant. However, when RACM mechanism with new photolysis rate constant is adopted the averaged daily maximum ozone concentration increase up to 28 ppb (about 23 %) in the same area compared to the result of original version.

According to the detail comparison among the four cases simulation results, the major effect of updating photolysis rate constant for ozone is the production rate in daytime, while the change of gas-phase chemical mechanism can lead to increase ozone concentration for all time. It is also the reason why ozone concentration can be different at nighttime in different gas-phase chemical mechanism.

Shunliu Zhao (1), Amir Hakami* (1), Matthew Turner (2), Daven Henze (2), Shannon L. Capps (3), Ted Russell (3), Athanasios Nenes (3)

(1) Carleton University

(2) University of Colorado at Boulder

(3) Georgia Institute of Technology

Adjoint models can efficiently evaluate the derivatives of a metric (function) of model outputs with respect to all inputs, and hence play an essential role in sensitivity analysis, optimization-based simulations, uncertainty analysis and model tuning. As a widely used comprehensive regional air quality model, CMAQ has been equipped with an adjoint model for gas-phase processes. With the increased recognition of the impacts of aerosols on climate and human health, an extension of the adjoint model to aerosol processes is desired. The aerosol processes include aerosol microphysics, aerosol thermodynamics, heterogeneous chemistry and cloud processes. Our presentation will focus on the development of an adjoint model of aerosol dynamics.

This work adopted the automatic differentiation tool, TAPENADE, to facilitate the adjoint development. To enhance the computational efficiency of the adjoint code, manual modifications were made from time to time. The developed adjoint was evaluated against the second-order accurate complex variable method (CVM). Compared to the central finite difference method, the CVM is not subject to subtractive errors and can hence provide more accurate results by allowing a much smaller step size. The benchmark scenario of CMAQ4.7 was used to provide a variety of conditions for the evaluation after a 12-h spin-up period. After the process-by-process comparisons between the adjoint and the CVM, comparisons were conducted for all the aerosol processes of aerosol dynamics. It was observed that the adjoint and CVM sensitivities were compatible for all the cases for 1- and 12-h model runs. Further experiments will be performed after the adjoint models of aerosol thermodynamics and cloud dynamics are incorporated into the evaluation framework.

The developed adjoint model is in progress of being integrated to the CMAQ-ADJ framework. Real-life implementations will be carried out to demonstrate the accuracy, efficiency and robustness of the model under different environmental conditions. One potential application of the full adjoint model is to assess the relative impacts on a metric such as aerosol induced mortality from various emissions sources in support of policy making.

K. Wyat Appel, Charles Chemel, Shawn J. Roselle, Kenneth Schere, Thomas Pierce, Rohit Mathur, S.T. Rao and Stefano Galmarini

The Air Quality Model Evaluation International Initiative (AQMEII) is a model evaluation effort involving numerous research groups from North American and Europe with the goal of advancing the methods for evaluating regional-scale air quality modeling systems.  As part of the AQMEII project, the CMAQ model has been applied to simulate air quality over North America (NA) and Europe (EU) for the year 2006.The CMAQ simulations utilized the same meteorological driver (WRF) and model for lateral boundary conditions (GEMS) for the NA and EU domains.  However, the emissions used are specific to each continent, and are based on national emission inventories for 2005 and 2006. The simulation results are evaluated for daily daytime (8am – 8pm LST) average ozone and daily average particulate matter (PM) against available surface observations.  For NA, the primary sources of observed ozone and PM2.5data are the AQS (U.S) and NAPS (Canada) networks, while for EU the AirBase, AURN and EMEPS networks are the primary sources of ozone and PM10observations.

Michael G. Barna, Marco A. Rodriguez, Kristi A. Gebhart, Bret A. Schichtel, Tom Moore, John Vimont, and William C. Malm

The CAMx regional air quality model was used to simulate nitrogen deposition and ozone formation in the western US, with a focus on impacts at national parks. This simulation used the 2005 emission inventory and meteorology developed for the Four Corners Air Quality Task Force (4CAQTF). CAMx was run for the entire year using a 12 km domain that covered most of the western US. Although air quality in many areas of the western US is typically regarded as pristine, current ozone monitoring at numerous national parks shows concentrations exceeding the current 75 ppb standard, with many more parks likely to be in violation of the proposed 60 - 70 ppb standard. Nitrogen deposition is also an issue at several western national parks, where sensitive ecosystems can be altered when nitrogen deposition exceeds a critical threshold. Recent wet- and dry-deposition estimates for total nitrogen in the western US range between 0.8 kg ha-1 yr-1 in western Washington to 4.2 kg ha-1 yr-1 in California's Central Valley. At Rocky Mountain National Park, located in northern Colorado and the site of extensive research on the effects of nitrogen deposition to sensitive alpine ecosystems, the measured total wet deposited nitrogen for 2005 was 1.9 kg ha-1 yr-1, which is greater than the projected 'critical load' of 1.5 kg ha-1 yr-1. It is notable that current monitoring efforts do not consider reduced gaseous nitrogen (namely ammonia), even though significant ammonia sources exist throughout the western US that could potentially have large impacts at downwind national parks. These sources are primarily composed of agricultural operations such as fertilizer application and animal feedlots, and include California's Central Valley, Idaho's Snake River Valley, and northeastern Colorado, and all are examples of ammonia sources near national parks that contain sensitive high-alpine ecosystems. CAMx estimates of deposited ammonia will be considered along with the routinely reported values of nitric acid, ammonium and nitrate to provide a more complete picture of the overall nitrogen deposition budget.

A. Chtcherbakov, R. Bloxam, Y. Hall, S. Wong

Ontario Ministry of the Environment, Toronto, Canada

There were a number of time periods with mostly clear skies and temperatures above 30°C during the 2010 ozone season (May to September) however there were not many days where the Ontario Ministry of the Environment's Air Quality Index (AQI) reached poor levels due to ozone concentrations as compared to 2005 when there were many more cases when the AQI reached poor levels.

Since meteorological conditions in both 2005 and 2010 appeared to be conducive to periods of high ozone concentrations, an assessment of ozone based on regional air quality modelling with CMAQv4.6 and analysis of observed ozone concentrations was performed. The purpose of this study was to assess the relative impacts of emissions and meteorological conditions on ozone formation in 2010. To separate these two factors, the assessment was performed in several stages.

In the first scenario, emissions were kept exactly the same as for the 2005 base case except for biogenic emissions which were recalculated using 2010 meteorological input data. Plume rise for major point sources was also recalculated to reflect the 2010 meteorological conditions. This scenario assesses only the meteorological differences between the 2005 and 2010 ozone concentrations.

In the second scenario emissions from major point sources were updated to 2010 or 2009 values (depending on the availability of the data- NPRI data from 2009 for Canada and 2010 EGU sources from the US). Emissions from on road mobile sources were recalculated to represent meteorological conditions and fleet distribution in 2010.

Modelled CMAQ results showed the combined effect of changed emissions and meteorological conditions on ozone concentrations and were compared with the monitoring data. The results of the analysis are presented in this study.

B. Czader, X. Li, and B. Rappenglueck

Photochemical modeling of ozone formation in the Houston area generally underestimates the concentrations of free radical precursors contributing to ozone formation. There are unresolved questions about the quantitative contribution of direct emissions and of atmospheric reaction products to the formation of compounds that are related to the free radical chemistry, and ultimately, its role in Houston's ozone production.

Here we present modeling results of a few selected SHARP 2009 episodes using the Weather Research Forecast (WRF) model and the Community Multiscale Air Quality model (CMAQ) along with an evaluation of modeled results with special emphasis on using the Moody Tower measurements to assess how well modeled representation of atmospheric chemistry can simulate observed chemistry. The relative importance of photolysis of ozone (O₃), nitrous acid (HONO), formaldehyde (HCHO) and hydrogen peroxide (H₂O₂) as radical sources in the Houston atmosphere is also addressed.

George Delic, HiPERiSM Consulting, LLC, P.O. Box 569, Chapel Hill, NC

27514

This presentation continues a decade-long study of CMAQ behavior when compiled and executed with vendor-supported compilers on commodity hardware platforms. In the past CMAQ has been ported to compilers from the Portland Group and (more recently) the Intel Corporation. We propose that the time has come to add other compilers to the CMAQ fold: specifically new results are presented for Absoft Fortran from the Absoft Corporation. A great deal has changed over ten years in compiler development, and while compilers from different vendors tend to leap frog each other in performance, all have undergone ground-breaking evolution in following hardware developments. These developments are not reflected in the limited choice of compilers and the options chosen in the standard releases from the CMAS Center download site. One purpose of this report is to rectify this limitation by concurrently evaluating the latest compiler versions from Intel (12.0), Portland (11.5) and Absoft (11.1) for CMAQ 4.7.1 on 64-bit Linux operating systems using top-of-the-line performance hardware from Intel and Advanced Micro Devices (AMD). These processors are, respectively, the Intel Nehalem quad-core (W5590) and the AMD 12-core (6176SE). For fastest MPI runtimes all tests are performed locally (on-node) to avoid any interconnect latency issues, using 8 to 48 MPI processes. In each case compiler options have been chosen after extensive research into three areas fundamentally significant for CMAQ: the memory model, numerical precision, and time to completion. These areas have been explored in the standard download of the model for the small benchmark model case contained therein, as well as full 24 hour episodes on a 279 X 240 Eastern US domain at 12 Km grid spacing with 34 vertical layers with multiple gas chemistry solvers.

Antara Digar, Xue Xiao, Daniel S. Cohan, Rice University,

and Kristen Foley, U.S. EPA.

Air quality models predict pollutant concentrations and responsiveness to controls based on simplified representation of the complex nonlinear physical and chemical processes in the atmosphere, yielding results that are inherently uncertain. Probabilistic estimates of pollutant responses to emission changes considering model uncertainties can help inform the formulation of robust air pollution abatement strategies. Bayesian analysis allows these estimates to be adjusted based on the performance of each model formulation. Previous studies have examined uncertainties in model input parameters (parametric) and formulations (structural) separately. We present a framework that uses actual observations at ozone monitors for evaluating the relative performance of the model under various structural and parametric settings using a Bayesian Monte Carlo (BMC) approach. Model ensemble method and direct sensitivity analysis using the High-order Decoupled Direct Method (HDDM) in the Comprehensive Air quality Model with extensions (CAMx) are applied to characterize ozone sensitivities. Considering a summertime air pollution episode in Texas, we explore how various factors like alternative anthropogenic and biogenic emission inventories, chemical mechanism, boundary conditions, dry deposition scheme, and photolysis rates may influence ozone sensitivities to NO_X and VOC emissions from the Dallas-Fort Worth region. This study showcases how Bayesian probabilistic analyses via an ensemble approach can supplement deterministic estimates of ozone response. Methodological issues regarding the metrics for comparing model results with observations and for assessing the Bayesian results are also discussed.

Chuen-Meei Gan^1,2, Y.H. Wu¹, B.Gross¹and F. Moshary¹

¹Optical Remote Sensing Lab, 140^th Street at Convent Ave., T553, CCNY, New York, NY 10031, USA

²EPA, 109 T. W. Alexander Drive, Research Triangle Park, NC 27709

Email: Gan.Meei@epamail.epa.gov

Air quality model forecasts from Weather Research and Forecast (WRF) and Community Multiscale Air Quality (CMAQ) are often used to support air quality applications such as regulatory issues and scientific inquiries on atmospheric science processes. In urban environments, these models become more complex due to the inherent complexity of the land surface coupling and the enhanced pollutants emissions. Thus, if the surface parameter forecasts such as PM_2.5 are not accurate, this makes it difficult to diagnose model performance issues. Because of this reason, getting accurate boundary layer dynamic forecasts is as important as quantifying realistic pollutants emissions. In this study, we explored the usefulness of vertical sounding measurements on assessing meteorological and air quality forecast models. In particular, we focused on assessing the WRF model (12km x 12km) coupled with the CMAQ model for the urban New York City (NYC) area using multiple vertical profiling and column integrated remote sensing instruments (e.g. lidar, ceilometers and sunphotometer). This assessment was helpful in probing the possible root causes for WRF-CMAQ overestimates of surface PM_2.5 occurring both predawn and post-sunset in the NYC area during the summer of 2007 and 2010. Moreover, we found that the significant underestimates in the WRF Planetary Boundary Layer (PBL) height forecast was a key factor in explaining this anomaly. On the other hand, the model predictions of the PBL height during daytime when convective heating dominates were found to be highly correlated to lidar derived PBL height with minimal bias. In addition, this study includes a mathematical method using a direct Mie scattering approach to convert aerosol microphysical properties from CMAQ model into optical parameters (e.g. aerosol optical depth and Angstrom coefficient) for direct comparisons with multi-wavelength (1064-532-355 nm) lidar and sunphotometer measurements, located in City College of New York. This multispectral information may provide better insight into aerosol speciation and production inconsistencies within the model.

James Godowitch, Robert Gilliam, and Golam Sarwar

Due to the lack of routine concentration profile measurements, the evaluation of photochemical model concentrations aloft has been rather limited. Therefore, in-situ observations made during airborne flights from intensive field studies have been relied upon in efforts to investigate a model's ability to simulate species concentrations in the daytime planetary boundary layer (PBL). In particular, concentrations collected aloft by instrumented research aircraft during the Second Texas Air Quality Study (TexAQS II) are valuable observational data sets for assessing the ability of CMAQ to reproduce ozone (O₃) and other pollutant species within and above the afternoon PBL. Model simulations with CMAQv4.7.1 using WRF-generated meteorology with four-dimensional data assimilation (FDDA) were performed on a domain encompassing the greater Houston, TX metropolitan region with a 4-km grid cell size and 34 vertical layers spanning the TexAQS II experimental study period of August 1 through October 15, 2006.

Emphasis in this model evaluation is on the high resolution (in the horizontal and vertical) O₃ measurements obtained by a downward-looking lidar system on-board the NOAA Twin Otter aircraft at various distances downwind of Houston. Comparisons of modeled O₃ concentrations will be made to investigate how well the horizontal distribution and vertical structure of ozone generated by CMAQ compare to the observed O₃ patterns across the urban plume. In addition, CMAQ model results using refined meteorological fields generated from WRF simulations which incorporated supplemental wind profiler measurements into the FDDA procedure will also be evaluated against the observed vertical O₃ cross-sections to determine if these modeled ozone patterns more accurately replicate the spatial ozone pattern than the base case results. Selected species profile measurements obtained during flights of the NOAA WP-3 research aircraft in the region from selected case study days will also be compared against modeled concentration profiles.

Syuichi ITAHASHI, Itsushi UNO, Soontae KIM

We found that the variability in the fine-mode (submicron) aerosol optical depth (AOD) over the oceans adjacent to East Asia increased (4-8 %/yr) to a peak around 2005-2006 and subsequently decreased (4-7 %/yr) during 2000-2010, as revealed based on MODIS aerosol sensor and WRF-CMAQ. Such fluctuations in AOD are thought to reflect the widespread installation of fuel-gas desulfurization (FGD) devices in power plants in China because aerosol sulfate is a major determinant of the AOD in East Asia. In this next-step study, we analyzed the sensitivity of SO₂ emissions to aerosol sulfate, and AOD through one of the useful source-receptor analysis tools of Direct Decoupled Method (DDM) which was implemented in CMAQ version 4.7.1

James T Kelly¹, Christopher G. Nolte², and Prakash V. Bhave²

¹Office of Air Quality Planning & Standards, US EPA, RTP, NC 27711

²National Exposure Research Laboratory, US EPA, RTP, NC 27711

Associations between atmospheric particulate matter (PM) and adverse health effects are widely reported. A large fraction of atmospheric PM is secondary inorganic matter formed following the emission and reaction of gas-phase pollutants. The processes of secondary inorganic aerosol (SIA) formation must be simulated with sufficient accuracy in regional air quality models (AQMs) to develop pollution control strategies to meet federal PM standards. Compared with secondary organic aerosol, the chemical and physical processes leading to SIA formation are well understood. However, representing the understanding of inorganic aerosol processes in AQMs is computationally challenging. A number of studies have reported computationally efficient approaches for simulating the thermodynamic and dynamic processes governing SIA formation. However, due to the lack of availability of observations and of different SIA modules in the same host AQM, photochemical simulations based on different SIA treatments are rarely if ever inter-compared and evaluated against observations of size-composition distributions and gas-particle partitioning. Here, simulations of SIA formation in the Southeast US are conducted for the Bay Regional Atmospheric Chemistry Experiment (BRACE) period using CMAQ with the standard aerosol module and with the sectional, fully-dynamic UC-Davis aerosol module. Model results are inter-compared and evaluated against detailed observations from the BRACE campaign, which include cascade-impactor observations for inorganic particle components. Model performance for total mass concentrations of inorganic PM components is similar for CMAQ and CMAQ-UCD. However, differences in modeled size distributions and gas-particle partitioning are evident. This work attempts to understand these differences in terms of the different approaches for aerosol process modeling and evaluates their impact on predictions of dry deposition of sulfur, oxidized nitrogen, and reduced nitrogen.

James T Kelly and Kirk R Baker

Office of Air Quality Planning & Standards, US EPA, RTP, NC 27711

Chenxia Cai, Jeremy Avise, Bruce Jackson, and Ajith Kaduwela

Planning and Technical Support Division, California Air Resources Board, Sacramento, CA 95812

Despite much progress in recent decades, many regions of California suffer from poor air quality due to ozone and particulate matter (PM) pollution.  To develop optimal emission control strategies, air quality modeling systems are needed that can accurately predict the response of pollution concentrations to emission reductions.  However, simulating air quality in California is challenging due to the complex terrain, diverse emissions, and other complicating factors.  In May-June 2010, the Calnex field study was conducted to answer important scientific questions about emissions, chemical transformations, climate processes, transport, and meteorology in California.  The study provides a rich dataset of observations from aircrafts, ships, and supersites, which are supplemented by California’s routine monitoring networks.  This dataset is being used by the California Air Resources Board (CARB) and US EPA in diagnostic evaluations of their photochemical modeling systems to improve understanding of the model’s capabilities and to identify areas for model development.  In this study, results of preliminary simulations by CARB and US EPA for the Calnex 2010 study are inter-compared and evaluated against observations.  Note that the work is considered preliminary because emission inputs for the 2010 period are not yet available.  However, the study represents the first time that US EPA and CARB modeling systems have been directly inter-compared.  Simulation results for cases where both modeling systems use the same emissions are considered along with an EPA simulation based on the 2005 National Emission Inventory to identify potential similarities and differences in the systems.  Future work will involve a more comprehensive inter-comparison and evaluation where model simulations are based on emissions representative of 2010.

Daegyun Lee(1), Sangkyun Kim(1), Daewon Byun(2), Hyuncheol Kim(2), Fong Ngan(2), Soontae Kim(3), Chongbum Lee(4), Chankgrae Cho(4)

(1) National Institute of Environmental Research, S. Korea

(2) NOAA Air Resources Lab., USA

(3) Ajou University, S. Korea

(4) Kangwon National University, S. Korea

Regional air quality models such as the Community Multiscale Air Quality (CMAQ) model, have been widely used to study and simulate multi-scale air quality issues. Although they are capable of providing high quality atmospheric chemistry profiles through the utilization of high resolution inputs relating meteorology and emissions with chemical reactions, they cannot simulate air quality accurately if other input data are not appropriate and reliable. There have been few studies on the importance of chemical initial conditions (ICs) as it is considered that the impact of concentration fields specified at the beginning of simulation wears off quickly. This study demonstrates significant errors during the early part of the simulation can damage model predictions and conversely if the ICs are prescribed appropriately with available observations, they can compensate shortcomings of the air quality prediction system especially when episode-based emissions inputs representing real-life emission variations such as forest fires as well as effects of long-range transport events that are not reflected in the basic model inputs. The key hypothesis of the present study is that prediction of aerosols can be improved by the initialization of the aerosol fields with the satellite-derived Aerosol Optical Depth(AOD). We compare effects of using fine-mode and total AOD for the initialization in terms of regional bias characteristics.

Note: The manuscript of this study has been accepted by Atmospheric Environment and in press now.

Peng Liu, Alexandra P. Tsimpidi, Yongtao Hu, Armistead G. Russell, Athanasios Nenes
Georgia Institute of Technology, Atlanta, GA, 30332

In order to investigate how future air quality will respond to a variety of global change, which includes not only the climate change, but also the changes in emissions, technology and land-use pattern, etc, it is important to provide reasonable predictions for meteorological conditions related to air quality of the present year. Therefore, 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.

In this work, we will evaluate the downscaling results over the U.S. of GISS ModelE by WRF of the year 2006 to 2010, the modeling period of which serves as the present year when comparing the results of the future year. Five-year modeling period is chosen because, to evaluate the results, statistical distributions of key outputs of the meteorological conditions (e.g., temperature, rain, humidity, winds, planetary boundary height, precipitation and storm track locations) will be compared to corresponding distributions of observations. The need to evaluate model results using statistical distributions arises because the present years modeled are climatically representative, not actual years. In other words, regional meteorological fields provided by WRF are driven by output from simulation of global climate models (GCMs). The WRF modeling domain has a 36-km horizontal grid-spacing, and 35 layers in vertical with the top at 5000pa.

Yongqiang Liu, Scott Goodrick, Gary Achtemeier

Center for Forest Disturbance Science, USDA Forest Service

The performance of Daysmoke and a number of other smoke plume rise models including those used in SMOKE are evaluated in this study using field measurements. The measurements were made with a ceilometer for 20 prescribed burns in the Southeast from mid-winter to early summer during 2009-2011. About half of the burns had burned areas over 1000 acres. The averaged smoke plume height over all burns is about 1 km. Smoke plume heights have the same magnitude for the burns in spring, and are lower (higher) for the burns in winter (summer) by about 0.2 km. The evaluation results will be presented.

To understand the atmospheric behavior of radioactive materials emitted from the Fukushima Daiichi nuclear power plant after the nuclear accident that accompanied the great Tohoku earthquake and tsunami on 11 March 2011, we simulated the transport and deposition of iodine-131 and cesium-137 using a chemical transport model. The model roughly reproduced the observed temporal and spatial variations of deposition rates over 15 Japanese prefectures (60-400 km from the plant), including Tokyo, although there were some discrepancies between the simulated and observed rates. These discrepancies were likely due to uncertainties in the treatment of emission, transport, and deposition processes in the model. A budget analysis indicated that approximately 13%of iodine-131 and 22%of cesium-137 were deposited over land in Japan, and the rest was deposited over the ocean or transported out of the model domain (700 x 700 km²). Radioactivity budgets are sensitive to temporal emission patterns, and thus accurate estimation of emissions to the air is important for estimation of the atmospheric behavior of radionuclides and their subsequent behavior in land water, soil, vegetation, and the ocean.

Robert Nissen, Paul Makar, Colin di Cenzo, Andrew Teakles, Harry Yau, Junhua Zhang, Qiong Zheng

The Community Multiscale Air Quality (CMAQ, v4.6) and A Unified Regional Air-quality Modelling System (AURAMS, v1.4.2) models were run at 12-km resolution over coastal British Columbia for a one-month period in the winter (January 28-February 28) and the summer (July 15-August 15) of 2005. The Global Environmental Mesoscale (GEM) Regional model provided the driving meteorology for both air quality models. The same emissions inventory and the same horizontal grid projection were employed for both models. In this presentation, the ozone and PM2.5 fields of the two models will be compared to ground observations for coastal and interior locations. Both spatial and temporal aspects will be considered. Verification statisics will be reported, and discrepancies between the two models and observations will be explained where possible in terms of numerical simulation techniques, parametization schemes, chemical schemes and other model variables.

Stephen B. Reid, Erin K. Pollard, Yuan Du, and John C. Stilley - Sonoma Technology, Inc.

Douglas R. Lawson - National Renewable Energy Laboratory

This study involved a top-down emissions inventory evaluation that compared gridded emissions data to ambient monitoring data collected at four urban areas in the Upper Midwest: Chicago, Milwaukee, Gary, and Detroit. This evaluation was performed on 2005 and 2008 emissions inventories being used by the Lake Michigan Air Directors Consortium (LADCO) to conduct regional air quality modeling in the Great Lakes area. Special attention was given on the on-road mobile source component of the inventories, as LADCO used the MOBILE6.2 model to estimate on-road emission for the 2005 inventory and EPA's new MOVES model to estimate on-road emissions for the 2008 inventory.

Analyses of emissions and ambient data included pollutant ratio and hydrocarbon composition comparisons by season, day of week, and wind quadrant. Results were used to identify areas of agreement and differences between the ambient data and emission inventories, to identify aspects of the emission inventories or emissions modeling processes (e.g., spatial allocation, temporal allocation, and chemical speciation) that may need improvement, and to demonstrate the utility of top-down emission inventory evaluation techniques. In general, emissions and ambient data compared most favorably on weekdays in areas dominated by mobile source emissions.

Golam Sarwar, Rohit Mathur, Daiwen Kang, George Pouliot, Rob Gilliam, David Wong

In this study, the Community Multiscale Air Quality (CMAQ) model (version 4.7.1) was applied to simulate air quality in Houston during the 2006 Texas Air Quality Study (TexAQS) period. Meteorological data were obtained from the Weather Research and Forecasting(version 3.2) model. Both the meteorological and air quality model contained 35 vertical layers with a surface layer height of 20 meter. Nested air quality model simulation was conducted: the larger modeling domain covered the eastern United States using 12-km grid resolution while the smaller domain encompassed the central and eastern Texas using 4-km grid resolution (fine-scale).Anthropogenic emissions were obtained from the National Emissions Inventory developed by the United States Environmental Protection Agency. However, all point source emissions in the National Emissions Inventory for Texas were replaced by the specialized emissions inventory developed during the TexAQS study. Biogenic emissions were estimated using the Biogenic Emission Inventory System version 3.14. Model results are compared with observations from the Air Quality System network and surface measurements from the TexAQS study. Predicted carbon monoxide concentrations compare reasonably well with the observed data [overall Normalized Mean Bias (NMB) H10%]. Predicted nitrogen oxides concentrations appear to be high compared to the observed data (NMB > 100%). The over-predictions occur both during the day and night. The overall NMB for daily maximum 8-hr ozone is less than 10%. However, the model overestimates daily maximum 8-hr ozone at low observed values and underestimates at high observed concentrations. The underestimation is particularly severe when observed daily maximum ozone levels are greater than 80 ppbv. The model overestimates daily mean PM_2.5 concentrations compared to the observed data (NMB H34%).

Huy N.Q. Tran and Nicole Mölders

University of Alaska Fairbanks, Geophysical Institute & College of Natural Science and Mathematics, Department of Atmospheric Sciences, 903 Koyukuk Dr., Fairbanks, AK 99775-7230, USA

The Weather Research and Forecasting (WRF) offline coupled with the Community Multiscale Air Quality (CMAQ) model system (WRF-CMAQ) is used to simulate the PM_2.5 concentrations in the Fairbanks nonattainment area for an episode in winter 2009/10. Performance of WRF-CMAQ will be evaluated with observed values from stationary stations, and with high temporal and spatial resolution temperature and PM_2.5 data collected by instrumented vehicle. The Alaska adapted CMAQ model version will be used to evaluate contributions of mobile sources to the elevated PM_2.5 in the Fairbanks non-attainment area, and to calculate the gradient of concentration dilution along the distance from he traffic road. The later application can be use to interpolate future measurements from instrumented vehicles onto maps of concentration dilution.

Kazuyo Yamaji (Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Japan)

Jie Li (Institute of Atmospheric Physic (IAP), China)

Itsushi Uno (Kyushu University, Japan)

Yugo Kanaya (Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Japan) 

FumikazuTaketani  (Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Japan) 

We investigated concentration level and temporal and spatial distributions of atmospheric aerosols over Northeast Asia based on both regional model (WRF/CMAQv4.7.1) and observations. First of all, performances of simulated aerosols were evaluated using in detailed observational data of intensive filed campaign at remote site (Fukue island) in the East China Sea. The simulated total PM2.5 was 12 micro g / m3 for monthly average in May 2009, that was reasonable as compared with observation, 13 micro g / m3. On the simulated and observed PM2.5, SO₄^2- was largest contributor, approximately 60 % of total PM2.5. NH₄⁺ (20 %) and OC (10 %) are next major species. Each MB of SO₄^2- and NH₄⁺ were -0.2 and -0.3 micro g / m3, respectively. On the other hand, simulated OC was approximately a half of observation. As for continuous observation in 2009, this model could capture seasonal change of observed PM2.5, but simulated PM2.5 was underestimated. Although observed PM2.5 might include H₂O, model seemed to underestimate.   

Johannes Bieser, Armin Aulinger, Volker Matthias, Markus Quante

Benzo[a]pyrene (BaP) has been proven to be toxic and carcinogenic. Since 2010 the European Union officially established target values for BaP concentrations in ambient air. In this study BaP concentrations over Europe have been modelled using
a modified version of the Chemistry Transport Model CMAQ which includes the relevant reactions of BaP. CMAQ has been run using different emission datasets for  the years 1980, 2000, and 2020 as input data. In this study the changes in BaP  concentrations between 1980 and 2020 are evaluated and regions which exceed the European annual target value of 1 ng/m³ are identified, i.e. the PoValley, the Paris metropolitan area, the RhineRuhr area, Vienna, Madrid, and Moscow. Additionally the impact of emission reductions on atmospheric concentrations of BaP is investigated. Between 1980 and 2000 half of the BaP emission reductions are due to lower emissions from industrial sources. These emission reductions, however, only contribute to one third of the total ground level BaP concentration reduction. Further
findings are that between 2000 and 2020 a large part (40%) of the BaP concentration reduction is not due to changes in BaP emissions but caused by changes in
emissions of criteria pollutants which have an impact on the formation of ozone.

Jaemeen Baek, Charles O. Stanier, Gregory R. Carmichael

University of Iowa

Epidemiological studies have shown that elevated concentrations of fine particulate matter with aerodynamic diameter less than 2.5 micro-meters (PM_2.5) are related with cardiovascular and respiratory related hospital admissions or mortality rates.  Among PM_2.5 species, black carbon has been linked to cardiovascular mortality, lowered heart rate variability, and blood pressure.  To better understand the relationship between human health and particulate exposure, it is necessary to understand the source contributions that attribute to atmospheric concentrations as well as total concentrations.  Carbonaceous PM_2.5 measurements are important information to understand relationships between human health and black carbon, but they are limited by time (one measurement for every three to six days) and space (discrete point locations).  The chemical transport model, such as CMAQ, is another way to obtain air pollutants exposure that is based on state-of-the-art knowledge.    However, air quality modeling is resource consuming and source apportionment using a 3D model often requires heavier computations.  The CMAQ tracer method to apportion major sources of primary PM_2.5 is applied for black carbon source apportionment for 2000-2004 over Chicago with 4km resolution.  The CMAQ tracer method requires less computational resources than a brute force method, and tracks primary PM_2.5, which includes black carbon, primary organic carbon and unspecified PM_2.5.  Comparison of total and CMAQ-source apportioned carbonaceous PM with monitored values from the STN network will be presented.  Performance analysis of meteorology simulations using WRF and CMAQ simulations will be analyzed together to estimate uncertainties in PM2.5 simulations. 

Elizabeth M. Blayney and J. Jason West

Current methods of estimating the health burden associated with air pollution use atmospheric model output to drive health impacts, ranging from local scales to global scale analyses conducted at coarse grid resolutions.  Here, we aim to quantify how the grid cell resolution of concentration affects estimates of mortality attributable to air pollution in the United States. Specifically, we learn how health outcome estimates driven with course resolution concentrations differ from results produced with finer resolution conducted in the same geographical area. This work addresses important unresolved uncertainties in studies using global chemical transport models to estimate health impacts.

We first estimate the total premature mortalities attributable to ozone and PM_2.5 in the continental US by using the output of the Community Multi-scale Air Quality Monitoring System (CMAQ) model for 2005 at 12 km resolution. Total and respiratory-related premature mortalities are estimated using the Environmental Benefits Mapping and Analysis Program (BenMAP). Ozone, total PM_2.5, and particulate species’ concentrations have been manipulated to calculate metrics consistent with concentration-response relationships in epidemiological studies.

Second, we evaluate the importance of grid resolution by artificially averaging over 12 km grid cells to coarser resolution, for grids ranging in size from 12 km to >100km, approaching the resolution of global chemical transport models. This averaging method isolates the influence of grid resolution in subsequent analysis.  Pollutant concentrations at these coarser resolutions are then evaluated in BenMAP. This multi-step evaluation is performed separately for ozone, PM_2.5, and the components of PM_2.5, as these species have different spatial distributions that may influence the error caused by grid cell resolution. Estimates of total mortality attributable to air pollution in the United States, and the dependence of these estimates on grid resolution are significant for epidemiology, environmental science, and environmental policy. 

Shannon Capps, Daven Henze, Armistead Russell, and Athanasios Nenes

Aerosols affect climate and are associated with degradation of human health; therefore, effective emissions control scenarios must address aerosols and aerosol precursor gases. The regulation of particles with diameters less than 2.5 micrometers (i.e., PM_2.5) brings this challenge to the forefront of State Implementation Plan development. A significant fraction of this dry, fine mode aerosol mass is made up of inorganic species including SO₄⁼, NO₃^-, and NH₄⁺. Secondary inorganic aerosol formation is described in chemical transport models (CTMs) by thermodynamic equilibrium models. The resulting nonlinear relationships between aerosol precursor emissions, such as SO₂, NO_x, and NH₃, and inorganic aerosol concentrations confound emissions control decision-making. Receptor-oriented sensitivity analysis provides a scientific-basis for decision-making by efficiently elucidating relative impacts of a variety of sources on a particular location with spatial, sectoral, and temporal resolution.

The GEOS-Chem adjoint has been employed to reveal receptor-oriented sensitivities of inorganic components of PM_2.5 in selected, densely populated areas of the U.S. to aerosol precursor emissions sources both regionally and throughout the world. Specifically, the sensitivities derived from a previously implemented adjoint of an adapted MARS-A equilibrium thermodynamic model^1,2are compared with those from the first implementation of the adjoint of ISOYYOPIA (ANISOYYOPIA) in a CTM adjoint. Additionally, the adjoint is used to quantify which emissions sectors lead to inorganic aerosol concentrations that would likely cause exceedance of the National Ambient Air Quality Standard for PM_2.5 in different seasons.

1. Henze, D., Hakami, A. & Seinfeld, J. Development of the adjoint of GEOS-Chem. Atmos. Chem. Phys. 7, 2413-2433 (2007).

2. Henze, D., Seinfeld, J. & Shindell, D. Inverse modeling and mapping US air quality influences of inorganic PM_2.5 precursor emissions using the adjoint of GEOS-Chem. Atmos. Chem. Phys. 9, 5877-5903 (2009).

Xin Qiu¹, Louise Aubin², Ron Haley¹

¹ Novus Environmental Inc.
² Region of Peel, Department of Public Health
Contact: xinq@novusenv.com

The epidemiological evidence clearly shows that existing levels of urban air pollution is adversely affecting human health. The CMA estimates that by 2013 nearly 90,000 people will die from the acute effects of air pollution and there will be more than 700,000 deaths due to long-term exposures, with the economic cost accumulating to over $250 billion . Sensitive populations including children, the elderly, people with allergies, asthma or heart and lung conditions are more vulnerable and face greater risk.

While municipalities are aware that poor air quality affects health and are exploring ways to improve local air quality, they do not have the authority to regulate air emissions. However, they can influence many local emission sources and how communities are designed. Municipalities have the opportunity to reduce air pollution related health impacts through land use planning strategies that reduce exposure such as land use compatibility guidelines and set-back's based on air quality considerations. Improved understanding of the local factors that contribute to air quality and how future land use development and transportation strategies influence air quality is needed to help municipalities develop strategies to improve local air quality.

This presentation identifies the factors that influence air quality at a municipal level, and presents an improved forecasting approach that can predict regional and local scale air quality impacts at high resolution. How this system can be used to predict changes in air quality resulting from new development and changing patterns of emissions is discussed. A municipal perspective on how this information can be used to guide development where there are sensitive land uses in close proximity to major sources of air pollution is presented.

Adam Reff, Jerry Davis, Bryan Hubbell

U.S. Environmental Protection Agency Office of Air Quality Planning and Standards Research Triangle Park, NC 27711

In 2005, the U.S. Environmental Protection Agency studied the response of ambient fine particulate matter (PM2.5) to emissions changes during its regula- tory impact assessment (RIA) using a response surface model (RSM) based on the community multiscale air quality (CMAQ) model. In this work, we use the RIA RSM to calculate PM2.5 sensitivities across the U.S. and create explana- tory statistical models for their spatial variability. Statistical models are con- structed using sensitivities as dependent variables and pollutant concentrations, meteorological variables, and emissions levels as independent covariates using geographically weighted regression (GWR). GWR is a technique that effectively accounts for spatial variation in statistical relationships and reduces autocor- relation by generating separate coefficients for each point in the domain under study. A self-organizing map (SOM) of the nearly 4000 resulting coefficient maps is then used to concisely visualize results. Preliminary analyses suggest such linkages as PM2.5 sensitivities to SOx EGU emissions positively contribut- ing to concentrations of oxidants (O3+H2O2+HOX) in the eastern U.S., and VOCs and water vapor mixing ratio negatively contributing to sensitivities of NH3 emissions from area sources. We anticipate that results will provide infor- mation to improve the effectiveness of air quality management plans, as well as provide new insights about atmospheric chemical dynamics inherent in CMAQ for future users and developers of the model.

Raquel Silva, Zac Adelman, Meridith Fry, Susan C. Anenberg, J. Jason West

Short-term and long-term exposure to outdoor air pollution are associated with increased cardio-respiratory mortality and morbidity, as shown by a number of epidemiologic studies. Anthropogenic emissions have caused a considerable rise in the global concentrations of ground-level ozone (O₃) and fine particulate matter (PM_2.5) in the atmosphere, not only in urban areas but also in rural regions. In this study, we will estimate the contribution of emissions from different sectors to the global burden of human mortality due to outdoor air pollution.  We will simulate O₃ and PM_2.5 concentrations using the MOZART-4 global chemical transport model. MOZART-4 will be run at a 0.6°x0.5°horizontal resolution using GEOS-5 meteorological fields, to estimate concentrations at a finer resolution than in previous global health assessments. Input emissions for the present day (2005) will be characterized by the IPCC AR5 Representative Concentration Pathway 8.5 (RCP8.5) global emissions inventory.  First, we will estimate anthropogenic contributions to air pollutant concentrations as the difference between MOZART-4 simulations of present-day (2005) and preindustrial O₃ and PM_2.5, and will estimate global premature human mortality by applying health impact functions derived from epidemiologic studies and methods developed previously by our group.   Second, we will quantify the contributions of individual anthropogenic emissions sectors (e.g. transportation, industry, residential & commercial emissions) to current O₃ and PM_2.5 concentrations and premature human mortality using brute-force simulations in which we zero-out single emissions sectors.

Chris Werner, Bok Haeng Baek, Mohammad Omary, Zachariah Adelman, Saravanan Arunachalam, J. Jason West

Diesel engines are a major source of emissions of fine particulate matter (PM_2.5) and its precursors in the U.S., with non-road engines contributing a significant portion (2.5 - 3% of national primary PM_2.5 emissions, for example). Newly available emissions control technologies that nearly eliminate PM from diesel exhaust are included on all new equipment currently sold. Since regulations accelerating fleet turnover and mandating retrofit of these technologies onto existing engines do not exist outside of California, the adoption of the CARB program (which affects primarily construction & mining equipment) on a nationwide basis represents an opportunity for reducing emissions of PM that would otherwise continue through the long service lifetime of in-use diesel equipment, with co-benefits of reducing emissions of NO_x, CO, and hydrocarbons.

Here we investigate the air quality benefits of this hypothetical national program, applied to non-road diesel engines. Using a base case previously developed from the EPA 2005 National Emissions Inventories (NEIs), SMOKE is employed to simulate emission reductions of multiple precursor species from the nonroad diesel, construction & mining sectors. CMAQ is run to demonstrate the effects of these reductions on air quality parameters including ozone and PM_2.5. We simulate two 3-month periods (winter and summer) across the continental United States at 36 km resolution, using 2005 model inputs. We aim to evaluate the national benefits for improved air quality, and to identify regions with potentially strong improvements in air quality. These results will support further analysis to assess the benefits of the program for human health.

Jeff Lundgren, Wayne Boulton, Martin Gauthier, Nancy Chan

Potential changes to near surface ozone concentration resulting from proposed onshore and offshore oil exploration facilities in the Caspian Sea Region were examined. CMAQ version 4.6 was applied in a nested 36 km, 12km and 4km configuration with the 36 km nest covering much of central Asia and the 4km nest centered on the area of the development in the northern Caspian Sea. WRF version 3.1 was used to develop meteorological fields. The model was run for a three month summer period from June through August, when highest ozone concentrations are expected. Model baseline year was 2007.

The background emission inventory for the region was compiled by the University of North Carolina based on previous work done in the region. Oil exploration facilities emissions scenarios were based on 1) peak emissions during standard operation expected to occur in the year 2030, and 2) on maximum short term emissions that might occur during emergency flaring. Biogenic emissions were modeled using MEGAN and model boundary conditions for the three month period were extracted from global GEOSCHEM results for the same model period.

Resulting ozone concentrations were examined in terms of one-hour and eight-hour maximum onshore and offshore concentration. In addition, the AOT40 metric for cumulative exposure of vegetation to daytime ozone over 40ppb was calculated for both the maximum five-day exposure and over the full three month period. Model results highlight increase in potential for overwater boundary layer ozone transportation and resulting ozone exposure to vegetation at the sea shore.

Alison Eyth, Rich Mason, Alexis Zubrow

EPA's Office of Air Quality Planning and Standards is working to upgrade its emissions modeling platform to use 2007 and 2008 base years. The new "Version 5" platform will be based on the 2008 NEI, and augmented with data being brought in from other sources as appropriate. We anticipate updated versions of the platform becoming available as improvements to inventories and methodologies are implemented and tested. Planned improvements include the use of SMOKE-MOVES for onroad emissions, improved latitude-longitude coordinates for point sources, updated information on installed control devices, improved modeling of oil and gas emissions, an updated approach to average fire emissions, temporal allocation of residential wood and agricultural ammonia emissions based on meteorological conditions, updates to spatial surrogates, and more finely detailed speciation of particulate matter needed to drive CMAQ version 5. Expected uses of the new platform will be discussed. The selection of 2007 or 2008 for the base year will be made according to the requirements of each specific modeling project.

Rob Gilliam, Jon Pleim, and Jim Godowitch
Atmospheric Modeling and Analysis Division, NERL, USEPA, Research Triangle Park, NC

Air quality models, such as the Community Multiscale Air Quality (CMAQ) model, rely on meteorology models to provide wind, temperature, humidity, air density, radiation, cloud hydrometeors (cloud water, ice, rain, and snow), surface fluxes, and other surface parameters on a computational grid. While the air quality modeling community generally uses numerical models developed for weather prediction, such as the Weather Research and Forecasting (WRF) model, meteorology modeling for air quality applications has distinctly different requirements and emphasis. For example, meteorology modeling for air quality requires accurate wind speed and direction at the surface and also throughout the PBL both day and night and in the nocturnal residual layer for realistic advective transport of point source and urban plumes. Another particular characteristic of air quality modeling is that applications are often retrospective which allows for continuous use of four dimensional data assimilation (FDDA). Recent improvements in FDDA techniques, including the use of new observation data such as from radar wind profilers and Velocity Azimuth Display (VAD) wind profiles from NWS Doppler radar, have significantly improved meteorological and air quality model simulations. Air quality applications also emphasize the diurnal evolution of PBL height and simulation of nocturnal low level jets. Recent improvements in PBL modeling yield more realistic simulations of both stable and convective boundary layers. Examples of WRF and CMAQ modeling experiments with improved FDDA and PBL components will be presented.

David Wong, Rohit Mathur, Jonathan Pleim, Francis Binkowski, Tanya Otte, Rob Gilliam, Aijun Xiu, Shawn Roselle, Jeffrey Young

¹ Atmospheric Modeling and Division, NERL, U.S. EPA, RTP, NC 27711
² Institute for the Environment, University of North Carolina, Chapel Hill, NC 27599

Traditionally, atmospheric chemistry-transport and meteorology models have been applied in an off-line paradigm, in which archived output on the dynamical state of the atmosphere simulated using the meteorology model is used to drive transport and chemistry calculations of atmospheric chemistry transport model (CTM). A modeling framework that facilitates coupled on-line calculations is desirable since it (1) provides consistent treatment of dynamical processes and reduces redundant calculations, (2) provides ability to couple dynamical and chemical calculations at finer time-steps and thus facilitates consistent use of data, (3) reduces the disk-storage requirements typically associated with off-line applications, and (4) provides opportunities to represent and assess the potentially important radiative effects of pollutant loading on simulated dynamical features. A coupled on-line atmospheric modeling system is developed based on the Weather Research and Forecasting (WRF) meteorological model and the Community Multiscale Air Quality (CMAQ) air quality modeling system. This 2-way coupled WRF-CMAQ system is being made publicly available as part of the 2011 CMAQ modeling system release. The flexible design of the system which facilitates configurations for the both on-line and off-line modeling paradigms will be described. The impacts of including direct radiative forcing of simulated aerosol distributions on modeled dynamical and chemical features on regional to hemispheric scales will be discussed. Initial evaluation of the model against measurements from routine networks and specialized campaigns will be presented.

Heather Simon(1), Sharon Phillips(1), Norm Possiel(1), George Pouliot(2), John Koupal(3), Harvey Michaels(3)

1) Office of Air Quality Planning and Standards, EPA, RTP, NC

2) National Exposure Research Laboratory, EPA, RTP, NC

3) Office of Transportation and Air Quality, EPA, Ann Arbor, MI

Recently EPA's Office of Transportation and Air Quality (OTAQ) released a new model to estimate emissions from onroad mobile sources. This new model, MOVES, is being phased in to replace the existing MOBILE6 model and has already been used in air quality modeling in support of several EPA rules. Nationwide, MOVES estimates 70% higher NOx emissions than MOBILE6 for a 2005 annual inventory. In this presentation, we focus our analysis on the Northeast states in which 2005 annual NOx emissions from MOVES are 50% greater than emissions estimates from the MOBILE6 model. Using onroad emissions inputs from both MOBILE6 and MOVES, we assess 2005 CMAQ model performance of NOx, NOy, and ozone. Model bias for each CMAQ run is evaluated on hourly and monthly temporal scales for rural and urban sites separately. This analysis shows that CMAQ run with MOBILE6 emissions tends to slightly underestimate NOx concentrations while CMAQ run with MOVES emissions tends to slightly overestimate NOx concentrations. NOx bias is highest during evening hours and during summer months. Rural NOy concentrations are overestimated by CMAQ for both sets of emissions inputs. Finally, MOVES emissions lead to slight improvements in one-hour average ozone performance, especially at higher observed concentrations. The reasons for this improved performance are explored.

Christoph Knote(1,2), Dominik Brunner(1,2)

1) Laboratory for Air Pollution / Env. Technology, Empa - Materials Science and Technology, Duebendorf, Switzerland
2) C2SM, Center for Climate Systems Modeling, ETH, Zurich, Switzerland

Wet scavenging is a major sink for atmospheric trace gases and aerosols. However, this process is not only a sink, but also a transformation procedure (e.g. processing of aerosols in clouds), and can even represent a source (e.g. production of sulphate by aqueous-phase oxidation of SO2). Chemistry transport models sometimes represent this process only as a sink, or assume independent columns in their parameterizations. With the advent of kilometre-scale modeling it becomes clear that several assumptions made are not applicable anymore: evaporation and condensation processes in clouds need to be considered explicitly. Precipitation has to be treated as prognostic quantity, with horizontal transport across column borders and the possibility to evaporate before reaching ground. Wet scavenging parameterizations also need to account for that.

In our presentation we give an overview of the development of a comprehensive wet scavenging scheme for the online-coupled regional chemistry transport model COSMO-ART (Vogel et al., 2009, ACP), recently developed KIT-Karlsruhe (Germany). This modeling system is based on state-of-the-art components and is comparable to WRF/Chem. However, a detailed wet scavenging scheme has been missing. We have coupled the MESSY submodel SCAV (Tost et al., 2006, ACP) to COSMO-ART. This scheme features an explicit description of the transfer processes between gas- / aerosol-phase and cloud droplets / rain. A comprehensive aqueous-phase chemistry mechanism based on MECCA (Sander et al., ACP, 2005) is included via KPP, allowing for easy modifications. We have extended the SCAV scheme by an explicit consideration of the condensation and evaporation of cloud droplets in a manner consistent with COSMO microphysics. Furthermore, the chemical composition of rain drops is now also consistent with the treatment of rain in the meteorological part, allowing for 3D transport and evaporation. We will briefly outline our adaptations, then show results from idealized 2D flow over hill simulations for evaluation purposes and finally present the impact of the new coupling on realistic 3D simulations over the European domain.

References:

Sander, R. Kerkweg, A., Jockel, P., Lelieveld, J. (2005). Technical note: The new comprehensive atmospheric chemistry module MECCA. Atmos. Chem. Phys., 5, pp. 445-450.
Tost, H., Jockel, P., Kerkweg, A., Sander, R., Lelieveld, J. (2006). Technical note: A new comprehensive SCAVenging submodel for global atmospheric chemistry modelling. Atmos. Chem. Phys., 6, pp. 565-574.
Vogel, B., Vogel, H., Bumer, D., Bangert, M., Lundgren, K., Rinke, R., Stanelle, T. (2009). The comprehensive model system COSMO-ART - Radiative impact of aerosols on the state of the atmosphere on the regional scale. Atmos. Chem. Phys., 9, pp. 8661-8680.

Dan Loughlin, Farhan Akhtar, Bill Benjey, Chris Nolte and Rob Pinder

U.S. EPA Office of Research and Development

Daven Henze

University of Colorado

The MARKet ALlocation (MARKAL) model is an energy system optimization model. The U.S. EPA has developed a database that allows MARKAL to be applied to examine alternative energy scenarios for the U.S. through 2055. The database covers the resource supply, electricity production, industrial, residential, commercial, and transportation sectors. It includes characterizations of current and future energy demands and energy-related technologies. MARKAL uses this information to select the technology and fuel pathways that meet energy demands at least cost. At the same time, the model tracks a wide range of pollutant emissions, including NOx, SO2, PM2.5, PM10, CO, CO2, CH4, N2O, BC and OC. MARKAL also allows constraints can be applied to one or more of these pollutants and accounts for the emission constraints in selecting an optimal pathway.

The primary goal of this presentation is to introduce the MARKAL modeling framework as a tool for developing long-term emission projections. We also briefly describe two applications of MARKAL, providing a flavor of how the model is being used in ongoing research activities.

In the first application, the Global Change Air Quality Assessment, MARKAL is being used to investigate the challenges that will be faced in managing future air quality. Alternative future scenarios are being modeled, providing a set of long-range emission projections. These scenarios, which extend from 2005 through 2055, can incorporate different assumptions about population growth and migration, economic growth and transformation, energy resources, land use, climate change, technology development, and air quality, climate and energy policies. MARKAL emission outputs are converted into regional- and technology-specific emission growth factors and used within the Sparse Matrix Operator Kernal Emission (SMOKE) model to produce a future-year inventory. Air quality implications are then evaluated using the Community Multiscale Air Quality (CMAQ) model.

The second MARKAL application is within the GLIMPSE project (GLIMPSE stands for GEOS-Chem/LIDORT adjoint radiative transfer model Integrated with MARKAL for the Purpose of Scenario Exploration). In this project, MARKAL is being used to explore multi-pollutant control strategies that simultaneously address criteria pollutants, greenhouse gases, and regional impacts of short-lived climate forcers such as black and organic carbon and sulfate aerosol. This framework is being used to examine tradeoffs and synergies related to various air quality and climate goals. For example, we are identifying which technology pathways for climate change mitigation also provide major health co-benefits.

Shaocai Yu, Rohit Mathur, Jonathan Pleim, David Wong,

Steve Howard, and S.T. Rao

Atmospheric Modeling and Analysis Division,

National Exposure Research Lab, U.S. EPA, RTP, NC 27711

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 simulated with the newly developed two-way coupled WRF-CMAQover the Gulf of Mexico (4-km domain) and eastern United States (12-km domain). The cloud droplet number concentrations are diagnosed from the activation of CMAQ-predicted aerosol particles. 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 and Morrison cloud microphysics scheme) and an existing radiation scheme (CAM and YYTMg) in the WRF. With the satellite observation data such as CERES, MODIS and CALIPSO and field experimental data such as 2006TexAQS/GoMACCS, we will evaluate the cloud properties and indirect aerosol forcing.

Lisa Silverman, Sheryl Ehrman, Dale Allen, Kenneth Pickering, Heidy Plata, and Thomas Pierce

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 compared spring 2005 and 2006 tropospheric NO2 columns from CMAQ with satellite-retrieved columns from the Ozone Monitoring Instrument (OMI). After filtering out periods with excessive clouds, biomass burning, lightning-NO production, and / or anthropogenic emissions, we examined the contribution of pulsing to soil NO emissions following 26 unique precipitation events. The change in columns was compared to biogenic NO emissions from BEIS3 and to the changes in CMAQ columns for simulations with and without soil-NO emissions. As part of this evaluation, we divided the events into high and low precipitation categories and examined the change in modeled and satellite-retrieved columns and emissions for each of these cases. As expected, the high precipitation cases resulted in a much stronger satellite response than the low precipitation cases. The satellite NO2 observations were also plotted against the CMAQ model estimates for NO2 as well as soil moisture data to determine if a correlation could be observed. It was observed that the NO2 emissions estimates for the cases in the low precipitation rate category had a strong linear correlation on the day of the precipitation event with soil moisture and CMAQ model NO2 estimates. In addition, for the cases in the high precipitation rate category, the OMI DOMINO satellite product had a strong correlation with soil moisture and with the CMAQ model. Additional analysis of these episodes is underway and will be discussed.

Bonyoung Koo¹, Prakash Karamchandani¹, Greg Yarwood¹, Eladio Knipping²

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

2. EPRI, 2000 L Street NW Suite 805, Washington, DC 20036

The Community Multiscale Air Quality (CMAQ) modeling system Version 5.0 (CMAQv5.0) is scheduled for release in September 2011. This release differs from previous releases in that external scientists and developers can contribute alternative science treatments that will be included as part of the official release after these contributions are tested and evaluated by EPA. This paper presents two such contributions to CMAQ 5.0. The first contribution is the Advanced Plume Treatment (APT) for Plume-in-Grid (PinG) applications to resolve sub-grid scale processes, such as the transport and chemistry of point source plumes. This treatment has been available in previous versions of CMAQ, but has usually been 1 or more releases behind the official CMAQ release. The second contribution provides an alternative framework for organic particulate matter (PM) formation using a volatility basis set (VBS) approach. Of the various components of atmospheric PM, organic aerosol (OA) is among the most abundant and the least understood and is usually under-predicted substantially in current air quality models. Traditionally, primary organic aerosols (POA) have been treated as non-volatile and non-reactive. However, environmental chamber studies that vary temperature and aerosol dilution have shown that POA can be semi-volatile and undergo gas-particle partitioning in the atmosphere similar to secondary organic aerosols (SOA). Also, it is now known that further oxidation of semi-volatile organic compounds (SOA and POA) can create products with lower volatility leading to increased OA formation downwind. The organic aerosol (OA) module in CMAQ 5.0 uses a two-product model for SOA with polymerization decreasing SOA volatility over time and assumes non-volatile POA with oxidative aging increasing the mass of POA over time. The VBS approach provides a unified framework for gas-aerosol partitioning of both POA and SOA including chemical aging. A new organic PM module based on the VBS approach is implemented in CMAQ 5.0 and its effect on modeled OA formation is discussed. The VBS implementation for CMAQ 5.0 can differentiate OA from biogenic, biomass burning and anthropogenic sources.

Ellen Cooter, USEPA, RTP, NC

Jesse Bash, USEPA, RTP, NC

Verel Benson, Benson Consulting, Columbia Missouri

LiMei Ran, UNC Center for the Environment, Chapel Hill, NC

Bidirectional CMAQ advances the estimation of ammonia emissions from agricultural cropland beyond the temporally static off-line NEI inventory to a dynamic, spatially and temporally resolved in-line method based on the compensation point approach. Pilot study results presented previously have indicated that moving in this direction can result in substantial changes, i.e., improvement in CMAQ characterization in the partitioning of N removal between wet and dry deposition, and spatial sphere of influence of point and area sources. Movement forward from the pilot study approach to estimate agricultural management influences on CMAQ ammonia flux requires substantial modifications to ensure national consistency, reasonable spatial resolution of farm management behaviors, proper representation of biogeochemical processes leading to soil emissions and to achieve the flexibility needed to support the exploration of alternative future societal- and climate-driven emissions scenarios. This presentation will describe the methodological foundation and practical implementation required to provide the input information required by bidirectional CMAQ. Input files and preliminary evaluation of those estimates will be presented. These new inputs are provided as part of the bidirectional science option in CMAQv5.0 and are required to execute this science option.

Aika Yano

M. Talat Odman

Biomass burning plumes are not well resolved in current air quality modeling systems due to insufficient grid resolution and/or inadequate sub-grid treatments. We developed a system that couples sub grid plume model and air quality model to enhance the ability to predict air quality impacts from biomass burnings. Adaptive Grid CMAQ (AG-CMAQ) is capable of increasing grid resolution that can track the biomass burning plumes at the regional grid scales. AG-CMAQ still requires precise burning emission inputs such as the plume's spatial spread and vertical profile at the sub-grid scale. Here we use a sub-grid plume model called Daysmoke to predict short range dispersion of prescribed burning plumes. The plume concentrations are carried over to AG-CMAQ when a biomass burning plume travels downwind and the plume resolution becomes compatible with the air quality model grid resolution. The interface where the two models share their information is time dependent, and the location of the interface is updated frequently in an analysis called "handover". Since the size of the grid cells is a key factor in determining how well a plume is resolved, we performed a case study on different parameters that affect grid resolution and handover process. The Atlanta smoke event of 2007 was used to compare and evaluate the performance of the updated adaptive grid Daysmoke CMAQ (AGD-CMAQ). The handover method is evaluated in both AGD-CMAQ and standard CMAQ (uniform grid) to distinguish the effect of adaptive grid resolution. In simulating prescribed burn plumes, the adaptive grid is refined where the plume emissions are incorporated. Therefore, we assessed different methods of inserting plume emissions into AGD-CMAQ and compared how the girds adapted and how well the plume distributions are captured by the grids. Some of the parameters in the weight function that determine grid refinements are varied and compared as well. By combining the best setting for each parameter that was tested in this case study, AGD-CMAQ that is optimal for plume modeling is determined.

Gill-Ran Jeong¹, Jesse Bash², Daven Henze¹, and Rob Pinder²

¹Department of Mechanical Engineering at University of Colorado, Boulder, CO 80309 ²Atmospheric Modeling and Analysis Division, US EPA, Research Triangle Park, NC 27711

NH₃ flux measurements indicate that agricultural and natural surfaces can either be a source or sink depending on the ammonium concentrations and pH in the soil and vegetation, and ambient NH₃ concentrations. Our recent studies show that the NH₃ concentrations of CMAQ peak sharply at the surface, at levels higher than observed surface measurements, though satellite observations aloft show higher concentrations than the model predicts. Preliminary results indicate that modeling bidirectional exchange may partially rectify these discrepancies. Since the NH₃ bidirectional exchange model in CMAQ alters emissions and deposition simultaneously, we will use process analysis to investigate the governing processes that lead to model improvements. CMAQ (v 5.0) was run for a 12 km Continent United States (CONUS) domain from June 1^st ~ August 31^st, 2009 with and without the new NH₃ bidirectional exchange option, bidi and base cases respectively. Soil ammonium and pH used by the bidirectional exchange model to estimate NH3 fluxes from fertilizer applications were estimated from crop specific fertilizer application and management practices output from a USDA agro-ecosystem model, Environment Policy Integrated Climate (EPIC). Base case annual fertilizer NH₃ emissions were estimated as a function of fertilizer sales with seasonal adjustment factors and an assumed diurnal profile applied. CMAQ simulations with NH₃ bi-directional exchange were evaluated against retrievals from the Troposphere Emission Spectrometer (TES) instrument aboard NASA's Aura satellite and surface based Ammonia Monitoring Network (AMoN) observations on a regional scale, and against TES special observation in Eastern North Carolina and Carolina Ammonia Monitoring Network (CAMNet) at a local scale.

Matthew Woody, Jeffrey Rissman, Frank Binkowski, J. Jason West, Saravanan Arunachalam

We investigated the impacts of aircraft emissions on fine particulate matter (PM_2.5) at the Hartsfield-Jackson Atlanta International airport for June and July, 2002 using an adaptation of CMAQ called the Advanced Modeling System for Transport, Emissions, Reactions, and Deposition of Atmospheric Matter (AMSTERDAM). Aircraft emissions during the landing and takeoff cycle (LTO) and below 3,000 meters were represented as plume-in-grid (PInG) emissions using AMSTERDAM's PInG treatment. Our previous work using CMAQ-AMSTERDAM focusing on impacts from aircraft emissions to inorganic PM_2.5 and total PM_2.5 indicated aircraft increased average total PM_2.5 concentrations by up to 235 ng m^-3 near the airport and by 1-7 ng m^-3 throughout the Atlanta metro area. However, aircraft reduced concentrations by 0.5-1 ng m^-3 downwind of the airport, attributable to reductions in sulfate aerosol. In another of our previous studies, aircraft emissions were modeled by CMAQ as traditional point sources within the ATL airport grid cell, and we showed that modeled SOA concentrations increased by 2% due to primary organic aerosol (POA) emissions from aircraft, which provided additional surface area for SOA to partition onto. We will present results from additional modeling work that is being performed to a) determine computational efficiency and model sensitivity to the number of points used to represent aircraft PInG emitter locations around a major airport, and b) examining aircraft's impacts on secondary organic aerosol concentrations using the volatility basis set (VBS) within CMAQ-AMSTERDAM to represent the formation and aging of organic aerosols. Parameterization for the VBS components will be determined using current and ongoing field study measurements, chamber studies, and box models specific to aircraft SOA formation.

George Pouliot

Atmospheric Modeling Division, National Exposure Research Laboratory, Environmental Protection Agency, Research Triangle Park, NC 27711

pouliot.george@epa.gov

Jessica McCarty, Carie Ernst, Tyler Kerr University of Louisville, Louisville KY

Amber Soja Institute of Aerospace (NIA), NASA Langley Research Center, Hampton, VA

Biomass burning has been identified as an important contributor to the degradation of air quality because of its impact on ozone and particulate matter. Two components of the biomass burning inventory, rangeland and crop residue burning, are currently poorly characterized in the National Emissions Inventory. Recently, a crop residue burning inventory based on field measurements and remote sensing information has been developed. We will focus on the both post harvest and pre-harvest burning that takes place with bluegrass, corn, cotton, rice, soybeans, sugarcane and wheat. This paper summarizes the processing and assumptions needed to incorporate this new emission inventory of crop residue burning into EPA's National Emission Inventory. We will focus on temporal allocation, speciation, and how to distinguish crop residue burning from other forms of biomass burning. Additionally, preliminary results using this new inventory within a chemical transport model will be provided. Finally, some early results in creating a rangeland burning inventory will also be presented.

Jaemeen Baek, Charles O. Stanier, Pablo E Saide Peralta, Jameson Schoenfelder, Gregory R. Carmichael, and Annmarie G. Carlton

University of Iowa, Rutgers University

In the presence of clouds, gas and aerosol phase species can dissolve into water droplets and participate in aqueous phase chemical reactions, changing the concentrations of those species significantly. These important processes oxidizes S(IV) to S(VI), converts gaseous species to aerosol phase species, and removes both aerosol and gas phase components via deposition. The Community Multiscale Air Quality (CMAQ) model simulates aqueous chemistry as a part of cloud dynamics. The aqueous chemistry module includes wet deposition, scavenging, gas-aqueous phase partitioning, dissociation of acidic species, formation of secondary organic aerosol and aqueous phase oxidation reactions that form sulfate and secondary organic aerosol. As part of a broader effort to develop adjoint applications of CMAQ for aerosol, our objective is to create the adjoint models for the aqueous chemistry processes. The current design of the forward code that uses the forward Euler and bisection method makes automatic calculation of the adjoint infeasible. A more elegant forward solution involves coding the thermodynamic, kinetic, and deposition processes within the framework of the Kinetic PreProcessor, which allows for easier updating of mechanisms and, importantly for our application, automatic generation of the adjoint model. The set of the chemical and physical equations in CMAQ creates a sparse and stiff Jacobian matrix with reaction rates ranging from 10^-2 to 10⁸ and the skill and efficiency of KPP's solutions to this are evaluated. Simulation results and computational efficiency of the KPP generated forward model will be compared with those of CMAQ aqueous chemistry module. Evaluations of tangent linear and adjoint model will be presented as well.

Karsten Baumann¹, Jamie Schauer², Don Blake³, Steve Mitchell⁴, Mike Fort¹, Eric Edgerton⁵

^{1Atmospheric Research & Analysis, Inc., 1000 Perimeter Park Dr, Ste G, Morrisville, NC 27560}

^{2Civil & Environmental Engineering, University of Wisconsin, 660 N Park St, Madison, WI 53706}

^{3Dept of Chemistry, University of California, 570 Rowland Hall, Irvine, CA 92697-2025}

^{4Nicholas School of the Environment, Duke University, Box 90328, Durham, NC 27708}

^{5Atmospheric Research & Analysis, Inc., 410 Midenhall Way, Cary, NC 27513}

Mechanical thinning of the forest understory prior to Prescribed Burning (PB) is believed to be an effective management practice to reduce wild fire risk and restore long-leaf pine savannas in the fire-dependent forest ecosystems of the South-Eastern United States. As part of the DoD-sponsored Defense Coastal/Estuarine Research Program, in situ measurements of PB emissions from the combustion of pine dominated forest understory were conducted in conjunction with detailed before/after fuel inventory surveys, yielding actual fuel consumption. Innovative mobile aerosol composition monitors were employed to measure and distinguish emissions from untreated (control) fuels with mechanically thinned (hydro-axed) fuels. Measured compounds include reactive gases (NH3, HONO, HNO3, HCl, SO2, light organic acids) and particle phase organic compounds (POC), water-soluble ionic species, organic and elemental carbon (OC and EC), and total PM2.5 mass. More than 100 POC species, including key molecular markers were quantified and over 40 non-methane hydrocarbon (NMHC) species measured, including certain aromatics and biogenics that are important PM precursors. Metallic and mineral emission components were determined via energy dispersive Xray fluorescence and inductively coupled plasma mass spectrometry. The measured emissions are being related to ecosystem-specific parameters describing the forest fuel types and conditions of the investigated forest areas. Applying the carbon mass balance, emission factors (EF in g or mg species per kg biomass burnt) were calculated for the suite of aerosol species measured. Preliminary results indicate that more of the available material is being consumed when the fuel is hydro-axed. Furthermore, particulate species emissions from drier fuels are systematically higher than for wetter fuels, with dry hydro-axed fuel causing the highest EF for almost all PM species. OC is the dominant PM2.5 constituent in emissions from all investigated fuels, followed by EC, nitrate, potassium, and chlorine. Emissions of PM precursors like ammonia, aromatics (esp. benzene and toluene), and biogenic organic compounds (alpha-pinene and isoprene) are significant. EF values from this study are compared with corresponding data published by the Emission Factor and Inventory Group in the U. S. Environmental Protection Agency's Office of Air Quality Planning and Standards (AP-42, Fifth Edition, Vol. I (13): Miscellaneous Sources, 13.1 Wildfires and prescribed burning, Suppl. B, October, 1996). For PM2.5, EF values from dry fuel compare well with AP-42, while wetter fuels generate systematically lower EF values (less than half) for both treated and untreated fuels. Both carbon monoxide and methane EF values are less than a third of the corresponding AP-42 levels, with treated fuels producing generally less emissions per mass fuel burned. The same trend exists for NMHC, except that control fuel EF are within 10 % of the AP-42 levels.

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

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

Receptor models (RM) approaches have been the traditional approach for particulate matter (PM) source apportionment analyses. RMs rely on using observed concentrations of the PM composition at a receptor(s) to solve a species balance equation to estimate the source impacts. They require limited knowledge of emission sources such as source profiles but have mathematical and data limitations. Here we develop a hybrid, iterative PM source apportionment approach based on species balances that utilizes a chemical transport model (CTM, e.g. CMAQ) equipped with a sensitivity analysis tool (e.g. DDM-3D) to provide physically and chemically consistent relationships between sources and receptors. This hybrid approach enhances RM results by providing initial estimates of source impacts from a much larger number of sources than are typically used in RMs, and provides source-receptor relationships for secondary species. Further, the method addresses issues of source colinearities. Hybrid results also provide information on uncertainties in the source impact quantification.

We apply this hybrid modeling approach to conduct PM source apportionment at Speciation Trend Network sites nationwide for a winter month across the U.S., with attention to six major cities. The simulated ambient PM concentrations at these receptor sites are apportioned to thirty two separate sources during measurement winter days. Hybrid method source apportionments results are re-grouped to lesser number of categories to match up and compared with the Chemical Mass Balance model (CMB) results. Source impact uncertainties of the approach are estimated in a CMB manner 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 and the associated uncertainties.

Wildland fires can significantly contribute to detrimental air pollution levels. For this reason, their emissions are frequently integrated into multidimensional air quality simulations. Attempts to model the impacts of wildland fires on air quality and simulate the progression of fire pollutant plumes have been previously reported. However, little attention has been given to the vertical distribution of fire emissions included in photochemical models. Commonly, injection of fire emissions into gridded domains is carried out only considering a maximum plume rise estimate. Emissions are then all introduced at this height or uniformly distributed from this level down to the surface. These practices can misrepresent the vertical structuring of wildland fire plumes, which is typically more complex. The simplifications may also affect simulated ground level pollutant levels.

In this work we analyze the sensitivity of surface concentrations to the vertical layering of fire emissions in CMAQ. Several techniques used to develop vertical plume profiles are compared, including empirical estimates, Briggs equations, and detailed profiles produced by DAYSMOKE, a plume rise model developed by the U.S. Forest Service. Model results are evaluated by simulating selected fire episodes in the Southeast for which measurement data are available. Additionally, sensitivity analyses of pollutant concentrations to emissions layering were performed applying the decoupled direct method (DDM). The results of this study provide valuable information on the importance of detailed vertical profiles for wildland fire emissions in air quality modeling and the significance of vertical resolution in effectively simulating elevated plumes using three-dimensional air quality models.

Prakash V. Bhave, Golam Sarwar, Havala O. T. Pye, George A. Pouliot, Heather Simon, Jeffrey Young, Chris G. Nolte, Rohit Mathur

ISOYYOPIA is a computationally efficient thermodynamic equilibrium module used in many air quality models to distribute semi-volatile inorganic compounds between the gaseous and particulate phases. Version 1.7 of ISOYYOPIA treats the SO₄/NO₃/Cl/NH₄/Na/H₂O system and has been incorporated in previous releases of the CMAQ model. The recently developed ISOYYOPIA v2.1 includes new treatment of three crustal cations - calcium, potassium, and magnesium - and optimizes some activity coefficient calculations to reduce the computational burden of the thermodynamic calculations (Fountoukis & Nenes, ACP 2007). In this study, we incorporate v2.1 into CMAQ and evaluate its impact on air quality model predictions through a series of simulation pairs that use v1.7 and v2.1. When the crustal cation concentrations are zeroed out in v2.1, CMAQ results are very similar to those obtained with v1.7. However, emission-perturbation tests indicate that v2.1 is more numerically stable than v1.7. CMAQ simulations are also conducted by turning on the thermodynamics of crustal cations in v2.1 and enabling mass transfer of gaseous HNO₃, NH₃, and HCl to both fine and coarse particles. In these simulations, coarse-mode nitrate increases in inland regions with substantial dust emissions while fine-mode nitrate decreases relative to v1.7. This is due to the preferential transfer of HNO₃ to the coarse particles that contain Ca, Mg, and K. ISOYYOPIA v2.1 also increases fine-mode nitrate in areas with appreciable emissions of anthropogenic fugitive dust (e.g., from road surfaces and agricultural tilling). Overall, incorporation of v2.1 improves the numerical stability of the CMAQ modeling system and improves estimates of particle-phase nitrate across the entire size spectrum. This update is scheduled for public release as part of CMAQ v5.0.

• William Koshak, Earth Science Office, NASA/MSFC-VP61, 320 Sparkman Drive, Robert Cramer Research Hall, Huntsville, AL 35805.
• Harold Peterson, NASA Postdoctoral Program, 320 Sparkman Drive, Robert Cramer Research Hall, Huntsville, AL 35805.

Abstract

The NASA Marshall Space Flight Center introduced the Lightning Nitrogen Oxides Model (LNOM) a couple of years ago to combine routine state-of-the-art measurements of lightning with empirical laboratory results of lightning NOx production. The routine measurements included VHF lightning source data [such as from the North Alabama Lightning Mapping Array (LMA)], and ground flash location, peak current, and stroke multiplicity data from the National Lightning Detection Network" (NLDN). Following these initial runs of LNOM, the model was updated to include several non-return stroke lightning NOx production mechanisms, and provided the impact of lightning NOx on an August 2006 run of CMAQ. In this study, we review the evolution of the LNOM in greater detail and discuss the model's latest upgrades and applications. Whereas previous applications were limited to five summer months of data for North Alabama thunderstorms, the most recent LNOM analyses cover several years. The latest statistics of ground and cloud flash NOx production are provided.

Yunseok Im and Myoseon Jang

A predictive model for the secondary organic aerosol (SOA) formation, including heterogeneous reaction and partitioning [Partitioning Heterogeneous Reaction Consortium Secondary Organic Aerosol Model (PHRCSOA)], has been developed. The PHRCSOA model was tested for toluene SOA produced from the photochemical reaction of toluene in presence of different concentrations of NO_x and SO₂ using the UF Atmospheric Photochemical Outdoor Reactor (UF-APHOR) dual chambers. The SOA model mainly consists of three parts: a gas-phase kinetic model, a partitioning SOA model, and a heterogeneous SOA model. In the gas-phase kinetic model, 140 toluene oxidation products were predicted using an explicit MCM mechanism. Gas-phase products are classified into 20 lumping groups based on their vapor pressures (5 levels) and reactivity (4 levels) for heterogeneous reactions in aerosols. The organic mass (OM) in SOA is decoupled into partitioning OM (OM_P) and heterogeneous OM (OM_H). The OM_P is estimated using the partitioning model used in CMAQ, which was developed by Schell et al (J. Geophys. Res. 2001, 106, 28275-93). For the OM_H calculation, the model developed by Jang (Environ. Sci. Tech. 2006, 40, 3013-22) has been modified in this study to dynamically predict OM_H over the course ofSOA chamber experiments. The modification of OM_H allows for the application of the SOA model to the regional scale model. In the PHRCSOA model of this study, NO_x effects on toluene SOA formation was also estimated through stoichiometric coefficients ( ) related to 20 lumping species. In order to study effects of sulfuric acids on toluene SOA formation, SO₂ was photochemically oxidized with toluene and applied to the PHRCSOA model of this study. The toluene SOA in the absence of SO₂ was compared to the toluene SOA in the presence of SO₂ using dual chamber experiments and simulated with the PHRCSOA model. The outdoor chamber study showed that SO₂ increases the toluene SOA formation. For example, SOA from the toluene experiment(Tol = 54ppb, NO_x =18ppb, SO₂ = 28 ppb) showed higher OC (1.61ug/m³) than OC (0.72ug/m³) from the SOA in the similar condition without SO₂. The toluene SOA yield with SO₂ increases by 137% compared to toluene SOA without SO₂.

A.G. Carlton¹, K.R. Baker², T.E. Kleindienst³, J.H. Offenberg³, M. Jaoui⁴

¹Rutgers Universiry, New Brunswick, NJ

²U.S. Environmental Protection Agency, OAQPS, Research Triangle Park, NC

³U.S. Environmental Protection Agency, ORD, Research Triangle Park, NC

⁴Alion Corporation, Research Triangle Park, NC

Organic aerosol (OA) accounts for a substantial fraction of the U.S. atmospheric particulate matter (PM) burden. While a substantial fraction, 50-80%, of OA is non-volatile and will not evaporate under atmospheric conditions, the amount of semivolatile gas-phase organic compounds in equilibrium with OA ranges from 20-400% of the OA mass. Atmospheric models generally describe this phase transfer (i.e., secondary organic aerosol (SOA) formation) by applying the absorptive partitioning theory of Pankow with either a two-product or volatility basis set (VBS) approach. Chemically explicit OA predictions from CMAQv4.7.1, based in large part on two-product parameterizations, are evaluated with tracer measurements made during CALNEX. Further CMAQ's SOA predictions are mapped according to the VBS approach and OA volatility distributions are qualitatively compared to previously published thermodenuder-aerosol mass spectra.

W. T. Hutzell and J. Streicher

U.S. Environmental Protection Agency, 109 T.W. Alexander Dr., Research Triangle Park, NC, 27709

In CMAQ version 5.0, the in-line method for calculating photolysis rates removes shortcomings that found in version 4.7 of the method. Regarding model physics, the most significant change alters the surface albedo used to calculate photolysis rates. An algorithm calculates albedo based on wavelength, land use category, season, solar zenith angle, snow, and ice coverage. Previously, albedo had the same value over all wavelengths except for the one longest one. Model simulations were used to assess how the upgraded method compares to the off-line method based on look-up tables. The assessment compared observed concentrations of ozone to predictions from each method. Predictions and observations represent concentrations during summer and winter conditions over the continental United States. For the entire domain, the in-line method slightly improves performance for one hour averages of ozone during the summer conditions. Improvement takes place because bias of predictions decreases from observations over states in the Ohio Valley and south of the Great Lakes. During the winter conditions, the in-line method deteriorates performance by increasing bias from observations over Central Atlantic and Southeastern states. A diagnosis showed that NO₂ photolysis correlates with changes in ozone predictions between the two methods. Further analysis investigated how these results depend on the changed albedo. Albedo appears to have the strongest effects on predictions during winter when snow and ice coverage increase the albedo over large areas. We present further results from investigating the role of the surface albedo in predictions. Our conclusions suggest how the albedo algorithm may change to improve predictions.

Greg Osterman, Jessica Neu and Annmarie Eldering¹
Jeff McQueen, Youhua Tang²
Rob Pinder³

¹ Jet Propulsion Laboratory/California Institute of Technology
² National Oceanic and Atmospheric Administration/National Weather Service
³ Environmental Protection Agency

To help protect human health and ecosystems, regional-scale atmospheric chemistry models are used to forecast high ozone events and to design emission control strategies to decrease the frequency and severity of ozone events. Despite the importance of nighttime aloft ozone, regional-scale atmospheric chemistry models often do not simulate the surface nighttime ozone concentrations well and nor do they sufficiently capture the nighttime ozone transport patterns. Fully characterizing the importance of the processes has been hampered by limited measurements of the vertical distribution of ozone and ozone-precursors. The main focus of this work is to begin to utilize remote sensing data sets to characterize the impact of nighttime aloft ozone to air quality events.

We will provide initial results and describe our plans to use NASA satellite data sets, transport models and air quality models to study ozone transport, focusing primarily on nighttime ozone. We will use satellite and ozonesonde data to help understand how effective the air quality models are in simulating ozone in the lower free troposphere. Our primary goal for the work will be to begin to utilize remote sensing data sets to characterize the impact of nighttime aloft ozone to air quality events.
Our specific objectives are:

• Characterize nighttime aloft ozone using remote sensing data and sondes.
• Evaluate the ability of the Community Multi-scale Air Quality (CMAQ) model and the National Air Quality Forecast Capability (NAQFC) model to capture the nighttime aloft ozone and its relationship to air quality events.
• Analyze a set of air quality events and determine the relationship of air quality events to the nighttime aloft ozone.

To achieve our objectives, we will utilize the ozone profile data from the NASA Earth Observing System (EOS) Tropospheric Emission Spectrometer (TES) and other sensors, ozonesonde data collected during the Aura mission (IONS), EPA AirNow ground station ozone data, the CMAQ continental-scale air quality model, and the National Air Quality Forecast model.

Jesse O. Bash

U.S. EPA NERL

Annmarie G. Carlton

Rutgers University

The characterization fate and transport of mercury by photochemical models requires parameterization of the reduction of Hg^II to Hg⁰. Removal of this pathway introduces model wet deposition biases and ambient concentrations. Typically models include aqueous phase reduction of divalent oxidized mercury (Hg²⁺) by the hydroperoxyl radicals (HO₂^.), or use an empirical aqueous phase reduction rate scaled to the estimated lifetime and seasonality of total gaseous mercury. However, the HO₂ reduction mechanism is improbable under environmental conditions and scaled rates do not represent real atmospheric chemical processes. Recent laboratory work demonstrates that C2-C4 dicarboxylic acids (DCA) can complex with Hg^II to form Hg⁰, providing a plausible aqueous phase photochemically driven reduction mechanism. The aqueous phase HO₂ reduction reaction was replaced with a plausible DCA pathway in CMAQ4.7.1 and model predictions were evaluated with wet deposition measurements from the Mercury Deposition Network. Model sensitivities were run for HO₂^. , DCA, and no condensed phase reduction cases. The removal of a condensed phase reduction reaction resulted in over predictions of summertime and wintertime wet deposition (124% and 73% respectively), the HO₂^. case resulted in under predictions of summertime and wintertime wet deposition ( -24% and -33% respectively), and the DCA case resulted in an under prediction of wet deposition in the summertime and an over prediction in the wintertime (-10% and 16% respectively). In addition, CMAQ 4.7.1 with the updated aqueous phase reduction reaction is evaluated against an entire year of wet deposition measurements and qualitatively to ambient speciated mercury measurements taken at AMNet sites in 2009.

Jerry L. Gorline, NOAA/NWS/Meteorological Development Laboratory, Silver Spring, Maryland, and Jeff McQueen, NOAA/National Centers for Environmental Prediction, Camp Springs, Maryland

NOAA is building the National Air Quality Forecasting Capability (NAQFC) in collaboration with the U.S. EPA. The NAQFC links the National Centers for Environmental Prediction's (NCEP) North American Mesoscale (NAM) model with EPA's Community Multiscale Air Quality (CMAQ) modeling system to produce predictions of gridded ground-level ozone and aerosol concentrations. To aid air quality model development and assess air quality predictions, the Meteorological Development Laboratory (MDL) provided the NAQFC implementation team verification metrics for ozone and aerosol predictions produced by the NAQFC.

In 2009, the Meteorological Development Laboratory(MDL) concluded a multi-year study of CMAQ model performance, providing verification results for 2007-2009. For the daily maximum of 8-h averaged experimental ozone predictions, we compared verification results using an exceedance threshold of 75 parts per billion (ppb). For the daily maximum of 1-h averaged developmental aerosol predictions, we compared results using a threshold of 35 ug/m³. MDL provided verification scores for the CONUS domain and statistics for six geographic regions in the CONUS. Small seasonal increases in bias for ozone predictions in 2007-2009 were noted. Biases ranged from slight under-prediction in June to slight over-prediction by July. For developmental aerosol predictions, there were strong seasonal bias changes in 2007-2009, and in earlier testing years, from under-prediction in the warm season, April-September, to over-prediction in the cool season, October-March. While these biases are consistent with missing source contributions (e.g. wildfires), in the summer months, additional complexity of the aerosol test predictions are contributing to large prediction errors, and are the subject of ongoing investigation.

This year MDL will present the results of CMAQ model performance in urban areas compared to performance in rural areas. The EPA has provided MDL with a list of urban and rural classifications and elevation information for air quality observing sites over CONUS. MDL will provide urban vs. rural comparisons, taking particular note of rural sites downwind of major urban areas. MDL will compare model performance as a function of elevation and compare model performance in coastal areas compared to inland sites. If possible, we will investigate high elevation sites to assess performance for sites that are under stronger influence of air masses from free troposphere, if nocturnal Planetary Boundary Layer (PBL) height data are available. These comparisons will be performed over CONUS and for six geographical regions in CONUS. We will also provide an update on the performance of air quality predictions since 2009 for both ozone and aerosols.

J. Richmond-Bryant¹, A. Reff²

¹National Center for Environmental Assessment, U.S. Environmental Protection Agency, 109 TW Alexander Drive, B243-01, Research Triangle Park, NC 27711 USA

²Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, 109 TW Alexander Drive, C304-04, Research Triangle Park, NC 27711 USA

Uncharacterized microscale spatial and temporal variability in urban air pollutant concentration dynamics may potentially contribute uncertainty or bias to epidemiological model results. In this study, methods for quantifying this variability are explored by treating urban buildings as a matrix of bluff bodies to determine the retention of air pollution in their wakes, which are bounded by the street canyons. This method was based primarily on bluff body theory that derived a functional relationship between nondimensional contaminant residence time (H) within a wake and the following fluid properties of the air: 1) Reynolds Number (Re), 2) street canyon height (D) to width (W) aspect ratio (D/W), and 3) turbulence intensity, defined as the square root of turbulence kinetic energy (k) divided by the freestream wind speed (U). H is given by the characteristic time of exponential concentration decay in a bluff body wake nondimensionalized by D and U. Empirical relationships were built from sulfur hexafluoride (SF₆) concentration and meteorological data collected during the Midtown Manhattan 2005 (MID05) Study held in August, 2005 in Manhattan, NY, along with geographical information system (GIS) data describing the building topography. Results were then compared with results from a similar previous analysis using data collected during the Joint Urban 2003 (JU2003) study in Oklahoma City, OK.

For the MID05 data, Re ranged from 1.65x10⁶ to 7.74x10⁷, with a median of 1.13x10⁷. The range of Re was consistent with earlier observations from the JU2003 study, although the measured winds tended to be more turbulent (median k = 2.2 m²s^-2) compared with JU2003 (median k = 0.45 m²s^-2). Values for H ranged from 7.2 to 1186, with a median of 80.9. The distribution of H was substantially wider for MID05 than for JU2003, with values exceeding observations of H by an order of magnitude for single obstacle wind tunnel studies with Re ~ 10⁴. Inverse relationships were validated between H and Re and between H and D/W for the MID05 data and for a pooled data analysis from the MID05 and JU2003 studies. The pooled model of H vs. Re provided a good fit to all of the data but provided a biased estimate of the Oklahoma City model results. The pooled model of H vs. D/W did not provide a good fit, suggesting that the building topographies of the two cities are too different to produce a reasonable comparison. These inter-study comparisons suggest that the relationships may contain underlying site-specific features that would require elucidation prior to generalizing to other urban sites. Overall, results from this work present a foundational method for providing estimates of H based on readily available sources of data. Future work is planned for applying the method to more realistic and reactive ambient pollutants. Other potential applications include development of a module based on the H model to analyze sub-grid scale variability in urban settings within CMAQ or to analyze neighborhood and urban scale spatial variability around central site monitors for use in epidemiological studies. (This presentation does not necessarily reflect the policy of the U.S. EPA.)

K. Pickering, A. Prados, S. Napelenok

EPA's NOx Budget Trading Program has led to substantial emission reductions at major point sources in the eastern and central United States. Surface measurements are insufficient to track the resulting changes in NOx concentrations. Therefore, we use the tropospheric column NO2 data from the OMI instrument on NASA's Aura satellite to monitor these changes. The regions considered for the NO2 trend analysis have been defined using the CMAQ model with the Decoupled Direct Method (DDM) which provides an indication of the relative sensitivity of ambient NO2 to NOx emissions from specific sets of point sources. The sensitivities are calculated for each layer of the CMAQ-DDM model, and tropospheric column-integrated sensitivities were subsequently computed from the output. Clusters of major point sources were considered for this analysis in 8 regions in a Summer 2005 CMAQ-DDM simulation. We then examine changes in OMI tropospheric column NO2 in the summers of 2005 to 2009 in the regions of strong sensitivity.

Naresh Kumar

Scientific and computational advancements in regional air quality models over the past decade have led to their adoption and widespread use by regulators, states, industry, and other agencies for a variety of applications, including the development of regulations, design of air quality control strategies and implementation plans, and short-term operational air quality forecasting. Researchers have also extended their use to new applications ranging from health studies to integration with global-scale models for climate studies. The models have improved significantly over the years in their representation of chemical and physical processes; however, many areas still remain where further improvements are needed. Although there is a great deal of air quality model development taking place at academic, government and private institutions, there is little coordination amongst different groups. A workshop was conducted to bring together researchers from academia, government and private institutions, industry, and other stakeholders to brainstorm on various model development needs. A comprehensive research agenda has been developed that can be used by the community to help guide research plans and to promote collaboration amongst researchers.

The focus of this workshop was limited to the use of atmospheric models to support regional-scale air quality applications, such as forecasting or control strategy development. As a result, this workshop did not address model development needs that pertain specifically to interfacing them with local-scale or global-scale models. Future workshop may define their scope to provide adequate discussion to those and other important research questions.

In this paper, a summary of key messages from the workshop will be presented in four areas of research that were the focus of the workshop. These areas include: Homogeneous-Phase Chemistry; Heterogeneous-Phase Chemistry (including Inorganic and Organic Aqueous-Phase Chemistry); Organic Particulate Matter: Formation and Aging of Secondary Organic Aerosol; and Meteorological Processes Affecting Air Quality.

Shao-Hang Chu

Did the recession impact recent decreases in observed sulfate concentrations

Shao-Hang Chu

USEPA/OAQPS/AQAD

Observed sulfate aerosols are mainly produced from the oxidation of sulfur dioxide (SO₂) in the atmosphere. Most of the SO₂ emissions, however, result from fossil fuel burning. For this reason, most of the aerosol produced from the oxidation of SO₂ is considered to be from anthropogenic origins. Significant drops in the observed sulfate concentrations in the last two years nationwide draw an interest to examine what may be the major driving forces behind this unusual event.

In this study, 2002-2009 sulfate and meteorology data from 30 major cities nationwide were analyzed to examine the possible impact of the economic recession on observed sulfate concentration decreases. The results indicate that, on the average, over half of the sulfate concentration drops observed in these urban areas during 2008 and 2009 cannot be explained by regulatory emission controll effort alone. This is reflected in the meteorologically adjusted sulfate trend. The remaining sulfate concentration decline is likely to come from additional SO₂ reductions resulting from reduced economic activities, industrial productions and utility demands which was commonly realized almost everywhere in the country. Thus, it is determined that the economic recession must have a significant impact on the observed decreases of sulfate concentration in the last two years.

K. Wyat Appel, Shawn Roselle, Kristen Foley, Rohit Mathur

A new version of the CMAQ model, v5.0, will be released in the fall 2011 and includes a number of significant changes to the model. This presentation will cover the operational and diagnostic evaluations of an annual (2006) CMAQ v5.0 simulation. Results from the simulation with the new version of the model will be compared to results from the previous version of the model to highlight areas where the model performance changed significantly between the two versions.

Chris Emery, Jaegun Jung, Jeremiah Johnson, Michele Jimenez and

Ralph Morris

ENVIRON International Corporation

773 San Marin Drive, Suite 2115

Novato, California 94998

Policy Relevant Background (PRB) for ozone is defined by the US Environmental Protection Agency (EPA) as the minimum surface ozone concentration that could be achieved through North American emission controls. Specifically, PRB is the concentration that would occur in the absence of all North American anthropogenic emissions (including Canada and Mexico). As reductions in the US ozone standard approach background levels, PRB takes on increased importance as it determines the feasibility and cost of compliance. Historical PRB estimates have been based upon observational and global modeling analyses. Observational research has contradicted the relatively low levels estimated by older low-resolution modeling, while presenting compelling evidence that ozone entering the US west coast is increasing at a rapid pace. Since PRB is not directly measureable everywhere and at all times, modeling is a useful tool. We summarize an updated comprehensive modeling analysis of PRB ozone for the year 2006, using both low-resolution global (GEOS-Chem) and high-resolution regional (CMAQ and CAMx) chemical transport models from which to elucidate the range in model predictive performance (full emissions) and resulting differences in temporal and spatial background variability over the US (PRB emissions). We also investigate model sensitivity to various components of natural background and non-US emission sources.

D. J. Allen¹, K. E. Pickering², R. W. Pinder³, B. H. Henderson⁴, K. W. Appel³, and A. Prados⁵

[1]{Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD, USA}

[2]{Atmospheric Chemistry and Dynamics Branch, Code 613.3 NASA-Goddard, Greenbelt, MD, USA}

[3]{Atmospheric Modeling and Analysis Division, U.S. EPA, Research Triangle Park, NC, USA}

[4]{University of North Carolina Chapel Hill, NC, USA}

[5]{Joint Center for Earth Sciences Technology (JCET), University of Maryland Baltimore County, Baltimore, MD, USA}

A lightning-nitrogen oxide (NO) algorithm is developed for the regional Community Multiscale Air Quality Model (CMAQ) and used to evaluate the impact of lightning-NO emissions (LNO_x) on tropospheric photochemistry over the eastern United States during the summer of 2006. The scheme assumes flash rates are proportional to the model convective precipitation rate but then adjusts the flash rates locally to match monthly average observations.

Over the eastern United States, LNO_xis responsible for 20-25% of the tropospheric nitrogen dioxide (NO₂) column. This additional NO₂reduces the low-bias of simulated NO₂columns with respect to satellite-retrieved Dutch Ozone Monitoring Instrument NO₂(DOMINO) columns from 41 to 14%. It also adds 10-20 ppbv to upper tropospheric ozone and 1.5-4.5 ppbv to 8-hour maximum surface layer ozone, although, on average, the contribution of LNO_xto surface ozone is 1-2 ppbv less on poor air quality days. Biases between modeled and satellite-retrieved tropospheric NO₂columns vary greatly between urban and rural locations. In general, CMAQ overestimates columns at urban locations and underestimates columns at rural locations. These biases are consistent with in situ measurements that also indicate that CMAQ has too much NO₂in urban regions and not enough in rural regions. However, closer analysis suggests that most of the differences between modeled and satellite-retrieved urban to rural ratios are likely a consequence of the horizontal and vertical smoothing inherent in columns retrieved by the Ozone Monitoring Instrument (OMI).

Within CMAQ, LNO_xincreases wet deposition of oxidized nitrogen by 50% and total deposition of nitrogen by 11%. This additional deposition reduces the magnitude of the CMAQ low-bias in nitrate wet deposition with respect to National Atmospheric Deposition monitors to near zero.

In order to obtain an upper bound on the contribution of uncertainties in chemistry to upper tropospheric NO_xlow biases, sensitivity calculations with updated chemistry were run for the time period of the Intercontinental Chemical Transport Experiment (INTEX-A) field campaign (summer 2004). After adjusting for possible interferences in NO₂measurements and averaging over the entire campaign, these updates reduced 7-9 km biases from 32 to 17% and 9-12 km biases from 57 to 46%. While these changes lead to better agreement, a considerable NO₂low-bias remains in the uppermost troposphere.

Prakash Karamchandani¹, Greg Yarwood¹, Steven Brown², Shu-Yun Chen¹

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

2. NOAA ESRL Chemical Sciences Division, 325 Broadway, R/CSD7, Boulder, CO 80305

An advanced reactive Lagrangian puff dispersion model is used to understand the influence of nighttime chemical transformations, including heterogeneous reactions, and transport processes on the formation of next-day ozone downwind of source regions. Our study takes advantage of nighttime plume measurements collected by the NOAA WP-3D research aircraft during the second Texas Air Quality Study in 2006 and detailed analysis of this data by NOAA to supplement the plume modeling presented here. SCICHEM, the plume model used in this study, is a state-of-the-science puff model with complete chemistry treatment, including the CB05 gas-phase chemistry mechanism, aerosol chemistry, and chemistry in clouds. The modeling focuses on plume intercepts of the Oklaunion power plant near Wichita Falls, TX on the night of October 10, 2006. Oklaunion is a coal-burning electric generation unit with one stack and is relatively isolated from other sources. The Oklaunion transects range from 18 to 73 km downwind of the power plant, with transport times ranging from 0.5 to 2.5 hours. Aircraft measurements at the edges of the plume intercepts are used to characterize the background concentrations for SCICHEM. Hourly and date-specific emissions from Oklaunion are obtained from CEM data. The meteorology is characterized by surface and upper-air measurements in the vicinity of the power plant. Model performance is evaluated both by comparing modeled concentrations at downwind plume transects with aircraft measurements and comparing predicted chemical conversion rates to results of the NOAA data analyses. Sensitivity tests are performed to evaluate and improve the representations of plume chemistry, focusing on the rate of NO_x oxidation and the products that are formed.

S. Morteza Mesbah^a, Stephan Schott^b, Amir Hakami^a

^aDepartment of civil and environmental engineering, Carleton University

^bDepartment of public policy and administration, Carleton University

Cap-and-trade emission programs have proven to be an effective regulatory tool compared to the traditional command-and-control approach. In cap-and-trade programs, the right to pollute the environment is defined as a property right that can be traded among polluters. The nature of the pollutant, however, affects the design of the program. NO_x as one of the precursors of ozone is a non-uniformly mixed pollutant with a short atmospheric lifetime. The ozone formation potential of NO_x emissions depends on a number of factors such as location, time of release, and the chemical regime of the trajectory. NO_x SIP call trading program (2003 to 2008) caused a significant reduction in total NO_x from SIP sources in the US. It was successful in reducing total NO_x emissions but may have not optimally reduced ozone as the location-specific ozone formation potentials were neglected. In a previous work, the emission-based cap-and-trade systems (with exchange rate of unity, i.e. all emissions are considered to be of equal importance), and sensitivity-based exchange rates were contrasted. A hypothetical NO_x trading program was constructed that could more effectively reduce surface ozone alongside NO_x emission reductions (Mesbah et al., CMAS, 2009).

This paper is an extension of the previous work for the SIP call trading sources, with season-long adjoint simulations, CEM-augmented emission inventory for the SIP call, and a new methodology to estimate unit level marginal abatement cost curves for power plants in different locations based on real-life operational data and considering the opportunity cost. A marginal abatement cost curve shows the per-unit emission reduction cost as a function of emission levels. The estimation of marginal abatement cost curves is important since a power plant can use its marginal abatement cost curve information to decide whether it is worthwhile to buy emission permits from other facilities or not. Also, different marginal abatement cost curves change the behavior of the market and its effect on ambient air quality.

The unit level NO_x marginal abatement cost is calculated for SIP call sources. The abatement can occur by installation of a control technology or a reduction in electricity generation. Therefore, marginal abatement is the summation of the marginal control technology cost and marginal output reduction cost (opportunity cost). The control technology cost is estimated using The Integrated Planning Model of the US EPA for two common post-combustion control technologies: selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR). To estimate the opportunity cost, regional electricity and fuel price data from CIA and the data of fuel consumption, electricity generation, and NO_x emissions from the clean air market are used. Using the estimated unit level marginal abatement costs, different approaches including a) command-and-control, b) trading system with no exchange rate (SIP call trading), and c) trading system with adjoint exchange rates are compared.

Yunsoo Choi^1,2, Hyuncheol Kim^1,2, Daniel Tong^1,2, Pius Lee¹, Fantine Ngan³, Yunhee Kim^1,2, Rick Saylor⁴, Ariel Stein^1,2, Jeff McQueen⁵, Ivanka Stajner⁶

¹NOAA Air Resources Laboratory, Silver Spring, MD

²Earth Resources Laboratory, Inc

³University of Corporation for Atmospheric Research, Boulder, CO

⁴NOAA Air Resources Laboratory, Atmospheric Turbulence and Diffusion Division, Oak Ridge, TN

⁵NOAA/NWS/National Centers for Environmental Prediction, Camp Spring, MD

⁶Noblis, Inc., Falls Church, VA

Simulation results from the Community Multiscale Air Quality (CMAQ) model version 4.7 over the Conterminous United States (CONUS) for August 2009 are analyzed to evaluate the daily deviation of observed and modeled ground-level concentrations of nitrogen oxides (NO_x) and ozone (O₃) at stations classified according to a geographically-based land use designation or a O₃-NO_x-VOC chemical sensitivity regime. The physical land use designations are derived from the Advanced Very High Resolution Radiometer (AVHYY) global land cover characteristic data representing three categories: urban, forest, and other. The O₃ sensitivity regimes (NO_x-saturated, mixed or NO_x-sensitive) are inferred from low to high values of photochemical indicators based on the HCHO to NO₂ ratio column data from the Global Ozone Monitoring Experiment 2 (GOME-2) and model. The weekly cycles of ground level NO_x and O₃ concentrations from the U. S. EPA's Air Quality System (AQS) and CMAQ are examined over the AVHYY regions and GOME-2 based chemical regimes. Interestingly, the AQS-observed weekly cycles of NO_x over the AVHYY urban, other and forest region are similar to those over the GOME-2 NO_x-saturated, mixed, and NO_x-sensitive regimes, respectively. However, the AQS-observed O₃ weekly cycle over the AVHYY urban region is significantly different from that over the GOME-2 NO_x-saturated regime. Whereas the weekend high O₃ anomaly is clearly shown in both AQS and CMAQ over the NO_x-saturated regime, the weekend effect is not captured over the geographically-based urban region, suggesting that the urban region includes other chemical regimes. This study suggests that chemical classification into NO_x-saturated, mixed and NO_x-sensitive regime stations gives a more accurate picture for the weekly O₃ cycles than the geographically-based urban, forest, and other classification.

Turner, M.; Henze, D.; Hakami, A.; Zhao, S.; Resler, J.; Carmichael, G.; Stanier, C.; Baek, J.; Saide, P.; Sandu, A.; Russel, A.; Jeong, G.; Nenes, A.; Capps, S.; Percell, P.; Pinder, R.;Napelenok, S.; Pye, H.; Bash, J.; Chai, T.; Byun, D.

In 2008 the World Health Organization estimated that air pollution is the cause of approximately 2 million premature deaths worldwide per year. Long-term exposure to fine particulate matter, including black carbon (BC), has been associated with adverse health effects including cancer, cardiovascular disease, and premature mortality. Quantifying the role of emissions from different sectors and different locations in governing the total health impacts is critical towards developing effective control strategies. To answer such questions, an adjoint model can provide sensitivities of concentrations with respect to emissions at a highly resolved spatial and sectoral level of specificity. This tool can be used to determine the sensitivity of mortality in a region with respect to emissions in various locations. From a single simulation of the CMAQ Adjoint model, we are able to obtain sensitivities of regional BC concentrations with respect to BC emissions from any and all sectors of the continental United States. We can use this data to calculate the effect of BC emissions in these sectors on regional mortality.

Roberto San Jose¹, Juan Luis Perez¹, Enzo Magliulo² and Nektarios Crysoulakis³

¹Technical University of Madrid (UPM), Boadilla del Monte, 28660 Madrid, Spain

² CNR ISAFoM p.o. box 101 S. Sebastiano (Na) - Italy

³Foundation for Research and Technology - Hellas (FORTH) IACM, 100 N. Plastira Str., Vassilika Vouton, P.O. Box 1385, GR-71110, Heraklion, Crete, Greece.

A contribution to Session 7 "Model evaluation and Analysis"

An oral presentation is preferred

WRF and CMAQ models are used in very high spatial resolution (200 m) to simulate urban metabolism (WRF-UCM meteorological fluxes) and air pollution concentrations (CMAQ model). The work was developed into the BRIDGE EU funded project to analyze and study the impact of urban metabolism with different alternatives in urban planning scenarios.

In order to study the urban metabolism (turbulent fluxes) in a very detail way, we decided to use WRF-UCM and CMAQ in a very high resolution mode. However, due to the huge computational time involved in the simulations in order to respect the rate of three between the spatial resolutions up to 200 m, we decided an intermediate configuration approach.

We used the GFS boundary conditions with one degree and an intermediate model domain with 5.4 km spatial resolution with 37 x 37 grid cells over each city and a rate of 27 to go down up to 0.2 km spatial resolution. Urban Canopy Model (UCM) was activated by producing the proper additional landuse types: a) high urban density; b) low urban density and c) commercial / industrial landuse type.

These information was produced from very high spatial resolution landuse and topography files. 3D urban information was available for the cities to produce the average height of the buildings for each of the additional urban landuse types, width of the average canyon street for each of the additional urban landuse types and the average width of the roof for the additional urban landuse types.

The runs show that increase in the high spatial resolution produces a decrease on the quality of the comparison between observational monitoring data and modeling data. In some cases - like in London - during some specific periods of time the wind speed results and not realistic. However, in general the model - when running with very high spatial resolution - produces good results although the quality in slightly reduced. Some numerical stability could be involved in the very high spatial resolution runs. CMAQ results show results according to those obtained with WRF-UCM results. CMAQ ozone correlation coefficient in Athens is higher than 0.7 for the whole 2008 year. Correlation coefficients for meteorological variables using WRF-UCM very high resolution are found to be 0.983 for temperature with 200 m spatial resolution in Helsinki, however, sensible heat fluxes are correlated with 0.595 for 200 m spatial resolution. Further investigation related to the behavior of WRF-UCM and CMAQ using very high spatial resolution is needed.

Greg Yarwood¹, Yvonne Scorgie², Nick Agapides³, Edward Tai¹, Prakash Karamchandani¹, Kelsey Bawden³, Jason Spencer^3, Toan Trieu³

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

2. ENVIRON Australia, 100 Pacific Highway, North Sydney NSW 2060

3. New South Wales Office of Environment & Heritage, 59-69 Goulburn St, Sydney NSW 2000

An efficient and accurate screening methodology has been developed for evaluating ozone impacts of proposed new emission sources in the New South Wales Greater Metropolitan Region (GMR) which includes greater Sydney, Newcastle and Wollongong. Photochemical grid models, such as the Comprehensive Air quality Model with extensions (CAMx), typically are used to evaluate ozone impacts because they account for non-linear ozone chemistry and other important processes (transport, dispersion, etc.) However, applying CAMx for every proposed new source would impose an unnecessary burden that can be avoided by using an ozone screening method (OSM) to determine whether more detailed evaluation is needed. We developed an OSM based on high-order sensitivity analysis of CAMx simulations for the GMR. CAMx simulations were performed for four summer months from 2003/2004 and 2004/2005. Prototypical new sources were introduced at five locations within the GMR and the ozone plumes that formed downwind of each new source were analyzed. Ozone sensitivity to the new source emissions of NOx, VOCs and CO were computed at first and second-order by using the higher-order decoupled direct method (HDDM) in CAMx. Ozone sensitivity coefficients were implemented in the OSM (a spreadsheet) to calculate the incremental increase in ozone concentration that results from adding a new emission source in any one of the five sub-regions of the GMR. VOC reactivity was accounted for by calculating source specific reactivity factors using DDM. Thus, the OSM accounts for non-linear interaction between NOx and VOC emitted by the new source, differences in the reactivity of VOCs emitted by the new source, and sub-regional variation in ozone sensitivity to VOC and NOx within the GMR airshed. The OSM accurately predicted ozone impacts of new sources over a wide range of VOC/NOx emission ratios and for sources with emissions up to an order of magnitude larger than used in the CAMx simulations to develop the method.

Wei Zhou, Daniel Cohan

Dept of Civil and Environmental Engineering, Rice University, Houston, TX, USA

The credibility of air quality models in predicting the impacts of future emission reduction plans is built upon the evaluation of their performances to reproduce historical air quality improvements. Previous studies indicate that model predictions underestimate the observed decrease of O₃ concentration after substantial NO_x emission controls from point and mobile sources. The underestimate of O₃ decrease may be caused by the uncertain estimation of emission and its trend, insufficient response of model chemical mechanisms to the emission change, or errors in meteorological simulations. In this study, CMAQ is applied to simulate the air quality change after the NO_x SIP Call and mobile emission controls significantly reduced NO_x emissions between 2002 and 2006. Firstly, we examine whether the meteorological simulation has influence on the prediction of ozone change. The same meteorological model and parameterizations were used for the two years, resulting in similar performance for daily minimum and maximum temperatures relative to observations. The sensitivity of the daily maximum ozone to the daily temperature is similar in both modeling and observation in both years. We then focus on the impact of uncertainties of the emission inventories and chemical reaction rates on simulations of the O₃ trends between these years. Monte Carlo analysis is applied to quantify how sensitive the O₃ trend is to each uncertain factor and identify the key uncertain factors that may have led to the underestimate of O₃ reduction from 2002 to 2006.

Carey Jang, Sharon Phillips, Tyler Fox, Bryan Hubbell, and Dale Evarts

Office of Air Quality Planning and Standards, USEPA

Air quality models can be a powerful regulatory tool for comparing the efficacy of various emissions control strategies and policy decisions. However, due to the often enormous computational costs and the complication of the required emission inputs and processing, using complex air quality models to generate outputs to meet time-pressing requirement of policy analysis always present a challenge and is typically inefficient, if not ineffective. To address this issue, we have recently developed an innovative policy analysis tool, the Response Surface Model (RSM), to provide a real-time assessment of efficacy of various emissions control strategies and extend the RSM applications in the USA and China. RSM is a reduced-form modeling tool which utilizes advanced mathematical and statistical techniques to characterize the relationship between model input parameters (i.e., pollutant emissions) and outputs (i.e., pollutant concentrations) and in a highly economical manner. The RSM technique has recently been successfully tested and stringently validated (through "out-of-sample validation" and "cross validation") for PM 2.5, ozone, and deposition, respectively. The development of multi-pollutant RSM using CMAQ and its ongoing applications both in the USA and China will be discussed and presented.

As a continued effort to build a bridge between scientists and policy makers, we are currently developing an "Air Benefit And Control Assessment System" (ABACAS) which is a user-friendly modeling framework to provide a common platform for conducting an integrated assessment of emissions control cost and their associated air quality and health benefits. The ABACAS will include three key modules, a control estimate and analysis module (AirControlNet), a real-time control/response module (RSM/CMAQ) to assess the emissions control and resulting air quality benefit, and a health benefit assessment module (BenMAP). Initial efforts toward the development of ABACAS as a joint bilateral collaborative effort between the USA and China will be discussed and presented.

Volker Matthias(a), Armin Aulinger(a), Markus Quante(a), Charles Chemel(b), Juan L. Perez(c), Roberto San Jose(c), Ranjeet Sokhi(b)

(a) Helmholtz-Zentrum Geesthacht, Institute of Coastal Research, Max-Planck-Strasse 1, 21502 Geesthacht, Germany
(b) University of Hertfordshire, College Lane, Hatfield, AL10 9AB, United Kingdom
(c) Technical University of Madrid, Boadilla del Monte, 28660 Madrid, Spain

An intercomparison of CMAQ model results for a high particulate matter episode in winter 2003 in Germany has been performed. This episode was connected with stagnant conditions under high pressure and low wind speeds. It has already been studied in a first
intercomparison experiment with five different participating model systems (Stern et al., 2008). It was found that most of the model systems significantly underestimated the high PM10 values observed in large parts of Germany.
The aim of this subsequent study was to look in more detail at the different effects that influence the particle concentrations at ground but also in higher altitudes. For this purpose, three European research groups, namely University of Hertfordshire, Technical University of Madrid and Helmholtz-Zentrum Geesthacht (formerly named GKSS Research Center), which all apply the CMAQ model in Europe, took part in the study. It has been undertaken in the framework of COST 728.
The intercomparison was done in three rounds. All groups simulated the concentrations of different aerosol components and their precursors for the 16-days period between 24 February and 11 March 2003.
In the first round all groups used their own setup in terms of meteorological input files, emission preparation, grid spacing, boundary conditions, and computing platforms. Although all groups used CMAQ 4.5 in the same configuration, the results were very diverse.
In the second round, the input fields were exchanged between the groups and the results were reproduced on different high performance computers. The influence of compilers and computing platforms was investigated and found to be rather small.
In the final step, sensitivity runs with different meteorological and emission input fields were performed. Meteorological fields were calculated with both, MM5 and WRF. Additionally, different parameterizations were chosen for MM5. It could be shown that for this period the meteorological fields had the largest impact on the results. The temporal disaggregation of the emission data fields could play an important role for some anthropogenic aerosol components. Nitrate concentrations for example were found to be strongly dependent on ammonia emissions, because in the model ammonium nitrate is only formed if free ammonia is available. Ammonia emissions that vary on a monthly basis only may lead to large steps in nitrate aerosol concentrations connected with drastic increases in ammonia emissions on the first day of a month.
In conclusion it was found that even if well established research groups use the same chemistry transport model in the same configuration the model results for a given period might be very diverse.
Our work showed the importance of the sensitivity of PM concentrations to emissions and meteorological inputs and why models, based on seemingly similar configurations and setup conditions, can lead to differing results. This highlights the need for common approaches and frameworks for applying mesoscale modelling to air pollution problems.

Kirk Baker, U.S. Environmental Protection Agency

A national scale air quality modeling analysis was performed to estimate the impact of county specific anthropogenic NOX and VOC emissions on model estimated ozone concentrations. Source contribution is estimated for the primary form of the ozone National Ambient Air Quality Standard (NAAQS). Air quality impacts are estimated with the Comprehensive Air Quality Model with Extensions (CAMx) model. CAMx simulates the numerous physical and chemical processes involved in the formation, transport, and destruction of ozone, particulate matter and air toxics. Emissions of precursor species are injected into the model where they react to form secondary species such as ozone and then transport around the modeling domain before ultimately being removed by deposition or chemical reaction. Photochemical model source apportionment tracks the formation and transport of ozone from emissions sources and allows the estimation of contributions at receptors.

Scott Spak, Jaemeen Baek, Gregory Carmichael, Abigail Fontaine, Mark Janssen, Yoojung Kim, Michael Koerber, Michael Majewski Nicole Riemer, Charles Stanier

Regional fine particle episodes driven by synoptic meteorology affect the Upper Midwest several times each winter, leading to daily NAAQS exceeedances across the Great Lakes. We employ 12 km simulations with WRF, CMAQ and CAMx along with measurements taken during the 2009 LADCO Winter Nitrate Study, a three-month intensive field campaign at a pair of urban (Milwaukee) and rural (Mayville) sites in Wisconsin, to:

- develop a conceptual understanding of these episodes;

- evaluate model skill in simulating urban and rural bulk and speciated fine particle concentrations during episodes;

- quantify the contributions of nitrate formation pathways and transport processes to particle concentrations using CMAQ process analysis; and

- estimate the efficacy of NOx, SO2, and ammonia emissions controls on reducing episode intensity

Daven K. Henze

Kateryna Lapina

University of Colorado, Boulder, CO USA

Ozone (O3) and fine particulate matter (PM2.5) contribute to an array of environmental concerns, from air quality to ecosystem impacts. Formed largely from secondary processes in the atmosphere, these species have complex relationships between their distributions and their underlying precursor emissions, which hinders design of efficient mitigation control strategies. Receptor-based modeling provides a means of tracking back pollutant formation to sources while accounting for the transport and chemistry giving rise to their distribution. Here we use adjoint model sensitivities to examine specific air quality metrics and relate the non-attainment of these metrics to emissions from individual emissions species, sectors, and locations. We focus on metrics related to health impacts of PM2.5, and on background concentrations of O3 (the AOT40 index) related to ecological impacts. Calculations are performed over the U.S. with the GEOS-Chem adjoint model, the results of which quantify the contribution of specific source sectors and locations to these environmental issues. These findings are presented in the context of designing efficient pollution control strategies wherein the variability of the relationship between emissions and impacts is explicitly accounted for.

Ariel F.Stein¹, Manuel Santiago², Marta G. Vivanco², Yunsoo Choi¹, and Rick Saylor³

1 ERT on assignment of NOAA's Air Resources Laboratory, Silver Spring MD

2CIEMAT (Research Center for Energy, Environment and Technology). 28040 Madrid. SPAIN

3 NOAA's Air Resources Laboratory, Atmospheric Turbulence and Diffusion Division, Oak Ridge, TN

In this work we test the representation of secondary organic aerosol (SOA) formation from the photo-oxidation of a-pinene and limonene in the Community Multiscale Air Quality modeling system (CMAQ). A set of experiments performed at the European PHOto REactor (EUPHORE) smog chamber to study SOA formation from the photo-oxidation of a mixture of biogenic compounds is simulated with a box model version of CMAQv4.7. Parameterizations calculated by Carlton et al. (2010) to simulate SOA formation from these species are tested, in conjunction with the parameters obtained experimentally from this study based on theoretical calculations of the partitioning coefficients (K_om,i) of the expected particle phase oxidation products. The parameterizations are tested first against data from an experiment where only a-pinene and limonene were introduced in the chamber and then against additional experiments, where more complex mixtures (isoprene, a-pinene and limonene) were studied. The parameters incorporated in CMAQ4.7 to simulate SOA formation from terpenes clearly overpredict the experimental SOA by 50 - 90 %, while the parameters recently recalculated by Carlton et al. (2010), implemented in the new release of CMAQ5.0, show a clear reduction in this overestimation (20 - 50 %). The experimental parameterizations calculated in this work represent the closest approximation to the EUPHORE smog chamber data, with deviations as much as 20 %. To test the impact of the use of these three different parameterizations on the formation of SOA over the US, 3D CMAQ simulations for the entire month of August 2009 were performed. Comparison of model outputs with surface measurements shows that all of the parameterizations result in underpredictions of PM2.5 organic carbon (OC), possibly due to insufficient SOA formation. Future work will investigate hypotheses for the pervasive OC underprediction.

Amanda Pappin, B.Eng.

Amir Hakami, Ph.D.

Department of Civil and Environmental Engineering

Carleton University

Ottawa, Ontario

Acute and chronic exposure to ambient air pollution has been directly linked with adverse human health effects, resulting in significant social and economic burdens on society. Monetary valuation of these health effects is common in the decision making process for evaluating the economic feasibility of prospective air quality improvement measures. Forward sensitivity analysis of numerical air quality models has been used before to evaluate the response of concentration-based metrics (e.g. mortality, morbidity, and overall health costs) to changes in precursor emissions. Adjoint sensitivity analysis, on the other hand, allows air quality modelers to calculate spatiotemporal contributions from emission sources while conserving source specificity, making it a viable option for evaluating emissions control policies in a straightforward manner. Furthermore, the adjoint approach provides a platform for creating a direct linkage between air quality benefit assessment models on one hand and air quality models on the other. In this work, Health Canada's Air Quality Benefits Assessment Tool (AQBAT) is linked to the adjoint of CMAQ. We present the sensitivities of Canada-wide acute mortality benefits with respect to NO_x emissions reduction across our domain. We will also extend our analysis to source apportionment of health benefits in the US (based on relevant epidemiological studies). Our preliminary results for Canada show significant spatial and temporal variability in health benefit impacts of NO_x emissions reduction for Canada. More importantly, those results suggest that potential health benefits of many abatement options are vastly undervalued in current benefit-cost analysis frameworks that lack source specificity. For example, we estimate that health benefits due to avoided acute mortality (only a fraction of total health benefits) of subway systems in Toronto and Montreal, often considered expensive options, easily exceed $1B/yr in each city.

Fernando Garcia-Menendez, Yongtao Hu, and M. Talat Odman

School of Civil and Environmental Engineering,
Georgia Institute of Technology, Atlanta, GA, 30332-0512, USA

The 2011 forest fire season in the United States has been especially severe. Large wildfires have spread across several states, including Texas, Arizona, New Mexico, and Georgia, burning millions of acres. Wildland fires can significantly contribute to unhealthy air pollution levels at urban areas. Frequently, the air quality impacts from wildfires extend from local scales into larger regional scales. Air quality models capable of tracking regional scale emissions by simulating the physical and chemical processes that affect their ultimate fate are well suited to quantify the effects of wildfires on air quality. However, it is important to systematically evaluate the ability of existing air quality models to simulate the transport and transformation of pollutants emitted by wildfires, and better understand the models’ strengths and limitations.    

            In this study we evaluate the performance of CMAQ in simulating air quality impacts from wildfires. A series of wildfires that occurred in southeast Georgia and northern Florida over a 30 day period in 2007 were modeled. Evaluation was performed using observations from monitoring networks that recorded significant increases in pollutant concentrations during this period at metropolitan areas including Atlanta, Savannah, Macon, and Birmingham. Measured pollutant levels were systematically compared to modeled concentrations to generate statistical performance indicators quantifying model bias and error. Additionally, simulations using an adaptive grid version of CMAQ were carried out to explore the significance of increased resolution on plume development. Visualization and detailed analysis of results were undertaken to identify the most important research needs towards successful fire simulation. The findings of this work provide insight into CMAQ’s current ability to reproduce fire impacts and can be applied to determine the uncertainties associated with predicted pollutant levels in future work.

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

Climate change, especially the extreme events such as heat wave, flooding and drought, is responsible for large impact on human health and natural resources. Global climate models are typical used to predict the impact of future climate under projected emission scenarios. We have conducted climate simulations under Intergovernmental Panel on Climate Change (IPCC) Representative Community Pathways (RCP) 4.5 and RCP 8.5 using global climate model Community Earth System Model (CESM v1.0). The global mean temperature is projected to increase 1.7 and 4.2 °C by the end of the century compared with current climate in RCP 4.5 and 8.5 scenarios, respectively. In addition to global model simulations, we also downscaled CESM outputs to regional climate model Weather Research and Forecasting (WRF). The main purpose of downscaling is to take advantage of local high resolution land use and study local detailed climate change. Fossil fuel intensive scenario (RCP 8.5) was chosen for the downscaling study and the downscaled results will be used to compare with global outputs.

S.L. Napelenok, P.V. Bhave, H. Simon, G.A Pouliot, M. Lewandowski, T. Kleindienst, J. Offenberg, M. Jaoui, R.J. Sheesley

Traditional monitoring networks measure only total elemental carbon (EC) and organic carbon (OC) routinely. Diagnosing model biases with such information is difficult. Measurements of organic tracer compounds have recently become available and allow for more detailed diagnostic evaluation of CMAQ modeling results, which allow for more explicit representation of secondary organic aerosols. For additional carbon evaluation opportunity, the Carbon Apportionment (CA) version of the model makes it possible to track contributions from various sources of primary organic aerosols and elemental carbon. An ambient PM2.5 measurement campaign conducted in five Midwestern U.S. cities in March 2004 February 2005 allows for direct comparison of modeled and measured organic carbon concentrations by primary and secondary source category. With the more extensive measurement data and modeling capabilities, it is possible to evaluate model calculations of secondary organic aerosols produced from oxidation of aromatics, isoprene, monoterpenes, and sesquiterpenes, as well as primary contributions from diesel exhaust, gasoline exhaust, fires, food cooking, and other categories.

Serena H. Chung¹, Rodrigo Gonzalez Abraham¹, Brian Lamb¹, Jeremy Avise², Eric P. Salathe³, ⁴Yongxin Zhang, ⁵David G. Streets, ⁶Chris Nolte, ⁶Dan Loughlin, ⁴Alex Guenther, ⁴Christine Wiedinmyer, ⁴Tiffany Duhl

¹Washington State University

²California Air Resources Board

³University of Washington

⁴National Center for Atmospheric Research

⁵Argonne National Laboratory

⁶Environmental Protection Agency

As part of an ongoing analysis of the effects of global change on regional air quality in the US, we report results for current and future simulations in terms of the sensitivity of future conditions to changes in meteorology, global anthropogenic emissions (chemical boundary conditions), land use and biogenic emissions, and US anthropogenic emissions. Meteorological fields, downscaled from the results of the ECHAM5 global climate model using WRF, were used to drive the MEGAN biogenic emissions model, the SMOKE emissions processing tool, and the CMAQ chemical transport model for five representative summers within each of the current (1995-2004) and a future decade (2045-2054). CMAQ simulations employed two nested domains covering most of the Northern Hemisphere from eastern Asia to North America at 220-km horizontal resolution (semi-hemispheric domain) and covering the continental US at 36-km resolution (CONUS). For the current decade semi-hemispheric domain simulation, year 2000 global emissions of gases (ozone precursors) from anthropogenic, natural, and biomass burning sources from the POET and EDGAR emission inventories were used. Global emissions inventories for black and organic carbon from Bond et al (2004) were applied. For the future decade semi-hemispheric domain simulations, current decade emissions were projected to the year 2050 following the Intergovernmental Panel for Climate Change (IPCC) A1B emission scenario. For the CONUS simulations, US anthropogenic emissions for the current decade were based on the 2002 National Emissions Inventory prepared by the Environmental Protection Agency. For the future decade simulation, these emissions were projected to 2050 using growth factors from the U.S. EPA MARKAL nine-region database and energy system model following a scenario that assumes baseline criteria pollutant policies, including the application of the Clean Air Interstate Rule (CAIR). Results on ozone and PM2.5 concentrations and nitrogen deposition will be presented. Policy-relevant results, such as the relative response factor, policy-relevant background concentration, and critical loads, will be discussed.

Marco A. Rodriguez, Michael G. Barna, Kristi A. Gebhart, Jennifer L. Hand, Bret A. Schichtel, Catherine Benedict, Jeffrey L. Collett Jr., and William C. Malm

Excess wet and dry deposition of nitrogen-containing compounds is a concern at a number of national parks. The Rocky Mountain Atmospheric Nitrogen and Sulfur Study (RoMANS) II campaign was conducted from November 2008 to November 2009 in order to characterize the atmospheric concentrations of sulfur and nitrogen species at Rocky Mountain National Park (RMNP), in north-central Colorado. The Comprehensive Air Quality Model with extensions (CAMx) is used to simulate the fate of gaseous and particulate species subjected to multiple chemical and physical processes for this study. This work presents a detailed operational model evaluation with special emphasis on the reduced nitrogen species measured during RoMANS II. The model emissions and meteorology are further examined as a result of the operational evaluation in order to improve the photochemical model predictions. The source apportionment of ammonia emissions is indirectly investigated with the Particulate Source Apportionment tool (PSAT) in CAMx and compared with source apportionment results from a "hybrid" approach using Empirical Orthogonal Function (EOF) analysis and a conservative tracer simulation

By fossil fuel combustion and fertilizer application, the US has substantially altered the nitrogen cycle, with serious effects on climate change. Using the global temperature potential (GTP) as a common metric, we seek to quantify these impacts. Because the climate effects can be short-lived, by impacting the chemistry of the atmosphere, or long-lived, by altering greenhouse gas fluxes, we examine the GTP at 20 and 100 years in units of CO2 equivalents. At both time-scales, nitrogen enhancement of CO2 uptake is the largest contributor (-500 Tg/yr), because in the eastern US, areas of high nitrogen deposition are co-located with forests. In the short-term, the GTP20 due to NOx altering ozone and methane concentrations is also substantial (-300 Tg/yr). Because these effects are short-lived, they have little affect on the 100 year time scale. Finally, the GTP of N2O emissions also are substantial at both time scales (+200 Tg/yr). We have also attributed these impacts to combustion and agricultural sources, and quantified the uncertainty in these calculations. The net effect of combustion sources is to decrease GTP. The impacts of agricultural sources tend to cancel each other out, and the net effect is uncertain. Recent trends show decreasing reactive nitrogen from US combustion sources, while agricultural sources are increasing. Fortunately, there are many mitigation strategies currently available to reduce the climate change impacts of US agricultural sources.

J. M. Baldasano^a,b, M. T. Pay^a,b, S. Mailler^a, P. Jimenez-Guerrero^c,

^aEarth Sciences Department, Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS). Barcelona, Spain.

^bEnvironmental Modelling Laboratory, Technical University of Catalonia. Barcelona, Spain

^cPhysics of the Earth, University of Murcia, Spain

The development of comprehensive emission inventories constitutes a fundamental step for the advance in air quality modeling. Great efforts have been made in the last decades in order to improve emissions estimations and quantify the associated uncertainties.

In order to provide emissions as an input to mesoscale air quality models two different approaches are usually followed: top-down or bottom-up methodologies. Top-down emission inventories are derived from global or national estimations of pollutant emissions, then disaggregated spatially and temporary to the resolutions needed to be used by air quality models, e.g. adopting the EMEP inventory for Europe. In bottom-up emission inventories, on the contrary, each activity sector is treated separately with a specific methodology, considering specific emission factors and activity indicators, and emissions are aggregated thereafter, e.g. HERMES (Baldasano et al., 2008) or SMOKE (Houyoux et al., 2000).

Both approaches are compared here by means of the CALIOPE modeling system (Baldasano et al., 2010) for the Iberian Peninsula. WRF-ARWv3.2 (Michalakes et al., 2004; Skamarock and Klemp, 2008) and CMAQv4.7 (Byun and Ching, 1999; Foley et al., 2010) are used together with HERMES and EMEP emission inventories for the 26-28 November, 2008 period, with high spatial resolution, 4x4 km², and identical setup for initial and boundary conditions. Model performances in terms of NO₂ predictions are assessed against EMEP background air quality observations, providing comparable statistical indicators (Mean Normalized Biases and Gross Errors). However, the spatial distribution of emissions plays a fundamental role in final NO₂ modeled levels. Simulations performed using EMEP top-down emissions as inputs do not reproduce specific features that are captured when using HERMES bottom-up, such as the dispersion of point source emissions. Variations are observed both in the extent of affected areas and final pollutant levels.

We conclude that the use of bottom-up emissions inventories can provide substantial changes in the derived air quality levels when compared to top-down approaches. Therefore, the increase of resolution in meteorological and chemical transport models must be done together with an accurate spatial allocation of emissions, which can be easily achieved by means of bottom-up approaches.

Tanya L. Otte, Martin J. Otte, Christopher G. Nolte, and Jared H. Bowden (U.S. EPA, Research Triangle Park, NC)

In the regional climate modeling community, there is a philosophical divide over whether or not a constraint toward larger-scale driving fields (i.e., nudging) should be used on the interior of the limited-area domain. Prior research has demonstrated that nudging techniques can increase the skill of the regional climate model predictions. However, one of the primary arguments against using nudging for regional climate modeling is that nudging can inhibit the regional climate model's ability to properly simulate mesoscale detail. In this presentation, we will explore the effects of both analysis nudging and spectral nudging in multi-decadal, continuous WRF simulations over the continental United States. The evaluation will emphasize effects of nudging on temperature and precipitation distributions, as well as spectral representations of column-integrated fields and fields aloft.

Barron H. Henderson, Rob W. Pinder, Farhan Akhtar, Havala O.T. Pye, William Vizuete

Ozone (O3) in the troposphere is produced via chemical cycling of nitrogen oxides (NOx = NO + NO2). Reactions of peroxy radicals oxidize nitric oxide to produce nitrogen dioxide. The nitrogen dioxide (NO2) then photolyzes to produce ozone. The NOx cycle is primarily terminated via nitric acid production (NO2 + OH = HNO3). Nitric acid (HNO3) is not photochemically labile, so its production inhibits O3 production by removing NO2 and OH. The rate of HNO3 production is determined experimentally, but has proved difficult (Donahue, 2011). Recent papers suggest a downward revision at 298K (Mollner et al., 2011) and a decrease in the temperature sensitivity (Henderson et al., in manuscript). The influence of these updates is best evaluated in a hemispheric or global model, where the influence of boundary conditions is limited. We have simulated 2004 with suggested revisions for comparison with INTEX-A observations. This re-evaluation allows us to address the previously observed model biases for NO2 (Napelenok, 2007; Hudman, 2008) and HOx (Olson et al., 2004; Ren et al., 2008), as well as to examine the influence on ozone predictions. Initial results show that nitrogen partitioning is much improved, but that the model has a high bias for total oxidized nitrogen (NOy = NOx + PAN + HNO3). The high biased NOy causes an ozone high bias in the base model that is exacerbated by these updates. One potential cause is the estimation of NOx emissions with inverse models. The inverse models may be sensitive to the model chemistry that was removing NOx too quickly.

Dr. Marc Carreras-Sospedra,⁽¹⁾⁽²⁾ Michael MacKinnon,⁽¹⁾ Prof. Jacob (Jack) Brouwer,⁽¹⁾ Prof. Donald Dabdub⁽²⁾

⁽¹⁾Advanced Power and Energy Program, University of California, Irvine (UCI),

⁽²⁾Computational and Environmental Sciences Laboratory, UCI

The main goal of this project is to evaluate the air quality impacts of climate change and greenhouse gas (GHG) mitigation strategies in 2050. Detailed air quality model sensitivity analyses and model modifications are required to adapt the simulation capabilities to account for climate change impacts on air quality. In addition, previously developed methodologies for predicting future spatially and temporally resolved emissions fields are adapted to accurately account for air pollutant emissions impacts of likely GHG mitigation strategies. This is followed by simulations of atmospheric chemistry and transport in a set of air quality models to determine air quality impacts of GHG mitigation strategies. The project focuses on two representative regions of the United States: California and the Northeastern US. These two regions are likely to lead the nation in limiting GHG emissions and are already developing plans that delineate potential GHG mitigation strategies. In addition, the various mitigation strategies that are likely to be adopted in these regions will affect pollutant emissions fields in these regions that are typically plagued by poor air quality.

The project assesses the baseline air quality in the selected US regions in the year 2050, accounting for expected (and unexpected) changes in climate and increases in commercial and industrial activity globally, and in particular in South East Asia, which can affect background pollutant concentrations reaching the US. Sensitivity analyses that account for various changes that can be forced by climate change (e.g., temperature, evaporative emissions, and biogenic emissions) are used to identify the most important considerations for model adaptation to account for the effects of climate change. In addition, this study develops a broad set of future scenarios for greenhouse gas mitigation strategies that account for the spatial and temporal distribution of all major emissions sources. The foci of the mitigation strategies used in the scenario development are (1) transportation and (2) electrical power generation, since these two sectors account for the majority global GHG emissions and are featured prominently in current and proposed GHG mitigation strategies. Finally, air quality impacts of GHG mitigation strategies are evaluated using state-of-the-art air quality models. The ultimate goal of this project is to develop air quality simulation strategies that accurately account for climate change and to determine the most effective GHG mitigation strategies that can concurrently improve air quality.

Halil Cakir¹, Lin Zhang², Pat Dolwick¹, Joseph Pinto³

¹US EPA, OAQPS, RTP, NC 27711
²Harvard University, Div. of Appl. Sci. and Eng., Cambridge, MA 02138
³US EPA, NCEA, RTP, NC 27711

In evaluating model performance, comparisons are most often made between observations at monitoring sites and grid averaged output and bias is calculated as the difference between the two. Issues arise in adopting this commonly used approach. First of all, there is the 'change of support' problem in which the simulated values and the monitored values do not necessarily represent the same spatial and temporal scales. Often, bias is reported as a single number for the entire model domain, which tacitly assumes that the bias is the same everywhere and over the simulation period. However, terms in the species continuity equations and their errors vary spatially and temporally, and as a result this procedure ignores the spatial and temporal variability in the bias estimates and also their dependence on other variables. As a case study, bias estimates are generated for GEOS-Chem over the continental U.S. for 2006-2008. Regionally and seasonally dependent patterns are found. In particular, relatively large positive bias is found over Florida and nearby areas. Results such as these are used to provide clues as to the sources of the error in the model.

The views expressed are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency.

University of Tennessee, Knoxville¹

Harvard University²

Office Air Quality Planning and Standards (OAQPS), USEPA³

Yun-Fat Lam, Joshua S. Fu¹, Daniel J. Jacob², Carey Jang and Pat Dolwick³

Recent years, the term, Policy Relevant Background (PRB) ozone, has been used to quantify surface background ozone in the United States. The source of PRB ozone includes productions of ozone from natural and foreign emissions, and stratospheric ozone intrusions. A large uncertainty of stratospheric contribution in PRB ozone is caused by the distinct pattern of diurnal and seasonal variations of ozone intrusion. In this study, three years of GEOS-Chem simulations from 2006-2008 were used for generating initial (IC) and boundary (BC) conditions for CMAQ to investigate the effects of stratospheric ozone influence to surface PRB ozone. Additional GEOS-Chem scenarios from 2006 were simulated to quantify the PRB ozone and to evaluate the ozone contributions from foreign sources. For ozone intrusion study, we identified several intrusion events based on the evaluation of potential vorticity from meteorological inputs. We used those events to examine the potential impacts of stratospheric intrusion on PRB ozone. The impacts of using different layer structures on vertical transport in CMAQ and how it affects the overall surface PRB ozone will also be presented in this study.

An increase in the aviation activity in the past few years has led to increase in aircraft emissions as well as air quality impacts in the vicinity of airports. Among all air pollutants, hazardous air pollutants (HAPs) are a small proportion of air toxics that accumulate in the atmosphere. HAPs exhibit local impacts, as some of them have longer lifetime, lower decay and smaller scales of transport than other air pollutants, and result in adverse human health effects. The goal of this study is to assess potential impacts of HAPs due to emissions from the T.F. Green airport (PVD) - a mid-sized airport near Providence, Rhode Island. We used the Community Multiscale Air Quality (CMAQ) model with the Carbon bond mechanism (CB05) extended for air toxics, at 4-km resolution to predict 7 major HAPs (4 VOCs, 3 Carbonyls) for the year 2005. Airport emissions for PVD were developed at a high resolution using the Emissions Dispersion and Modeling System (EDMS) and processed through SMOKE to provide enhanced 4-D variability in the model. Non-airport emissions inventories were based upon the EPA's National Emissions Inventory (NEI) for 2005. We will present results that focus on a) CMAQ model evaluation using measurements from the Rhode Island Department of Environmental Management (RIDEM) at 9 monitoring sites (5 near airport, 4 permanent sites further away from the airport) to indicate CMAQ's ability in predicting the airport-related HAPs, b) Spatial and temporal trends of airport-related HAPs at the monitoring sites, and c) Comparison of CMAQ predictions with the recently released National Air Toxics Assessment (NATA 2005) at county- and census tract-levels to compare the performance of a grid-based chemistry-transport model with a dispersion model in predicting the HAPs in Rhode Island, in particular near the airport. Our findings indicate that HAPs concentrations tend to be high at the far away urban sites than those near the airport, except for formaldehyde that shows high concentrations on the airport site located near the runway. Furthermore, VOC HAPs (such as 1,3-butadiene, benzene, toluene, and xylene) are high during winters, and carbonyls (such as formaldehyde and acetaldehyde) are high during summers, highlighting the contribution from photochemical production. CMAQ results showed general underprediction for all HAPS, but demonstrate the general ability of the model to replicate the temporal and spatial trends seen in the observations.

Aijun Xiu, Zachariah Adelman, Francis Binkowski, Adel Hanna
Institute for the Environment, UNC-Chapel Hill

Rohit Mathur, Jon Pleim, David Wong
US EPA

In the Weather Research and Forecast (WRF) model, both the CAM and YYTMG radiation schemes are used for air quality and regional climate applications. The CAM radiation scheme was developed originally for the Community Climate System Model (CCSM) Version 3 and implemented in the WRF model. The new version of the Rapid Radiation Transfer Model (YYTMG) with a random cloud overlapping method is included in the WRF model since Version 3.1 and has been widely used in meteorological simulations for air quality applications. These two radiation schemes are also applied to regional climate research and applications, especially in the coupled WRF-CMAQ modeling system.

We will present results of comparing these two radiation schemes as well as their behaviors under the same default profiles, such as background aerosol profile. WRF simulations using both schemes will be evaluated and analyzed against observations. We will also compare the coupled WRF-CMAQ modeling system using the CAM and YYTMG radiation scheme with the radiative feedbacks of aerosols.

Heather Simon, Sharon Phillips, Norm Possiel, Kristen Foley, Shawn Roselle, Prakash Bhave

Photochemical models, such as the Community Multi-scale Air Quality Model (CMAQ) are critical to the rulemaking/NAAQS setting process as well as to the implementation of existing rules and the design of air pollution control strategies by both EPA and state agencies. The model is used in several ways for these assessments including 1) to predict the relative response of pollutant concentrations to emissions perturbations to evaluate the monetized benefits of control strategies and to project NAAQS attainment/nonattainment and 2) to quantify source contributions in order to understand the culpability of individual sources and source areas on pollutant concentrations . Evaluation of model performance is important to build confidence in the modeling analysis performed for these regulatory efforts and to gain understanding into model predictions. In addition, model performance evaluations help to identify potential weaknesses in the modeling system and guide future model research and development. In this talk, we describe model performance relevant to PM2.5, ozone, and selected precursor species from recent model simulations for 2005. In addition we discuss current EPA efforts to advance improvements in model performance.

Yun Hee Park, Arastoo Pour Biazar, Richard McNider, Kevin Doty, Bright Dornblaser

Errors in location and timing of model simulated clouds are a major source of uncertainty in air quality modeling practices. Thus, improving the simulation of clouds would be essential to reducing the uncertainty of weather forecasting and air quality models. The purpose of this research is to improve model clouds in time and space by assimilating Geostationary Operational Environmental Satellite (GOES) observations in Weather Research and Forecasting (WRF) model. The techniques in this study use GOES data to identify errors in simulated clouds, and then use the same observations to improve the clouds' representation in the WRF model. The basic approach is to estimate the maximum vertical velocity and moisture needed in an atmospheric column to support and sustain cloudy/clear areas (as indicated by GOES) and then adjust the model horizontal wind components to support/sustain the target vertical velocity. The assimilation of GOES cloud data requires the statistical evaluation of dependent model variables so that the model run remains stable as new data is added. Previously the technique used threshold values for target vertical velocity that were based on model statistics. In this study we used an analytical technique to estimate the target vertical velocity needed to clear or create clouds where the model is over-predicting or under-predicting clouds. Results from simulations during summer of 2006 will be presented.

Daiwen Kang, Sarwar Golam, Rohit Mathur

CMAQ simulations using a 12-km and a 4-km grid-spacing were performed during August - October, 2006 period. The 12-km domain covers the eastern United States, while the 4-km nested domain covers the east Texas and surrounding areas with a focus on the Houston area. Predictions from the simulations over these two grid resolutions are compared with surface observations. Preliminary analysis has shown that the model performance over the two grid resolutions varies from one species to another and dependent on the evaluation metrics being used; discerning performance changes is difficult based solely on traditional global performance metrics. To discern the benefits that can be gained using the finer grid resolution compared to the coarser grid resolution, we conducted detailed analysis in terms of grid-to-grid variability employing various temporal and spatial metrics for different pollutants. In this study, we assess the performance of the CMAQ model configured using these two grid resolutions utilizing available observations within the 4-km domain. The focus of this presentation will be on the spatial variability of these two grid resolutions for locations within the Houston airshed.

Air quality models are one of the main policy drivers providing the evidence needs for intergovernmental, national and local authorities to control and manage air quality for compliance with air quality legislation. The model is often required to be able to forecast air quality, impacts of policy changes or climate changes on air quality, human health and ecosystems.

Various temporal and spatial scales models were developed to address problem at specific scale. Often the case, the local scale models are developed independently of the regional scale models hence are unable to capture long range transport pollutants that undergo non-linear atmospheric chemical reaction. In the other hand, regional scale models do not solve small scale atmospheric features in order to compromise its model runtime hence the output only represent background air quality.

In order to bridge the gap within these two scale models and to enhance the model capability in modelling air quality and its impacts at fine spatial scale, a coupling regional scale model (CMAQ) and local scale mode (ADMS) were developed and applied over London areas to demonstrate their use in site specific modelling of hotspots and fine scale mapping of hourly NO₂and O₃concentrations. The system performance was assessed against the measurements from over 100 sites within London. Its application in quantifying the impacts of climate on soil water status and hence O₃deposition and ambient O₃concentration during the 2006 summer plant growing season over London area are discussed.

Ying LI, Alexis Lau, Jimmy Fung

The study simulates ozone air pollution by CMAQ model for Hong Kong and the Pearl River Delta region in November 2006 which covers a typical high ozone episode period. The observations data of Pearl River Delta (PRD) observation sites network are used to evaluate model performance. Ozone simulation by CMAQ Model has a good agreement with observation data at most sites. The good performance for ozone demonstrates the reliability of latest PRD local emission inventory integrated in the modeling system. In this study, to understand and evaluate the ozone sensitivity to NOx and VOC emission reduction by source categories, five different ozone control strategies are designed based on the efficiency and feasibility of available control measures. We found that reducing the NOx emission from mobile or point sources will lead to more or less ozone increase at the NOx-titration area. However, since the ozone increase generally happens at the lower ozone locations, and is caused by the NOx-titration effect but not new fresh ozone generated from these sources with NOx release, reducing the NOx emissions from the point source is also a feasible control strategy as well as reducing NOx and VOC emission from on-road mobile sources. Generally speaking, the more NOx and/or VOC emission reduction, the larger amount of ozone peak decrease. However, in the case of HK and the PRD, to avoid the possible ozone increase of HK and the PRD due to the reduced NOx-titration effect, the NOx emission controlling of large point sources is better implemented after, or together with, the effective control of other major sources with less NOx-titration effect (e.g. on-road mobile sources or possible VOC reduction).