Jonathan Pleim and the CMAQ Development Team

A new version of the Community Multiscale Air Quality (CMAQ) Model (version 5.0.2) was released in May 2014, which includes extended capability for diagnostic output including source apportionment for particulate matter and ozone, the Direct Decoupled Method (DDM) for tracking and output of sensitivities to input perturbations, and the sulfur tracking tool to account for the production of sulfate via the various gas and aqueous phase processes. The CMAQv5.0.2 model also includes a community contributed model component for an alternative secondary organic aerosol (SOA) model based on the Volatility Basis Set (VBS). Another community contributed component for plume-in-grid modeling (PinG), known as the Advanced Plume Treatment (APT) which is based on a second-order turbulence plume model with chemistry SCICHEM is still being tested for release in the near future.

Model development is ongoing toward the next major version of CMAQ scheduled for release in Fall 2015. The new version will include updates to three chemical mechanisms (CB05, SAPRC07, and RACM2) particularly to improve nitrogen cycling and provide additional sources for secondary organic aerosols. There will also be improvements to biogenic and sea salt emissions, updated aerosol nucleation, gravitation settling of coarse aerosols, bidirectional soil NO, and several improvements to meteorology modeling especially for high resolution applications. The 2-way coupled WRF-CMAQ will also be upgraded for both its direct and indirect aerosol feedback effects.

We are in the initial planning stage for the development of the next generation air quality modeling system. We envision that the new system will have multiple configurations including a global model with integrated chemistry-physics-dynamics in an unstructured grid with seamless mesh refinement capabilities. A prototype global model with gas-phase chemistry has already been developed based on the Ocean Land Atmosphere Model (OLAM). Global, regional, and offline regional configurations could be developed using common interoperable components for all 0-D processes such as gas, aqueous, and aerosol chemistry.

Jenny Stocker, Christina Hood, David Carruthers, Martin Seaton, Kate Johnson
(Cambridge Environmental Research Consultants, Cambridge, UK)

Jimmy Fung
(Division of Environment and Department of Mathematics, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong )

Within the regional air quality modelling community there is continual pressure to reduce the model grid scale, so that local variations in pollutant concentrations can be modelled. Whilst the scales have been reduced to as little as 1 km or even 500 m, such models will not, in the foreseeable future, be able to resolve high concentration gradients such as those adjacent to roads. An alternative approach is to nest models which explicitly represent sources within the regional models. Such an approach can be used with grid-based models operating at larger grid scales, for example 5-10 km.

The concept of nesting the Gaussian-type local dispersion model ADMS-Urban in a regional photochemical model, so as to avoid double counting and exploit the advantages of each model type, is described in Stocker et al. (Int. J. Environment and Pollution, 2012). The method has now been developed into a user-friendly and flexible automated system for nesting ADMS-Urban within CMAQ, CAMx or EMEP4UK using meso-scale meteorological data from WRF.

The system automates a series of utility programs and ADMS-Urban model runs, which extract ADMS-format meteorological data from WRF; extract local background concentrations from the regional photochemical model; perform a preliminary ADMS-Urban run and process the results to calculate nesting background concentrations; perform main ADMS-Urban runs with gridded and explicit emissions; and combine the results to obtain nested output concentrations. Nesting domains which cover multiple regional model grid cells are automatically divided into separate runs for each grid cell, with meteorology and background from the corresponding grid cell. CERC's Run Manager software is integrated into the nesting system to allow multiple simultaneous runs to be spread across multiple PCs.

The presentation will introduce the nesting concept, system structure and user interface. Example inputs and results from an implementation of the system to model Hong Kong with regional model data from CAMx will be shown, together with comparisons of the model results with observed data from rural, urban background and roadside monitoring sites.

Fernando Garcia-Menendez¹, Yongtao Hu² and M. Talat Odman²

1) Center for Global Change Science, Massachusetts Institute of Technology

Adaptive grid methods can be used to advance the multiscale capabilities of chemical transport models and attain unprecedented grid resolution in regional-scale air quality modeling. Previous efforts to explore the use of adaptive grids in air quality models have been limited to horizontal adaptation or simplified models. Here a three-dimensional fully adaptive algorithm designed for grid-based chemical transport models is presented. The mesh-moving algorithm builds on a two-dimensional horizontal adaptive grid method previously applied in CMAQ to allow vertical refinement. The new algorithm is capable of simultaneously refining horizontal and vertical grid resolution, yet retains a grid's original structure, enhancing compatibility with existing air quality models. Extremely high levels of grid resolution can be achieved through the grid refinement method. Using the CMAQ modeling framework, advection tests were carried out to evaluate the algorithm's functionality and demonstrate its potential to better capture concentration gradients lost in fixed grid simulations. Atmospheric simulations featuring concentrated plumes in the free atmosphere, as well as plumes near inversions or the top of the boundary layer, could especially benefit from three-dimensional adaptation. Furthermore, the algorithm may aid in better capturing small-scale processes with chemical transport models. Research needs and recommendations for complete implementation into operational air quality models will also be discussed.

Donna B. Schwede¹, Deborah Luecken¹, John Walker², Wyat Appel¹, James Kelly³, Kirk Baker³

¹National Exposure Research Lab, US EPA, RTP, NC

²National Risk Management Laboratory, US EPA, RTP, NC

³Office of Air Quality Planning and Standards, US EPA, RTP, NC

Excess atmospheric nitrogen deposition can cause significant harmful effects to ecosystems. Organic nitrogen deposition can be an important contributor to the total nitrogen budget, contributing 10-30%, however there are large uncertainties in the chemistry and deposition of these compounds. Recent measurement studies provide the opportunity for the improvement and evaluation of the treatment of organic nitrogen compounds in the Community Multiscale Air Quality (CMAQ) model. We examine updates to the CB05 chemical mechanism to improve the treatment of organic nitrate compounds and the resulting impact on atmospheric dry and wet deposition as well as the impact on air quality, including ozone concentrations. Our results show that previous versions of CMAQ greatly underestimate the deposition of organic nitrates. Improvements to the speciation as well the solubility of the organic nitrate species are important to correctly predicting the deposition. Comparisons of CMAQ concentration and deposition with measured values are provided to evaluate the model updates.

Limei Ran¹, Robert Gilliam², Frank Binkowski¹, Aijun Xiu¹, Larry Band¹, Jonathan Pleim²

¹Institute for the Environment

University of North Carolina at Chapel Hill, NC USA

²Atmospheric Modeling and Analysis Division

ORD NERL/USEPA, Research Triangle Park, NC

Accurate description of surface characteristics is important in meteorology and air quality (AQ) modeling systems such as WRF/CMAQ for simulating the exchange of heat, moisture, momentum, and trace atmospheric chemicals between the land surface and the atmosphere. In WRF/CMAQ, surface characteristics including vegetation parameters and surface albedo are specified in the land surface model (LSM) look-up tables by land use category and plant phenological dynamics are modeled using simple time and temperature dependent functions. Land use data used in these LSMs are often out of date and these LSMs have simple canopy treatment with a big-leaf empirical stomatal conductance functions as well as simple hydrology and snow processes. For year-long retrospective WRF/CMAQ simulations, these LSMs using simple canopy treatment with table prescribed surface representations from out of date land use data clearly show limitations in capturing seasonal landscape changes (e.g. phenology and albedo) and disturbances (e.g. fires, storm damages). The goal of this research is to improve land surface modeling in WRF/CMAQ by incorporating MODIS temporal LAI and albedo products for faithful surface representation. Albedo and LAI are two important parameters in meteorology and air quality modeling because albedo affects not only the surface energy budget and fluxes but also photolysis rates in the air quality model. LAI is important not only for scaling leaf level fluxes to the canopy level but also controlling deposition of various atmospheric gases and particles. The LAI and albedo from the current WRF/CMAQ configuration will be first evaluated against observation data from MODIS products and from FLUXNET and SURFRD surface networks. This presentation will focus on 2006 WRF meteorology simulations for the CONUS 12km domain. Simulated meteorology (e.g. temperature, moisture, wind speed and direction, surface radiation budgets, and PBL height) will be compared and evaluated among different simulations with table prescribed LAI and albedo and with MODIS LAI and albedo. The simulated results will also be analyzed with measurement data to demonstrate benefits and issues in using satellite LAI and albedo data in the air quality modeling system.

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 HiRes2 air quality forecasting system is beginning operation at Georgia Tech in 2014, and uses WRF3.6 and CMAQ5.0.2 with SAPRC07TC mechanism, AERO6 aerosol module, inline BEIS biogenic emissions and inline 3-D point source emissions processing. In addition to those features, and other capabilities of the original HiRes system, HiRes2 provides not only the air quality concentration e.g. O₃ and PM_2.5 forecasts, but also forecasting of air quality impacts from sources such as power plants and on-road vehicles by utilizing the DDM-3D technique that has been integrated into the CMAQ5.0.2. Further the unique air quality forecasting products that HiRes2 provides are continuously "corrected" by near real-time ground level measurements and NASA satellite retrievals. The DDM-3D sensitivities are used in an inverse modeling context to "correct" errors in emissions inputs by utilizing near real-time observations during the last 7 days. The "corrections" of emission biases adjust gridded emissions inputs for forecasting the next 7 days to improve forecasting performance. Currently, the ground level observations utilized are PM_2.5 measurements averaged to daily values, and the satellite retrievals utilized are NASA MODIS collection 6 AOD products that are "fused" into forecasted fields along with PM_2.5 measurements.

Beata H. Czader

Peter Percell

Daewon Byun

Yunsoo Choi

A hybrid Lagrangian-Eulerian modeling tool has been developed using the CMAQ code. It is a moving nest model that resides inside Eulerian CMAQ structure. It may follow the trajectory defined by the mean mixing layer wind or any other user defined trajectory. At each time step and location iIt obtains boundary conditions from CMAQ output concentrations. Inside the window, chemical transformation, advection and diffusion are taken into account. We will present details about develpemnt, testting and applications of this moving nest model.

Christos Efstathiou¹, Jana Matejovicova², Gerhard Lammel^1,3

¹ Masaryk University, Research Centre for Toxic Compounds
in Environment, Kamenice 5, Brno, Czech Republic

² Slovak Hydrometeorological Institute,
Jes niova 17, Bratislava, Slovakia

³ Max Planck Institute for Chemistry,
Hahn-Meitner-Weg 1, Mainz, Germany

This work presents the development of a multi-compartment version of the Community Multiscale Air Quality (CMAQ) model to account for emission and transport of persistent organic pollutants (POPs) between environmental compartments. This system takes full advantage of CMAQ's capability to simulate atmospheric advection, diffusion, gas-phase chemistry, cloud/precipitiation, and aerosol processes. The modifications reported here include the addition to the CMAQ system of three particle partitioning models options: an adsorption model (Junge-Pankow), an organic matter (OM) absorption model (Koa), and a "dual" OM absorption and black carbon (BC) adsorption model. In addition, chemical transformations and surface-atmosphere exchange processes were included for these semi-volatile organics. Simulations for the purpose of model testing and evaluation were conducted for the year 2006 on a domain covering the entire Europe.

Beata Czader

Bernhard Rappenglueck

We will present CMAQ results for 1,3-butadiene simulated with an extended SAPRC99 mechanism for Houston area. The simulated mixing ratios are compared to spatially and temporally resolved measured values at several stations. Relative contributions of different types of emissions are analyzed. The highest contribution was found to be from industrial sources exceeding contribution from mobile emissions. Because of that only little weekday - weekend effect was found in 1,3-butadiene concentrations in Houston. Episodical, very high emissions from industrial sources are not reflected in the inventory leading to underprediction of mixing ratios. Mixig ratios of 1,3-butadiene at regional and urban stations were properly simulated showing that mobile emissions are accurately accounted for in the inventory.

Jingqiu Mao, Larry Horowitz, Peter M. Edwards, Kyung-Eun Min, Steve Brown, Ilana B. Pollack, Thomas B. Ryerson, Martin Graus, Carsten Warneke, Jessica B. Gilman, Brian M. Lerner, Andy Neuman, John B. Nowak, Patrick R. Veres, James M. Roberts, Felipe Lope-Hilker, Ben H. Lee, Joel A. Thornton, Jennifer B. Kaiser, Frank N. Keutsch, Glenn M. Wolfe, Thomas F. Hanisco, Joost A. De Gouw, Kenneth C. Aikin, Kelley C. Wells, Dylan B. Millet, Vaishali Naik, Fabien Paulot, Meiyun Lin, Daniel J. Jacob

We implement a new isoprene oxidation mechanism in two global chemistry models (GEOS-Chem and GFDL AM3). GEOS-Chem model results are evaluated with observations for ozone, isoprene oxidation products, and related species from the International Consortium for Atmospheric Research on Transport and Transformation aircraft campaign over the eastern United States in summer 2004. The GEOS-Chem model successfully reproduces the observed concentrations of organic nitrates (ANs) and their correlations with HCHO and ozone. ANs in the model is principally composed of secondary isoprene nitrates, including a major contribution from nighttime isoprene oxidation. The correlations of ANs with HCHO and ozone then provide sensitive tests of isoprene chemistry and argue in particular against a fast isomerization channel for isoprene peroxy radicals.

More recently, we evaluate this mechanism in a high resolution global chemistry-climate atmosphere model (GFDL AM3) with the Southeast Nexus (SENEX) field observations in summer 2013. Nighttime observations suggest that abundant ozone and VOCs in the residual layer may provide an efficient sink for nighttime NOx. We show from model results that injecting nighttime NO_x plumes in residual layer instead of surface layer has a significant impact on afternoon surface ozone by 5 - 10 ppb. Our results suggest that accurate representation of nighttime chemical and physical processes in the model are critical for predicting afternoon surface ozone. We will also discuss the implications of these nighttime processes on exporting NO_x from continental boundary layer.

Evan Couzo¹, Barry Lefer², Jochen Stutz³, Greg Yarwood⁴, Prakash Karamchandani⁴, Barron Henderson⁵, and William Vizuete⁶

¹Massachusetts Institute of Technology

²University of Houston

³University of California-Los Angeles

⁴ENVIRON

⁵University of Florida

⁶University of North Carolina

The chemical mechanisms used in regulatory air quality models typically allow for only homogeneous formation of nitrous acid (HONO), an important radical precursor. This study adds heterogeneous formation on surfaces as a HONO source to the Comprehensive Air quality Model with extensions (CAMx). A surface sub-model was developed that uses fluxes of nitric acid (HNO₃) and nitrogen dioxide (NO₂) to accumulate a surface inventory via dry deposition that can be converted to HONO. Modeling was performed for the Houston, Texas, region on April 21, 2009. Comparisons to measurements made at Moody Tower during the Study of Houston Atmospheric Radical Precursors (SHARP) show that adding heterogeneous formation increases HONO concentrations, particularly in the early morning and at night. Heterogeneous HONO formation reduces normalized mean error for daytime (nighttime) HONO concentrations from 62% to 36% (96% to 65%). Maximum daily 8-hr O₃ concentrations were up to 3.5 ppb greater as a result of heterogeneous HONO formation. Direct HONO emissions equal to 0.8% of NO_x emissions were also added to the model, but they were considered separately from heterogeneous HONO formation. Increases over the base case of HONO and O₃ were seen, though the magnitudes are not as great as with heterogeneous HONO formation. Maximum daily 8-hr O₃ concentrations were up to 0.4 ppb greater than with homogeneous HONO formation alone. Significant early morning and nighttime HONO under predictions were seen compared to SHARP measurements in the direct emissions scenario. Direct HONO emissions led to local increases of HONO and O₃ in areas with high NO_x emissions, but heterogeneous HONO formation led to regional increases of both. Process analysis was used to determine the effect on O₃ chemistry in downtown Houston. Daily total hydroxyl radical (OH) production from HONO photolysis was 3.94 ppb/day with heterogeneous HONO formation, 2.40 ppb/day with direct emissions, and 1.40 ppb/day in the base case. A 10% increase of hydrocarbon and carbon monoxide oxidation by OH was seen with heterogeneous HONO formation, while the increase in the direct emissions scenario was 3%. Nitric oxide (NO) to NO₂ conversion increased by 8% with heterogeneous formation, while the increase was only 3% with direct HONO emissions. Radical sources, radical propagation, and oxidant production were enhanced at each step in the chemical cycle - particularly just after sunrise - by the addition of heterogeneous HONO formation and direct HONO emissions, but the effects were greater with heterogeneous formation.

Yuzhi Chen, Jason Surrat, Kenneth Sexton, Roger Jerry, William Vizuete

Isoprene, the most emitted non-methane hydrocarbon, is known to influence ozone (O₃) formation in urban areas rich with biogenic emissions. To keep up with the recent advance on isoprene oxidation chemistry including the identification of isoprene epoxydiols (IEPOX) as a precursor to secondary organic aerosol (SOA), Xie et al. (2013) updated the SAPRC (Statewide Air Pollution Research Center)-07 chemical mechanism. It is currently unknown how the Xie modification of SAPRC07 impacts the ability of the model to predict O₃. In this project we evaluated the Xie mechanism with simulations of 24 isoprene experiments from the UNC gas-phase chamber. Our results suggest that the new mechanism accelerates the NO_X(nitrogen oxides) inter-conversion and produces more O₃ than SAPRC07 for all experiments. In lower NO_X experiments, the new mechanism worsens O₃ performance towards the wrong direction, increasing bias from 8.9% to 15.8%. We found increased NO_X recycling from PANs accounts for that. This could be explained by more PANs made due to increased first generation VOC products and OH production.

Yasuyuki Akita

Marc Serre

J. Jason West

Human health impacts from exposure to ambient air pollutants are often assessed using air quality models. Recent air pollution epidemiologic studies have attempted to account for intraurban exposure gradients, recognizing the importance of local scale spatial variability in ambient concentrations. Global scale air quality models, however, necessarily use a coarse grid resolution due to limits on computational power. These coarse resolution models are expected to underestimate urban peak concentrations, particularly for PM2.5, leading to underestimates of human exposure. In this study, we evaluate the influence of grid resolution on modeled PM2.5 concentrations over the continental US using the outputs of satellite estimates of ground-level PM2.5 and a chemical transport model (CTM). The satellite-based concentrations at a 0.1 degree grid were aggregated to multiple coarse resolutions from 0.2 to 4 degrees. Then, population weighted concentration (PWC) was computed at each resolution. The PWC and the grid size were inversely correlated (Pearson's correlation coefficient r is -0.925) and the PWC at coarse global model resolution (2.5 deg) was approximately 10% lower than the original scale (0.1 deg.) estimate. The magnitude of the bias due to the model resolution for satellite data was, however, smaller than that estimated in a previous study based on the CTM. The spatial variation of each modeling approach was further explored by comparing the modeled concentrations with ground observations and computing the spatial autocorrelation of concentrations from the observations, CTM, and satellite datasets. Quantifying the uncertainty associated with both the modeling approach and grid cell size provides insight to correct for sub-grid variations in pollutant concentrations for global health assessment.

Marina Astitha*, Emmanouil Anagnostou, Jaemo Yang, Xinxuan Zhang, Maria Frediani

University of Connecticut, Department of Civil & Environmental Engineering, 261 Glenbrook Road, Storrs-Mansfield, 06269-3037, CT, US

*astitha@engr.uconn.edu

Extreme weather events associated with high wind speed and precipitation (rain, snow, ice) have severe impacts in human lives and the environment and have become more frequent in the Northeastern part of the United States. The storm type, strength and duration dictates the severity of the consequences to infrastructure and every-day life as well as human life itself. The linkage between weather prediction, weather impacts and resiliency has become a major target for legislators and scientists in the recent years. In this work, we present the framework under which we use numerical weather prediction models in-house to produce real-time operational forecasts during extreme weather events for NE U.S. and further analyze past storm cases that affected the region. Approximately 100 storms have been analyzed so far covering the period from 2001 to 2013 ranging from thunderstorms, snow/ice storms to typical winter storms and hurricanes. The numerical weather prediction models used are WRF and RAMS/ICLAMS in the view of assessing the uncertainty of atmospheric variables by implementing two different and, at the same time, similarly configured modeling systems. To accomplish that, we use the two models to create an ensemble that informs other research activities related to infrastructure resiliency and adaptation. Going one step forward, we assess the effects of natural aerosols (sea salt, desert dust) in atmospheric dynamics by including the radiative feedback and the explicit treatment of cloud condensation nuclei in RAMS/ICLAMS modeling system. The consistent coupling between atmospheric physical and chemical processes allows the investigation of aerosol effects on meteorological conditions and vice versa in an interactive way. The same modeling system was applied in forecast mode during the Integrated Precipitation and Hydrology EXperiment (IPHEX) that supports the Global Precipitation Measurement (GPM) mission, through daily weather forecasts for hydrological forecasting and post-experiment activities (IOP: May-June 2014, North Carolina). Results from both activities related to ensemble weather prediction and aerosol effects will be presented and discussed.

George Delic, HiPERiSM Consulting, LLC, PO Box 569, Chapel Hill, NC 27514

When discussing topics such as energy, climate, and pollutant transport from regional to global scales, it is appropriate to also review some algorithms common to both the Community Multi-scale Air Quality Model (CMAQ) and the Global Modeling Initiative (GMI), as respectively developed (or supported) by the U.S. EPA and NASA Goddard Space Flight Center. Some motivations of new strategies for improvement in these models include issues such as increasing model complexity, changes in computer architectures, and developments in compiler technology. The memory model in both CMAQ and GMI maps the atmospheric domain onto a regular grid of cells to enable multiple physical processes to work in tandem and to synchronize at discrete transport time steps. One such physical process is in the gas phase chemistry-transport model (CCTM). This is one of the most computationally intensive modules, and currently employs a memory model that is not well suited to the commodity architectures now in use because it requires repeated mapping between the regular grid and linear arrays mapped into blocks of cells passed to the CCTM. Both CMAQ and GMI allow a choice of various gas-phase solvers that implement different algorithms for integration of a stiff system of ordinary differential equations, with sparse Jacobians, to describe production and loss of chemical species. One such solver is the Gear algorithm and its use is common to both CMAQ and GMI. The implementation in both models is a method based on the work of Jacobson and Turco [1]. This legacy algorithm was developed on vector register architectures on Cray computers over 20 years ago, and relies on hardware features of that era that are not available in current computer architectures. Thus, a logical first step in a new strategy is to replace the existing CCTM algorithm with one better suited to today's architecture. Such a proposal was previously described in [2]. This replaces the legacy sparse matrix methodology of [1] by a more modern one. During the course of the work in [2] a revision of the memory model suggested itself and would significantly improve the numerical precision possible for the chemistry solver in both CMAQ and GMI models. The blocking of cells into groups required in the method of [1], before invoking the CCTM, means that the convergence criterion uses an average error over all cells in a block. This assumption places limits on the numerical precision possible for chemical species concentrations. Therefore, a second logical strategic step is to change the memory model described above, to one that traverses the existing grid domain, and applies the Gear solver to each individual cell, and not blocks of cells as proposed in [1] as is the currently implementation in both CMAQ and GMI. This presentation describes how these two strategic modifications were successfully applied to CMAQ. For the Gear algorithm the legacy JSparse solver method [1], is replaced by the FSparse solver based on procedures from the CSparse library of Davis [3]. The revised Gear solver adds both instruction level parallelism and thread level parallelism to the standard EPA CMAQ release. Results with 8 threads are ~1.8 times faster than the standard EPA release of CMAQ. This speedup represents 90% of the theoretical maximum of 2. Similar improvements are in progress on the Gear solver implementation in the GMI code and results will be reported at the time of the meeting.
[1] M. Jacobson and R.P. Turco (1994), Atmos. Environ. 28, 273-284.
[2] G. Delic, 8th International Workshop on Parallel Matrix Algorithms and Applications (PMAA14), July 2-4, 2014, Universite della Svizzera italiana, Lugano, Switzerland.
[3] T.A. Davis, Direct Methods for Sparse Linear Systems, SIAM, Philadelphia, 2006.

Kan Huang, Joshua S. Fu*, Xinyi Dong, Jian Sun, Cheng-En Yang

Department of Civil & Environmental Engineering, University of Tennessee, Knoxville, 37996-2313

As the air quality standards tighten and the global emissions increase, regional/long-range transported air pollutants have become more and more important. Thus, reasonable representation of initial and boundary conditions are crucial for regional air quality simulations. This study stands as part of the Model Inter-Comparison Study - Asia Phase III (MICS-Asia III), aiming to obtain common understanding of model performance and uncertainties in Asia. In addition to meteorology and emissions inputs, IC/BCs are one of the most important factors affecting the model performance. Asia is currently the continent suffering most severe air pollution, thus using "clean air" boundary profiles originally generated from the CMAQ model is inappropriate. Recently, demands from global simulations to provide IC/BC for regional air quality models are growing with the advantages of providing relatively high temporal and vertically resolved gas and aerosol information.

In this study, we intend to use multiple global models to provide IC/BC for CMAQ simulations for the MICS-ASIA project. Four global models are selected, i.e. GEOS-Chem (http://wiki.seas.harvard.edu/geos-chem/index.php/Main_Page), HCMAQ (Mathur et al., 2012), CHASER (Sudo et al., 2002), and GFDL-AM3 (Donner et al., 2011). Spatial mapping and vertical interpolation from GCMs to CMAQ are developed based on different horizontal and vertical resolutions of various global models (Lam and Fu, 2010). Also, species mapping are developed between each GCM and CMAQ due to inconsistence of output species.

Boundary conditions at the four directions (North, South, West and East) are evaluated against satellite sensors with capability of measuring the vertical profiles, e.g. O₃ from Tropospheric Emission Spectrometer (TES) and the Ozone Monitoring Instrument (OMI). Simulations using different IC/BCs are inter-compared and evaluated by using ground-based observations near the boundary, e.g. EANET network in Japan near the eastern boundary and in Southeast Asia near the southern boundary. This study demonstrates the feasibility of global simulation in the CMAQ modeling system and impacts of different IC/BCs on convection, urban photochemistry and long-range transport.

References:

Donner, L. J. et al. The Dynamical Core, Physical Parameterizations, and Basic Simulation Characteristics of the Atmospheric Component AM3 of the GFDL Global Coupled Model CM3. Journal of Climate 24, 3484-3519 (2011).

Lam, Y. F. and J. S. Fu (2010). Corrigendum to "A novel downscaling technique for the linkage of global and regional air quality modeling". Atmos. Chem. Phys., 10, 4013-4031.

Sudo, K., Takahashi, M., Kurokawa, J., and Akimoto, H., CHASER: A global chemical model of the troposphere - 1. Model description, J Geophys Res-Atmos, 107, doi:10.1029/2001jd001113, 2002.

Mathur, R., et al., Extending the Applicability of the Community Multiscale Air Quality Model to Hemispheric Scales: Motivation, Challenges, and Progress, Air Pollution Modeling and its Application XXI NATO Science for Peace and Security Series C: Environmental Security 2012, pp 175-179.

Ajith P. Kaduwela^1,2, Sarika Kulkarni^1,*, Jianjun Chen¹, Jin Lu¹, Daniel Chau¹, Jeremy C. Avise^1,3, and John A. DaMassa¹,

¹Air Quality Planning and Science Division, Air Resources Board,

California Environmental Protection Agency, Sacramento, CA 95814, USA

²Air Quality Research Center, University of California, Davis, CA 95816, USA

³Department of Civil and Environmental Engineering, Washington State University, Pullman, WA 99164, USA

The U.S EPA modeling guidance recommends using the YYF approach to project current Design Values into future for the NAAQS attainment demonstration of ozone (O₃) and PM_2.5. However, it is widely recognized that higher O₃ mixing ratios are, in general, more responsive to emission controls of limiting precursors than lower mixing ratios are. The current form of the YYF concept does not allow for this enhanced response to emissions controls at the high end of the simulated/measured O₃ distribution and uses a single YYF value to represent a broad range of O₃ values in the baseline and future years. We have developed segmented YYF approach termed band-YYF that takes into account the varied model responses for different ranges of O₃ mixing ratios. The "band-YYF" approach was previously demonstrated for the now revoked 1-hour O₃ NAAQS in the San Joaquin Valley (SJV). In this study, we present the application of the band-YYF concept to the 8-hour O₃ NAAQS. We will also discuss the applicability of the band-YYF concept to the 24-hour and annual PM_2.5 NAAQS.

Maudood Khan^{*, 1} and Arastoo Pour Biazar²

^*Corresponding author email: maudood.n.khan@comsats.edu.pk

¹COMSATS Institute of Information Technology, Islamabad, Pakistan

²University of Alabama in Huntsville, Huntsville, Alabama, USA

Advanced models of the earth system and publicly available data from space-borne systems are tools that can enable nation-states to build and efficiently operate reliable, affordable, and clean electricity generation system. Nowhere are these tools needed more than in developing countries which have limited resources at their disposal, ever expanding needs, and are highly vulnerable to the changing climate.

The Government of Pakistan has embarked on an ambitious plan to install 82 Giga Watt of additional electricity generation capacity by 2030 through investments in coal, nuclear, and hydroelectric power generation. India - its next-door neighbor and regional competitor is currently generating 59 percent of its electricity through coal, and is planning to build an additional 800,000 MW of coal-fired power plants by 2030.

We present the results of sensitivity simulation over Pakistan aimed at quantifying the impact of these investments on air quality using the WRF-CMAQ modeling system. Anthropogenic emissions prepared as part of Atmospheric Chemistry and Climate Model Inter-comparison Project (ACCMIP) project were augmented by local data from the Pakistan Environmental Protection Agency, Asian Development Bank, and others. The performance of the system was evaluated by comparing modeling predictions against in-situ data recorded at meteorological stations operated by the Pakistan Meteorological Department, Water and Power Development Authority, and satellite observations from NASA's Earth Observing System (EOS). WRF simulations were conducted using both 24-category USGS and MODIS Land Use/Land Cover (LULC). Preliminary results indicate significant variation in performance between meteorological stations. Air temperature is generally over predicted at elevated sites and under-predicted in urban areas. Use of LULC information from MODIS sensor improves model performance, especially at urban sites.

Sensitivity simulations have focused on ten thermal power plants that are currently operating, and four new facilities that are expected to come on-line over the next eight years. The Baseline scenario assumes continued use of High Sulfur Fuel Oil (HSFO) without emission controls. In Scenario 1, Flue Gas Desulfurization technology is installed on all units providing substantial reductions in emissions of criteria pollutants. In Scenario 2, each unit currently operating regardless of its size is replaced with a 660 MW unit, with higher stack height, control technologies (FGD, Electrostatic Precipitators, and Scrubbers), and mandatory use of blended coal (80 percent sub-bituminous imported coal with 20 percent domestic coal from Thar fields). In Scenario 3, twelve new 660 MW units are built at designated sites in Punjab and Baluchistan provinces. Preliminary results indicate substantial benefits of emission control technologies on air quality. The analysis is the first step towards assessing the impact of coal-fired power plants and emission control technologies on public health and agriculture productivity, which we hope to present at subsequent meetings.

Kumar Saurabh, Goyal P. and Sindhwani Rati

Center for Atmospheric Sciences, Indian Institute of Technology,

Hauz Khas, New Delhi, India-110016

The presence of alarmingly high levels of Respirable Suspended Particulate Matter (PM₁₀), has become a point of major concern for the regulatory authorities in megacity Delhi and world over. In order to have a better insight of the air quality of megacity Delhi, a regional chemical transport model WRF/Chem has been applied during winter season from 24^th to 30^th, December 2008. During these 6-days the air quality monitoring stations in Delhi recorded extremely high values of the daily averaged PM₁₀ concentration, which reached up to 308 mg/m³, nearly 5-times higher than the National Ambient Air Quality Standards (NAAQS). In order to assess the model performance several statistical parameters have been evaluated. The simulated meteorological conditions namely temperature, wind speed and relative humidity showed very good correlation with their observed values during the study period. The WRF/Chem model very well simulated the 8-hourly variations of the chemical species concentrations over the megacity, with best correlation coefficients of 0.52, 0.41 and 0.42 w.r.t observed O₃, NO₂ and PM₁₀ respectively. The study used emissions data from Emission Database for Global Atmospheric Research (EDGAR) having (0.1 x 0.1) resolution. Due to coarser resolution of EDGAR inventory and lack of good accountability of regional emission sources resulted in consistent under prediction of criteria pollutants concentration. To overcome this drawback, a spatial inventory of 2 km x 2 km has been further used to simulate pollutant concentration over megacity Delhi.

The long-range transport of air pollutants from other continents has been found to have great effects upon regional air pollution.  To account for the influence of continental transport, boundary conditions (BC) from a Global Chemistry Model (GCM) have been used in a limited-domain air quality model (i.e., CMAQ and CAMx) to provide real-time chemical flux through the boundaries.  In the recent past, various GCMs have been available for regional air quality studies, including GEOS-chem, WRF-Chem, CHASER and IFS-CB05 MACC.  In this study, we cross-compared three GCM models in the year of 2010 for what we labelled the “East Asia domain” using CMAQ under the MISC-ASIA, and HTAP Jointed experiments; the afore mentioned were provided by Seoul National University for GEOS-Chem; Nagoya University for CHASER and ECWMF for IFS CB05 MACC.   All CMAQ results are tabulated to provide insight about the model comparison of these GCMs for regional application and to demonstrate how the influence of boundary condition selection affects the regional simulation in East Asia.

Xiangting Hou, Chih-Yuan Chang and Kuo-Jen Liao

Department of Environmental Engineering, Texas A&M University-Kingsville

Identification of regional air qualiy management strategies is a difficult task because formation of air pollutants is interdependent and air quality at different locations may have different responses to common emission sources. In this study, we develop a mathmatical programming model, Optimal Emission Reduction Alternatives (OPERA), which finds cost-minimized control strategies for attaining prescribed multipollutant air quality targets at multiple locations simultaneously using sensitivities of pollutants to emissions and cost functions of emission reductions. The Community Multiscale Air Quality Model (CMAQ) with the High-order Decoupled Direct Method (HDDM) is applied to calculate sensitivities of ambient ozone and fine particulate matter (i.e., PM2.5) concentratons to changes in emissions from different sources. Cost functions of air pollutant emission reductions are estmated using an U.S. EPA's control technology analysis tool, AirControlNET.

In this presentation, we will demonstrate how to use OPERA to develop cost-minimized control measures for simultaneously reducing ozone and PM2.5 concentrations in major Metropolitian Statistical Areas (MSAs with more than 5 million people) in the eastern U.S. Overall, OPERA is expected to improve the capability of state (e.g., Texas) and regional (e.g., northeastern U.S.) air quality planning organizations developing multipollutant air quality management strategies.

Elena C. McDonald-Buller^1*, Yosuke Kimura¹, Melissa P. Sulprizio², Jaegun Jung³, Greg Yarwood³, Daniel J. Jacob², Jeremiah Johnson³, Chris Emery³, Gary McGaughey¹, and David T. Allen¹

¹University of Texas at Austin, Austin, Texas 78758, United States

²Harvard University, Cambridge, Massachusetts 02138, United States

³ENVIRON International, Novato, California 94998, United States

North American Background (NAB; formerly known as Policy Relevant Background or PRB) ozone concentrations have been defined by the U.S. Environmental Protection Agency as those that would occur in the absence of anthropogenic emissions in continental North America. NAB ozone concentrations are a construct of chemical transport models that have been used to inform decisions about National Ambient Air Quality Standards (NAAQS). At the state-level, understanding the contributions of natural and transboundary anthropogenic sources can be important for effective air quality planning, especially under standards that may become increasingly stringent in the future. An analysis of NAB ozone concentrations in Texas was conducted using a regional air quality model with a high-resolution global simulation. Texas has highly diverse climatology, land cover, and topography. Across all of Texas, NAB ozone concentrations increased with altitude above ground level. Differences in NAB maximum daily 8-hour averaged (MDA8) ozone concentrations exist between urban areas in eastern Texas and those located in the intermountain west. The variability in NAB MDA8 ozone concentrations during multi-day episodes with distinct transport patterns can differ substantially from monthly or seasonal averages. This suggests the need for careful understanding of time periods selected for regulatory air quality modeling and planning efforts.

Golam Sarwar1, Brett Gantt1, Donna Schwede1, Rohit Mathur1, Alfonso Saiz-Lopez2

1 Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, Environmental Protection Agency, Research Triangle Park, North Carolina

2 Atmospheric Chemistry and Climate Group, Institute of Physical Chemistry Rocasolano, CSIC, Madrid, Spain

Iodide present in sea-water can interact with atmospheric ozone leading to enhanced ozone deposition over the oceans. The interaction of iodide in sea-water with atmospheric ozone, along with other processes, leads to the release of halogens into the atmosphere which participate in atmospheric chemistry to reduce atmospheric ozone concentrations over marine environments. Most chemical transport models, however, do not include these air-sea processes. In this study, an enhanced ozone deposition scheme based on the interaction of sea-water iodide and atmospheric ozone and the detailed chemical reactions of organic and inorganic halogen species are incorporated into the hemispheric Community Multiscale Air Quality model (CMAQ). Model simulations are performed without and with the enhanced ozone deposition scheme and the halogen chemistry over the Northern Hemisphere for the summer months of 2006. Preliminary results suggest that both the enhanced ozone deposition and the halogen chemistry reduce ozone over marine environments. The enhanced ozone deposition scheme reduces monthly mean surface ozone by up to 4 ppbv while halogen chemistry reduces monthly mean surface ozone by up to 8 ppbv. However, their spatial impacts are different and together they reduce monthly mean ozone concentrations by up to 10-11 pbv. Model predictions obtained with the inclusion of these processes agree better with observed ozone over the marine environments. Enhanced ozone deposition and halogen chemistry are two important atmospheric processes that improve model performance especially over marine and coastal environments.

Raquel Silva, J. Jason West

Outdoor air pollution has increased significantly due to anthropogenic emissions of air pollutants and their precursors. Exposure to air pollution from ozone and fine particulate matter (PM_2.5) can cause adverse health effects, particularly cardiovascular and respiratory morbidity and mortality. We previously estimated 2.1 million PM_2.5- related deaths annually from cardiopulmonary diseases and lung cancer, and 470,000 ozone-related deaths annually from respiratory diseases, at a global scale. This and other recent studies have quantified global air pollution mortality but they do not estimate the contribution of different emission sectors or they focus on a specific emissions sector.

Here we estimate the impact of five anthropogenic emissions sectors on present-day global mortality due to exposure to ozone and PM_2.5. We perform simulations at fine (0.67x0.5) and coarse (2.5x1.9) resolutions with the MOZART-4 global chemical-transport model (CTM), driven by GEOS-5 meteorological fields and with input emissions from the RCP8.5 global emissions inventory for the present day (2005). To estimate the contribution from each sector, we use brute-force simulations, zeroing-out single emissions sectors (all transportation, land transportation, energy, industry, and residential & commercial).

We estimate cause-specific mortality due to exposure to air pollution using health impact functions (a log-linear function for ozone and the GBD 2010 Integrated Exposure-Response model for PM_2.5), and taking into account the spatial distribution of the exposed population and country-level baseline mortality rates. Estimates are obtained both at fine and coarse resolutions to ascertain the potential bias from using coarse resolution global models. Understanding how different sectors impact global and regional human mortality can help prioritize air pollution control strategies at the international and national levels.

Wen Xu¹, Fuquan Yang², Nick Walters², Xin Qiu²

¹Alberta Environment and Sustainable Resource Development, Edmonton, Alberta, Canada

²Novus Environmental Inc., Guelph, Ontario, Canada

In situ oil sands development is expected to dominate bitumen production in the coming decades and much of it will be located in the south Athabasca oil sands area (SAOS). The Weather Research and Forecasting (WRF) model is run for SAOS in 2010 to provide meteorological input for air quality modelling, improve model performance and reduce model biases. The SAOS WRF modelling uses a fine temporal input resolution (i.e., 3-hourly interval NAYY) and local observation data for nudging to generate WRF meteorological fields for use in CMAQ and CALPUFF modelling. Based on developed emission inventories and modelling inputs for 2010, this study applies the CMAQ model to simulate the ground level concentrations of ozone, PM_2.5, PM₁₀, NO₂, SO₂, CO, and acid deposition in SAOS in 2010. CALMET and CALPUFF models are also applied to a relatively smaller domain, to compare CALPUFF modelling results of 2010 acid deposition in SAOS with the CMAQ modelling output.

The CMAQ modelling results for 2010 are evaluated against observations from 10 ambient monitoring stations. The model overestimates ground level ozone concentrations by 0 to 10 ppb monthly, but is robust enough to capture monthly patterns and high percentile values from observed ozone concentrations. It also illustrates a larger underestimation of PM_2.5 concentrations during summer months mainly due to counting out wild fire events during wildfire season in 2010. Furthermore, the model overall underestimates NO and NO₂ concentrations, except in some summer time overestimations. Nevertheless, the model captures the monthly patterns of NO and NO₂ concentrations in the observations throughout the year.

The WRF model full-year simulation is also evaluated for application in the SAOS area for 2010. Weather parameters, including temperature, wind speed and direction, etc., are compared and evaluated with observations from 9 surface weather stations in the SAOS modelling domain. The evaluation indicates the model performs reasonably well and that the WRF model is sufficient to support CMAQ and CALPUFF modelling in this study.

Yong Zhang^{1, 2}, Zhongyuan Mi¹, Ting Cai ^{1, 3}, Leonard Bielory^{1, 3}, Yang Gao^{4, 5}, L. Ruby Leung⁵, Joshua Fu⁴, Panos G. Georgopoulos^{1, 2, 3}

¹Environmental and Occupational Health Sciences Institute (EOHSI), Rutgers University, 170 Frelinghuysen Road,Piscataway, NJ 08854, USA

²Department of Chemical and Biochemical Engineering, Rutgers University, 98 Brett Road, Piscataway, NJ 08854, USA

³Department of Environmental Science, Rutgers University, New Brunswick, NJ 08901, USA

⁴Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, TN 37996, USA

⁵Atmospheric Sciences and Global Change Division, Pacific Northwest National Laboratory, Richland, WA 99352, USA

Abstract: Many diseases are linked with climate trends and variations. In particular, climate change is expected to alter the spatiotemporal dynamics of allergenic airborne pollen and potentially increase occurrence of allergic airway disease. A comprehensive prognostic modeling system, combining climate models and anthropogenic and biogenic emission models with an expanded version of the Community Multiscale Air Quality (CMAQ) Model has been developed to support integrated studies of the impact of climate change on Airway Allergic Disease (AAD). The present work focuses on description of the mechanistic pollen emission module and the transport module, and their applications to model the spatiotemporal distributions of allergenic pollen from representative trees, weeds and grasses.

The pollen emission module incorporated major physical processes such as direct emission and re-suspension of pollen particles and considering effects of meteorological parameters such as ground surface temperature, friction velocity and humidity, etc. Season start and length of birch, oak, ragweed, mugwort and grass were simulated using species specific observation-based and processed-based methods. Daily and hourly flowering fractions of trees, weeds and grasses were parameterized using the observed airborne pollen counts, phenology and relevant meteorology factors such as humidity, temperature and sunrise time. Vegetation coverage of allergenic plants were derived using observed airborne pollen counts and Land Use Land Cover data from Biogenic Emissions Land use Database, version 3.1 (BELD3.1). Using the emission module, spatiotemporal distributions of pollen due to transport was simulated via the combined application of the Weather Research and Forecasting (WRF) model and an adapted version of the CMAQ model.

The modeling system were used to simulate the allergenic pollen season timing and airborne levels of representative trees, weeds and grasses for multiple historical and future years. The estimated mean start dates and season length for birch, oak, ragweed, mugwort and grass pollen season in 1994-2010 are mostly within 0 to 6 days of the corresponding observations for the majority of the National Allergy Bureau (NAB) monitoring stations across the contiguous United States (CONUS). The simulated spatially resolved maps for onset and duration of allergenic pollen season in the CONUS are consistent with the long term observations. Changes of pollen season timing and airborne levels depend on latitude, and are associated with changes of growing degree days, frost free days, and precipitation. These changes are likely due to recent climate change and particularly the enhanced warming and precipitation at higher latitudes in the CONUS.

Keywords: Climate change, Allergenic pollen, CMAQ, Allergic airway disease, Emission model

Yuqiang Zhang¹, Jared Bowden² , Zachariah Adelman^1,2, Vaishali Naik³, Larry W. Horowitz⁴, J. Jason West¹

¹ Environmental Sciences and Engineering Department, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599

² Institute for the Environment, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599

³ UCAR/NOAA Geophysical Fluid Dynamics Laboratory, Princeton, NJ 08540

⁴ NOAA Geophysical Fluid Dynamics Laboratory, Princeton, NJ 08540

Actions to mitigate greenhouse gas (GHG) emissions will slow climate change, which will influence air quality in the long term; reductions in co-emitted air pollutants, such as SO₂, NO_x and CO which usually share same sources as GHGs, can affect air quality immediately. We previously used a global model (MOZART-4) to show that global GHG mitigation will have significant co-benefits for air quality and human health. In doing so, we contrasted the Representative Concentration Pathway Scenario 4.5 (RCP4.5), treated as a GHG mitigation scenario, with its associated reference case scenario (REF). Here we aim to estimate the air quality co-benefits in the US at fine resolution using dynamical downscaling techniques, comparing RCP4.5 and REF in 2050. We also quantify the co-benefits due to reductions in co-emitted air pollutants versus slowing climate change and its influence on air quality. Finally, we investigate the co-benefits due to domestic GHG mitigation in the US alone at fine resolution, and compare these co-benefits with those resulting from GHG mitigation outside of the US. This work focuses on downscaling the meteorology and air pollutant chemistry to the US domain at a finer resolution. To do that, we use the latest Weather Research and Forecasting (WRF) model as a Regional Climate Model (RCM) to dynamically downscale the GFDL AM3 General Circulation Model (GCM) over the US at 36 km resolution, in 2000 and 2050. The 2000 simulation is then compared with multi-year surface observation data and all simulations with the GCM. These simulations are used as inputs for the newest Community Multiscale Air Quality (CMAQ) modeling system. Anthropogenic emissions for the REF and RCP4.5 scenarios are directly processed through SMOKE to prepare temporally- and spatially-resolved CMAQ emission input files, instead of the traditional method of mapping the ratio of total air pollutants in future RCP scenarios to present day US NEI emissions, as in other studies. Initial conditions (IC) and dynamic boundary conditions (BC) for CMAQ are derived from the global MOZART-4 simulations. We will evaluate the co-benefits of GHG mitigation by changing the meteorological and air pollutant emissions inputs for RCP4.5 and REF, as well as the BCs and fixed global methane.

Chul Yoo*, Hyung-Cheon Kim, Dae-Il Kang, Ji-Hyung Hong
National Institute Environmental Research, Ministry of Environment in Korea

Soon-Tae Kim, Chang-Han Bae,
The University of AJOU, Suwon city, Kyonggi-Do, Korea

The Government had devised legislation of Special Act and established guidelines to  improve air quality of Seoul Metropolitan area in Korea. Local governments of Seoul, Incheon and Gyeonggi made detailed plans to implement this Special Act would perform effectively. In 2013 MOE revised 1st Basic plan to improve the atmospheric conditions for the next 10 years and announce 2nd Basic plan of Seoul metropolitan area. The goal of the 2nd Basic Plan is to decrease PM10, PM2.5 and O3 down to 30 /m3, 20 /m3 and 60ppb respectively. This 2nd Basic Plan include management plan for mobile sources, point emissions as well as household emissions, etc. Based on the 2nd Basic Plan, Local governments must reduce the emissions of air pollutants to meet the goal.
More detailed descriptions of 2nd Basic Plan and its implementation plan will be presented in the conference.

Deborah Luecken, Donna Schwede

Atmospheric Modeling and Analysis Division

U.S. EPA, Research Triangle Park, NC 27709

The interconversion of oxidized nitrogen species plays an essential role in the formation of ozone and particulate matter, and is used in models as a direct way to modify ozone formation.   However, organic nitrogen species can also act as an intermediary in this process by trapping NOx and either removing it from the system or releasing it at another location downwind, where chemical or meteorological conditions might be different.

In this study, we look at the sensitivity of ozone concentrations to changes that are made to the assignment of organic nitrates to representative model species and to the physical properties assigned to these nitrates.  We use CB05 as our base mechanism and classify secondary nitrates by features that have the largest effect on nitrate decay and removal.  We will detail assumptions and uncertainties in the modifications that are made.  All sensitivity studies result in representation of organic nitrates with generally increased reactivity and NO2 release, which in turn increases ozone formation.  While these changes slightly improve performance of CMAQ at high ozone levels and slightly degrades it at lower levels, it improves some other measures of air quality.

Jerold A. Herwehe, Kiran Alapaty, and O. Russell Bullock Jr.

Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, USA

Several enhancements have been recently developed for the Kain-Fritsch (KF) convection parameterization in the Weather Research and Forecasting (WRF) model. To allow the smooth application of KF across scales down to about 1 km grid spacing, new formulations and science updates were developed for several cumulus cloud processes, including the convective adjustment timescale, the entrainment of environmental air, and the fallout of condensates from updrafts. To evaluate the performance of these multiscale KF developments, WRF v3.6 was used to simulate the 2006 summer season (JJA) over the eastern United States using 12 km grid spacing. Downscaled NCEP-DOE Atmospheric Model Intercomparison Project reanalysis (R-2) data drove coarser WRF simulations on a 108 km grid with a nested 36 km grid, which then provided initial and lateral boundary conditions and four-dimensional data assimilation inputs for the 12 km grid. Individual simulations on the 12 km grid were produced with each of these new KF formulations for a sensitivity study of the relative impacts of each on regional meteorological parameters such as cloudiness, temperature, surface fluxes, and grid- and subgrid-scale precipitation. Analysis of the results obtained from these summer simulations and an evaluation of the performance of the multiscale KF scheme, along with its potential implications for air quality modeling, will be presented

Eleana M. Little, Thien Khoi V. Nguyen, Annmarie G. Carlton

Author Affiliations: Department of Environmental Sciences, Rutgers University, New Brunswick, New Jersey 08901, United States

Atmospheric fine particulate matter (PM_2.5) is hazardous to human health, primarily contributing to pulmonary and cardiovascular diseases. Although secondary organic carbon (SOC) is a significant contributor to organic PM_2.5 mass, the Community Multi-Scale Air Quality (CMAQ) model persistently underpredicts SOC mass concentrations in a wide array of urban, rural, and pristine areas. Broadly, SOC is produced by the tropospheric chemical oxidation of volatile organic carbon (VOC) precursor compounds. This study aims to better characterize potential sources of error in SOC prediction - namely, the influence of aerosol water and precursor compound mixing ratios. Normalized mean bias (NMB) is computed for SOC for each day of the week, with data further separated by both season and geographical region from a year-long CMAQv.4.7.1 simulation of the continental U.S. Overall NMB for SOC mass for all regions and seasons is -66%. Seasonal NMBs range from -22% (winter in the Northeastern U.S.) to -83% (winter in the Mountain Western U.S.). SOC NMBs are then compared to both semi-empirical estimates of water concentration and measured concentrations of selected VOCs that are SOC precursor compounds. NMB for SOC mass in dry areas is more correlated with predicted aerosol water than in humid regions. Geographic differences in mean and median aromatic VOC relationships to SOC NMB suggest varying regional influences of outlier VOC measurements. We anticipate that these results will help guide development of CMAQ's SOA module.

Advancements in Operational CMAQ MODIS AOD data-assimilation at Baron Advanced Meteorological Systems during Forecast Year 2013

Advancements in Operational CMAQ MODIS AOD data-assimilation at Baron Advanced Meteorological Systems during Forecast Year 2013

George Pouliot,Atmospheric Modeling Division, National Exposure Research Laboratory, Environmental Protection Agency, Research Triangle Park, NC 27711pouliot.george@epa.gov

Venkatesh Rao,Office of Air Quality Planning and Standards, Environmental Protection Agency, Research Triangle Park, NC 27711,Rao.Venkatesh@epa.gov

Jessica McCarty,Michigan Tech Research Institute,Ann Arbor, MI,jmccarty@mtu.edu

Amber Soja,Institute of Aerospace (NIA), NASA Langley Research Center, Hampton, VA,Amber.J.Soja@nasa.gov

Biomass burning has been identified as an important contributor to the degradation of air quality because of its impact on ozone and particulate matter. One sector of the National Emission Inventory, crop residue burning, has been difficult to characterize. Efforts have been made in previous national inventories to estimate this sector but uncertainties remain. In this paper, we will review this sector in the current and previous national inventories, discuss current uncertainties in the current estimates for the 2011 NEI, and explore possible future work and improvements. The main areas of uncertainty include identification and location of crop residue fires; estimating acres burned, and temporal allocation. Emission factors and fuel loadings for emission calculations also contribute further uncertainty. We will provide rationale for the methods used by EPA to develop inventories for this sector.

K. Wyat Appel, Chris Nolte, James Kelly, Prakash Bhave and Kathleen Fahey

At the 8^th Annual CMAS Conference initial results from the comparison of CMAQ4.7 model results against size-resolved measurements of aerosol SO₄^2-, NO₃^-, NH₄⁺, Na⁺, Cl^- were presented for rural sites across North America. Since that time the work has advanced to include model data from the latest version of CMAQ (v5.0), observations from additional sites across North America and additional species that were not available in previous versions of CMAQ. This work utilizes measurements from Micro-Orifice Uniform Deposit Impactor (MOUDI) to evaluate estimates of particle size composition from the Community Multiscale Air Quality (CMAQ) modeling system for rural sites in North America (i.e. U.S. and Canada). Measurements of particle SO₄^2-, NO₃^-, NH₄⁺, Na⁺, Cl^-, Mg²⁺, Ca²⁺ and K⁺ size distributions from MOUDI instruments are compared to the same values from the CMAQ model for time periods in 2002 through 2005. The MOUDI measurements were predominantly made in remote areas (e.g. National Parks) in the United States and Canada, and measurements were typically made for a period of roughly one month. For particle SO₄²- and NH₄⁺, the model performance was consistent across both the U.S. and Canada sites, with model slightly overestimating the peak particle diameter and underestimating the peak particle density compared to the observations. The model performance for NO₃^- was much more mixed than that of SO₄²- and NH₄⁺, with the model ranging from an underestimation to overestimation of both the peak diameter and peak particle density across the U.S. and Canada sites. Model performance continued to be mixed for the remaining species, with the best overall model performance occurring at the MOUDI sites located in Florida. An additional analysis of the impact of applying a sharp cutoff at 2.5 μm to the model values (as opposed to simply summing the Aitken and Accumulation modes) showed that the difference in model values between the two methods is typically 1 μgm^-3 or less, while the difference in seasonal and annual model performance compared to observations from the IMPROVE, CSN and AQS networks is very small.

Lijun Diao¹, Yunsoo Choi¹, Beata Czader¹, Xiangshang Li¹, Mark Estes²

¹Department of Earth and Atmospheric Sciences, University of Houston, Houston, TX 77204, USA

²Texas Commission on Environmental Quality, Austin, TX, 78753, USA

Aircraft data from Texas 2013 DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) field campaign over Houston during September 2013 is subject to PCA (Principal Component Analysis) analysis to characterize isoprene sources. Biogenic isoprene signature is evident in the third PC (Principal Component), while anthropogenic signals are from two other PCs. Evaluations of Community Multiscale Air Quality (CMAQ) model simulations of isoprene with airborne measurements are better in suburban areas than in industrial areas. Further comparisons of model outputs to eight surface automated gas chromatograph measurements near the Houston ship channel industrial area during the nighttime reveal that modeled anthropogenic isoprene can be underestimated by as high as >10. A new simulation with modified anthropogenic emission inventory (constraining using the ratios of observed values versus simulated ones duirng the nighttime) gives better isoprene predictions at night with significant reduction of mean bias. This study implies that the model-estimated isoprene emissions from the 2008 NEI (National Emission Inventory) are underestimated in the city of Houston, which shed light on that other climate models or CTMs (Chemistry and Transport Models) using the same emission inventory might have the similar issue in the other urban areas over the US.

Chris Emery, Jaegun Jung, Tan Sakulyanontvittaya, Greg Yarwood

ENVIRON International Corporation

The US EPA is reviewing the current ozone National Ambient Air Quality Standard (NAAQS) and is anticipating lowering the standard to a range of 60-70 ppb in the form of a three-year average of the annual 4^th highest daily maximum 8-hour concentration. Photochemical model simulations employing the High Order Decoupled Direct Method (HDDM) have been used to estimate ozone response from US anthropogenic emissions reductions that just attain alternative standards in several US cities (Simon et al., 2012; EPA, 2014; Downey et al., 2014). These past studies simulated historical years ranging from 2005 to 2007. Specifically, the EPA's Health Risk Assessment (EPA, 2014) relied on HDDM simulations of 2007 to project ozone frequency distributions that meet alternative ozone standards in 2006-2010 for use in health risk models. We report on new applications of the HDDM technique developed by Yarwood et al. (2013) using the Comprehensive Air quality Model with extensions (CAMx) with meteorological and emissions inputs from the EPA's 2011 and 2018 modeling platform. Emission projections that meet alternative ozone standards in the 60-70 ppb range are developed for a more recent historical year (2011) as well as a future year (2018), and results are compared to those reported by Downey et al. (2014) from their CAMx HDDM simulations of 2006. These results demonstrate the level of additional emission controls needed in the four cities analyzed by Downey et al. (2014) beyond the reductions that have occurred since 2006 and that are expected to occur out to 2018.

Fahey, K.M., Sarwar, G., Appel, K.W., and C.G. Nolte

Atmospheric Modeling and Analysis Division, National Exposure Research Lab, U.S. EPA, RTP, NC 27711

While CMAQ often reliably simulates particulate mass, previous studies have indicated that it can significantly underpredict particle number concentrations (Elleman and Covert, 2010; Kelly et al., 2011). Traditional regulatory modeling applications have focused on particulate mass, but as applications expand to include cloud-aerosol interactions and climate change or the health impacts of ultrafine particles, it may be important to well-characterize not only the aerosol mass but also the number size distribution.

Here we investigate the impacts of two processes important to the evolution of the aerosol size distribution: nucleation and the size distribution of particulate emissions. Updated "urban" PM modal emissions fractions developed by Elleman and Covert (2010) and alternative nucleation parameterizations, including the corrected version of CMAQ's current binary nucleation parameterization (Vehkamaki et al., 2002), are added to CMAQ v5.0.2. Monthly simulations are performed over the continental U.S. at 12 km resolution to determine the spatial and seasonal impacts of these updates. In addition to comparing modeled and observed mass concentrations, we also evaluate number distributions with measurements taken during the Pittsburgh Air Quality Study (Stanier et al., 2004). Finally we examine the potential radiative impacts of these updates with the WRF-CMAQ two-way model and make recommendations for future development.

References:

Elleman, R. and D. Covert (2010) Aerosol size distribution modeling with the Community Multiscale Air Quality modeling system in the Pacific Northwest: 3. Size distribution of particles emitted into a mesoscale model. J. Geophys. Res., v115, D03204.

Kelly, J.T., Avise, J., Cai, C., and A.P. Kaduwela (2011) Simulating Particle Size Distributions over California and Impact on Lung Deposition Fraction. Aerosol Science and Technology, 45:148-162,

Stanier, C., Khlystov, A., and S.N. Pandis (2004) Ambient Aerosol Size Distributions and Particle Number Concentrations Measured during the Pittsburgh Air Quality Study", Atmospheric Environment, v38, 3275-3284.

Vehkamaki, H., Kulmala, M., Napari, I., Lehtinen, K.E.J., Timmreck, C., Noppel, M., and A. Laaksonen (2002) An improved parameterization for sulfuric acid-water nucleation rates for tropospheric and stratospheric conditions. J. Geophys. Res., v107(22).

Christian Hogrefe¹ and Stefano Galmarini²

¹ Atmospheric Modeling and Analysis Division, U.S. EPA

² European Commission Joint Research Center

We present highlights of the results obtained in the second phase of the Air Quality Model Evaluation International Initiative (AQMEII) that was completed in May 2014. Activities in this phase were focused on the application and evaluation of coupled meteorology-chemistry models over both North America and Europe using common emissions and boundary conditions for all modeling groups. Roughly twenty modeling groups from both continents participated in this activity. Among the highlights presented are examples of operational, diagnostic, and dynamic model performance that were performed during this phase. We also present the objectives for a new phase of regional model intercomparisons that is coordinated with the ongoing work under the Task Force on Hemispheric Transport of Air Pollution (TF-HTAP) framework. The objectives of this new activity are to evaluate the performance of global, hemispheric and regional modeling systems over Europe and North America and to compare source / receptor relationships simulated through perturbation experiments reflecting emission changes in upwind continents.

Anikender Kumar¹, Alexander Rincon¹, Nestor Rojas¹

¹ Department of Chemical and Environmental Engineering, Universidad Nacional de Colombia, Bogota, Colombia

The fully coupled WRF-Chem (Weather Research and Forecasting with Chemistry) model is used to simulate air quality over Bogota. Bogota is a tropical South-American megacity located over a plateau in the middle of very complex terrain. An extensive sensitivity analysis to model gas phase chemistry schemes was performed. The WRF-Chem model was adopted for simulating the hourly ozone concentrations for three episodes, representative of dry, intermediate/transition and wet periods in 2010. The computational domains were chosen of 120x120x32, 121x121x32 and 121x121x32 grid points with horizontal resolutions of 27, 9 and 3 respectively. The model was initialized with real boundary conditions using NCAR-NCEP's Final Analysis (FNL) and a 1^ox1^o (~111 km x 111 km) resolution. Boundary conditions were updated every 6 hours using reanalysis data. The emission rates were obtained from global inventories, namely the REanalysis of the TROpospheric (RETRO) chemical composition and the Emission Database for Global Atmospheric Research (EDGAR). Comparisons between estimated and observed ozone concentrations were carried out through a series of common statistics. In this study, the atmospheric concentrations of ozone over Bogota were calculated to be 40-60 ppb during the simulation periods. Overall, the present case study shows that the model has an acceptable performance over Bogota. This study provides a general overview of WRF-Chem and can constitute a reference for future mesoscale air quality modeling exercises over Bogota and other Colombian cities.

Gary Moore, AECOM, 250 Apollo Drive, Chelmsford, MA 01824
William Leatham, AECOM, 250 Apollo Drive, Chelmsford, MA 01824
Jeffery Connors, AECOM, 250 Apollo Drive, Chelmsford, MA 01824
Robert Paine, AECOM, 250 Apollo Drive, Chelmsford, MA 01824

The Weather Research Forecasting (WRF) version 3.5.1 model was run over 3 one year periods for coastal British Columbia, CA. The modeling was a retrospective exercise to provide a source of meteorological data for dispersion and transport modeling, for sources near Prince Rupert, CA. The traditional (Tesche/Complex)¹ statistical evaluation metrics were computed and analyzed utilizing the METSTATS² program. According to traditional evaluation parameters the model performs within acceptable limits the majority of the time. However, in most model evaluations tests for degrees of significance and confidence intervals are not discussed. Stagnation with low ventilation tends to cause the largest observed concentrations. Generic metrics often provide little discriminatory capability for such conditions. In addition, the model evaluation literature has little to report on factors like the autocorrelation of residuals which accumulate model defects in trajectories and other meteorological parameters important to dispersion. In this study we explore visually and quantitatively model performance for specific percentile ranges of surface meteorological parameters. We utilize joint distributions like wind roses, cumulative trajectory drift plots, confidence intervals on statistical evaluation metrics and time series to provide better diagnostics to judge model performance.

Robert Nedbor-Gross, Barron Henderson, Justin Davis, Jorge Pachon

Unhealthy air pollution is common in developing cities where economic growth outpaces the government regulation. In South America, population centers and economic growth occur in mountainous regions. The mountainous terrain has been documented as problematic for the meteorological component of air quality models. This presentation will focus on the evaluation of meteorology for Bogota Colombia. The meteorology is one component of a larger collaboration between the Universidad de La Salle and the University of Florida to prioritize regulatory strategies for Bogota using air pollution simulations. Accurate ranking relies on representative pollutant predictions, which require high quality meteorological predictions.

This project utilizes the Weather Research and Forecasting Model (WRF) for meteorological development for the Community Multi-scale Air Quality model (CMAQ). Various physics configurations have been tested based on previous work to determine which is most applicable. The configurations were analyzed using the Atmospheric Model Evaluation Tool (AMET). AMET was used to evaluate the model against Bogota's air quality monitoring system RMCAB. The configurations were checked to see if the thresholds for wind speed, wind direction, temperature and relative humidity were met. Detailed analysis of the model was performed against the RMCAB monitors that included meteorological data. Certain stations were weighted more heavily than others based on their locations and reports of poor air quality.

From the evaluation performed thus far, the model is adequate for regulatory purposes. However, the evaluation showed that wind direction error was consistently above the desired threshold, possibly due to the mountainous topography. This threshold however was set for a previous project in a different environment and is therefore arbitrary. Future work will involve the acquisition of more data for evaluation and to potentially be used for observational nudging.

Li Pan ^1,2, Pius Lee ¹, Hyun Cheol Kim ^1,2, , YouHua Tang^1,2, Daniel Tong ^1,2 ,Rick Saylor ³, Ivanka Stajner⁴, Weiwei Chen^1,5, Tianfeng Chai ^1,2 and Barry Baker¹

¹ NOAA/Air Resources Laboratory, College Park, MD

² UMD/Cooperative Institute for Climate and Satellites, College Park, MD

³ NOAA/ARL/Atmospheric Turbulence and Diffusion Division, Oak Ridge, TN

⁴ NOAA/NWS Office of Science and Technology, Silver Spring, MD

⁵ Northeast Institutes of Geography and Agroecology, Chinese Academy of Sciences, Changchun, China

Wildfires are a common on the North American continent during all times of the year, but predominantly occur during spring and summer months. They pose a significant risk to air quality and human health. Compared to anthropogenic emissions, emissions from wildfires are largely uncontrolled and unpredictable. Therefore, quantitatively describing wildfire emissions and their contributions to air pollution remains a substantial challenge for air quality forecasting efforts. In this study, we test the wildfire calculation algorithm used by the National Air Quality Forecasting Capability (NAQFC), which includes Hazard Mapping System (HMS) – wildfire detecting algorithm, BlueSky-wildfire emission estimation algorithm and SMOKE-wildfire plume rise algorithm, by comparison with ground, satellite and flight measurements during the Southeast Nexus (SENEX) field experiment. Southeast Nexus (SENEX) is a NOAA field study conducted in the Southeast US in the summer of 2013. To support this field campaign, NOAA/Air Resources Laboratory provided high horizontal grid resolution (4km) CMAQ model simulations covering the Southeast U.S and 12km resolution CMAQ simulations over the entire continental United States.  Fire emissions derived from HMS-BlueSky-SMOKE calculation were included in both CMAQ simulations. One focus of this analysis is to address if observed fire signals are reproduced by model simulations using this fire emission algorithm and to determine algorithm limitations.  The second focus is on quantitatively estimation of wildfire emissions and evaluation of model performance in an effort to improve NAQFC treatment of wildfire pollutant emissions. 

Roberto Perea¹, Duanjun Lu², Rosa M. Fitzgerald¹, Dongchul Kim³ and William R. Stockwell⁴

Modes are found in measured and modeled aerosol distributions and they illuminate processes affecting aerosol properties. However there has been much less examination of modes in tropospheric ozone distributions. In this study, the California region is used as a test-bed because ozone values can be high during episodes and because of the availability of ozone measurements. The Community Multi-scale Air Quality Model and the WRF-Chem models are used to perform ozone simulations. The objective of this study is to examine differences by modes in the measured and simulated ozone distributions. Although there are differences in the error and bias of the simulated ozone mixing ratios due to choices of mechanisms, boundary conditions, emissions and other factors, these differences can have a strong effect on the simulated ozone variability, distribution and modes. Furthermore the distributions of ozone mixing ratios are examined for both the measurements and the simulated ozone. These differences may show systematic problems in the chemical mechanisms for urban and regional air quality models. This study illustrates the potential utility of the examination modes in ozone data for the evaluation of air quality models.

Perry C. Shafran*, Jeffery McQueen, Jianping Huang*, Binbin Zhou*

NOAA/NWS/NCEP Environmental Modeling Center, College Park, MD

* I.M. Systems Group, Rockville, MD

Verification is essential in numerical modeling systems. Validation allows scientists to determine biases, monitor the model's execution, and find errors in various parts of the forecast. Rigorous verification is performed at that National Centers for Environmental Prediction's (NCEP's) Environmental Modeling Center (EMC).

Two air quality models are executed at NCEP. First, the Community Multi-Scale Air Quality (CMAQ) model, predicts, among other things, levels of ground-level ozone and particulate matter. CMAQ uses a combination of meteorology, chemistry, and emissions in order to predict levels of ozone and PM2.5.

The second model is the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model. This model predicts, among other things, levels of smoke and dust in the air and forecasts a parcel's trajectories.

The verification systems used validate the output for these items will be described in full. The CMAQ model uses a verification system against point observations, while the HYSPLIT is verified against gridded analyses. Also, the meteorological input for both of these models, the North American Model (NAM), also is verified with the point-observation verification system. Verification results for operational vs. parallels will be presented.

Heather Simon, Chris Misenis, Golam Sarwar, Norm Possiel

Multiple studies have documented over-prediction of ozone concentrations in coastal areas during period of low ozone concentration in regional photochemical modeling applications. These time periods often occur during periods of onshore transport when coastal ozone concentrations are predominantly impacted by marine ozone levels rather than local formation or continental transport. Some studies have suggested that this model over-prediction could be ameliorated with improved representation of marine ozone loss pathways including enhanced wet deposition due to surface reactions with marine iodine and gas-phase loss reactions with halogens originating from marine sources. In this work, we investigate a potential confounding factor, the accuracy of modeled advection and vertical mixing patterns. We use the hysplit tool to compare back trajectories calculated using WRF model output with those predicted using EDAS analyses. We use this analysis to more fully characterize any meteorological uncertainties that may contribute to this coastal ozone over-prediction.

Youhua Tang^1,2 (youhua.tang@noaa.gov), Tianfeng Chai^1,2 (tianfeng.chai@noaa.gov), Li Pan^1,2 (Li.Pan@noaa.gov), Pius Lee¹ (pius.lee@noaa.gov), Daniel Tong^1,2,3 (daniel.tong@noaa.gov ), and Hyun-Cheol Kim^1,2 (hyun.kim@noaa.gov)

1. NOAA Air Resources Laboratory, 5830 University Research Court, College Park, Maryland, MD 20740, USA

2. Cooperative Institute for Climate and Satellites, University of Maryland, College Park, Maryland, MD 20740, USA

3. Center for Spatial Information Science and Systems, George Mason University, Fairfax, Virginia, VA 22030, USA

A CMAQ modeling system with optimal interpolation method is built to make analysis for summer 2011 over continental USA in 12km horizontal resolution. A global model provided the dynamic aerosol boundary conditions for this regional study. The CMAQ model yields the first-guess field and the data assimilation method combines AIRNOW surface PM2.5 and Satellite total AOD (MODIS) to make correction in certain time interval. The assimilation can significantly improve the model's PM2.5 prediction, though its effect could fade out eventually depending on locations and events. So the assimilation has to be made every certain time, eg. 6 hours, to keep its effect. The issue of combining PM2.5 and MODIS AOD in the assimilation is also discussed. The assimilated product not only provides the product that is nearest to the realistic world, but also gives the correction factor etc which can be used for other further studies, like emission verification. The MODIS AOD data used in this study also help correct the influence of lateral boundary condition near the regional domain's border.

Marco Rodriguez¹, Courtney Taylor¹, Chao-Jung Chien¹, Caitlin Shaw¹, Tiffany Samuelson¹, Stephen Reid², Kenneth Craig², Josh Nall³, Leonard Herr⁴

¹AECOM Inc.

²Sonoma Technology, Inc.

³Wyoming Department of Environmental Quality

⁴Bureau of Land Management, Utah State Office

The Uinta Basin, in northeastern Utah, and the Upper Green River Basin (UGRB), in southwestern Wyoming, are areas where oil and gas activities are important and expected to continue development in the foreseeable future. Both areas have periods in the late winter and early spring with elevated 8-hour ozone concentrations. Given the complexity and emerging understanding of ozone formation during wintertime episodes, two regional air quality models were run to replicate the ozone events in each Basin: the Community Multi-Scale Air Quality (CMAQ) Model and the Comprehensive Air Quality Model with Extensions (CAMx) Model. Model-predicted concentrations of ozone and its precursors were compared to available surface and above surface observations. The model sensitivity to the parameterization of oil and gas emissions sources also was investigated, including the effect of temporalization and vertical distribution of oil and gas emissions sources. A model inter-comparison was completed to assess the model-performance differences within each basin. A subsequent inter-comparison of CMAQ's model performance between the Uinta Basin and the UGRB investigates the similarities and differences in the two airsheds.

Teng YAO¹, J.C.H. Fung^1,2,*, S.Y. Wang³, Y.J. Li¹, J.Z. Yu^1,3, A.K.H. Lau^1,4, C.K. CHAN^1,5

¹Division of Environment, Hong Kong University of Science & Technology, Hong Kong, China.

²Department of Mathematics, Hong Kong University of Science & Technology, Hong Kong, China

³Department of Chemistry, Hong Kong University of Science & Technology, Hong Kong, China.

⁴Department of Civil & Environmental Engineering, Hong Kong University of Science & Technology, HK, China.

⁵Department of Chemical & Biomolecular Engineering, Hong Kong University of Science & Technology, HK, China

Secondary organic aerosol (SOA) is one major pollutant in the Pearl River Delta (PRD) region of China. However during air quality studies it is largely under-estimated by Chemical Transport Models (CTMs), due to the oversimplified treatment of primary organic aerosol (POA) in CTMs. Based on dilution sampler's measurement, currently CTMs treat POA emitted from sources such as vehicles and biomass burning as nonvolatile. However recent studies show that POA is semi-volatile, and after POA is emitted into the ambient atmospshere, POA experiences dilution and evaporation that are missed in the CTMs. Subsequently chemistry reactions between gas-phase semi-volatile POA and oxidants are also overlooked. Thus one major reason for under-estimation of SOA formation is due to these missing physical and chemical processes of POA.

To evaluate contribution of SOA formation from oxidation of semi-volatile POA in the PRD region, we implemented Volatility Basis-Set (VBS) model into CAMx (Comprehensive Air Quality Model with Extends), assigning POA into 9 lumped species with different saturation concentrations spaning from 0.01 ug/m³ to 100 mg/m³. After POA emission is released to ambient atmosphere, the ratio between gas-phase and particle-phase is decided by the saturation concentration of POA and concentration of ambient sorptive matter. Gas-phase POA reacts with OH radical, becomes less volatile, and finally turns into SOA.

With VBS model implemented into CAMx, simulation result shows that CAMx with VBS model provides a better SOA simulation compared with the traditional CAMx with nonvolatile POA emission, which predicts the average SOA concentration around 2 ug/m³ over Pearl River Delta region, under-estimateing by 2-7 ug/m³ (around 60% - 90% of the observed SOA concentration). After the implementation, CAMx with VBS model is relieved from the SOA under-estimating problem. Moreover, simulation result also shows that CAMx with VBS model provides temporal variations of SOA concentration with better agreement with the high-resolution time-of-flight aerosol mass spectrometry (HR-ToF-AMS) observation.

Yadong Xu;

Dr. William Vizuete;

Dr. Mark Serre;

The uncertainty of local-scale ambient ozone concentrations has been an important factor that hinders our ability to study the associations between the exposure levels of ozone and a wide range of adverse health outcomes.   Accurate estimates of ambient ozone concentrations are needed for many epidemiological studies to obtain a better assignment of personal exposure.

Spatiotemporal estimation approaches based on only observational data recorded in the U.S. Environmental Protection Agency’s Air Quality System (EPA AQS), which were often used for many previous epidemiological studies, suffer from missing data issues due to the scarce monitoring network across space and the inconsistent recording periods at different monitors.  Although chemical transport models (CTMs) provide good spatial and temporal coverage for ambient ozone concentrations, direct use of the modeled outputs from CTMs can be problematic because of the substantial predictive bias in their model performances. 

This study investigates the geo-statistical model approach using the Bayesian Maximum Entropy (BME) Framework that can combine these techniques and take advantage of the strength from both data sources, the accuracy of the observational data and the good spatial /temporal coverage of CTM outputs. 

We first investigate the spatial heterogeneity and temporal variability of ozone model performances in the CTMs across the U.S.   We also investigate the methods to obtain soft information from CTM model outputs based on how well the CTM model predictions reproduce the observed values.  We demonstrate this approach in the United States, using hourly ozone measurements from EPA AQS network in combination with outputs from Community Multiscale Air Quality (CMAQ) Modeling System.  

Through mapping ozone concentrations visually and also quantitative assessment with cross-validation statistics, we demonstrate how the integration of soft information by the BME method can effectively increase the estimation accuracy.  The improved exposure estimates for ambient ozone concentrations can provide crucial information for studying associations between exposure to air pollution and adverse health effects in the United States.  

Hamish Hains, M.A.Sc., P.Eng.; Nick Walters, M.A.Sc., EIT.; Dr. Xin Qiu, Ph.D., ACM, EP

This research develops a method to evaluate the area effectively represented by a meteorological measurement station and quantify the impact of diminishing meteorological accuracy on air dispersion modelling.

Using the Weather Research and Forecasting Model (WRF), Toronto, Ontario, was modelled in 4km grid resolution for a 1-year period. Recommended model benchmarks were used to evaluate measured meteorological data against the WRF modelled results. Using WRF's gridded data, the area adequately represented by met stations typically used in regulatory applications was defined.

Dispersion modelling was performed using the AERMOD model using both measured (typical) and modelled (WRF) meteorological data for common source parameters in areas not adequately represented by a meteorological station. The difference between these modelling results was used to evaluate the sensitivity of the dispersion model to meteorological inputs and quantify the uncertainty in a model's results based on proximity to a meteorological data source.

Andy Hawkins¹, Dawn Froning², Chris Misinis³, James Thurman³, Roger Brode³, George Bridgers³

¹US EPA, Region 7, Lenexa, KS 66219

²Missouri Department of Natural Resources, Jefferson City, MO 65102

³US EPA, Office of Air Quality Planning & Standards, RTP NC 27711

The New Source Review (NSR) and Prevention of Significant Deterioration (PSD) programs require that new sources or existing sources with proposed modifications must demonstrate that additional emissions emitted to the atmosphere will not cause or contribute to a violation of the National Ambient Air Quality Standards (NAAQS). The US EPA's Guideline on Air Quality Models (Appendix W to 40 CFR Part 51) currently specifies AERMOD as the preferred model for projecting near-field dispersion of emissions for most NSR/PSD applications. One of the key modeling inputs to AERMOD is representative meteorological data which is generally derived either from National Weather Service (NWS) data or onsite meteorological data processed through AERMOD metrological preprocessor (AERMET).

The Mesoscale Model Interface Program (MMIF) is a program used to convert prognostic meteorological model output fields to the parameters and formats required for direct input into dispersion models such as AERMOD. MMIF derived AERMOD inputs could potentially be used in project areas where representative meteorological data are not available. In this study we investigate using MMIF derived inputs from the Weather Research and Forecasting (WRF) model and compare key meteorological parameters (WS, WD, etc.) against both onsite and NWS AERMET meteorological data, as well as investigate the consequences of using a MMIF derived meteorological dataset with AERMOD (i.e. Q-Q plots, other quantitative metrics).

Jianping, Huang^1,2*, Jeff McQueen², Perry Shafran^1,2, Pius Lee³, Daniel Tong^3,4, Li Pan^3,4, Youhua Tang^3,4 Sikchya Upadhayay^5,6, Geoff DiMego², and Ivanka Stajner⁶

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

²NOAA National Centers for Environmental Prediction, College Park, MD

3 NOAA Air Resources Laboratory, Silver Spring, MD

4 University of Maryland, College Park, MD

5 Syneren Technologies Corporation, Lanham, MD

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

The NOAA National Air Quality Forecasting Capability (NAQFC) provides operational forecast products of surface ozone and experimental products of surface particulate matter with diameter less than 2.5 m (PM_2.5) nationwide. Hourly meteorological inputs to the NAQFC are provided by NOAA NCEP regional operational weather forecasting model, the Non-hydrostatic Multi-scale Model on the Arakawa staggered B-grid (NMM-B) and the chemical fields (e.g., O₃ and PM_2.5) simulated by the Community Multiscale Air Quality (CMAQ) model. Previous evaluation of the NMMB-CMAQ forecasting system revealed consistent seasonal forecasting biases, including wintertime over-prediction and summertime under-prediction. The impact of meteorological inputs on PM_2.5 predictions still remains large uncertainty. In this study, two-month NMMB and CMAQ simulations in January (i.e., winter time) and July (i.e., summer time) 2014 are analyzed, with a focus on linkage CMAQ PM_2.5 biases with NMMB outputs. First, we present statistical evaluations of meteorological inputs such as planetary boundary layer height, cloud, solar radiation, wind, and air temperature, and their impacts on PM_2.5 predictions. Second, we present several sensitive studies on the impacts of later boundary conditions of NEMS (NOAA Environmental Modeling System) Global Aerosol Component (NGAC), fire/smoke emissions, and recent changes in NMM-B model on PM_2.5 predictions during several PM_2.5 episodes. Finally, possible suggestions are provided for further improvement of PM_2.5 predictions.

Eunhye Kim¹, Soontae Kim¹, and Hyun Cheol Kim ^2,3

¹ Ajou University, Dept. of Environmental Engineering, Suwon, Korea

² NOAA/Air Resources Laboratory, College Park, MD

³ UMD/Cooperative Institute for Climate and Satellites, College Park, MD

High-resolution spatial and temporal sea surface temperature (SST) data from polar-orbiting and geostationary satellites are utilized for regional air quality simulation in Korea. Korea has one of most complicated coastal lines and shallow bathymetry in the world, so coarse spatial and temporal SST easily fails to provide proper meteorological conditions, such as land-sea breeze and thermal forcing. Proper representation of local circulations is crucial in simulating local pollutant events, especially in high ozone episodes. We utilized Weather Research and Forecasting (WRF)-Community Multi-scale Air Quality (CMAQ) modeling framework with multi domains; 27-km East Asia, 9-km S. Korea, and 3-km Seoul Metropolitan Area (SMA). Three sensitivity simulations are conducted, using, (1) default SST inherited from National Oceanic and Atmospheric Administration (NOAA) Global Forecast System (GFS), which is used to initiate WRF, (2) Real-Time, Global SST high resolution (RTG_SST_HR) available from NOAA Environmental Modeling Center (EMC), and (3) RTG_SST_HR + diurnal variation from Multi-functional Transport Satellites (MTSAT) of Japan Meteorological Agency. The RTG_SST_HR has daily and 1/12 degree spatial resolution, and we also added diurnal variations in SST using 4-km, hourly MTSAT SST. Results show clear enhancement in ground and 2-m temperature simulation by using spatially- and temporally- enhanced SST data, resulting in considerable changes in surface ozone concentration, up to 10~20 ppb in the SMA during summer, 2013. Surface observations from Korea Meteorological Administration sites and AirKorea are utilized for the evaluation of meteorological and chemical simulations, respectively.

Awkash Kumar*

Research Scholar, Centre for Environmental Science and Engineering, Indian Institute of Technology, Bombay, Mumbai - 400 076, India,

Rashmi S. Patil

Emeritus Fellow, Centre for Environmental Science and Engineering, Indian Institute of Technology, Bombay, Mumbai - 400 076, India

Anil Kumar Dikshit

Professor, Centre for Environmental Science and Engineering, Indian Institute of Technology, Bombay, Mumbai - 400 076, India

Air pollution from vehicles is increasing rapidly in almost all cities around the world due to increase in population. Vehicular sources are directly proportional to population because everyone needs vehicles to travel. In this paper, a very interesting result has been obtained about vehicular pollution modeling for different time periods. Chembur is the most populated area in Mumbai city due to industrial and vehicular sources, has been selected for vehicular pollution modeling with period of one day, one month and one year respectively of 2011. AERMOD, which was developed by USEPA, was used as a tool of vehicular pollution modeling. It requires many kinds of input data like meteorological parameters, land use surface characteristics and source emission profiles. Generally, temporal and spatial interpolated meteorological data is used, which is collected from nearby meteorological station. In this paper, the Weather Research and Forecasting (WRF) model has been used to generate nine meteorological parameters which give real time on site meteorological data. This approach gives good result of traffic modeling. The modeling of six roads of Chembur has been performed. The results of AERMOD show the variation of pollution level varying with time period and season for the domain. The results show interesting behaviors of the model for different averaging time. In this study, WRF model has been simulated for a year successfully which saves enormous time and resource of collecting meteorological data from a station and this method gives accurate result.

Chris Misenis, James Thurman, Tyler Fox

Typically, most dispersion models use either site-specific meteorological data or nearby National Weather Service data available at airports to drive simulations. However, in some instances, site-specific data are cost prohibitive and the airport data fail to be representative of the modeling domain. Recent advances in prognostic meteorology lead to favorable comparisons of simulations to observational meteorological data. With that in mind, U.S. EPA is examining the possibility of allowing the use of prognostic meteorological data as a part of the Guideline on Air Quality Models (Appendix W to 40 CFR Part 51).

As a part of this process, the Meteorological Model Interface (MMIF) tool was developed to extract the necessary fields from prominent meteorological models (e.g., WRF) for use in dispersion models (e.g., AERMOD, SCICHEM, etc.). MMIF is used here to extract data at several locations across the U.S., from 12km WRF simulations conducted annually in 2007-08 and 2010-2012. The prognostic data are compared with observational data to assure the suitability of their use. The extracted data are then used as input to AERMOD simulations at the aforementioned locations. A comparison of the AERMOD results using observed meteorology and the prognostic meteorology is presented.

Kiran Alapaty¹, John S. Kain², Jerold A. Herwehe¹, O. Russell Bullock Jr., ¹

and Megan S. Mallard¹

¹National Exposure Research Laboratory, U.S. Environmental Protection Agency,

Research Triangle Park, North Carolina, USA.

²National Severe Storms Laboratory, National Oceanic Atmospheric Administration,

Norman, Oklahoma, USA.

The focus of our presentation is mainly on the development and testing of a seamless version of the Kain-Fritsch (KF) convection parameterization scheme that works across all spatial scales down to 1 km. To update the KF scheme for multiple scales in the WRF model, first, we proposed a scaling parameter to introduce horizontal scale-dependence in the KF scheme for use with various convection parameters, and then we developed new formulations for: (1) the convective adjustment timescale; (2) the entrainment of environmental air; (3) the fallout of condensates from updrafts; and (4) the stabilizing capacity. These scale-dependent formulations make the KF scheme operable at all scales down to about sub-kilometer grid resolution. Additionally, we have introduced methodologies for (5) impacting grid-scale vertical velocity with convective updrafts and downdrafts and (6) eliminating double counting of precipitation due to concurrent usage of grid-scale and subgrid-scale cloud formulations for any grid cell. Results obtained from our regional weather and climate simulations using the WRF model will be presented to demonstrate the effects of these science updates to the KF scheme.

K.R. Baker, U.S. Environmental Protection Agency

A.G. Carlton, Rutgers University

T.E. Kleindienst, U.S. Environmental Protection Agency

J.H. Offenberg, U.S. Environmental Protection Agency

M.R. Beaver, U.S. Environmental Protection Agency

J.T. Kelly, U.S. Environmental Protection Agency

M. Jaoui, Allion Science and Technology, Inc.

The State of California currently has numerous counties designated as non-attainment for the PM2.5 National Ambient Air Quality Standard. Organic carbon is one notable component of PM2.5 in California. Air quality modeling is typically needed to support the development of emissions control strategies. It is important that CMAQ and other photochemical models used for regulatory applications accurately characterize the atmospheric organic aerosol burden from precursor species to secondary organic aerosol formation so that emission control strategies elicit model response similar to the actual atmosphere. This enables accurate source attribution of atmospheric aerosol and development of effective control strategies to manage air quality.

Here we present an evaluation of both gas and aerosol carbon model estimates compared with observation data at two surface locations in California: Pasadena and Bakersfield. Ambient estimates of SOA based on measured chemical tracers and the fossil and modern composition of particulate carbon are presented and compared with model estimates to provide unique information about the nature of organic carbon at these locations in California. In addition, total VOC and speciated VOC measurements from the CalNex sites are compared with CMAQ estimates to provide a comprehensive overview of model performance of gas and aerosol carbon in California.

Xiangshang Li, Yunsoo Choi, and Beata Czader

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

Four WRF-SMOKE-CMAQ simulations for the DISCOVER-AQ 2013 Texas campaign period are performed to characterize the issues in the simulated meteorological and chemical conditions. These four simulations differ in meteorological inputs, domain setup, and in performing observation nudging. Our objective is to develop a best-effort WRF simulation as the basis for subsequent CMAQ modeling. Two retrospective WRF simulations based on NAYY input using objective analysis (OA) with observation nudring are carried to see how they improve over existing WRF forecast run with the standar NAYY input. Also, the WRF run including NAYY is compared with the WRF names "AQF" that uses NAM forecast data as input, which is used for an air quality forecasting. The two NAYY cases with observation nudging provide better temperature and wind statistics with around 9% index of agreement (IOA) gain in temperature and 10-14% boost in U-wind and V-wind. Interestingly, "AQF" simulate the surface ozone slightly better than other three cases as measured by IOA. Also, the impact of a nesting scheme (4km domain is nested inside of 12km domain) on the simulated meteorological fields near the center of the 4km WRF domain is marginal. This study show that the high ozone episodes during the project periods were associated with the uncertainties of the simulated cold front passage, chemical boundary condtion and small-scale temporal wind fields. Even with some irmprovements in the simulated meteorological fields by observation nudging performed every three hours, the simulations fail to capture small-scale wind shifts in the industrial zone that likely cause the observed high ozone exceedance, which results in the simulated underestimation of observed high ozone on September 25. This study also shows that overestimated background ozone from the southerly chemical boundary is a critical source for the model to provide the general overpredictions of the ozone concentrations from CMAQ during September 2013.

Matthew Woody, Kirk Baker, Havala Pye

Regional air quality models, including CMAQ, have traditionally underpredicted ambient organic aerosols (OA). In an effort to improve OA model predictions, the volatility basis set (VBS) was developed and provides an alternative to the traditional representation of OA in air quality models (non-volatile primary organic aerosols (POA) and secondary organic aerosols (SOA) modeled using the Odum 2-product model). The VBS framework considers semi-volatile POA, represented by semi-volatile organic compounds (SVOCs), and intermediate volatility organic compounds (IVOCs), gas phase species believed to be missing from most emission inventories which form SOA upon oxidation. In this study, we evaluate OA predictions from CMAQ v5.0.2 with VBS against measurements obtained during the CalNex-2010 Field Study. Model simulations are performed over California in May and June of 2010 using a 4-km horizontal grid resolution.

Jeffery T. McQueen, Geoffrey Manikin, Manuel Pondeca, Stan Benjamin, Geoffrey DiMego, Steven Levine, Matthew Pyle and Dana Carlis

The NWS National Centers for Environmental Prediction (NCEP) North American Model (NAM) downscaling (DNG) system developed by NOAA/Global System Division and NCEP/ Environmental Modeling Center (EMC) downscales meteorological model fields to a National Digital Forecast Database (NDFD) high resolution grid at either 5 or 2.5 km over the Continental U.S., Alaska, Hawaii, Puerto Rico and recently Guam. One use of the downscaled fields is to provide first-guess background grid for NCEP's Real-Time Mesoscale Analysis (RTMA) and Updated Real-time Mesoscale Analysis (URMA) where fields are further adjusted to the nearby observations using the two dimensional version of the NCEP variational Global Statisical Interpolation (GSI) assimilation system. NAM parent (12 km) hourly predictions are downscaled out through 84 hours whereas NAM nests (6-, 4- and 3-km) are downscaled through 60 hours. 2 m temperature and dewpoint temperature, 10 m winds and surface pressure are adjusted to a high resolution topographical database. Terrain effects are also introduced for winds while temperatures are also downscaled to the high resolution topography using similarity theory. Such diagnostic downscaled meteorological analyses and predictions that account for complex terrain and coastal effects are often used to drive air pollution models. This presentation overviews the NCEP mesoscale analysis and prediction systems (eg: NAM and its high resolution nests, RTMA) while reporting on improvements to the NCEP DNG especially for improved wind analyses and predictions. An evaluation of meteorological fields important to air quality modeling (eg: near surface winds, temperatures, moisture and boundary layer heights,cloud cover) will be reported on for the NAM, NAM nests and DNG predictions.

The DNG system also sharpens gradients of all fields around coastal areas using the land mask dataset. Basically, temps, winds, and dew point from the downscaled NAM land-water point are replaced by downscaled winds from the nearest neighbor point that agrees with the Analysis (NDFD) land-water point. Snow depth is then adjusted depending on the interpolated temperature. Chance of wetting rain, probability of precipitation from the NCEP Short Range Ensemble System (SREF) and mixed layer depth winds, temperature and humidity are also computed from the downscaled fields.

In the past year, the system has been unified form more direct application for a user defined grid domain as well as ensuring consistent downscaling approaches for all domains. Application of the DNG have been extended to downscale the NCEP Downscaled Global Extended (DGEX) to 192 forecast hours, the Global Forecast System (GFS) and the High Resolution Window (based on WRF and NEMS/NMMB models) models.

A more representative way to represent terrain effects on winds could be accomplished for the case when the NAM first guess topography lie below the output analysis topography. A spatially constant terrain scaling term could be replaced by using high resolution analysis land-use roughness (e.g.: De Rooy and KOK, 2004, W&F) or similarity theory (Troen and Mahrt, 1986). Recently, a diagnostic mass consistent wind downscaling scheme was tested to include terrain effects on winds. Winds are adjusted by computing terrain gradients and then velocity potential using the Poisson equation (Sherman, 1977; Calmet, Scire, 1986; Ratto, 1996). In this study, the impact of these approaches for improving the downscaled winds will also be presented and evaluated against observations.

Daniel Goldberg,Tim Vinciguerra, Samantha Carpenter, Linda Hembeck, Tim Canty, Ross Salawitch & Russell Dickerson

A comprehensive set of atmospheric trace gas observations is available for July 2011 in the Baltimore-Washington metropolitan region as part of NASA's DISCOVER-AQ air quality campaign. Observations of trace gases during the campaign include: ozone, volatile organic compounds (VOCs), reactive nitrogen species (NOx and NOy), carbon monoxide, and methane, among other trace gas species. This dataset is extremely valuable due to the high temporal resolution of observations over a concentrated area. Observations of trace gases have been compared to a 12-km simulation of the Comprehensive Air-Quality Model with Extensions (CAMx) v6.10. Using the ozone source apportionment tool (OSAT) in CAMx, we can begin to grasp at how much ozone is generated locally versus transported from upwind locations. Initial results show that up to 70% of the surface ozone in Maryland during the summer can be attributed to pollution from outside of the state's borders. Modifications to the model, supported by literature recommendations and which improve agreement with DISCOVER-AQ observations, can further increase this percentage. Findings show that surface ozone can no longer be identified as a pollutant generated by only local emissions.

Rui Zhang¹, Daniel S. Cohan¹, Arastoo Pour Biazar², and Erin Chavez-Figueroa¹

¹Department of Civil and Environmental Engineering, Rice University, Houston, TX

²The National Space Science Technology Center, University of Alabama in Huntsville, Huntsville, AL

Biogenic volatile organic compounds (BVOC) comprise approximately 75%-80% of national VOC emission inventory and can affect regional and urban air quality by contributing to ozone and particulate matter (PM) formations. The current BVOC emission models (e.g. MEGAN) have large uncertainty to gauge the amount of insolation reaching the canopy (i.e. Photosynthetically Active Radiation (PAR)) due to the frequent underestimation of grid resolved clouds in meteorological models. In this study, PAR estimates were generated by scaling the principle insolation retrievals from high-resolution Geostationary Operational Environmental Satellite (GOES) imager with the consideration of water vapor, total overhead ozone and aerosol optical depth. The PAR estimates were evaluated using available ground radiation network data and show much stronger correlations than a meteorological model (WRF) with observations in terms of temporal variations as well as spatial patterns.

The impact of biogenic emission estimates using satellite PAR on air quality modeling was quantified by a WRF-MEGAN-CMAQ simulation platform over the southern US with 4km horizontal resolution. The test simulation period coincides with the NASA DISCOVER AQ campaign in September 2013 at Houston Texas, where abundant meteorological, gaseous and aerosol pollutants measurements are available for comparisons. Using satellite PAR, the modeled isoprene and monoterpene emission rate over southern US during that simulation period were on average 10-20% less than that derived from the WRF model. The reduction occurs because the satellite data corrects the PAR over-prediction bias of WRF, which tends to under-predict cloud cover. Isoprene emissions tend to have a stronger response over the whole simulation domain to the reduced PAR from satellite retrievals than monoterpene emissions. The response of ground ozone and PM simulation due to the change of biogenic emission estimates were also evaluated using the observational data.

Daiwen Kang, Wyat Appel, Pat Dolwick, Norm Possiel, Shawn Roselle, James Godowitch, Jon Pleim, and Rohit Mathur

The Deriving Information on Surface Conditions from COlumn and VERtically Resolved Observations Relevant to Air Quality (DISCOVER-AQ) field campaigns have collected and maintained integrated and diverse data sets to understand the diagnosis of air quality conditions from surface to above. These data sets provide perfect opportunity to thoroughly evaluate air quality models for their performance for simulating air pollutants (such as ozone) and their precursors in space and time. However, it is also challenging on how to effectively use the data from different measurement platforms with different spatial and temporal resolutions in model evaluations both collectively and individually in the most appropriate way. In this study, we explore different ways to effectively utilize the rich data sets in the application to model evaluations. The data sets to be used include the aircraft measurements, ozone sonde data, ship measurements, and other available ground and upper air measurements. To facilitate model evaluation, programs are being developed to incorporate measurements from different platforms in different data formats and model outputs into one central venue to generate various statistics and graphics using the statistical R language. These procedures were utilized in separate evaluations of the CAMx and CMAQ modeling during the DISCOVER-AQ campaign at Baltimore-Washington, D.C. area in July, 2011 and are summarized in two separate presentations.

C. Nolte, T. Spero, N. Fann, S. Anenberg, P. Dolwick, S. Phillips

The U.S. Global Climate Research Program Interagency Workgroup on Climate Change and Human Health has identified understanding the potential climate impacts on human health as a high priority topic for a Special Report (SR) within the overall National Climate Assessment (NCA). The NCA-SR will be a peer-reviewed assessment of existing research on the impacts of observed and projected climate change on human health in the United States, with a strong focus on impact quantification. As part of a case study analysis to estimate potential climate impacts on U.S. air quality, we examine changes in seasonal air quality over the U.S. in the 2030 timeframe using CMAQ driven by regional climate fields from WRF. The WRF regional climate fields are created by downscaling data from two global climate models driven by two different climate forcing scenarios (i.e., RCP 6.0 and RCP 8.5) for the 11-year periods 1995-2005 and 2025-2035. The resultant air quality outputs from this analysis are used with BenMAP and projections of future population distribution from the Integrated Climate and Land Use Scenarios (ICLUS) project to estimate human health impacts attributable to climate change. Preliminary results indicate increases of up to 3.5 K in daily maximum temperature and up to 4 ppb in May-September average daily maximum 8-h ozone levels across most of the central and northeastern U.S., though there is substantial variability both between models and interannually for a particular model. Implications of these changes in air quality on mortality will also be shown.

Pat Dolwick, Daiwen Kang, Sharon Phillips, Norm Possiel, Heather Simon, and Brian Timin

Model performance evaluations are often limited by the amount of ambient data available for comparison against the simulated pollutant levels. The 2011 DISCOVER-AQ field study in the mid-Atlantic region employed a combination of aircraft, sondes, and ships to collect a robust, three-dimensional data set of ozone and ozone precursors in the Baltimore-Washington region during July 2011. This analysis compares the results of a 2011 base year CAMx 12-km model simulation against the field study data to investigate the ability of the model to properly simulate: a) the daily exchange of ozone between the residual layer and the surface boundary layer, and b) the local-scale ozone concentrations in and near the sea breeze influence of the Chesapeake Bay.

Uma Shankar, Jeffrey Prestemon, Aijun Xiu, Bok Haeng Baek, Kevin Talgo, Mohammad Omary and Dongmei Yang

Under the Forest and Rangeland Renewable Resources Planning Act (RPA) of 1974, the Forest Service and other federal agencies are required to generate national RPA Assessment reports on a ten-year cycle, describing projected conditions of the US forest and rangeland in the next 50 years. In support of this goal, analyses were conducted to project the impacts of climate change on annual areas burned (AAB) using global climate model projections of fire weather parameters corresponding to the latest projection time frame, 2010-2060. The climate data were remapped to a Southeastern U.S. modeling domain to develop statistical models of fire activity over this time frame. These theoretically based models recognize changes in driving climate variables as well as those in biophysical, socioeconomic, and land use variables that have been shown to explain historical variations of wildfires in time and space. These models have been applied to yield corresponding AAB projections, gridded over the spatial domain, at 5-year intervals from 2015 to 2060. The AAB projections have been used in a stochastic model to constrain estimates of daily burned areas for the eventual estimation of wildfire emissions for use in air quality assessments in selected years. The impacts of wildfires on ozone air quality, and the mass loadings of radiatively important aerosol species (organic and elemental carbon, sulfate and nitrate) are being examined through CMAQ simulations in these annual periods. Results are presented for available years on the ambient concentrations of aerosols and ozone, along with indicators of the direct radiative impacts of aerosols emitted in wildfires.

Melanie B. Follette-Cook (MSU/GESTAR), Christopher P. Loughner (ESSIC-UMD), Kenneth Pickering (NASA GSFC)

During the month of September, 2013 the Deriving Information on Surface conditions from Column and Vertically Resolved Observations Relevant to Air Quality (DISCOVER-AQ) campaign used two aircraft in tandem with an extensive suite of ground-based in-situ and remote sensing instruments to measure trace gases and aerosols over the Houston area. Ten flights and over 200 profiles of trace gases and aerosols were completed over the course of the month. The P-3B, one of two aircraft deployed, contained several in-situ sampling instruments. Nitrogen dioxide was measured by two different techniques, Laser Induced Fluorescence (TD-LIF) and 4-channel chemiluminescence, which also measured ozone. Formaldehyde was measured by the Difference Frequency Generation Absorption Spectroscopy (DFGAS) technique. We performed a month-long CMAQ simulation to compare with this unique dataset. The variety of meteorological flight conditions, air quality conditions, consistent flight patterns, and large sample make the DISCOVER-AQ dataset ideal for model evaluation.

Ivanka Stajner (1), Jeff McQueen(2), Pius Lee(3), Roland Draxler(3), Jinaping Huang (2,4), Perry Shafran (2,4), Daniel Tong (3,5), Li Pan (3,5),Phil Dickerson(6), and Sikchya Upadhayay (1,7)

(1) NOAA NWS/OST

(2) NOAA NWS/NCEP

(3) NOAA ARL

(4) IMSG

(5) Cooperative Institute for Climate and Satellite, University of Maryland

(6) EPA

(7) Syneren Technologies

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 surface ozone, surface smoke, and vertically integrated smoke from wildfires, as well as operational predictions of surface dust and vertically integrated dust for the contiguous 48 states (available at http://airquality.weather.gov). Predictions are produced beyond midnight of the following day at 12 km resolution and 1 hour time intervals and they are distributed in numerical and graphical formats.

Ozone predictions 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 emission estimates from the EPA and chemical processes within the Community Multiscale Air Quality (CMAQ) model. Predictions of wildfire smoke and dust storms, 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.

Developmental testing of aerosol predictions with NEI inputs using CB05 chemical mechanism and AERO-4 aerosol modules continues to show seasonal biases - overprediction in the winter and underprediction in the summer. Experimental ozone testing includes three updates in the CB05 system: lateral boundary conditions, dry deposition processes, and limiting the minimum planetary boundary layer height. A major emission update has been implemented since the summer of 2012, which relies on projected changes in emissions from mobile sources, reducing NOx emissions and ozone overestimation, especially in the eastern US.

Dust predictions were implemented in 2012, followed by an implementation of a longer time step to reduce run time for those predictions under a heavy dust loading. With the transition to a new computing platform in 2013, smoke predictions were updated to include an increase in maximum smoke plume rise (up to 1.25 times the planetary boundary layer depth), a decrease in wet removal of particles, and a modification in daily cycling of smoke emissions. Impacts of an automated algorithm for detection of locations of wildfire smoke emissions in Canada and Mexico are being evaluated in smoke predictions.

Recent testing has been focused on transitioning CB05 based system into operations for ozone predictions. Testing of PM2.5 predictions continues with efforts to include intermittent sources of smoke and dust together with an effort to develop linkages between national AQ predictions and global atmospheric composition predictions, as resources allow.

Michael Moran, Sylvie Gravel, Radenko Pavlovic, Paul Makar, Craig Stroud, Sylvain Menard, Paul-Andre Beaulieu, Balbir Pabla, Junhua Zhang, Qiong Zheng, Philip Cheung, and Samuel Gilbert

GEM-MACH is an on-line chemical transport model that is embedded within GEM, Environment Canada's multi-scale operational weather forecast model. A limited-area configuration of GEM-MACH has been used as Environment Canada's operational regional air quality forecast model since November 2009. The current operational version, GEM-MACH10, is run twice daily to produce 48-hour forecasts of hourly O3, PM2.5, and NO2 fields on a North American grid with 10-km horizontal grid spacing, 80 vertical levels from the surface up to 0.1 hPa, and a 300 s time step. This past summer an experimental version of GEM-MACH10 with near-real-time wildfire emissions was also run to predict wildfire impacts on North American air quality.

In addition, two other experimental GEM-MACH configurations are being used to make real-time air quality forecasts at higher spatial resolution (2.5-km horizontal grid spacing) for other applications. The first configuration is centered over northeastern Alberta and has been run since late 2012 to support air quality field studies in the Alberta Oil Sands region. The second configuration is centered over southern Ontario and is being developed to provide high-spatial-resolution air quality forecasts for the 2015 Pan Am and Parapan Am Games, which will take place in the Toronto area in summer 2015. This talk will provide an overview of these four different real-time applications of GEM-MACH.

Chris Emery, Sue Kemball-Cook, Jaegun Jung, Ed Tai, Greg Yarwood

ENVIRON International Corporation

Bright Dornblaser

Texas Commission on Environmental Quality

As the National Ambient Air Quality Standard (NAAQS) for ozone becomes more stringent, understanding regional ozone transport becomes increasingly important. Because their lifetimes are relatively long in the upper troposphere, ozone and a substantial fraction of total oxidized nitrogen (NOy) can be transported for long distances and potentially mixed downward into the planetary boundary layer, where they can influence surface ozone concentrations. It is therefore necessary that regional air quality models used for ozone air quality planning accurately simulate the transport and fate of ozone and NOy in the upper troposphere and lower stratosphere. We describe a set of modeling analyses using the Comprehensive Air quality Model with extensions (CAMx) to improve our understanding of the transport of stratospheric ozone and NOy into the troposphere and its effect on surface ozone in Texas during the summer of 2006.

Leading up to this work, we found that CAMx-modeled NOy in the free troposphere above 8 km was under estimated relative to regionally-averaged aircraft data from the INTEX-A field experiment. A similar low bias for NOx in the upper troposphere occurs in other global and regional models, and the addition of aircraft and lightning NOx emissions reduces this bias but does not eliminate it. Sensitivity tests performed with CAMx showed that stratosphere-to-troposphere transport is an important source of upper tropospheric NOy and ozone that must be represented correctly in order to accurately model their respective vertically integrated column masses, which is critical for model/satellite intercomparison. In this project, we improved the CAMx simulation of the upper troposphere by implementing top boundary conditions derived from a global model, so that explicitly defined stratospheric concentrations of NOy and ozone are advected into the top layer of CAMx. Results were compared against routine ozonesonde data available from four launch sites in the continental U.S. We used the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) and CAMx models in tandem to study the origin and fate of air transported from the stratosphere to the troposphere. Sensitivity tests evaluated the effect of layer collapsing versus using all layers available from the driving meteorological model.

Hyun Cheol Kim ^1,2, Pius Lee¹, Soontae Kim³, Changhan Bae³, Byeong-Uk Kim⁴, and Eunhye Kim³

¹ NOAA/Air Resources Laboratory, College Park, MD

² UMD/Cooperative Institute for Climate and Satellites, College Park, MD

³ Ajou University, Dept. of Environmental Engineering, Suwon, Korea

⁴ Georgia Environmental Protection Division, Atlanta, GA

In this study, we present that wintertime pollutant transport patterns in East Asia are visible from multiple satellite observations when inspected with corresponding synoptic weather analysis charts. Transport pathways of pollutants and anthropogenic emissions in East Asia are investigated using satellite images, surface weather chart, and chemical transport model simulation in the context of conceptual categorization of synoptic weather pattern. We combined daily distributions of Moderate Resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Depth (AOD) and Community Multi-scale Air Quality (CMAQ) simulated particulate matter (PM) to represent aerosol distribution; and Global Ozone Monitoring Experiment 2 (GOME-2) and Ozone Monitoring Instrument (OMI) NO₂ column density as a proxy for fresh anthropogenic emission flux; and Korean Meteorological Administration surface weather analysis chart to understand synoptic weather pattern using GIS geo-referencing technique. We identified a periodic extension of the Siberian high to the south China and its associated migratory systems are important to understand transport patterns in this region by investigating a two year data between June 2012 and May 2014 of synoptic weather chart and clustering information from previous researches. Based on the relative location and strength of high pressure system over south China and low pressure system over Manchuria, we classified three types of synoptic pattern that might affect high surface PM events in China and Korea. Classified patterns include (1) Expansion of Siberian high as a result of cold surge, (2) Cold front passage associated with migratory northern low pressure system, and (3) Stagnant high pressure system near Yellow Sea. In all cases, the development of high pressure system in south China is essential for development of pollutant event associated with favorable meteorological conditions. We demonstrated that observed and simulated surface PM show very good agreement, not only with MODIS AOD but also with NO₂ column density, implying the possible contributions of transported anthropogenic emissions to regional high surface PM events. Particularly, in the type 2 pattern, we demonstrate unique narrow-band-shaped high PM plumes are well reproduced in CMAQ, MODIS AOD and NO₂ column density observations, showing excellent performances of current modeling system. We also demonstrate these PM plumes are originated from northeastern China, pushed southward by cold front passage. All 3 types of transport patterns are shown to be important in regional air quality and global meteorology, in terms of intensity (e.g. type 3), frequency (e.g. type 1 & 2), and vertical lifting (e.g. type 2). We conclude the general placement of synoptic patterns in East Asia is strongly favorable for the transport of emissions and pollutants from China to Korea and Japan. Synoptic flow patterns work as a very efficient pumping machine, associated with periodic Rossby wave propagation and baroclinic instability, to shift East Asian anthropogenic emission into Pacific, resulting in direct and indirect impact on global scale meteorology and climate change.

Chuen-Meei Gan¹, Jonathan Pleim¹, Rohit Mathur¹, Christian Hogrefe¹, Charles N. Long², Jia Xing¹, David Wong¹, Shawn Roselle¹ and Chao Wei¹

(1) Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, US Environmental Protection Agency, Research Triangle Park, North Carolina, USA

(2) Climate Physics Group, Pacific Northwest National Laboratory, Richland, Washington, USA

Multi-decadal simulations with the coupled WRF-CMAQ model have been conducted to systematically investigate the changes in anthropogenic emissions of SO₂ and NO_x over the past 21 years (1990-2010) across the United States (US), their impacts on anthropogenic aerosol loading over North America, and subsequent impacts on regional radiation budgets. In particular, this study attempts to determine the consequence of the changes in tropospheric aerosol burden arising from substantial reductions in emissions of SO₂ and NO_x associated with control measures under the Clean Air Act especially on trends in solar radiation. Extensive analysis of available observations over the past two decades (Gan et al. 2014), indicate a shortwave radiation "brightening" over the past 16 years (1995-2010) in the US, particularly in the east. The relationship of the radiation brightening trend to decreases the aerosol burden is less apparent at the western US. Model and onservation comparisons for the annual trends and seasonal trends of AOD, aerosol concentration and radiation will be discussed. Discrepancies between trends inferred from the observations and models will be presented and possible causes for these differences will be discussed.

Reference

Gan, C.-M., Pleim, J., Mathur, R., Hogrefe, C., Long, C.N., Xing, J., Roselle, S., and Wei, C.: Assessment of the effect of air pollution controls on trends in shortwave radiation over the United States from 1995 through 2010 from multiple observation networks, Atmos. Chem. Phys., 14, 1701-1715, doi:10.5194/acp-14-1701-2014, 2014.

Chih-Yuan Chang, Eric Faust, Xiangting Hou, and Kuo-Jen Liao (kuo-jen.liao@tamuk.edu)

Department of Environmental Engineering, Texas A&M University-Kingsville

Lucas Henneman, David Lavoue, Armistead Russell

Implementing air pollution controls is costly. Given these costs, it is of interest to investigate the efficacy of regulations that require such controls in terms of providing the expected air quality benefits. This research, which is supported by the Health Effects Institute, investigates the effectiveness of these controls by examining the relationships between air pollution regulations, emission levels, and ambient air quality concentrations in Atlanta, Georgia.

Daily observations of ozone and particulate matter are from the SEARCH (SouthEastern Aerosol Research and Characterization) network's Jefferson Street monitoring station in downtown Atlanta (33.78° N, 84.42°W) from 2000 to 2012. Concentrations are meteorologically detrended using nonlinear filters and a statistical model to account for fluctuations in air pollutant levels associated with meteorological variability. Detrended concentrations are then related to emissions from on-road vehicles and power plants using a multiple linear regression model. The model includes terms that account for interactions between emissions and pollutant levels, and between emissions and meteorological variables. Regression coefficients are combined into sensitivities, which are compared to sensitivities modeled using CMAQ-DDM (the Community Multiscale Air Quality Model with the Decoupled Direct Method). Both the empirically derived and DDM-calculated sensitivities of ozone to anthropogenic NO_x emissions show dependence on ozone concentrations. However, empirically derived sensitivities have a higher slope.

Sensitivities are combined with estimates of emissions in the absence of controls to create counterfactual time series of air pollution concentrations. A comparison of counterfactual ozone with observed ozone shows that controls have had little effect on annual medians over the study period. However, the annual distribution has narrowed, with improvements seen in the upper range of ozone concentrations and increases in concentrations in the lower range. This is likely due to decreased NO_x on lower ozone days to react with ozone, and signifies a shift to the air in Atlanta being radical-limited on these days.

This study produces both empirical and model-based methods to estimate relationships between emissions, meteorology, and air quality, along with the associated uncertainties. Results will allow for the assessment of the health impacts of current and may aid the development of future air pollution regulatory programs.

Shaocai Yu¹, Qingyu Zhang¹, Pengfei Li¹, Bixin Chen¹, Yanqun Li¹, Weiping Liu¹, David Wong², Kiran Alapaty², Jon Pleim² and Rohit Mathur²

¹ Research Center for Air Pollution and Health, College of Environmental and Natural Resources, Zhejiang University, Hangzhou, Zhejiang 310058, P.R. China.

²Atmospheric Modeling and Analysis Division, National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA

Over the past three decades, the megacity areas in China has suffered from air pollution and heavy haze because of its decades-long burst of economic growth and rapidly expanding clout as an industrial giant. The Beijing-Tianjin-Hebei, Yangtze Delta and Pearl River Delta regions are three key areas with heavy haze pollution in China. In this study, impact of the thermal power industries from the Beijing-Tianjin-Hebei regions on Beijing haze formation in 2013 was studied by the newly-developed two-way coupled WRF-CMAQ model. The newly-developed two-way coupled WRF-CMAQ model has implemented indirect aerosol effects on grid-scale clouds by including parameterizations for both cloud droplet and ice number concentrations calculated from the CMAQ-predicted aerosol particles. The model simulations are carried out over the eastern China (12 km resolution) and Beijing-Tianjin-Hebei (4 km resolution) domains for 2013. Evaluations of model performance on PM_2.5, PM₁₀, O₃, SO₂, NO₂, CO, AQI and aerosol optical depth (AOD) are carried out by comparing to satellite observation data such as MODIS and surface monitoring networks over the eastern China.

Prakash Karamchandani¹, Jeremiah Johnson¹, Tejas Shah¹, Jaegun Jung¹, Susan Collet², Toru Kidokoro³, and Yukio Kinugasa³

¹ENVIRON, Novato, CA, USA

²Toyota Motor Engineering and Manufacturing North America, Inc., Ann Arbor, MI, USA

³Toyota Motor Corporation, Shizuoka, Japan

Source attributions and sensitivities to precursor emissions are of interest to policymakers considering emissions control strategies to meet current and future air quality regulations for ozone and fine particulates (PM_2.5). Photochemical modeling remains the best option to address these questions. Currently available regional air quality models that are widely used for regulatory applications, such as CMAQ and CAMx, have the capability to calculate the contributions of various source categories to air quality, either by using brute force methods, or by using available probing tools built into the model. This paper presents and compares the results of a variety of methods to calculate source culpabilities for current and future year ozone and PM_2.5 concentrations from global boundary conditions as well as from a number of anthropogenic and natural source categories, such as on-road mobile sources, off-road mobile sources, area sources, point sources, and wildfires and biogenic sources. The methods compared include brute force methods with both CMAQ and CAMx, as well as the CAMx source attribution tools (OSAT and PSAT) for ozone and PM_2.5 and the CAMx source sensitivity tool, HDDM, for ozone. In addition to looking at these broad source categories, the contributions of sub-categories of on-road mobile sources, such as light duty gasoline vehicles, light duty diesel vehicles, heavy duty gasoline vehicles, and heavy-duty diesel vehicles are also estimated to determine the relative importance of these sub-categories. A nested grid configuration is used for the simulations, with a 36 km grid covering the continental US, and two 12 km grids covering portions of the western and eastern United States. The simulations are conducted for a summer month and a winter month. The global boundary conditions are obtained from a global chemical transport model. Source contributions to future year design value concentrations in selected metropolitan areas are calculated as well as future year population exposures in non-attainment areas.

Kristen M. Foley, Christian Hogrefe, and Shawn J. Roselle

U.S. EPA

Dynamic model evaluation is concerned with comparing observed and modeled changes in air quality that are caused by changes in emissions and/or meteorology. In a number of previous studies, the focus was on evaluating the modeling system's response to emission reductions, and the approach taken to accomplish this goal was to select time periods characterized by large and quantifiable changes in emissions but similar meteorological conditions. In particular, many of these studies focused on 2002 and 2005 to perform dynamic model evaluation and often found that model predicted changes were smaller than observed changes, implying that the modeling system's response to emission reductions may be too weak if one were to make the assumption that the change between the two years is largely driven by well-quantified emission reductions. In this study, we present two approaches for better quantifying and accounting for the effects of interannual meteorological variability in dynamic evaluation studies. The first approach is based on performing model sensitivity simulations that isolate the effects of emission changes on simulated pollutant concentrations from the effects of meteorological changes by holding either emissions or meteorology constant. While no real world scenario exists against which these sensitivity simulations can be evaluated, statistical methods exist to estimate the effects of meteorology on ambient concentrations and we will utilize these methods in our analysis. The second approach is based on performing and analyzing long-term (1990 - 2010) air quality simulations driven by year specific meteorology and emissions. These simulations allow us to compare observed and simulated trends in air quality and to determine the robustness of dynamic evaluation studies that rely on individual years rather than multiple years. Results indicate that the robustness of dynamic evaluation studies aimed at evaluating the modeling system's response to emission changes can be improved by considering the average concentrations over several years before and after the emission change of interest. In addition, the results also reveal that the long-term model simulations analyzed in this study generally capture the trends in observed summertime ozone concentrations.

Patrick Campbell, Yang Zhang, Khairunnisa Yahya, Kai Wang, Christian Hogrefe, George Pouliot, Christoph Knote, Alma Hodzic, Roberto San Jose, Juan L. Perez, Pedro Jimenez Guerrero, Rocio Baro, and Paul Makar

Recent increases in oil and natural gas production throughout the western U.S. have come with scientific and public interest in emission rates, air quality and climate impacts related to the industry. We use a fully coupled meteorology-chemistry model WRF-Chem to simulate high ozone episodes during the winter of 2013 over the Uinta Basin in northeastern Utah, which is densely populated by thousands of oil and natural gas wells. The high resolution WRF simulations are able to qualitatively reproduce the wintertime cold pool conditions that occurred in 2013, allowing the model to reproduce the observed multi-day buildup of atmospheric. Two different emission scenarios for the oil/gas sector were employed in this study. The first emission scenario was based on the U.S. EPA National Emission Inventory (2011, version 1) for the oil/sector for the Uinta Basin. The second emission scenario (top-down) was based on the atmospheric measurements in the Uinta Basin. WRF-Chem simulations using the two emission datasets resulted in significant differences for concentrations of most gas-phase species. Comparison of simulations using the two emission datasets reveals that the top-down case captures the high O3 episodes. In contrast, the simulation case using the NEI-2011 inventory is not able to reproduce any of the observed high ozone concentrations in the Uinta Basin. A sensitivity analysis reveals that the major factors driving high ozone in the Uinta Basin in winter are shallow boundary layers with light winds, high emissions of volatile organic compounds compared to nitrogen oxides emissions from the oil and natural gas operations, and enhancement of photolysis fluxes and reduction of O3 loss from deposition due to snow cover. The model results show a disproportionate contribution of aromatic compounds to ozone formation relative to all other hydrocarbon emissions. We also present modeling results for winter of 2012, when high ozone levels were not observed in the Uinta Basin.

The Atmospheric Model Evaluation (AMET) is a leading software tool being used for evaluating air quality and meteorological models such as MM5, WRF and CMAQ. AMET relies on three software components - MySQL, R and perl. Use of AMET involves 2 steps - a pairing step where model predictions are paired with observations at monitoring locations, and an analysis step where the paired data (in time and space) are then used to generate multiple measures of model performance - both graphical and statistical. While AMET has very powerful analyses capabilities overall, the user community has reported a fairly steep learning curve associated with using AMET. Bulk of the challenges has been associated with ensuring that the appropriate software components are installed, and ensuring that interdependencies between software versions are addressed. Few challenges have also been reported related to the heterogeneous nature of observation networks (specifically for air quality), where not all pollutants are measured at all networks, and the nuances associated with systematically ensuring that all possible observations available for a given domain are used in getting a comprehensive model evaluation. To address these challenges, a new web-interface to AMET has been prototyped, where users can perform one of two steps: a) load CMAQ outputs through the web-interface to then populate the MySQL database, and then interactively select pollutants/networks/analyses to be performed, and b) rely on pre-loaded paired obs-model pairs using a Google Earth enabled map interface to assess model performance at specific sites or group of sites. The former uses canned AMET analyses scripts on the server side, while the latter uses dynamic client-side plotting capabilities, and does not use AMET analyses scripts. Key advantages of the latter option to model performance evaluation is that once the CMAQ simulations are complete and the obs-model pairs are loaded in the MySQL database using existing tools, users on the web can instantly perform explorative model evaluation for their choice of monitoring locations in an interactive manner. Possible extensions to this prototype are to provide capabilities for meteorological model performance, and to bring in corresponding emissions data used in CMAQ for concurrent analyses of emissions, meteorology and air quality fields. We will present this prototype, and discuss steps necessary and challenges with making this a potential community resource for the CMAS users community.

Lakshmi Pradeepa Vennam¹, Saravanan Arunachalam¹, Jared Bowden¹, B.H Baek¹, Mohammad Omary¹, William Vizuete¹, Seth Olsen²

¹University of North Carolina, Chapel Hill, NC

²University of Illinois at Urbana-Champaign, IL

Understanding the present-day impacts of aircraft emissions on surface air quality is essential to plan potential mitigation policies for future growth. Stringent regulation on mobile source-related emissions in the recent past coupled with anticipated rise in the growth in aviation activity can increase the relative impacts of aviation-attributable surface air quality if adequate measures for reducing aviation emissions are not implemented. Though aircraft emissions during in-flight mode (at upper altitudes) contribute a significant (70 - 80%) proportion of the total aviation emissions, landing and takeoff (LTO) related emissions can have immediate impact on surface air quality, as most of the large airports are located in urban areas, specifically those that are designated in nonattainment for O₃ and/or PM_2.5. In this study, we modeled impacts of aircraft emissions during LTO cycles on surface air quality using the latest version of the CMAQ model for two contemporary years (2005, 2010) and one future year (2018). For this regional scale modeling study, we used highly resolved aircraft emissions from the FAA's Aviation Environmental Design Tool (AEDT), meteorology from NASA's Modern-Era Retrospective Analysis for Research and Applications (MEYYA) downscaled with the WRF model, dynamically varying chemical boundary conditions from the CAM-Chem global model (which also used the same AEDT emissions but at the global scale), and spatio-temporally resolved lightning NOx emissions estimated using National Lightning Detection Network (NLDN) flash density data. We evaluated our model results with air quality observations from surface-based networks and in-situ aircraft observation data for the contemporary years. We will present results from model evaluation using this enhanced modeling system, as well as the trajectories in aviation-related air quality (focusing on O₃, NO₂ and PM_2.5) for the three modeling years considered in this study. These findings will help plan potential strategies to be considered for overall reduction in aviation-related air quality and health impacts on U.S. wide basis in future years.

Yu Morino, Tatsuya Nagashima, Seiji Sugata, Kei Sato, Kiyoshi Tanabe, Akinori Takami, Hiroshi Tanimoto, and Toshimasa Ohara (National Institute for Environmental Studies)

Model evaluation of PM2.5 species had been spatially and temporally limited in Japan because of limitation in observational data. Recently, simultaneous measurements of PM2.5 species were conducted. Using the observational data, we evaluated model performance of PM2.5 species throughout Japan in winter, spring, and summer in 2012.

We used three simulation models: CMAQ v4.7.1 with SAPRC99-AERO5 module (AERO5), CMAQ v5.0.2 with CB05-AERO6 module (AERO6), and CMAQ v5.0.2 with CB05-AERO6-VBS module (AERO6-VBS). Anthropogenic emission inventories are based on JATOP for Japan and REAS v2.1 for other regions. MEGAN v2.10 and GFED v3.1 are used for biogenic VOC and biomass burning emissions, respectively. Concentrations of SO42-, NO3-, and NH4+ were well reproduced by the all models in summer, while SO42- was underestimated in winter and spring and NO3- was overestimated in winter. Simulation better reproduces NO3- concentration when deposition velocities of HNO3 and NH3 are enhanced by a factor of five, as was done in a previous study. EC concentration was well reproduced over the all periods, while OA concentration was largely underestimated by AERO5 and AERO6 in the all periods. Simulation using AERO6-VBS better reproduced OA concentration than the other models in spring and summer, mostly due to higher anthropogenic SOA concentration in AERO6-VBS. However, OA concentration was underestimated also by AERO6-VBS in winter and spring. These comparisons are the important first step for model improvement in Japan.

Johannes Bieser (1), Volker Matthias (1), Beate Geyer (1), Ian Hedgecock (2), Francesco DeSimone (2), Christoph Gencarelli (2), Oleg Travnikov (3), Andreas Weigelt (1)

(1) Helmholtz-Zentrum Geesthacht, , Institute of Coastal Research, Max-Planck-Strasse 1, 21502 Geesthacht, Germany

(2) CNR - Institut Inquinamento Atmosferico, U.O.S. Di Rende, UNICAL-Polifunzionale, 87036 Rende, Italia

(3) Meteorological Synthesizing Center-East of EMEP, 2^nd Roshchinsky proezd., 8/5 Moscow 115419, Russia

Presenting author email: johannes.bieser@hzg.de

Mercury is a toxic substance that is ubiquitous in the environment. In the atmosphere mercury exists in three forms: Gaseous Elemental Mercury (GEM), Gaseous Oxidized Mercury (GOM), and Particle Bound Mercury (PBM). GOM and PBM make up only 1 percent of the total. But deposition, which is the only sink for atmospheric mercury, is dominated by these two species. Therefore, oxidation processes are key to understand the behaviour of mercury in the environment. Yet, only few continuous measurements of GOM and PBM are available. Because of this lack of data, many studies have used wet deposition as a proxy for the oxidized mercury species. However, recent studies have shown that good agreement of modelled and observed deposition is not necessarily an indicator for correct oxidation processes in the atmosphere.

This model study is part of the European Union FP7 Research Project GMOS (Global Mercury Observation System). GMOS focuses on the improvement and validation of mercury models to assist establishing a global monitoring network and to support political decisions. In the course of this study the CTM Community Multiscale Air Quality model (CMAQ v5.0.1) was used to simulate transport, oxidation, and deposition of mercury using different chemical oxidation mechanisms and emission inventories. For this, CMAQ has been set up with different reactants and reaction rates for the oxidation of GEM. Besides the reactions with ozone, OH, and chlorine which are already implemented in the multi pollutant version of CMAQ, special emphasize was on the implementation of bromine reactions into the model.

The study focuses on a model domain covering the whole of Europe, as this is the region with the highest data coverage in the GMOS observation database for the year 2013. Additionally, a regional domain over China was set up, where two GMOS measurement stations are located. The model results are evaluated based on ground based long-term measurements of speciated mercury from the GMOS observation network. Furthermore, observations of GEM and GOM in the planetary boundary layer and the lower free troposphere from the ETMEP-2 (European Tropospheric Mercury Experiment) are used. Finally, the impact of the different reactants on regional deposition patterns is investigated by comparison with wet deposition measurements.

Michael G. Barna, National Park Service, Lakewood, CO

Tammy M. Thompson, Colorado State University, Fort Collins, CO

C. Thomas Moore, Western States Air Resources Council, Fort Collins, CO

Oil and gas extraction in the Intermountain West has rapidly increased in recent years, resulting in thousands of additional wells in basins such as the Uinta (Utah), Piceance (Colorado), Denver-Julesberg (Colorado), Green River (Wyoming) and San Juan (New Mexico). Much of this development has occurred on federal or tribal lands that are upwind of protected wilderness areas and national parks, several of which are already experiencing ozone concentrations near or in excess of current air quality standards. Oil and gas development results in significant ozone precursor emissions of nitrogen oxides (NOx) and volatile organic compounds (VOC), and represents a major and growing source sector in this region. To quantify the contribution of oil and gas emissions to regional ozone, and to help identify potential mitigation strategies (i.e., NOx and/or VOC controls on oil and gas development), the CAMx photochemical grid model was applied. For this study, two 'reactive tracer' mechanisms within CAMx were used to evaluate the response in ozone concentration relative to precursor emissions: the Anthropogenic Precursor Culpability Assessment (APCA), and the Higher-order Decoupled Direct Method (HDDM). These two approaches offer a more efficient approach to evaluating ozone dynamics as compared to more traditional (and computationally expensive) emission sensitivity simulations.

The Air Quality Model Evaluation International initiative has now concluded its second phase. The latter related to the evaluation of on line coupled models applied, like in phase one, to a EU and NA years of air quality simulations. The second phase has been particularly complex given the complex nature of these modeling systems and the necessity of defining clear situations in which the couple nature of the systems could be maximized and emphasized for model evaluation purposes. An interesting exercise has been the comparison of phase one against phase two in terms of overall performance of un-coupled/off-line vs coupled/on-line models. For this reason we have exploited the ensemble techniques developed at JRC over the years to compare the large amount of information pertaining to several models and several species analyzed in the two phases. The ensemble techniques used year strictly relate to the determination of the best members to be included and obtaining the best performing ensemble rather than simply averaging all model results. Correlation of errors and filtering or redundant results is taken into account so that the best information can be extracted by the abundance of information available. The result shows a relative improvement in the modeling capacity of some species (e.g. NO2 and PMs) and a substantially unchanged performance for ozone. The uniqueness of the AQMEII database remains an invaluable source of information that for both Phase 1 and 2 is still to be exploited. Phase 3 of the initiative will also be presented as continuation of the AQMEII international collaboration.

Thien Khoi V. Nguyen¹, Annmarie G. Carlton¹, Shannon L. Capps²

¹Department of Environmental Sciences, Rutgers University, New Brunswick NJ 08901, USA.

²United States Environmental Protection Agency through Oak Ridge Institute for Science and Education, Research Triangle Park, NC 27711, USA.

Particle-phase liquid water is a ubiquitous component of atmospheric aerosols and influences a variety of critical atmospheric processes, including light scattering, the hydrological cycle, aqueous chemistry, and particulate matter (PM) formation. Yet despite the abundance and importance of aerosol water, it is not routinely measured, and mass concentrations are not well known. Here we use the thermodynamic model ISOYYOPIA (v2.1) to estimate aerosol water mass concentrations from 2000-2010 in urban and rural locations using speciated ion and meteorological data from sites that are a part of the Southeastern Aerosol Research and Characterization (SEARCH) network and the Interagency Monitoring of Protected Visual Environments (IMPROVE) program. These estimations are coupled with aerosol scattering data from the Aerosol Robotic Network (AERONET) to better understand the regional and historical trends of aerosol water in the United States in the context of regional differences and improved air quality. We find that aerosol water is highest in the Northeast and Southeast and lowest in the West and West Pacific. Additionally, analysis in the region of the Southern Oxidant and Aerosol Study (SOAS) indicates decreases in aerosol water mass concentrations by 29%, 60%, and 67% over the last decade for June, July, and August, respectively. The observed trends in the Southeast are consistent with the hypothesis that decreases in aerosol water may explain recently noted reductions in organic mass concentrations despite no apparent decrease in biogenic volatile organic carbon precursor emissions. These results provide evidence for modulation of biogenically derived PM in the presence of anthropogenic perturbations.

Jia Xing, Rohit Mathur, Jonathan Pleim, Christian Hogrefe, Chuen-Meei Gan, David Wong

The U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA

Ozone in troposphere gains great attention from scientific studies because it's not only a key air pollutant that affects human health, crop productivity and natural ecosystems, but also a greenhouse gas that impacts radiations and global climate. An accurate model representation of upper tropospheric ozone become more necessary for regional models when the simulation is conducted over a expanding spatial domain with a long simulation period which provides sufficient opportunities for exchange between the boundary layer and free troposphere. It's also been well documented that lateral boundary conditions (LBCs) are essential for regional air quality simulations. Static LBCs cannot adequately represent variability in the free troposhpere. The LBCs derived from global models may also be influenced by simulation biases at global scale as well as uncertainties from downscaling methodologies. Ideally, the upper tropospheric ozone will be calculated simultaneously in the regional modeling.

In this study, O₃ in the upper troposphere was specified by a previously developed scaling of ozone in the upper troposphere with the model estimated potential vorticity (PV). The constant of proportionality in the linear regression between ozone and PV was derived from limited data derived from ozonesonde measurements from the IONS network and model PV estimates over the CONUS during summer 2006. Recently, a 21-year air quality simulation across the northern hemisphere from 1990 to 2010 has been conducted. The model estimated O₃ in upper troposphere is significantly underestimated compared to ozone sonde observations provided by World Ozone and Ultraviolet Radiation Data Centre (WOUDC). Such underestimations might be associated with uncertainties of current tropopause O₃ scaling from the static PV function. Further analysis of the relationship between 21-year record of PV estimates and the historical series of O₃ sonde observations suggests a seasonal-, spatial- dependent PV function which is expected to be established in CMAQ. Improvement and uncertainties of this new function will be discussed as well.

Jennifer Hegarty, Chantelle Lonsdale, Karen Cady-Pereira, John Nowak, Jennifer Murphy, Matthew Alvarado, Milos Markovic,Trevor VandenBoer

The California Central Valley is a major source region of ammonia (NH₃), which leads to the substantial formation of ammonium nitrate (NH₄NO₃), and ammonium sulfate ((NH₄)₂SO₄) aerosols which degrade air quality. Yet emission estimates of NH₃ for this region and most regions around the globe are highly uncertain and available emission inventories can be off by an order of magnitude. Here we evaluate CMAQ simulations of NH₃ during the NOAA CalNex field campaign of May - June 2010 using satellite observations of NH₃ from the NASA Aura Tropospheric Emissions Spectrometer (TES). The simulations are driven with emissions fields prepared for the CalNex period by the California Air Resources Board (CARB). We also compare the CMAQ simulations of NH₃, ammonium nitrate (NH₄NO₃), and ammonium sulfate ((NH₄)₂SO₄) with ground-based and aircraft measurements from the CalNex field campaign. We will discuss how this initial evaluation of the NH₃ emission estimates in the Central Valley will provide valuable guidance on our future application of the CMAQ-adjoint model to better quantify emissions of NH₃ in an intensive agricultural region.

Boris Galvis¹, Jorge E. Pachon¹, Oscar L. Lombana², Barron H. Henderson³

¹ Universidad de La Salle, Department of Environmental Engineering, Bogota, Colombia

² Universidad de La Salle, Centro Lasallista de Investigación y Modelación Ambiental CLIMA

³ University of Florida, Department of Environmental Engineering, Gainesville, FL

Resuspended dust is likely to play a substantial but still undetermined role in Bogota's particulate emissions. The city's Environmental Authority (SDA), the University of La Salle (ULS) and the University of Florida (UF) are developing a complete air quality modeling system to assess strategies to curb air pollution, which requires a comprehensive emissions inventory. We carried out measurement campaigns to estimate resuspended dust emission factors in the city and applied two approaches. The first involved collecting the loose surface material from paved and unpaved roads at approximately 40 locations around the city. Using these samples we calculated road dust emission factors based on the mass of fine particulate determined by a granulometric analysis. In the second approach we used a mobile sampling platform to measure resuspended coarse and fine particulate concentrations every second in front and behind of a small a vehicle in the same roads and calculated dust emission factors using the difference between the measurements. Both approaches were adapted from EPA's AP-42 methods and previous works. In this work we compare results obtained for the emission factors and estimate the emission of resuspended dust for the city, we also describe the campaigns and the data treatment for the mobile measuring approach. These results can be of interest for cities in developing countries where resuspended dust can be a major fraction of the particulate matter.

Jiaoyan Huang, Heather Holmes, Mae S. Gustin

Mercury (Hg) is classified as a toxic air pollutant and atmospheric deposition is an important pathway by which Hg enters ecosystems. However, current Hg model simulations are focused on Eastern Asia and the Eastern US. In this study, CMAQ will be used to understand gaseous oxidized Hg (GOM) concentrations and deposition in California and Nevada. First, results simulated from CMAQ 4.7.1 and CMAQ 5.0.1 will be compared to investigate differences in Hg concentrations due to the bi-directional exchange algorithms that replace the dry deposition scheme for Hg air-surface exchange. Next, because GOM is influenced by the air from Asian long range transport at high elevations in this area, and there are new Hg emission standards for electric generating units (EGUs) designated by USEPA, several scenarios (CMAQ 5.0.1) will be tested. These include 1) base case with current US anthropogenic emissions based on 2011 National Emissions Inventory (NEI), 2) modified boundary condition (BC) (zero out GOM concentrations at layers above the planetary boundary layer in the Western US), and 3) low emission case (normal BCs with 80-90% US anthropogenic emissions reduction). The base case simulation will be compared with field measured Hg dry deposition in this area to understand the gaps between atmospheric Hg measurements and modelling. Based on this comparison, the dry deposition/bi-directional exchange scheme and related reactions, such as Br and HNO₃, will be modified to better represent atmospheric GOM behavior in the Western US.

Cesunica Ivey¹, Heather Holmes², Yongtao Hu¹, James A. Muholland¹, and Armistead G. Russell¹

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

²Department of Physics, University of Nevada Reno, Reno, Nevada, USA

Spatially resolved daily source impact estimates are desired to aid in public policy decisions and for ecological and human health air pollution studies. A Python function has been developed that combines emissions inventories, CMAQ results, and observed species concentrations to calculate sensitivities to PM2.5 emissions. The function executes a novel hybrid source apportionment method, which uses a nonlinear optimization approach to minimize the error between observed concentrations of PM2.5 mass and components (e.g. carbon, ions, and metals) and modeled estimates of PM2.5 concentrations and sensitivities. First, CMAQ-DDM is applied for a 24-hour period. Then modeled sensitivities and observations (PM mass and species) are used as inputs to the hybrid optimization function. These are combined to provide a gridded field of hybrid adjustment factors for each source category, which are then applied to the gridded emissions to increase, decrease, or maintain initial estimates as determined by the optimization step. Two additional iterations of CMAQ-DDM and optimization are performed with the new emissions estimates to generate new sensitivity estimates. Methods are applied and evaluated for winter and summer 2006. PM mass and species concentrations are obtained from multiple monitoring networks including CSN, SEARCH, and IMPROVE.

Juan Li¹, Qingyan Fu¹, Wen Yang², Dongfang Wang¹, Juntao Huo¹,Liang Xian², Yihua Zhang¹, Qinggen Bian¹

1. Shanghai Environmental Monitoring Center, Shanghai, China, 200030

2. Chinese Research Academy of Environmental Sciences, Beijing, China, 100012

Using a tethered balloon with the maximum load of 200kg as a vertical observation platform, black carbon (BC) profiles in morning, noon, afternoon, and night were developed in Fengxian (N30°49'47",E121°30'04"), Shanghai on Dec. 13, 2013. It was found that in the morning BC concentrations decreased with height. At noon, changes in BC concentrations, associated with whole layers (i.e., around 4000ng/m³), was observed to be tied to active movement, ascending or descending, of air. In addition, mixing layer height is responsible for diluting BC concentrations. In afternoon hours, BC decreased with height. At night, BC profiles were closely linked to the formation of the residual layer, nocturnal boundary layer, and mixing layer. In the residual layer, trapped BC concentrations remained high. In the nocturnal layer, BC rarely varied with height. Below the mixing layer, high BC concentrations were ascribed to local emissions. However, because of the mixing layer, BC remained trapped and transport to an upper layer was prevented. A correct understanding and accurate prediction of BC behavior is critical for a correct understanding of potential health risks associated with air pollution in large cities.

Nina Randazzo, Daniel Tong, Pius Lee, Xiaozhen Xiong, Feng Ding, Juying Warner, Monika Kopacz

Prior studies have shown that the 2008 global economic recession, the worst one since the Great Depression, has caused noticeable reduction of NOx emissions (Russell et al., 2012; Tong et al., 2014). However, the corresponding emission projections were unable to reflect the drop in ambient NOx that was observed consistently by satellite and ground monitors during the Global Recession. This study aims to assess the relationship between economic activity and atmospheric CO trend, another byproduct of fossil fuel combustion, and to see if the CO trend is similar to the trend of NOx during and after the Global Recession. Interannual variability due to biomass burning complicates the quantification of the Recession's effects on CO levels. For this reason, this study will focus on ten urban areas in the United States: New York City, Los Angeles, Boston, Chicago, Dallas, Philadelphia, Houston, the District of Columbia, Miami, and Atlanta. The dominant sources of CO for these areas are anthropogenic, so the impact of biomass burning variability on ambient CO interannual variability is likely insignificant. The focus will be on July CO levels from 2005 to 2012. This study will use both ground observations from the EPA Air Quality System (AQS) during the morning rush hours and remote sensing data from the Atmospheric InfraRed Sounder (AIRS) instrument onboard NASA's Aqua satellite. The AIRS-Version 6 Level 2 product will be used because of its relatively fine spatial resolution and its potential to support near-real-time air quality forecasting operations. The satellite data will be screened based on quality indices and aggregated into monthly data. The annual rate of change of CO derived from AIRS data will be compared with AQS data during different stages of the recession.

Bernhard Rappengl ck^1,, Luis Ackermann¹, Sergio Alvarez¹, Julia Golovko¹, Martin Buhr², Robert Field³, Jeff Soltis³, Derek C. Montague³, Bill Hauze⁴, Adamson Scott⁴, Dan Risch⁴, George Wilkerson⁴, David Bush⁵, Till Stoeckenius⁶, Cara Keslar⁷

¹ Dept. of Earth and Atmospheric Sciences, University of Houston, Houston, TX, USA

² Air Quality Design, Boulder/CO, USA

³ University of Wyoming, Laramie/WY, USA

⁴ Meteorological Solutions Inc., Salt Lake City/UT, USA

⁵ T&B Systems, Santa Rosa/CA, USA

⁶ Environ, Novato/CA, USA

⁷ Wyoming Department of Environmental Quality, Cheyenne/WY, USA

During recent years, elevated ozone (O3) values have been observed repeatedly in the Upper Green River Basin (UGRB), Wyoming during wintertime. This paper presents an analysis of high ozone days in late winter 2011 (1-hour average up to 166 ppbv). Intensive Observational Periods (IOPs) were performed which included comprehensive surface and boundary layer measurements. Low windspeeds in combination with low mixing layer heights (~50 m agl) are essential for accumulation of pollutants. Air masses contain substantial amounts of reactive nitrogen (NO_x) and non-methane hydrocarbons (NMHC) emitted from fossil fuel exploration activities in the Pinedale Anticline. On IOP days in the morning hours reactive nitrogen (up to 69%), then aromatics and alkanes (each ~10-15%; mostly ethane and propane) are major contributors to the hydroxyl (OH) reactivity. This time frame largely coincides with lowest NMHC/NO_x ratios (~50), reflecting a relatively low NMHC mixture, and a change from a NO_x-limited regime towards a NMHC limited regime as indicated by photochemical indicators, e.g. O₃/NO_y, O₃/NO_z, and O₃/HNO₃ and the EOR (Extent of Reaction).

OH production on IOP days is mainly due to nitrous acid (HONO). On a 24-hr basis and as determined for a measurement height of 1.80 m above the surface HONO photolysis on IOP days can contribute ~83% to OH production on average, followed by alkene ozonolysis (~9%). Photolysis by ozone and HCHO photolysis contributes about 4% each to hydroxyl formation. High HONO levels (maximum hourly median on IOP days: 1,096 pptv) are favored by a combination of shallow boundary layer conditions and enhanced photolysis rates due to the high albedo of the snow surface. HONO is most likely formed through (i) abundant nitric acid (HNO₃) produced in atmospheric oxidation of NO_x, deposited onto the snow surface and undergoing photo-enhanced heterogeneous conversion to HONO and (ii) combustion related emission of HONO. HONO production is confined to the lowermost 10 m of the boundary layer. HONO, serves as the most important precursor for OH, strongly enhanced due to the high albedo of the snow cover.

Timothy Vinciguerra, Sheryl Ehrman, Simon Yao, Joseph Dadzie, and Alexa Chittams (Department of Chemical and Biomolecular Engineering, University of Maryland, College Park, Maryland)

Russell Dickerson (Department of Atmospheric and Oceanic Science, University of Maryland, College Park, Maryland)

Over the past decade, anthropogenic pollutants have been successfully reduced, providing improved air quality, but a new influx of emissions associated with hydraulic fracturing and natural gas operations could be counteracting some of the benefits that have been gained. Using hourly measurements from Photochemical Assessment Monitoring Stations (PAMS) in the Baltimore, MD and Washington, D.C. areas, it has been observed that following a period of decline, daytime ethane concentrations and the ratio of ethane to total measured VOCs have increased significantly in the most recent years. This trend appears to be linked with the rapid natural gas production in upwind, neighboring states, especially Pennsylvania and West Virginia. Furthermore, ethane concentrations failed to display this trend at a PAMS site outside of Atlanta, GA, a region where no such widespread operations exist.

XINXIN ZHAI¹, Mariel Friberg¹, Heather Holmes^{1, 2}, Yongtao Hu¹, James Mulholland¹, Armistead Russell¹

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

² Currently at University of Nevada^, Reno, NV

The spatial and temporal relationship between acute health effects and ambient air pollution is being assessed in Georgia from 2002 through 2010. Daily concentration fields of 12 pollutants, 1-hr max CO, NO2, NOx, and SO2, 8-hr max O3, and 24-hr PM10, PM2.5, and PM2.5 constituents SO4, NO3, NH4, EC and OC, are estimated by fusing measurements (observations) from the ambient air monitoring network in Georgia with chemical transport model (CMAQ) estimates at 12-km and 4-km resolutions. When used in population-based epidemiologic studies, measurements are limited spatially by a sparse monitoring network and by local source impacts at particular monitors; in addition, speciated PM measurements are limited temporally by sampling frequency. Emissions-based model estimates are limited both by bias and by accuracy in simulating temporal and spatial variation. The data fusion technique is developed to minimize bias over space while maximizing prediction of variance over time (i.e., Pearson R²). To evaluate the data fusion method, we applied a cross validation technique by withholding 10% of the observations and comparing predictions with the data withheld. Results indicate that bias in CMAQ predictions is substantially reduced by data fusion over the study domain (average bias decrease 30% to 80% for all species compared to CMAQ) and that the spatiotemporal correlation is high near monitors (R² ~ 0.80 to 0.98) and decreases with distance from nearest observation toward the level of the observation-CMAQ correlation (R² ~ 0.22 to 0.64). Errors are greater for primary pollutants than secondary pollutants. One limitation of this method is the degree of spatial resolution is limited by the model and model input (i.e., near-source gradients are not captured). Results will be incorporated into health risk assessment to provide better estimates of association of acute health effects and ambient air pollution.

Taciana Toledo de A. Albuquerque^1,3, Rita Yuri Ynoue², Maria de Fatima Andrade², Angel Vara², Sergio Ibarra², Ayres G. Loriato¹, Nadir Salvador ¹, Alexandre M. Santiago¹, Renato M. Sartorio¹, Igor Baptista¹, Erick G. S. Nascimento¹, Enzo Todesco², Neyval C. Reis Junior¹, Davidson Moreira¹.

¹Graduate Program in Environmental Engineering (PPGEA), Department of Environmental Engineering, Federal University of Esperito Santo - Brazil

²Department of Atmospheric Sciences, Institute of Astronomy, Geophysics and Atmospheric Sciences, University of Sao Paulo - Brazil.

³Department of Sanitary and Environmental Engineering, School of Engineering, Federal University of Minas Gerais - Brazil.

The urban population of Brazil is most affected by air pollution from vehicle emissions. The Brazilian light-duty fleet typically runs on ethanol (95% ethanol; 5% water), gasohol (75% gasoline; 25% ethanol), or compressed natural gas, and a small portion of the heavy-duty fleet runs on biodiesel rather than normal diesel. An important characteristic of air pollutant emissions in Brazil is that ethanol accounts for 50% of the fuel burned by the transport sector. This paper aims to evaluate the impact of the vehicular emissions sources in the air quality of Southeast area of Brazil using the integrated numerical modeling system WRF-SMOKE-CMAQ during an anomalous dry period in the summer of 2014. CMAQ model has been running to provide local air quality scenarios for the States of Esperito Santo, Sao Paulo, Porto Alegre and Rio de Janeiro, but it was never used in Brazil for a regional grid area. In this study, meteorological fields were modeled using the Weather Research and Forecasting model WRFv3.4.1, for a month period (from January 15^th to February 15^th), using a domain with 9-km grid resolution, from 19 to 28^oS and 39 to 51^oW, which covers the capital cities with their respective Metropolitan Areas in the Southeast of Brazil (Sao Paulo, Rio de Janeiro, Belo Horizonte and Vitoria). Meteorological initial and boundary conditions are given by the 12UTC Global Forecast System (GFS) 0.5^o x 0.5^o horizontal resolution model. CMAQ model version 4.6 is run with the same grid resolution of the WRF model and the chemical mechanism and aerosol modules chosen were CB5 and aero4. The SMOKE emissions model was applied to build a spatially and temporally resolved vehicular emissions inventory for Southeast area. For the vehicular emissions, we used information provided by the Sao Paulo State Environmental Protection Agency (Companhia de Tecnologia de Saneamento Ambiental - CETESB), Traffic Engineering Company (CET), Laborat rio de Processos Atmosf ricos (LAPAt-IAG-USP), and National Department of Transit (DENATRAN). Horizontal emissions were spatially distributed using total road length and width (from OpenStreetMap) as a proxy. Measurements inside road traffic tunnels of gaseous and particulate compounds were used to compute emission factors and vehicular traffic counts were used for the temporal emissions diurnal variation.

Changhan Bae¹, Soontae Kim¹ ,Hyun Cheol Kim ^2,3 and Byeong-Uk Kim⁴

¹Ajou University, Dept. of Environmental Engineering, Suwon, Korea

² NOAA/Air Resources Laboratory, College Park, MD

³ UMD/Cooperative Institute for Climate and Satellites, College Park, MD

⁴ Georgia Environmental Protection Division, Atlanta, GA

Springtime ozone simulation and its sensitivity to combinations of emissions inventories over

Northeast Asia are tested during May and June, 2014. We utilized regional air quality model outputs from the Integrated Multidimensional Air Quality System for Korea (IMAQS/K), which is based on the Weather Research and Forecasting (WRF)-Sparse Matrix Operator Kernel Emissions (SMOKE)-Community Multi-scale Air Quality (CMAQ) modeling framework with nested domains; 27-km northeastern Asia, 9-km South Korea, and 3-km Seoul Metropolitan Area (SMA). Two combinations of emission inventories utilized in this studay are: (1) Intercontinental Chemical Transport Experiment Phase B (INTEX-B) 2006 for Asia and Clean Air Policy Support System (CAPSS) 2007 for South Korea and (2) Model Inter-Comparison Study (MICS)-Asia 2010 for Asia and CAPSS 2010 for South Korea. The latter consists of updated emission inventories with 45% higher NOx emissions and % higher Volatile Organic Compound (VOC) emissions over China and 25 % lower NOx emissions and 5% lower VOC emissions over the SMA than emission inventories used for the former combination. Simulations are evaluated with surface observations from the AirKorea surface network. Results show the updated emission inventories produce considerable changes in both daytime (i.e. 11 to 18 Local Standard Time) hourly and daily max 8-hour average ozone concentrations; 4 ppb higher for averaging over S. Korea and 13 ppb higher for averaging over the SMA. We also notice that, with increased foreign emissions and decreased domestic emissions, simulated ozone changes are sensitive to regional-scale transport of precursors, implying the intricacy of ozone simulation in the SMA.

Colleen Baublitz¹, Dr. Barron Henderson¹, Alexander Rincon², Dr. Boris R. Galvis², Dr. Jorge E. Pachon²

1 University of Florida, Department of Environmental Engineering, Gainesville, FL

2 Universidad de La Salle, Department of Environmental Engineering, Bogota, Colombia

Bogota is implementing the Models-3 system to improve its capacity to assess strategies to manage and curb air pollution. The Universidad de La Salle in Bogota and the University of Florida are supporting the local environmental authority on this task. The project team implemented a thorough emission inventory using the Sparse Matrix Operating Kernel Emissions Modeling System (SMOKE). SMOKE was targeted for this project due to its capacity to integrate more detailed information in the future and its incorporation with the Community Multiscale Air Quality (CMAQ) modeling system.

The SMOKE-derived emission inventory includes biogenic, commercial, industrial, and mobile emissions. For each category, we will describe the data locally available and the methods for linking this data to the SMOKE system. For each category, we will also describe the methods for temporal and spatial distributions of emissions, the mapping of source classification codes and the allocation of speciation profiles for non-methane volatile organic carbons and particulate matter. In addition, we will describe key uncertainties including records for specific commercial sources, a lack of continuous monitoring technologies for point sources, and several obstacles to the characterization of the vehicle fleet. Despite these uncertainties, sufficient information is available to implement SMOKE and to provide model ready emissions to CMAQ.

The CMAQ model combines meteorology from WRF and emissions from SMOKE to predict the state of air pollution in Bogot and to provide a test environment for regulatory strategies. The base case SMOKE inventory will be processed using CMAQ and evaluated against the observations from the citys air quality monitoring network known as the Red de Monitoreo de Calidad de Aire de Bogota, or RMCAB. This will give a better picture of where room for improvement is most needed in the initial emission inventory. The final SMOKE-CMAQ system will produce concentration fields to test the most effective regulatory strategies proposed by the SDA. This presentation will show initial results based on population weighted concentration fields.

Jonathan Dorn and David Cooley, Abt Associates, Durham, NC

Roy Huntley and Jennifer Snyder, U.S. Environmental Protection Agency, OAQPS, RTP, NC

Residential wood combustion is an important contributor to emissions of particulate matter and other pollutants. The Residential Wood Combustion (RWC) Tool is an MS Access-based tool developed by the U.S. EPA in conjunction with Abt Associates to estimate emissions from residential fireplaces, woodstoves, and other wood-burning devices for the National Emissions Inventory (NEI). The tool uses survey data to estimate the number of each type of 11 different wood-burning appliances in each county and the amount of wood burned. These activity data are then multiplied by emission factors to estimate emissions. EPA updated the RWC Tool for the 2011 NEI with new local survey data for several states, including California, Oregon, and Washington, as well as a new user-friendly interface.

This paper will discuss the recent improvements to the RWC Tool and present a case study involving the estimation of the health impacts of residential wood smoke emissions in the Pacific Northwest (PNW). Abt Associates used the RWC Tool to estimate wood smoke emissions in the PNW study area, and the Co-Benefits and Risk Assessment Tool (COBRA) to estimate health impacts based on a simplified air dispersion model and built-in health impact functions. By monetizing the health benefits using EPA-approved valuation functions, Abt Associates concluded that a complete elimination of residential wood smoke in the Pacific Northwest could lead to health benefits of more than $3 billion per year in the PNW study area.

Shantha Daniel, Marvin Jones, and Ron Thomas, Texas Commission on Environmental Quality (TCEQ)

The TCEQ currently administers four cap and trade programs in Texas. As part of the attainment demonstration modeling for the State Implementation Plan, the TCEQ models these cap and trade programs by limiting the future year emissions of sources subject to these programs to the appropriate program's total future year program cap. The spatial representation of future year emissions of the sources subject to these cap and trade programs has typically been based on the source's future year allocation of allowances specified for the relevant cap and trade program(s). Since future year allocations are typically distributed many years in advance, the sources that received future year allowances in many cases may not be operational in the future year. This paper describes the procedure used by the TCEQ to spatially redistribute future year allocations to sources that are expected to be operational, using historical analysis of past compliance/trading trends for each of the programs. This redistribution provides a more realistic representation of emissions than simply modeling the regulatory allowances. This paper also describes how sources subject to multiple programs are treated.

Jonathan Dorn and David Cooley, Abt Associates, Durham, NC

Roy Huntley and Jennifer Snyder, U.S. Environmental Protection Agency, OAQPS, RTP, NC

Emissions from Industrial, Commercial, and Institutional (ICI) fuel combustion are often a significant portion of most areas' total emissions inventory. Unless all ICI combustion emission sources are covered in a geographic area's point source inventory, it is necessary for inventory preparers to estimate ICI combustion nonpoint source emissions. Because there are specific challenges associated with estimating ICI nonpoint source emissions activity/rates, the U.S. EPA in conjunction with Abt Associates developed an MS Access-based ICI Combustion Tool to assist State, Local, and Tribal agencies in estimating nonpoint source emissions from ICI fuel combustion for the 2014 National Emission Inventory.

The primary data source behind the ICI Combustion Tool is total state-level ICI energy consumption data released annually as part of the Energy Information Administration's State Energy Data System (SEDS). The ICI Combustion Tool processes the SEDS data and adjusts the data to account for the fraction of fuel consumed by nonroad mobile sources whose emissions are included in the nonroad inventory and by non-fuel combustion uses of energy, such as product feedstocks. Through a user-friendly interface, users can update the underlying assumptions in the adjustment methodology. The ICI Combustion Tool also includes a nonpoint source to point source crosswalk and allows the user to perform point source activity subtractions to avoid double counting of emissions between their point and nonpoint inventories. The ICI Combustion Tool generates outputs in EPA's Emissions Inventory System (EIS) format, ready for submission to the EIS. This paper will provide an overview of the ICI Combustion Tool and the procedures for developing a credible estimate of nonpoint emissions from ICI fuel-combustion sources.

Hyun Cheol Kim ^1,2, Pius Lee¹, Young-Kwon Lim³, Soontae Kim⁴, Jung-Hun Woo⁵, Byeong-Uk Kim⁶, Changhan Bae⁴ and Eunhye Kim⁴

¹ NOAA/Air Resources Laboratory, College Park, MD

² UMD/Cooperative Institute for Climate and Satellites, College Park, MD

³ NASA Goddard Space Flight Center, Global Modeling and Assimilation Office, Greenbelt, MD

⁴ Ajou University, Dept. of Environmental Engineering, Suwon, Korea

⁵ Konkuk University, Dept. of Advanced Technology Fusion, Seoul, Korea

⁶ Georgia Environmental Protection Division, Atlanta, GA

We report recent changes of tropospheric NO₂ vertical column density (VCD) in East Asia observed from multiple satellites, highlighting on the declining trend over China since 2012. Tropospheric NO₂ VCD data from Global Ozone Monitoring Experiment (GOME), Scanning Imaging Absorption spectrometer for Atmospheric CHartographY (SCIAMACHY), Ozone Monitoring Instrument (OMI), and GOME-2, retrieved from the Royal Netherlands Meteorological Institute (KNMI) and available from the Tropospheric Emission Monitoring Internet Service (TEMIS), are utilized to investigate annual trends of NO₂ VCD since 2001. Until 2011, changes of NO₂ VCD over East Asia countries agree well with previous researches including the impact of economic downturn during 2008-2009 and following quick recovery in China. After peaking at 2011, NO₂ VCD observations from active instruments (OMI and GOME-2) over China started to show a slow decreasing trend, mostly led by the rapid changes in Jing-Jin-Ji (JJJ, Beijing-Tianjin-Hebei) region in northern China. The decreasing trend continues as of spring in 2014. Trends over Korea are weaker, but similar to those in China, with a slight peak in 2011, and Japan shows continuous declining trend since early 2000. Possible explanations for the trend, including policy-driven emission change and inter-annual variance of meteorology and satellite retrieval uncertainties, are discussed. We suggest further investigations on anthropogenic NO_x emission changes using bottom-up approach and the climatological impact from inter-annual variance of natural condition.

Alex Macpherson (U.S. Environmental Protection Agency), Rich Mason (U.S. Environmental Protection Agency), Angel Aymond (The College of William & Mary)

Rich Mason, Jennifer Snyder, Ron Ryan, Roy Huntley, Laurel Driver, Sally Dombrowski, Alexis Zubrow, Alison Eyth, Rhonda Thompson, Venkatesh Rao

EPA has developed a version 2 of the 2011 National Emissions Inventory (2011v2 NEI) that reflects improved estimates for several source categories for both criteria air pollutants (CAPs) and hazardous air pollutants (HAPs). These improvements resulted from numerous projects, ranging from National Air Toxics Assessment preview comments, 2011 version 6 emissions modeling platform comments, state/local review and edits and refinements in EPA tools. The improvements in 2011v2 NEI impact all major sectors but on a national-level, are most significant for onroad mobile, oil and gas, nonpoint mercury, commercial marine vessels, agricultural field burning and residential wood combustion. However, corrections to several point sources have significant local impacts. We will discuss the nature of the changes for key sectors, provide comparisons of emissions for select source categories and discuss the broad processes by which many of these improvements were received. We will also discuss the lessons learned from the 2011 NEI development and ideas for how to improve our approaches for the 2014 NEI.

Ling Huang, Elena C. McDonald-Buller, Gary McGaughey, Yosuke Kimura, and David T. Allen

Center for Energy and Environmental Resources, The University of Texas at Austin, Austin, TX

Land cover is a key parameter in determining leaf area index (LAI) and emissions of biogenic volatile organic compounds (BVOCs). Misclassification of land cover could lead to inaccurate model predictions of biogenic emissions through over- or under-estimations in LAI values, incorrect biome-dependent based emission factors, and emission activity factors. A regional land cover dataset for eastern Texas with high spatial resolution (30 m) was developed for the Texas Commission on Environmental Quality (TCEQ). This study contrasts the regional land cover product (referred to as TCEQLC_2010) with a widely used global product (MODIS) over eastern Texas through a simplified set of seven categories: water, urban, non-vegetated, grasses/crops, forest, shrubs, and savanna. Biogenic emissions estimated by the Model of Emissions of Gases and Aerosols from Nature (MEGAN) based on the two land cover products are compared to characterize the effects of different vegetation inventories on emissions predictions.

Jorge E. Pachon^1,3, Boris R. Galvis^1,3, Barron H. Henderson², Colleen Baublitz², Alexander Rinc n³

¹ Universidad de La Salle, Department of Environmental Engineering, Bogota, Colombia

² University of Florida, Department of Environmental Engineering, Gainesville, FL

³ Universidad de La Salle, Centro Lasallista de Investigacion y Modelacion Ambiental, Bogota

Bogota, Colombia, has experienced rapid economic growth in recent years, which has led to a deterioration of air quality from increased vehicle, industrial and commercial emissions. The Secretar a Distrital de Ambiente (SDA), the local environmental regulatory agency, has created several strategies for mitigating air pollution and has recruited the Universidad de la Salle (ULS) of Bogota and the University of Florida (UF) to prioritize solutions. Prioritization will require a complete modeling system that includes a comprehensive emissions inventory. At this first stage of the project, we have updated emission inventories from industrial, commercial, mobile and biogenic sources and prepare in a format suitable for air quality modeling.

Our results indicate that emissions of CO, CO₂, NMOC, NO_x and SO₂ are predominantly from mobile sources whereas PM is almost equally emitted from mobile and point (industrial and commercial) sources. Chemical speciation for NMOC and PM emissions was performed. Emissions were temporal and spatially disaggregated by the hour and in cells of 1x1km covering the modeling domain (55x55km surrounding the city of Bogota). The methods used guarantee that emissions can be recalculated to assess different emission reduction scenarios.

This manuscript aims to show the activities performed to update emissions in Bogota. Emphasis will be placed on strategies to overcome the lack of information in the city, which may be common in developing countries. Following presentations will detail the implementation of emission inventories into SMOKE and CMAQ and preliminary evaluation of the modeling results.

Marc Guevara¹, Maria Teresa Pay¹, Jose Maria Baldasano^1,3

¹Earth Science Department, Barcelona Supercomputing Center, Jordi Girona 29, Edificio Nexus II, 08034 Barcelona, Spain

²Environmental Modeling Laboratory, Technical University of Catalonia, Barcelona, Spain

Fugitive dust emissions caused by agricultural activities (i.e. land preparation and harvesting) can contribute to the ambient particulate matter mass (i.e. coarse fraction), especially in Mediterranean countries like Spain, characterised by large agricultural and some semiarid regions as well as low amounts of precipitation. The present work describes the integration and modelling of fugitive dust emissions caused by agricultural operations within the CALIOPE air quality forecast system and over Spain (http://www.bsc.es/caliope/es). An estimation methodology based on the work of Schaap et al. (2009) was implemented inside the system in order to analyse the contribution of this source to Spanish PM₁₀ air quality concentrations. The CALIOPE air quality system was applied on a 4km horizontal resolution grid covering the whole Iberian Peninsula, with a temporal resolution of 1 hour and using two different chemistry transport models (i.e. CMAQv5.0.1 and CHIMEREv2013b) so the impact of different dry deposition schemes on the modelled PM₁₀ concentrations was also analysed. The concentration results obtained running the two simulations (one for each chemical transport model) were evaluated against observational data from AirBase stations.

Schaap, M., Manders, A.M.M., Hendriks, E.C.J., Cnossen, J.M., Segers, A.J.S., Denier van der Gon, H.A.C., Jozwicka, M., Sauter, F.J., Velders, G.J.M., Matthijsen, J., Builtjes, P.J.H., 2009 Regional modelling of particulate matter for the Netherlands: Technical report, PBL report nr 50009900, 60 pp, ISSN 1875-2322

Lawrence Reichle, Catherine Yanca, Rich Cook, and Cheryl Caffery

Exhaust emissions from nonroad engines and equipment vary based on engine/equipment type, control technology, fuel composition, and operating conditions. Characterizing the magnitude and chemical composition of these emissions is not only critical for developing air toxic inventories, but also for developing chemical speciation profiles needed for accurate modeling of ambient ozone and particulate matter. Nonetheless, speciation data for nonroad engines, especially for engines with emission controls running on gasoline/ethanol blends now in widespread use, is very limited. Improvements in the quality of these data could result in more accurate air quality modeling as well improved capability of EPA's MOVES model to predict the emissions of in-use engines.

In 2010, Southwest Research Institute, under contract to U. S. EPA, performed emission testing on a variety of nonroad spark ignition engines. Testing included hydrocarbon, aldehyde and ketone, and alcohol speciation, on 0% ethanol and 10% ethanol fuels from various all-terrain vehicles, non-road motorcycles, and <18 horsepower handheld and non-handheld lawn and garden equipment. We have evaluated these data for potential use in developing speciation profiles and emission factors.

In this paper, we discuss differences we have found in chemical composition between catalyzed and non-catalyzed engines, 2-stroke and 4-stroke engines, and E0 and E10 blends. Moreover, we discuss potential implications for air quality modeling. Finally, we discuss limitations in these data, and identify additional data needs.

Harvey Michaels, Darrell Sonntag, Alison Eyth, Alexis Zubrow, Cay Yanca, Rich Cook

We incorporated the speciation process, which previously occurred inside of SMOKE,
into MOVES2014 to better provide model-ready species for air quality modeling.
MOVES2014 uses speciation data, including speciation profiles from the SPECIATE database,
to estimate model-ready species. MOVES2014 conducts speciation according to important factors that impact speciation (vehicle classes, model year groups, fuel types, and emission processes). This improves the accuracy of the MOVES2014 speciated inventories compared to using source classification codes within SMOKE and MOVES2010b. MOVES2014 produces carbon bond chemical mechanism "CBO5" species, but has the model framework to incorporate additional chemical mechanisms in the future. We also updated PM2.5 speciation in MOVES2014 to estimate the 18 PM2.5 species consistent with CMAQ Aerosol Module version 6 "AE6." Updates to SMOKE are also made to be consistent with these changes.

Nick Walters, Xin Qiu, Fuquan Yang, Hamish Hains

Urban transportation emissions can impact significantly on local and regional air quality. A common approach is to apply road network spatial allocations with regional transportation total emissions through emission processing models, such as U.S. EPA's SMOKE, to obtain gridded, hourly, speciated emissions. Tools are available to help generate these emissions including the U.S. EPA's Spatial Surrogate Tool or Environment Canada's new Spatial Emissions Distribution Information System (SEDIS).

An updated road network is necessary to accurately estimate transportation emissions in an urban area. CanVec is a digital cartographic reference product produced by Natural Resources Canada (NRCan) that offers quality topographic information in vector format from the best available data sources in Canada. Implementing the CanVec road network into Canadian urban emissions studies showed significant improvements in emission quantity and spatial allocations by using the common surrogate approach in SMOKE.

In addition, emissions generated by CALMOB6, a region specific fuel economy and emissions model developed by the City of Edmonton, were integrated into SMOKE for comparison. This model uses the output of urban travel forecasting models based on city specific traffic counts to provide emissions tailored exclusively for the area. This provides highly accurate spatial resolution in link based emissions across 21 different vehicle types within the city core area.

Direct comparison of region specific emissions and the general road network spatial allocation method showed improved spatial resolution and lower daily total emissions from CALMOB6. This provides a way to evaluate and validate SMOKE road network surrogates with region specific data. Integrated emissions were adapted for use as an input to the regional air quality model CMAQ with 1.33km grid resolution to simulate winter high PM2.5 events in a Canadian city.

Jia Xing, Jonathan Pleim, Rohit Mathur, Christian Hogrefe, Chuen-Meei Gan, David Wong, George Pouliot, Chao Wei

The U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA

An accurate description of emissions is crucial for model simulations to reproduce and interpret observed phenomena over extended time periods. A consistent series of spatially resolved anthropogenic emissions of SO₂, NO_x, CO, NMVOC, NH₃, PM₁₀ and PM_2.5 in the United States from 1990 to 2010 was developed by using an approach based on several long-term databases containing information about changes in activity data and emission controls (Xing et al., 2013). The set of inventories developed in this study is internally consistent, constrained by activity trends, within reasonable range of emissions and controls, and comparable with previous studies. However, due to the lack of a detailed historic record of control information over such an extended time period (except for the power sector), our estimations on control efforts for other sectors highly depended on National Emission Inventories (NEI) data or NEI trends. Further improvement on those details is still necessary.

The 36km-CONUS WRF-CMAQ simulations driven by the newly-developed emission inventory provide an excellent basis for the evaluation of the newly-developed historical inventory. Trends in air quality across the northern hemisphere were also simulated using internally consistent historical emission inventories obtained from EDGAR. Thorough comparison against available observations from ground monitors and satellite remote sensors will be conducted. Gaseous and aerosol measurements taken from several routine monitoring networks including the Clean Air Status and Trends Network (CASTNET), the Interagency Monitoring of Protected Visual Environments (IMPROVE), the Aerometric Information Retrieval System (AIRS)-Air Quality System (AQS), as well as satellite-retrieved NO₂, SO₂ vertical column density from Ozone Monitoring Instrument (OMI), SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY), Global Ozone Monitoring Experiment (GOME) products are used to validate and improve the current understanding of emissions. Improvements of the newly-developed inventories will be suggested.

Fernando Garcia-Menendez¹, Erwan Monier¹ and Noelle E. Selin²

1) Center for Global Change Science, Massachusetts Institute of Technology

2) Engineering Systems Division and Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology

Several studies have attempted to estimate the impact of climate change on air quality by using meteorological fields derived from general circulation model simulations to drive chemical transport models and project pollutant concentrations. However, large uncertainties associated with climate simulations may propagate into the predictions of future air quality. Here we investigate the effects of uncertainty in climate projections on U.S. air quality estimates. Future O₃ and PM_2.5 concentrations are simulated with the Community Earth System Model driven by meteorological fields derived from an ensemble simulation of 21st century climate change carried out with the MIT Integrated Global System Model (IGSM), which considers uncertainties in emissions of climate forcers, climate sensitivity and natural variability. The influence of natural variability on modeling analyses aimed at quantifying the effect of climate change on air pollution is explored by carrying out multidecadal atmospheric chemistry simulations under present and future climates using multiple initial conditions. Through these simulations, we identify the magnitude of unforced interannual variations in air quality and adequate time scales for climate-related impacts assessments. The impact of uncertainty in climate parameters inherent to global climate models is investigated by perturbing climate sensitivity within the IGSM and quantifying the response in pollutant concentrations. Different greenhouse gas emissions scenarios are also included in the ensemble to assess the potential effects of climate policy on future air quality. In addition, the analyses further examine how variations in air quality projections due to climate uncertainty may propagate into health impacts estimates. By simulating changes to O₃ and PM_2.5 levels across the climate ensemble, we weight the uncertainties in the climate penalty on U.S. air quality and identify important considerations that must be accounted for in modeling assessments attempting to project climate change impacts on air quality.

Evan E. Johnson, Andrew Yates, Richard N.L. Andrews, Saravanan Arunachalam, Carol Lenox, Tyler Felgenhauer

Consumer behaviors such as conservation, adoption of more efficient technologies, and fuel switching represent significant potential for greenhouse gas mitigation. Current efforts to model future energy outcomes have tended to use simplified economic assumptions to represent complex behavioral choices, based on a lack of available data. Simulation of various mitigation scenarios could be improved by better capturing the behavioral and attitudinal influences on energy-related economic activity. Further, a better sense of how mitigation outcomes vary with con sumer behavior will afford policymakers a clearer understanding of how to target behavioral energy initiatives strategically. To address these issues, we estimate a set of discrete choice models designed to capture variation in the degree to which consumers hesitate to adopt new energy-saving appliances and vehicles. Our models predict consumer purchasing behavior for HVAC systems and fuel-efficient vehicles, making use of ten years worth of detailed data collected by the U.S. Bureau of Labor Statistics Consumer Expenditure Survey from 2003-2012.

We approximate variation in adoption probabilities based on consumer responses to changes in capital and operating costs, calculating specific levels of willingness to pay for savings through energy efficiency. These empirical results allow us to generate implicit discount, or "hurdle" rates, across a range of appliances and consumer groups. We exploit the large sample size and richness of this data set to stratify hurdle rate ranges by consumer and household characteristics such as income, education, region, and ownership status. We find that hurdle rates within narrowly defined choice sets for fuel efficient vehicles are low and exhibit minimal variation across different types of consumers. Widening the choice set to estimate the probability adopting new technologies with more dramatic efficiency gains changes this finding. We find that hurdle rate estimates for high efficiency HVAC systems can range up to 55% from as low as 19% for highly educated consumers. Results are sensitive to a number of assumptions about consumer characteristics and appliance operating costs. The results from this work will feed into the the MARKet ALlocation (MARKAL) Energy System model for further use in developing current and future energy scenarios in the U.S.

Courtney Taylor¹, Marco Rodriguez¹, Chao-Jung Chien¹, Caitlin Shaw¹, Tiffany Samuelson¹, Leonard Herr²

¹AECOM Inc.

²Bureau of Land Management, Utah State Office

The Uinta Basin is an area in northeastern Utah where oil and gas activities are important and expected to continue in the foreseeable future. The Air Resource Management Strategy (ARMS) Modeling Study was initiated by the Bureau of Land Management (BLM), Utah State Office to inform and support the BLM's air management strategy. One of the main air quality concerns in the Uinta Basin is the elevated ozone levels measured during winter. Three future year emissions mitigation scenarios were developed focusing on oil and gas activities in the Uinta Basin by estimating potential development, decline in the production rates of existing sources, and feasibility of controls. Potential future air quality impacts were simulated with CMAQ for all criteria pollutants, including ozone and other air quality related values. Model-predicted impacts of mitigation scenarios were compared to current air quality conditions as well as potential future conditions without additional control requirements. Since ozone precursors, namely non-methane hydrocarbons (NMHC) and nitrogen oxides (NOx), determine to a large extent the formation of ozone, a detailed examination of the modeled concentrations of both NOx and NMHC was performed. Indicator species such as formaldehyde to nitrogen dioxide ratio (HCHO/NO2) were also used to determine NOx or VOC-limited regimes within the modeling domain and the corresponding effects from the three future mitigation scenarios. Benefits and drawbacks of the control strategies on co-pollutants also were assessed.

Jeongran Yun¹, Mark Beauharnois¹, Christian Hogrefe², Jia-Yeong Ku³, Winston Hao³, Eric Zalewsky³, and Kenneth L. Demerjian¹

1. Atmospheric Sciences Research Center, University at Albany, Albany. NY

2. U.S. Environmental Protection Agency's Atmospheric Modeling and Analysis Division, Research Triangle Park, NC

3. New York State Department of Environmental Conservation, Albany, NY

Kirk Baker, U.S. Environmental Protection Agency

James Kelly,U.S. Environmental Protection Agency

Chris Misenis, U.S. Environmental Protection Agency

Ozone and PM2.5 formation in the atmosphere is the result of precursor emissions and generally conducive meteorological conditions. In the summer, elevated ozone and PM2.5 are typically associated with periods of high temperature, low wind speed, and high solar insolation. California is particularly challenging to represent as elevated air quality episodes are typically associated with a variety of physical processes including terrain blocking of air masses, complex ocean-land circulation, and a broad range of stationary and mobile emissions sources. These conditions can lead to elevated ozone and PM2.5 in the summer.

Special field campaign and routine meteorological measurements taken in California during the summer of 2010 (CalNex field campaign) provide an opportunity to evaluate a configuration of the WRF model used recently to support regulatory air quality modeling and an alternative WRF configuration shown to replicate central valley flows during the summer 2010 period. While PBL height and mixed layer heights are not always coincidental, this field campaign provides an opportunity to evaluate WRF estimates of the boundary layer height. In addition to mixing height, important meteorological output variables for air quality retrospective analysis include wind speed, wind direction, temperature, humidity, and solar radiation.

The CalNex field campaign also provides a rich dataset of speciated PM2.5 and important precursor measurements. CMAQ estimates of inorganic and organic PM2.5 are compared with special measurements made during the field campaign at Pasadena and Bakersfield. Model estimates of inorganic precursors including ammonia and nitric acid and gases known to yield SOA including isoprene, monoterpenes, toluene, and xylene are also compared with measurements taken at the main CalNex monitor locations.

Ozone and fine particulate matter, or PM2.5, are criteria air pollutants in the U.S. with well-studied health impacts. The aviation sector contributes a growing proportion of these pollutants to the atmosphere. One of the Federal Aviation Administration’s key policy objectives is to reduce aviation-attributable human health impacts across the United States, notwithstanding absolute growth in aviation activity. 

Because people may be exposed to concentrations of ozone and particulate matter formed from the emissions of several airports, it can be computationally expensive to calculate the proportion of exposure a population receives from each airport using finite difference (brute force) or regression-based methods. Using the Decoupled Direct Method in Three Dimensions (DDM-3D), we are able to calculate sensitivity coefficients from individual airports in the domain as well as predict changes in atmospheric O3 and PM2.5 concentrations based on variation in emission levels from each study airport. 

We present the design of experiments and results from a modeling exercise that significantly expands the temporal scope of our preliminary results, presented at CMAS 2013. Our domain includes sensitivity coefficients for six precursor species and 66 airports in the continental United States (CONUS). This group of airports accounts for about 61% of total flights and 77% of total aviation jet fuel consumed annually in the U.S. National Air Space (NAS). Our results include metrics to be considered for quantifying airport-specific impacts on air pollution using this framework. 

Results from our modeling exercise using the newly-released CMAQ v5.0.2 with DDM-3D for O3 and PM can be used to quantify the air quality and health impacts of air pollution burden from each individual airport’s emissions due to landing and takeoff (LTO) activity in the U.S., and characterize the spatio-temporal patterns in primary and secondary components of aviation-attributable PM2.5. As a result, future scenarios that take into account changes in policy, differential growth in aviation activity across multiple airports, and changing population characteristics can be accurately evaluated for sustainable growth. 

Fang-Yi Cheng¹, Zhih-Min Yang¹, Soontae Kim²

¹ Department of Atmospheric Sciences, National Central University, Taiwan

² Department of Environmental Engineering, Ajou University, Korea

In this study, the process analysis and sensitivity of precursor emissions to ozone (O₃) and particulate matters (PM) are studied using Community Multiscale Air Quality (CMAQ) modeling system to understand Taiwan's air pollution problem. Taiwan is an island with Central Mountain Range (with peaks as high as 3952m) running from the north of the island to the south, flanked by gently sloping plains on the western side that complicates the local flow patterns and affect the dispersion of pollutants; as a result, a fine scale air quality modeling is needed to understand the complex behavior.

The study objective includes: (1) to illustrate the complex boundary layer meteorological conditions that affects the air pollutants dispersions; (2) to investigate the sensitivities of O₃/PM with respect to its precursor emissions and to identify the source contributions; (3) to identify the processes that contribute to the high O₃/PM problems. The preliminary result indicates the O₃ control regime can be shifted from a NOx-limited into a VOC-limited regime even with the same emission releases at the same locations. The process analysis results show that the major contributions to the accumulation of O₃ and PM include horizontal transport, emissions, gas-phase chemistry and aerosol processes depending on meteorological conditions. The results indicate a need for a region-specific emission control strategy with dynamic response according to the meteorological conditions in order to effectively control both O₃ and PM pollutants in Taiwan. The detailed simulation results will be discussed during the conference.

Jim Smith, Shantha Daniel, Zarena Post, Weining Zhao, Jim McKay, Melissa Ruano, and Lindley Anderson, Texas Commission on Environmental Quality

The Texas Commission on Environmental Quality recently published guidance for using emission credits generated from reductions of one ozone precursor pollutant (e.g. NO_X) to offset emissions of a second precursor (e.g. VOC) for new or modified sources within a Texas ozone nonattainment areas (inter-pollutant, or IP, credit use). The commission also published guidance for using credits within a nonattainment area that were generated in other nonattainment areas within the state (inter-basin, or IB, credit use). The guiding principle for all such trades is that photochemical modeling must show that the emission credits sufficiently offset the emissions from new or modified sources and that there is an air quality benefit or no detriment to the area in which the credits will be used. This paper describes the steps required to assess the effects of a proposed use of IP or IB credits and presents examples of both types of trades.

Xinyi Dong¹, Joshua S. Fu¹, Jonathan J. Buonocore²

¹Department of Civil & Environmental Engineering, University of Tennessee, Knoxville, 37996-2313

²Center for Health and Global Environment, Department of Environmental Health, Harvard School of Public Health, Boston, MA, 02215

As one of the most important contributors of anthropogenic emission sources, coal-fire power plant can significantly affect local air quality and public health¹. Although many modeling studies have been conducted to investigate the control benefit of reducing emission from regional energy generating units, our understanding  about the contribution from single emission source is still limited. While responses of air pollutants could be non-linear to emission changes, altering an intensive point emission source may also change the chemical pathways occur in local region. Consequently, it is necessary to evaluate the impact from single emission source to probe into the chemical responses of air pollutants, and also provide the most cost-efficient and effective emission control strategy design². In this study, the EPA regulation model Community Multi-scale Air Quality (CMAQ) has been applied over Eastern US to assess the air quality impacts from individual power plant's emission. We mainly focused in the Great Lakes and Mid-Atlantic region which contains a large concentration of fossil energy facilities.

CMAQ (version 4.7.1) was driven by meteorological filed from MM5 (version 3.7) over Eastern United States domain for full year simulation in 2005. CMAQ simulation was performed with 36km and nested-down to 12km domain. Our analysis will mainly focus on northeastern US where most of the large coal-fire power plants are located. National Emission Inventory 2005 (NEI2005) provided by US EPA was utilized in this study, and is processed by Sparse Matrix Operator Kernel Emissions Model (SMOKEv2.7). Initial and boundary conditions (IC/BCs) are downscaled from global model GEOS-Chem v9-01-03³.

We use brute-force method which include a baseline simulation that include all emission sources in the domain, and a sensitivity scenario exactly same as the baseline scenario except with the emission from the targeted power plants set to zero. Totally 40 sensitivity scenarios are designed to investigate 53 power plants impacts over northeastern US. Due to limited article length, here we only presented an example as the result from one cases: the Keystone Generating Station located in Pennsylvania. 

C. R. Lonsdale, M. J. Alvarado, R. J. Yokelson, K. R. Travis, and E. V. Fischer

Forecasting the impacts of biomass burning (BB) plumes on regional air quality is difficult due to the complex photochemistry that takes place in the concentrated young BB plumes. The spatial grid of global and regional scale Eulerian models is generally too large to resolve BB photochemistry, which can lead to errors in predicting the formation of O₃, secondary organic aerosol (SOA) and the partitioning of NO_y species. Using AER's Aerosol Simulation Program (ASP v2.1), we have developed a sub-grid scale parameterization of the near-source chemistry of BB plumes for use in regional and global air quality models. The parameterization takes inputs from the host model, such as solar zenith angle, temperature, and fire fuel type, and calculates enhancement ratios of O₃, NO_x, PAN, aerosol nitrate, and other NO_y species, as well as organic aerosol (OA). Results from the ASP BB parameterization and its implementation into the global atmospheric composition model, GEOS-Chem, will be presented. We will also present initial results from the coupling of ASP v2.1 into the high-resolution Lagrangian plume dispersion model, STILT-Chem, in order to better examine the interactions between BB plume chemistry and dispersion.

R. Chris Owen¹, Andy Hawkins², Leland Villalvazo³, Chris Misinis¹, Roger Brode¹, George Bridgers¹, Doris Jung⁴, and Cleve Holladay⁵

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

²US EPA, Region 7, Lenexa, KS 66219

³San Joaquin Valley APCD, Fresno, CA 93726

⁴Colorado DPHE, Denver, CO 80246

⁵US EPA, Region 9, San Francisco, CA 94105

New Source Review (NSR), including the Prevention of Significant Deterioration (PSD) program, requires proposed new major sources and major modifications to existing sources to demonstrate that any additional emissions will not cause or contribute to a violation of the National Ambient Air Quality Standards (NAAQS). The US EPA's Guideline on Air Quality Models (Appendix W to 40 CFR Part 51) currently specifies AERMOD as the preferred model for projecting near-field dispersion of emissions for most NSR/PSD applications and further specifies a tiered screening approach for nitrogen dioxide (NO₂). Prior to the April 2010 revision of the NO₂ NAAQS, most NSR/PSD impact analyses estimated annual NO₂ impacts using the Tier 1 and 2 screening methods. However, with the new 1-hour NO₂ NAAQS, more facilities are using a Tier 3 approach. The Tier 3 AERMOD approaches, Ozone-Limiting Method (OLM) and Plume Volume Molar Ratio Method (PVMRM), have been considered appropriate for estimating 1-hour NO₂ concentrations, but the two approaches are only considered to be detailed screening methods rather than refined methods.

SCICHEM is a Lagrangian model that simulates transport and dispersion of emissions using three dimensional Gaussian puffs to represent a time-dependent concentration field. Version 3.0b2 has two modes of operation, a limited-chemistry and a full-chemistry mode. The full-chemistry mode requires detailed inputs of three dimensional wind and chemical speciation fields (e.g., gridded output from an Eulerian chemistry model), while the limited-chemistry version can be run with meteorological data at a single location and requires a limited set of spatially uniform background concentrations. The limited chemistry version also supports meteorological and emissions input files that are in AERMOD equivalent formats. Partly due to the simplified input requirements for the limited-chemistry mode, SCICHEM is potentially a suitable model for the purposes of NO₂ modeling as a new Tier 3 modeling approach for NSR/PSD applications.

To date, the limited-chemistry version has received little evaluation. As a result, its performance with NSR/PSD facilities and applications is generally unknown. To begin to address this deficiency, in this study, we evaluate SCICHEM for a number of field studies that have also been used in AERMOD NO₂ evaluations. SCICHEM performance is benchmarked against both the field measurement data and the tiered AERMOD NO₂ modeling techniques. The sensitivity of the models to various input parameters (e.g., the NO₂/NOx in-stack ratio and background ozone) are also tested. The results of this study will provide technical support in the determination of the suitability of SCICHEM to estimate NO₂ concentrations for NSR/PSD applications and help provide guidance for the appropriate application of the model for various source types.

Jorge E. Pach n^1,2, Boris R. Galvis^1,2, Germ n Cabrera², Saravanan Arunachalam³

¹ Universidad de La Salle, Department of Environmental Engineering, Bogota, Colombia

² Universidad de La Salle, Centro Lasallista de Investigaci n y Modelaci n Ambiental, Bogota

³ University of North Carolina, Institute for the Environment, Chapel-Hill, NC

Bogota, Colombia, has experienced rapid economic growth in recent years, which has led to a deterioration of air quality from increased vehicle, industrial and commercial emissions. It is estimated that most of the emissions of NO_x, CO, CO₂, SO₂ and VOCs in the city are from mobile sources, whereas PM is almost equally emitted by mobile and point sources. However, mobile sources are ubiquitous in the city and contribute to air pollution in the Bogota metropolitan area. The local environmental authority in Bogota (SDA) is interested to test emission reduction strategies for mobile sources in main vehicle corridors and assess the impacts on air quality and human exposure. Therefore, the application of a near-road air quality model is desirable.

R-LINE is a new research-level, line-source dispersion model being developed as a part of the ongoing effort to further develop tools for a comprehensive evaluation of air quality impacts in the near-road environment (Snyder et al, 2013). This model is being used in conjunction with traffic activity and primary mobile source emission estimates to model hourly exposures to traffic emissions for several ongoing studies in the U.S., the notable among them being the Near-Road Exposures and Effects of Urban Air Pollutants Study (NEXUS) (Vette et al, 2012). R-LINE has been evaluated against other dispersion models including AERMOD, ADMS-Roads and CALINE (Heist et al, 2013).

This abstract aims to present the application of R-LINE in Bogota, Colombia. We propose to use the methodology described in Cook, et al. (2008) that produces a spatially and temporally resolved mobile source emissions inventory (i.e., hourly emissions for all pollutants modeled, by vehicle class and road link). We will use information from traffic activity provided by the local government and emissions factors generated for every vehicle category in the city in previous studies. Meteorological information will be provided from local monitoring stations. As a result of the application of the dispersion model R-LINE we will have pollutant concentrations, both along the roads, as well as at various distances from the roads due to traffic activity. We will present results from this application as well as evaluation of R-LINE against air quality observations using both fixed-site and mobile monitoring in the region. Emission reduction strategies for mobile sources can then be tested using the R-LINE model to assess the impacts on air quality in Bogota.

References

Cook R, Isakov V, Touma JS, Benjey W, Thurman J, Kinnee E, et al. (2008) Resolving Local-Scale Emissions for Modeling Air Quality near Roadways. J. Air Waste Manage. Assoc. 58:451-61.

Heist D, Isakova V, Perry S, Snyder M, Venkatram A, Hood C, Stocker I, Carruthers D,

Arunachalam S, Owen C. 2013. Estimating near-road pollutant dispersion: A model inter-comparison. Transportation Research Part D: Transport and Environment, Vol 25, Pages 93-105

Snyder M.G., Venkatram A., Heist D.K., Perry S.G., Petersen W.B., Isakov V. (2013) RLINE: A Line Source Dispersion Model for Near-Surface Releases. Atmospheric Environment 77, 748-756.

Vette A, Burke J, Norris G, Landis M, Batterman S, Breen M, et al. (2013) The Near-Road Exposures and Effects of Urban Air Pollutants Study (NEXUS): Study design and methods. Science of the Total Environment 448: 38-47.

RWDI: Jeff Lundgren, Wayne Boulton, Greg Conley, Martin Gauthier, Akhila Wolfe, Carol McClellan

UNC: Zac Adelman, Mohamed Omary

Peel: Louise Aubin, Kim McAdam

The Regional Municipality of Peel (Peel Region) is has a population of more than one million people and is adjacent to the City of Toronto. RWDI was commissioned by the Region of Peel Public Health to develop a flexible and comprehensive modelling and monitoring system to study the impacts of potential emission scenarios to guide policy decisions relating to public health and the Region's growth and sustainability programs.

The system is based on year-long model simulations using WRF/SMOKE/CMAQ. The modeling system is configured with nested 36km, 12km, 4km and 1km resolution grids. The parent 36km domain covers most of Northeastern North America. The inner-most 1km resolution domain covers Peel Region along with the heavily urbanized region along western tip of Lake Ontario, referred to locally as the "Golden Horseshoe". Major urban centres within the 1km domain include the Cities of Toronto, Mississauga, Oakville, Burlington, Hamilton, and St. Catharines.

Meteorological inputs were generated using WRF and emissions processing was performed using SMOKE. Regional emissions inventories from both the US and Canada were adopted, along with MEGAN for biogenic sources. Inline processing was performed for point sources and high resolution spatial surrogates were developed to allocate emissions within the innermost model domain.

A passive monitoring program was also implemented to assist in the evaluation of model performance and inform potential areas of interest for future modelling scenarios.

Results for the first full year of model simulation showing both the positive and negative impacts of adopting a 1km grid resolution are presented, demonstrating the complexity involved in a modeling exercise of this scope and magnitude. Future work will incorporate various emission change scenarios designed to assess potential air quality health management strategies and to inform policy and development within Peel Region.

Yuyan Cui^1,*, Jerome Brioude^1,2, Wayne Angevine^1,2, Gregory J. Frost^1,2, Jeff Peischl^1,2, Thomas Ryerson¹, Steve C. Wofsy³, Gregory W. Santoni³, Eric Kort⁴, Michael Trainer¹

1.Chemical Sciences Division, Earth System Research Laboratory, NOAA, Boulder;

2.Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder;

3.Department of Earth and Planetary Sciences, Harvard University, Cambridge;

4.Department of Atmospheric, Oceanic and Space Sciences, University of Michigan, Ann Arbor.

* NRC postdoc (yuyan.cui@noaa.gov)

Methane (CH₄) is a primary component of natural gas and has a large global warming potential. Statewide missions of CH₄ in California have been found to be greater than expected from population-apportioned bottom-up state inventories. Therefore, there is a critical need to quantify CH₄ in this area to benefit policy, regulation and climate modeling. Previous studies have used trace gas measurements from the NOAA P-3 aircraft during the CalNex field study in 2010 to investigate emission ratios of CH₄/CO and CH₄/CO₂ in the L.A. Basin and to apportion the sources of these gases. In this work, we quantify emissions of CH₄ with an advanced mesoscale inverse modeling system, using the same CalNex measurements in the L.A. Basin as input to the inversion. To simulate the mesoscale atmospheric transport, we used the FLEXPART Lagrangian particle dispersion model driven by meteorological data from three different Weather Research and Forecasting (WRF) model configurations. To determine surface fluxes of CH₄ in the inverse modeling, we used a Bayesian least squares method to invert the observed concentrations measured over time, which we call a 4-D inversion. We used a cluster analysis to aggregate our emission grids in the inverse modeling. We analyzed daytime measurements of six aircraft flights to derive a posterior emission inventory which achieves much better correlation with the measurements (R=0.8) than calculations using the prior inventory (R=0.5). We calculate that the total emission of CH₄ from the L.A. Basin in the inverse system is ~200 Gg CH₄/yr higher than the bottom-up CH₄ state inventory for this region. The results are consistent with previous top-down studies, which indicates our inverse model is robust. Therefore, we used the inverse model to study emissions of CH4 in the Central Valley of California as well in which in which there are inadequate studies. In our inversions, we reported the total emissions for the Sacramento Valley (focusing on rice paddies), the San Joaquin Valley, and entire of the Central Valley. We found there was three times higher in our top-down method in total CH4 emissions for the Central Valley than values from EPA 2005 bottom-up inventory. This presentation will also estimate spatial distributions of CH4 surface fluxes and discuss the significant role of some sources that contributed to the discrepancies between our top-down calculations and bottom-up inventories.

M. J. Alvarado, K. E. Cady-Pereira, M. W. Shephard, J. D. Hegarty, C. R. Lonsdale, D. K. Henze, M. Turner

Atmospheric ammonia (NH₃) is an important precursor of inorganic fine particulate matter (PM_2.5), but the annual total and seasonal cycle of NH₃ emissions are highly uncertain. This uncertainty in the emissions of NH₃ can lead to uncertainty in air quality predictions at local to continental scales. While global SO₂ and NO_x emissions are expected to decrease due to air pollution control efforts, global emissions of NH₃ are expected to stay constant or increase due to increased agricultural production. Thus there is a clear need for global atmospheric NH₃ observations that can improve estimates of NH₃ emissions for air quality studies.

Here we present initial results of a satellite-based retrieval of NH₃ from the Cross-track Infrared Sounder (CrIS) aboard the Suomi-NPP satellite. The retrieval algorithm, based on the operational product from the Aura Tropospheric Emission Spectrometer (TES), will be presented, along with the results of validation studies with in situ NH₃ data from the 2013 NASA DISCOVER-AQ campaign in California's San Joaquin Valley and the NOAA SENEX campaign in the Southeast US. Initial ammonia retrieval results using both simulated and real observations show that CrIS is sensitive to NH₃ in the boundary layer, with peak vertical sensitivity typically around 800 hPa (~2 km). The NH₃ retrievals have a minimum detection limit of ~1 ppbv (peak profile value), and the information content is generally ~1 degrees-of-freedom for signal or less. We will then discuss the potential of these retrievals to constrain NH₃ emissions in these regions and globally.

¹Sindhwani Rati, ¹Goyal P, ¹Kumar Saurabh and ²Kumar Anikender

¹Center for Atmospheric Sciences, Indian Institute of Technology Delhi,

Hauz Khas, Delhi-110016

²National University of Colombia, Bogota, Columbia, USA

Long term exposure of human population to air pollutants and their impact on human health has gained much attention in the recent years. Delhi, a rapidly urbanizing agglomeration has been facing extreme atmospheric pollution mainly due to high levels of respirable suspended particulate matter (PM₁₀), sulphur dioxide and greenhouse gases (CO, NOx, CH₄). The main objective of this work was to estimate criteria pollutant emissions (CO, NOx, SO₂ and PM₁₀) from various sources during 2000-2010 and estimation of future emissions under Business-as-usual (BAU) and emission abatement policy scenarios (EAP) from 2011-2020. Road transport and power plants are still the major emission sources. During the decade 2000-2010, vehicular emissions majorly contributed CO (>70%) and NOx (>65%) whereas contribution of SO₂ and PM₁₀ has reduced due to implementation of EURO-III norms in this sector. In addition to this, under BAU scenario, SO₂ emissions from vehicles are expected to rise by 2.5 times up to 2020. Power plants remained the largest contributors of SO₂ (>75%) and PM₁₀ till 2009. However, due to the closing of Indraprastha (IP) power plant in December 2009, emissions of PM₁₀ and SO₂ have reduced by 30% in 2010. The CO, NOx, SO₂ and PM₁₀ emissions are predicted to reach 75.6 Gg, 90.05 Gg, 234.75 Gg and 57.49 Gg respectively in the BAU scenario whereas with the implementation of EAP scenario, emissions would be abated by 22%, 62%, 99% and 98% respectively by 2020 from the power plants. Domestic, Industrial and Waste sectors analyzed under similar scenarios also produced potentially viable results suggesting that effective implementation of abatement policies would help in making Delhi a better and cleaner megacity.

Melinda Beaver¹, Russell Long¹, Jim Szykman¹, Rachelle Duvall¹, Kirk Baker², James Kelly², Andy Weinheimer³, David Knapp³, Denise Montzka³, Sally Pusede⁴, and Ron Cohen⁴

1-US EPA, Office of Research and Development, RTP, NC

2-US EPA, Office of Air Quality Planning and Standards, RTP, NC

3-National Center for Atmospheric Research, Boulder, CO

4-University of California, Berkeley, CA

NASA's DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) campaign studied the air quality throughout California's San Joaquin Valley (SJV) during January and February of 2013. The SJV is a non-attainment area for EPA's particulate matter (PM2.5) and ozone National Ambient Air Quality Standards. A better understanding of the sources and processes leading to the valley's air pollution levels are needed to support effective emissions control strategy development. The wintertime study was designed to understand the vertical distribution of pollutants including PM, PM precursors, and other trace gases over routine, surface monitoring sites during meteorological conditions known to be conducive to PM2.5 formation. EPA's ORD supplemented an existing meteorological monitoring site at the Visalia municipal airport with O₃, NOx, and NOy measurements. NASA's P3B aircraft conducted 25 missed approaches (reaching altitudes as low as 32m) over this site throughout the campaign, providing unique opportunities to compare surface and aircraft results collected during different times of day and atmospheric conditions. Data from this surface site and others throughout the SJV will be presented and used to explore the performance of EPA's NOx monitoring methods (FEMs and FRMs) and the performance of EPA's CMAQ model during this study.

Cesunica Ivey¹, David Lavoue¹, Aika Davis¹, Yongtao Hu¹, Armistead Russell¹

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

In this work, fire emissions from the Fire Inventory of NCAR (FINN) are implemented in a one-year (2012) simulation of trace gas and fine particulate matter concentrations over continental US with 36-km grid spacing. FINN provides daily emissions for trace gases and aerosol species that are represented in the SAPRC-99 chemical mechanism. Emissions are available for fire events related to agricultural burning, prescribed burning, and wildfires for the years 2002-2013. The WRAP plume rise model is used to generate hourly maximum plume heights for each fire in the modeling domain. Emissions are formatted for use in CMAQ. FINN emissions replace National Emissions Inventory 2011 (NEI2011) emissions for fires, and emissions for all other sources in NEI2011 are retained. The CMAQ model is used to simulate pollutant concentrations, using WRF and NEI/FINN emissions as inputs, the SAPRC-99 chemical mechanism, and Aerosol Module 5. A non-linear optimization approach is used to assimilate modeled results and ground observations from the Chemical Speciation Network. Optimized results are evaluated by comparison to data collected from the Southeastern Aerosol Study, the 2012 Discover-AQ field campaign, and from ground monitoring networks (e.g. Interagency Monitoring of Protected Visual Environments and Southeastern Aerosol Research and Characterization). With these methods, biases and errors introduced by using monthly averaged fire emissions, as traditionally provided by the NEI, are addressed.

Mark Estes, Daniel Johnston, Fernando Mercado, and Jim Smith

Regional background ozone concentrations have been estimated for Houston and Dallas for a multi-year period. This study will examine the meteorological conditions associated with high and low ozone concentrations for the eastern half of Texas. Trends will be analyzed to determine whether background ozone trends differ for different transport patterns. Other meteorological factors such as stability, humidity, and recirculation will also be investigated to determine whether these factors are associated with high regional background ozone and estimated local ozone contributions. The implications of the findings will be discussed in light of recent proposed changes to the ozone standard.

Zhen Liu, Sandia Nat Lab

Joseph Pinto, US EPA

Ray P. Bambha, Sandia Nat Lab

Hope Michelsen, Sandia Nat Lab

Estimates of anthropogenic CH₄ emissions in the US have been largely inconsistent, particularly for oil and gas production (OGP) in the South-Central United States. In this work, we have quantified the contribution of OGP to the South-Central US (TX/OK/KS) CH₄ budget through atmospheric regional transport modeling driven by a process-based, spatially resolved OGP CH₄-emissions inventory sponsored by US EPA. We used the Community Multi-scale Air Quality (CMAQ) model and applied Bayesian inference to evaluate the model against CH₄ measurements at the DOE Southern Great Plains (SGP) central facility. Forward and inverse modeling results suggest that OGP emissions are the largest source of CH₄ observed at the DOE SGP site and the largest source of CH₄ in TX/OK/KS, constituting 2.7 Tg a^-1 (~45%) of total CH₄ emission in the region, or about one half (47%) of national total OGP emissions. Other sources, such as livestock, were found to be less important by comparison. We also found evidence of rapid nocturnal transport by the Great Plains low-level jet (LLJ) and sporadic oil and gas emissions. Our study demonstrates the importance of using improved knowledge for the spatial and temporal features of oil and gas emissions in top-down inversion studies that seek to constrain the CH₄ budget at regional and national scales.

Boris Galvis¹, Jorge E. Pachon¹, Edison Y. Ortiz² , Barron H. Henderson³

¹ Universidad de La Salle, Department of Environmental Engineering, Bogota, Colombia

² Universidad de La Salle, Centro Lasallista de Investigaci n y Modelaci n Ambiental CLIMA

³ University of Florida, Department of Environmental Engineering, Gainsville, FL

Bogota has achieved a significant improvement in its air quality during the last decade, however particulate matter concentrations in some areas of the city are still above WHO guideline values. In this work we analyze the data reported between 2002 and 2012 by Bogota's Air Quality Monitoring Network (RMCAB), using Openair tools. In general we have found that PM₁₀ concentrations have decreased or stabilized in most areas of the city since 2008, except in the industrial district were they are still growing and that mobile sources still play mayor role, especcially diesel vehicles. We have determined that O₃ highest concentrations are measured between 11:00 and 13:00, that they are higher on Sundays when there is much less traffic and are located north east of the city, close to the east mountain ridge. We also found that PM₁₀ and O₃ exceedances have become less common over the analyzed period and that they are more likely to occur on the firsts three months and the last two months of each year. We show that during the months of June, July and August concentrations of all pollutants drop largely because the of increase in wind speed caused by the change of the Inter-Tropical Convergence Zone (ITCZ). We were able to observe that in Bogota between July 2009 and 2010 during "El Ni o", when there was intense solar radiation and dry conditions, there was an increase in the O₃ and PM₁₀ exceedances all over the city. Furthermore, we see a correlated behavior of all pollutants except O₃. We observed a high peak between 07:00 and 11:00 and another softer peak between 18:00 and 22.00 for PM₁₀, PM_2.5, CO, SO₂ and NOx. At monitors that are mostly influenced by emissions from traffic it's possible to observe that Sundays are the days in which the lowest pollutant concentrations are measured.

Our analysis shows that the topography and meteorology of the area significantly benefit air quality conditions but are not the determinant factor in its improvement, which is likely linked to the introduction of lower sulfur fuels among other measures, and that the area is presumably VOC limited.

Weiwei Chen, Daniel Tong, Hang Lei, Li Pan

Soil particles emitted during agricultural activities is a major recurring source contributing to atmospheric aerosol loading. Emission inventories of agricultural dust emissions have been compiled in several regions. These inventories, compiled based on historic survey and activity data, may reflect the current emission strength, which introduces large uncertainties when such datasets are used to drive chemical transport models. In addition, there is no global emission inventory of agricultural dust emissions required to support global air quality and climate modeling. In this study, we present our recent efforts to develop a global emission inventory of PM₁₀ and PM_2.5 released from field tillage and harvesting operations using an emission factors-based approach. Both major crops (e.g., wheat and corn) and forage production are considered here. For each crop or forage, information of crop area, crop calendar, farming activities and emission factors of specified operations are assembled. The key issue of inventory compilation is the choice of suitable emission factors for specified operations over different parts of the world. Through careful review of published emission factors, we modified the traditional emission factor-based model by multiplying two coefficient factors to reflect the relationship between emission factors and soil texture and climate conditions. Then, the temporal (i.e., monthly) and spatial (i.e., 0.5 resolution) distribution of agricultural PM₁₀ and PM_2.5 emissions from each and all operations are estimated for each crop or forage. Finally, the emissions from individual crops are aggregated to assemble a global inventory from agricultural operations. The inventory is verified by comparing the new data with the existing agricultural fugitive dust inventory in North America and Europe, as well as the satellite observations of anthropogenic agricultural dust emissions.

Linda Hembeck, University of Maryland, College Park, MD

Christopher Loughner, NASA Goddard Space Flight Center

Timothy Canty, University of Maryland, College Park, MD

Russ Dickerson, University of Maryland, College Park, MD

Ross Salawitch, University of Maryland, College Park, MD

Elevated levels of tropospheric ozone caused by anthropogenic emissions of NOx and VOCs have a negative impact on human health and crops. Observations show surface ozone is enhanced at high temperatures due to a combination of meteorological and chemical factors. To make informed regulatory decisions on how to reduce surface ozone in the Baltimore-Washington region, a thorough understanding of urban plume chemistry is necessary. The ozone production efficiency (OPE), which is based on the observed ratio of O₃ and various nitrogen species, provides a mechanism for quantitatively assessing air quality representation of a key component of the photochemical evolution of urban plumes. A comprehensive set of atmospheric observations for which OPE can be found is available from NASA's DISCOVER-AQ campaign for July 2011 in the Baltimore-Washington region. We compare the abundance of ozone precursors and OPE inferred from these observations to values from the Community Multiscale Air Quality (CMAQ) model, to assess the representation ozone-related photochemistry within the model. Preliminary results show that the OPE as well as the NOx/NOy ratio in the Baltimore-Washington region derived from measurements is twice as high as within CMAQ, and that isoprene and formaldehyde are too low within CMAQ. Implications for policy will be briefly discussed.

Bernhard Rappengl ck¹, Graciela Lubertino², Sergio Alvarez¹, Julia Golovko¹, Beata Czader¹, Luis Ackermann¹

(1) Dept. of Earth and Atmospheric Sciences, Univ. of Houston, Houston, TX, USA

(2) Houston-Galveston Area Council, Houston, TX, USA

Nitrous acid (HONO) and formaldehyde (HCHO) are important precursors for radicals and are believed to favor ozone formation significantly. Traffic emissions data for both compounds is scarce and mostly outdated. A better knowledge of today's HCHO and HONO emissions related to traffic is needed to refine air quality models. Here we report results from continuous ambient air measurements taken at an Highway Junction in Houston/Texas from July 15 - October 15, 2009. The observational data was compared to emission estimates from currently available mobile emissions models (MOBILE6; MOVES). Obervations indicated a molar CO versus NO_x ratio of 6.01 0.15 (r² = 0.91), which is in agreement with other field studies. Both, MOBILE6 and MOVES, overestimate this emission ratio by 92% and 24%, respectively. For HCHO/CO an overall slope of 3.14 0.14 g HCHO / kg CO was observed. While MOBILE6 largely underestimates this ratio by 77%, MOVES calculates somewhat higher HCHO/CO ratios (1.87) than MOBILE6, but is still significantly lower than the observed ratio. MOVES shows high HCHO/CO ratios during the early morning hours due to heavy duty diesel off-network emissions. The differences of the modeled CO/NO_x and HCHO/CO ratios are largely due to higher NO_x and HCHO emissions in MOVES (30% and 57%, respectively, increased from MOBILE6 for 2009), as CO emissions were about the same in both models. The observed HONO/NO_x emission ratio is around 0.017 0.0009 kg HONO / kg NO_x which is twice as high as in MOVES. The observed NO₂/NO_x emission ratio is around 0.16 0.01 kg NO₂ / kg NO_x, which is a bit more than 50% higher than in MOVES. MOVES overestimates the CO/CO₂ emission ratio by a factor of 3 compared with the observations, which is 0.0033 0.0002 kg CO / kg CO₂. This as well as CO/NO_x overestimation is coming from light duty gasoline vehicles.

Hyun Cheol Kim ^1,2, Pius Lee¹, Li Pan ^1,2, Laura Judd ³, Daniel Tong ^1,2,3, Youhua Tang ^1,2, Tianfeong Chai ^1,2, Barry Lefer³, and Ivanka Stajner ⁴

¹ NOAA/Air Resources Laboratory, College Park, MD

² UMD/Cooperative Institute for Climate and Satellites, College Park, MD

³ University of Houston, Dept. of Earth and Atmospheric Sciences, Houston, TX

⁴ NOAA/National Weather Service, Silver Spring, MD

National Oceanic and Atmospheric Administration (NOAA) National Air Quality Forecast Capability (NAQFC) NO₂ surface concentration and tropospheric vertical column density (VCD) during 2009-2014 are compared with multiple observations including satellites, surface observations and field campaign in-situ measurements, to investigate the performance of model and current emission inventory, and multi-year trends of NO₂ fields over Contiguous United States. We conduct comparisons for (1) NO₂ VCDs between NAQFC and satellites, (2) surface NO₂ concentrations between NAQFC and US Environmental Protection Agency (EPA) Air Quality System (AQS), (3) surface NO₂ concentrations between AQS and satellite-NAQFC hybrid approach, (4) Continuous Emissions Monitoring System (CEMS) measurements at plant stacks and corresponding NAQFC NO₂ concentrations, and (5) NO₂ concentration and VCDs from field campaigns and NAQFC and/or satellites. Satellite-based NO₂ VCD measurements from the Scanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY), the Ozone Monitoring Instrument (OMI) and the Global Ozone Monitoring Experiment-2 (GOME-2), surface measurements from US EPA AQS sites, and California Nexus (CalNex, California, 2010) and Deriving Information on Surface Conditions from Column and VERtically Resolved Observations Relevant to Air Quality (DISCOVER-AQ, Baltimore-Washington D.C., 2011; California, 2013; Houston, 2013) campaigns are utilized. For satellite data comparisons, the vertical and horizontal information are greatly improved using Averaging Kernel and conservative downscaling method, respectively. We demonstrate satellite NO₂ measurements near urban cities always show serious underestimation due to coarse satellite footprint pixel sizes compared to typical urban NO₂ spatial gradients, and these biases can be resolved by using the conservative downscaling technique. Preliminary results show that NAQFC NO₂ VCD is considerably overestimated at several locations (e.g. Los Angeles, Houston and New Orleans) while good agreements are seen in other mega cities, especially in New York. OMI and P₃ aircraft NO₂ VCD comparison during 2010 CalNex also shows an excellent agreement when the resolution effect is considered, with the correlation improved from R=0.45 to R=0.88.

Darrell Sonntag, Harvey Michaels, David Brzezinski, David Choi, Catherine Yanca, Alexis Zubrow, Alison Eyth

MOVES is the Motor Vehicle Emissions model developed by the US EPA used to estimate emission inventories from on-road mobile emission sources. MOVES is used by state and local users for state implementation plans and conformity analysis, for EPA rule makings, and the National Emission Inventory. MOVES2014 is a significant update to the previous public version, MOVES2010b. This presentation will provide an overview of the impacts MOVES2014 has on mobile-source emission inventories.

Pius Lee 1*, Greg Carmichael2*, Brad Pierce3*, Arastoo Pour Biazar4, Dick McNider4*, Ted Russell5*, Yang Liu6*, Talat Odman5, Yongtao Hu5, David Edwards7*, and Edward Hyer8*
1. Air Resource Laboratory (ARL), NOAA, College Park, MD
2. College of Engineering, University of Iowa, Iowa City, IA
3. National Environmental Satellite and Information Service (NESDIS), Madison, WI
4. Department of Atmospheric Science, University Alabama, Huntsville AL
5. School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA
6. Department of Environmental Health, Emory University, Atlanta, GA
7. Corporation for Atmospheric Research, Boulder, CO
8. Naval Research Laboratory, Monterey, CA

This work reports the progress of the regional chemical reanalysis project over the Conterminous U. S. we announced to the CMAS community last year. It is a major undertaking NOAA/ARL is committing to sustain. The project leverages on the modeling, observation, and data assimilation expertise of the NASA Air Quality Applied Science Team (AQAST). It also leverages on our decade long forecasting experience with the National Air Quality Forecasting Capability (NAQFC). The reanalysis modeling system uses a much refined vertical layering than that used in NAQFC. It has added a total of 20 layers to near the surface, hundreds of meters below and above a fully developed planetary boundary layer (PBL) and near the spring time tropopause. We had tested and quantified statistically the incremental improvements in reproducing the chemical state of the atmosphere when observation data sets are incorporated into the reanalysis modeling system: (1) Microwave Lime Sounder retrieved stratospheric ozone and satellite observed AOD derived extra-domain wild fire plumes ingested via dynamic later boundary conditions derived from the RAQMS global model;(2) incorporation of intermittent emission from the NOAA Hazard Mapping System satellite-observed wild fires; (3) Satellite observed cloud transmissivity to adjust photolytic rates; (4) Satellite observed AOD to constrain particulate matter constituents; and (5) surface monitor observations on ozone and particulate matter concentrations to constrain ozone and particulate matter concentrations within the PBL. The incremental improvement of the statistics based on a standard metric across these 5 stages of the construction of the reanalysis system will be discussed.

William P. L. Carter

College of Engineering, Center for Environmental Research and Technology, University of California, Riverside, CA 92521, USA

The chemical speciation database previously developed for making model species assignments needed when processing organic gas emissions input for atmospheric models has been updated and enhanced to process speciation data from a wider variety of sources and for additional chemical mechanisms. Over 1700 chemical compounds are assigned to model species 12 different gas-phase chemical mechanisms, including versions of Carbon Bond, RADM2 and RACM2, versions of SAPRC and the MCM. It covers over 3000 chemical categories used in five different anthropogenic TOG profile databases, including Speciate 4.3 and the profiles maintained by California and Texas. In addition, the updated database now has assignments for biogenic emissions output by BEIS and Megan models, correcting errors and limitations in the undocumented biogenic assignments that are currently used. This involved developing a unified chemical classification system, assigning compounds to mixtures, assigning model species for the mechanisms to the compounds, and also making assignments for unknown, unassigned, and nonvolatile mass. The comprehensiveness of the assignments, the contributions of various types of speciation categories to current speciation profiles and total emissions data, inconsistencies with existing undocumented model species assignments, and remaining speciation issues and areas of needed work will be discussed.

It is recommended that this database, or something like it, always be used when processing speciation data for air quality models, especially when profiles or chemical mechanisms are updated or changed. Otherwise emissions input data may not be represented consistently using different mechanisms, or in a manner consistent with the intentions of the mechanism and profile developers. To facilitate its more widespread use and ongoing maintenance, it is proposed that such a database be added to the list of available CMAS modeling tools.

Christopher P. Loughner (Earth System Science Interdisciplinary Center - University of Maryland), Melanie Follette-Cook (Morgan State University), Kenneth E. Pickering (NASA Goddard Space Flight Center), and Mark Estes (Texas Commission on Environmental Quality)

While bay breeze circulations were a daily occurrence during the Texas DISCOVER-AQ field campaign, only toward the end of the month-long campaign did these local-scale circulation patterns contribute to extremely high surface ozone concentrations. The highest ozone air pollution episode in the Houston, TX region in 2013 occurred September 24-26. The maximum 8-hour average ozone peaked on September 25 at La Porte Sylvan Beach, reaching 124 ppbv, almost 50 ppbv above the current EPA standard of 75 ppbv. The September 24-26 air pollution episode was the only time during the DISCOVER-AQ field campaign when Houston was influenced by transport from the north, resulting in higher background ozone to be transported over the Houston metropolitan area instead of cleaner air from the Gulf of Mexico. In addition, the bay breeze circulation caused pollutants that were transported over the water in the morning to recirculate back inland and converge with pollutants over land at the bay breeze convergence zone. The highest surface ozone concentrations were reported near the bay breeze front at La Porte Sylvan Beach. This ozone episode will be presented using a CMAQ simulation and measurements made during the DISCOVER-AQ field campaign.

Alison Eyth, Rich Mason, Alexis Zubrow, Jesse Bash, Zac Adelman, Michele Jimenez, Regi Oommen, Martinus Wolf

EPA has developed an updated emissions modeling platform for 2011 based on Version 2 of the 2011 National Emissions Inventory (NEI). Aside from the incorporation of newer version of the NEI, a number of key updates to the modeling platform have been made. Updated emissions for Canada and Mexico have been obtained, along with updated biogenic emissions. Improved spatial surrogates for extended idling of long-haul trucks and oil and gas emissions were developed. New speciation profiles were incorporated along with improvements to speciation cross references. Temporal profiles for hourly emissions at small airports and Electric Generation Utility (EGU) emissions were refined. In addition, methods to correct Continuous Emissions Monitoring System (CEMS) data were incorporated, along with better handling of partial year reporters. Future-year projection methods for non-EGU emissions and mobile-source activity data were also improved.

Kenneth Pickering, James Crawford, Frank Flocke, Gabriele Pfister

The fourth deployment of the NASA Earth Venture-1 DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) field program took place from mid-July to mid-August 2014 in the Front Range region of Colorado (from south of Denver northward to Ft. Collins and Greeley). The primary DISCOVER-AQ objective was to collect data sets to allow improved interpretation of satellite data for air quality application. The Front Range Air Pollution and Photochemistry Experiment (FRAPP), led by NCAR, operated in collaboration with DISCOVER-AQ, but covered a somewhat larger region. The Front Range region often experiences high surface ozone concentrations in the summer. DISCOVER-AQ conducted lower tropospheric profiles of trace gases (e.g., NO2, O3, HCHO, CO) and aerosol types, optical, and microphysical properties on the P-3B aircraft three times each flight day over a set of Colorado Department of Public Health and Environment and NOAA air quality monitoring stations, which were upgraded to include Aeronet sun photometers and Pandora UV/Vis spectrometers. A King Air aircraft conducted remote sensing with the High Spectral Resolution Lidar (HSRL) for aerosols and the Airborne Compact Atmospheric Mapper (ACAM) instrument for trace gases. FRAPP operated the C-130 aircraft, instrumented for an extensive suite of trace gas and aerosol observations, to examine inflow and outflow to/from the Front Range, upslope and downslope pollutant transport, and emissions from oil/gas extraction activities. Both experiments contributed many additional ground-site instruments and platforms, such as lidars, tethered balloons, ozonesondes, and mobile vans. The two experiments offer copious data for evaluation of regional meteorological and air quality models.

Daniel Q. Tong^1,2,3*, Lok Lamsal^4,5, Li Pan^1,2, Charles Ding^1,6, Hyuncheol Kim^1,2, Pius Lee¹, Tianfeng Chai^1,2, Kenneth E. Pickering⁵, and Ivanka Stajner⁷

1 NOAA Air Resources Laboratory (ARL), NOAA Center for Weather and Climate Prediction, 5830 University Research Court College Park, MD 20740, USA

2 Cooperative Institute for Climate and Satellites, University of Maryland, College Park, MD 20740, USA

3 Center for Spatial Information Science and Systems (CSISS), George Mason University, Fairfax VA 22030, USA

4 Goddard Earth Sciences Technology and Research, Universities Space Research Association, Columbia, MD, USA

5 NASA Goddard Space Flight Center, Greenbelt, MD, USA

6 Now at University of California, Berkeley, CA, USA

7 NOAA National Weather Service, Silver Spring, MD, USA

Over the past decade, NO_x emissions in the United States have experienced dramatic changes, due to both emission control and the 2008 Global Economic Recession. The NOx trends over eight US most populous cities ("megacities") are examined using the NO2 observations from NASA Ozone Monitoring Instrument (OMI) (NO2 column) and the US EPA Air Quality System ground network (NO2 surface concentration). Both OMI and AQS observed consistent downward trends over these metropolitan areas from 2005 to 2012, with an average of -5.0% from OMI and -5.3% from AQS. During this period, the NOx changing rate is distinctly different at different stage of the recession. Satellite observed that the average reduction rates are -7.3%/yr, -9.2%/yr and -2.8%/yr prior to, during and after the recessions, respectively, which are comparable to the -6.0%/yr, -10.8%/yr, and -3.4%/yr for each period derived from AQS ground network. Finally, we will present preliminary results of using the OMI and AQS observed NOx trends to evaluate the NOx emission inventories and projection, including the newly released NEI2011s and the Cross-State Air Pollution Rule projection, as well as the operational emission data used to support the NOAA National Air Quality Forecasting Capability system.

Heather Simon, Adam Reff, Benjamin Wells, Neil Frank

In this work, we evaluate ambient ozone trends at urban, suburban, and rural monitoring sites across the United States over a period of decreasing NOx emissions (1998-2011). We find that at sites with statistically significant trends, 5^th percentile ozone concentrations are generally increasing by 0.5-1 ppb/year and 95^th percentile ozone concentrations are generally decreasing by 1-2 ppb/year. Trends in mid-range ozone concentrations are more varied, with more urbanized sites being more likely to show increasing trends and less urbanized sites being more likely to show decreasing trends. In heavily urbanized areas with high NOx emissions, ozone may be suppressed due to NOx titration, while the highest concentrations are often reported in downwind suburban and rural areas. No seasonal pattern was detected in these data. In general, these results indicate that as anthropogenic NOx emissions have decreased, the ozone distribution has been compressed, leading to less spatial and temporal variability in high and low percentile concentrations.

Evan Couzo^1,2, James McCann¹, Nick Johnson³, Clayton Barrows³, Seth Blumsack³, William Vizuete¹, J. JasonWest¹

¹University of North Carolina at Chapel Hill

²now at Massachusetts Institute of Technology

³Pennsylvania State University

Coal-fired power plants account for nearly 20% of all U.S. anthropogenic sources of nitrogen oxides (NOx), an important precursor for tropospheric ozone (O₃). Here, we illustrate how daily electrical grid management decisions can incorporate information from air quality forecasts in an attempt to avoid daily O₃ exceedances by shifting the location of electricity generation. Temporarily redispatching generation away from forecasted regions of high O₃ has the potential to reduce the costs associated with traditional emissions control strategies while still achieving pollution reductions. We use the direct decoupled method integrated sensitivity analysis tool in CAMx to estimate the sensitivity of 8-hr O₃ to NOx emissions from 80 individual power plants; those sensitivities are then used to estimate the least-cost means of reducing high O₃.

We illustrate this dynamic management system in the eastern US on two days in August, 2005, for which modeled ozone in Pittsburgh PA exceeded the 8-hr. standard. We examined several urban regions near Pittsburgh and showed that at most 27 individual power plants had a measurable impact on O₃ concentrations, and, in fact, there were instances where just one or two plants contributed the majority of the O₃ sensitivity. After identifying the plants whose NOx emissions contribute most to high 8-hr O₃, we used our calculated sensitivities in an electric power grid dispatch model to evaluate different decision rules that electric system operators might use to respond to predicted O₃ exceedances by shifting power generation (and thus NOx emissions) away from some power plants to others with spare capacity. One decision rule would be to shut down power plants with the highest absolute sensitivity, while another would be to take a cost-based approach and shut down power plants with the highest O₃ reductions per dollar of substitute electricity. An updated set of NOx emissions was developed to reflect the new spatial allocation of electricity generation. Since some power plants identified by CAMx's sensitivity tools were shut down, others needed to compensate for the reduced generation. CAMx simulations were then rerun with the updated power plant emissions. Rerunning the air quality model allowed us to (1) determine the reduction of peak O₃ in the location of concern as well as increases elsewhere, and (2) test how the integrated sensitivity tools in CAMx perform for a combination of decreases and increases among multiple power plants.

This coordinated modeling system provides a framework for an adaptive air quality management system that dynamically targets individual power plants. This, in effect, reduces emissions from critical sources without requiring costly investment in smokestack controls, thereby eliminating much of the cost associated with traditional pollution abatement strategies.

Brian Keaveny, NESCAUM; Jason Rudokas, NESCAUM

In this presentation we will present a multi-pollutant framework for examining the impacts of climate and air quality policies on emissions trends through 2050. The framework makes use of the EPA 9-region MARKAL energy system model (US9r), a technologically detailed multi-sector model of the nation's energy infrastructure. In our talk, we will detail how models such as US9r can directly inform multi-pollutant analysis conducted with CMAQ. This will be illustrated through a review of an analysis NESCAUM recently conducted using the modeling framework to examine the interaction between climate and air quality scenarios at the regional and national level.

The analysis investigates the projected impact of six climate mitigation measures on U.S. emissions of carbon dioxide (CO₂), sulfur dioxide (SO₂), and nitrogen oxides (NOx) associated with energy use in major sectors of the U.S. economy. Specifically, we analyze two carbon tax scenarios, two low carbon transportation scenarios, and two biomass fuel choice scenarios as climate mitigation measures. While these are climate strategies, they can also affect air quality through changes in precursor emissions that form ozone and particulate matter (PM) pollution in the atmosphere. Changes in emissions of CO₂, SO₂, and NOx under each climate mitigation scenario are compared to 2005 emissions and to a common projected reference case through 2050 that assumes a number of pre-existing rules and policies will be in place during this period. The projected air quality impacts of each scenario were also used to perform a broader photochemical modeling exercise that examined how each scenario impacted ambient air quality.

Alexander Cohan

Mark Janssen

The Eastern Regional Technical Advisory Committee (ERTAC) community has developed an emissions projection tool for electrical generation units (EGU). The ERTAC EGU model is designed to support photochemical modeling and development of state implementation plans (SIP). It has discrete handling of hourly CEM data, hourly projections, and hour specific applications of control and growth.

This study examines the application of the ERTAC EGU model to assess the effectiveness of future year control scenarios. 2018 EGU emissions projections from ERTAC and the Integrated Planning Model (IPM) are contrasted. Photochemical modeling results using CAMx will be presented.

Julia Gamas

Rebecca Dodder

Dan Loughlin

Cynthia Gage

The environment and its interaction with human systems(economic, social and political) is complex and dynamic. Key drivers may disrupt system dynamics in unforeseen ways, making it difficult to predict future conditions precisely. This kind of deep uncertainty presents a challenge to organizations faced with making decisions about the future, including those involved in air quality management. Scenario analysis is an important tool that can benefit decision-makers under these conditions. We propose the application of the future scenarios method to air quality management because it provides a structured means of sifting through and understanding an overwhelming number of driving forces and their dynamics, helping to identify robust policy choices within deeply uncertain systems. Clarity about relevant driving forces can assist in the choice of emission reduction strategies to anticipate cost-effective outcomes and avoid missed opportunities. The application of the scenarios method is a novel approach for air quality management for the U.S. We use the MARKAL model to illustrate the scenarios and gain further insights into the role of different drivers and their consequences for air quality, as well as to understand the robustness of different emission reductions strategies.

K. Wyat Appel, Robert C. Gilliam, Jonathan E. Pleim, George A. Pouliot, Daiwen Kang, Christian Hogrefe, Shawn J. Roselle, Rohit Mathur and David C. Wong

At the 12^th Annual CMAS Conference, initial results from the application of the coupled WRF-CMAQ modeling system to the 2011 Baltimore-Washington D.C. DISCOVER-AQ campaign were presented with the focus on updates and new methods applied to the WRF modeling for fine-scale applications. Simulations have been performed with an updated version of the coupled WRF-CMAQ modeling system at 12-km, 4-km and 1-km horizontal grid spacing over the Baltimore-Washington D.C. region for July 2011. For this presentation, a detailed evaluation of the air quality will be presented, with an emphasis on comparisons to the unique measurements (e.g., ship and aircraft) that were taken during the DISCOVER-AQ campaign. Specific measurements included measurements made by ship over the Chesapeake Bay, NASA P-3B and NASA UC12 aircraft, tethersonde, and non-routine ground-based measurements.

Jian Sun, Joshua S. Fu

A control cost analysis tool, named CoST-CE, is developed and serves as a component of the Air Benefit and Attainment Assessment and Cost Analysis System in China (ABaCAS-China), which is aimed at providing the policy makers and scientists in China with a user-friendly system framework for conducting integrated estimate of emission control cost and its associated air quality attainment and health benefits. Currently the CoST-CE tool is available for the assessment of pollutant (i.e. SO₂, NO_X and PM_2.5) emission abatement cost at the power sectors of Yangtze River Delta (YRD) region in China and the output can be linked to another component of ABaCAS, the Response Surface Model (resulting from RSM-CMAQ)- Visualization and Analysis Tool (RSM-VAT), to evaluate the impact of emission control on the air pollutant concentration like PM_2.5 and ozone in the atmosphere during January and August, 2010. The results shown that for the monthly averaged daily maximum ozone concentration, the control effect of NO_X mainly focuses on the northern part of Zhejiang Province, southern part of Jiangsu Province and the west of Shanghai city. However, the ozone concentration increases slightly along with the reduction of NO_X, due to the fact of VOCs limited situation in YRD region. For the monthly averaged PM_2.5, no evident reduction effect is observed for this region and it could possibly be contributed by the PM abatement equipment that has already been installed at the power plants. Therefore, the installation of even higher efficiency control technology may not influence the PM_2.5 concentration significantly. Another factor is that the dominant NO_X and PM_2.5 emission source in the YRD region is from industry, domestic life and transportation, according to the emission inventory. These results provide a general idea of benefit balance between emission control and air quality improvement. Policy and decision makers can utilize these tools to determine whether it is necessary to invest on certain pollutant abatement and which emission source should be controlled cost effectively.

Sharon Douglas, Jay Haney, Tom Myers, Yihua Wei and Belle Hudischewskyj, ICF

Iyad Kheirbek, New York City Department of Health and Mental Hygiene (DOHMH)

This air quality modeling study for New York City was designed to examine and quantify the effects of changes in heating oil and fuel use in the power sector on air quality in NYC. Key components of the assessment included the preparation of meteorological inputs for a base year of 2008, preparation of emission inputs for the base year and the various alternative emission scenarios, application of the Community Multiscale Air Quality (CMAQ) model, model performance evaluation, and assessment of the air quality impacts/benefits.

The air quality assessment focused on the following criteria pollutants: ozone, fine particulate matter (PM_2.5), nitrogen dioxide (NO₂), and sulfur dioxide (SO₂). The analysis was specifically designed to examine the benefits at the local scale for New York City (NYC) neighborhoods. To achieve this, the modeling was conducted using a high-resolution modeling grid with 1-kilometer (1-km) spacing.

Meteorological input fields for the CMAQ model for the NYC air quality assessment were prepared using the WRF meteorological model. Both the WRF and CMAQ models were applied for an annual simulation and for a modeling domain that includes four nested grids with approximately 45-, 15-, 5-, and 1-kilometer (km) horizontal resolution. Good model performance was achieved for both WRF and the CMAQ models.

CMAQ was also applied for a variety of scenarios reflecting changes in fuel-related emissions. The modeling results indicate that changes in heating oil use that have occurred since 2010 have resulted in improved air quality in New York City, with large reductions (on the order of 20 and 65 percent, respectively) in PM_2.5 and SO₂ concentrations. Changes in fuel use in the electric power generation sector since 2005 have also resulted in reductions of 2 to 4 percent in ozone, PM_2.5 and SO₂ concentrations.

This study was sponsored by the New York City Department of Health and Mental Hygiene (DOHMH) and Mayor's Office of Long Term Planning and Sustainability (OLTPS). The modeling results are being used to quantify the air quality and public health benefits attributable to recent changes in fuel use in the heating and power sectors.

J. Jason West, Steven J. Smith, Raquel A. Silva, Vaishali Naik, Yuqiang Zhang, Zachariah Adelman, Meridith M. Fry, Susan Anenberg, Larry W. Horowitz, Jean-Francois Lamarque

Actions to reduce greenhouse gas (GHG) emissions often reduce co-emitted air pollutants, with co-benefits for air quality and human health. Here we present the co-benefits of global GHG mitigation for global air quality and human health, estimated for the first time using a global atmospheric model, via two mechanisms: reducing co-emitted air pollutants, and slowing climate change and its influence on air quality. RCP4.5 is analyzed as a global mitigation scenario relative to its associated reference scenario, and we find that this mitigation avoids 0.5 0.2, 1.3 0.5, and 2.2 0.8 million air pollution related deaths in 2030, 2050, and 2100. When monetized, these health benefits are US$50-380 per tonne CO2, which exceeds the GHG mitigation costs in 2030 and 2050, and are within the low range of costs in 2100. We also find that the co-benefits of reducing co-emitted air pollutants far exceed the co-benefits via slowing climate change. Results from our continued work to downscale these co-benefits results to the United States will also be presented, allowing understanding of co-benefits at finer spatial resolution and of the co-benefits from domestic versus foreign GHG reductions.

James T Kelly¹, Kirk R Baker¹, and Sergey L Napelenok²

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

²US EPA, Office of Research and Development, RTP NC 27711

Precursor emissions from single sources can lead to the formation of secondary pollutants such as ozone and secondary PM_2.5. Estimates of single-source impacts on air quality are useful in a variety of regulatory applications related to New Source Review, Prevention of Significant Deterioration, and National Environmental Policy Act provisions. A range of models could potentially be applied to characterize single-source secondary impacts for regulatory applications. However, each has limitations that would need to be considered in a single-source assessment. While Eulerian gridded photochemical models may overly dilute a plume in its early stages, these models have the advantage of being thoroughly tested and vetted for use in regulatory studies focused on secondary pollutant formation. Eulerian photochemical model predictions of ozone formation have also been found to be in reasonable agreement with limited available aircraft observations (e.g., see Baker talk, this session).

In this study, we estimate the impacts of emissions from hypothetical sources on air quality for short time periods (~10 days) in winter and summer 2007 using the Community Mulitscale Air Quality (CMAQ) model version 5.0.2 with 4-km horizontal grid resolution. Three hypothetical sources are considered in each of two modeling domains, one over the South Coast Air Basin and one over the San Joaquin Valley of California. These regions were selected because of their conduciveness to secondary PM_2.5 and ozone formation. Emission levels of 100 tons/yr and 500 tons/yr are modeled for stack conditions designed for near-surface and aloft releases. The impacts of the sources are estimated using the brute-force (B-F) method of zeroing out emissions in a second simulation and by extrapolating direct decoupled method (DDM) sensitivities from the source emissions level to the zero-out emissions level. Emissions scenarios for individual releases of NOx, VOC, NH₃, SO₂, and EC as well as mixed releases of NOx-VOC, NOx-NH₃, and VOC-NH₃ are modeled.

Comparisons of the hourly impacts of source emissions on PM_2.5 and ozone were made for DDM vs. B-F approaches, 100 ton/yr vs. 500 ton/yr emissions levels, near-surface vs. aloft releases, and individual pollutant vs. mixed pollutant releases. The maximum impacts of emissions during the simulation periods on maximum daily average 8-hr ozone and 24-hr average PM_2.5 were also calculated domain wide and as a function of distance from the source. The presentation will focus on key findings with an emphasis on understanding the potential suitability of these techniques for use in typical regulatory applications.

Kristen Brown, Daven Henze, Jana Milford

This study uses the US EPA nine-region MARKAL model to examine the effect of internalizing externalities in the cost of energy. MARKAL represents energy use and emissions in the industrial, residential, commercial, electric, and transportation sectors across the US from 2005-2055, and determines the lowest cost way to satisfy demand for energy services. Fees are applied to lifecycle emissions in MARKAL across all sectors starting in 2015. These fees are based on estimated external costs, or damages, from the literature, with a range of values considered for greenhouse gases (CO₂, CH₄) and criteria pollutants (NO_x, SO₂, PM₁₀, PM_2.5, VOCs). Resulting changes in the energy system are compared to those in in a no-fee base case. Changes in fuel use, emissions controls, and emissions produced under various fees are examined. The total cost to society should be low because costs shift from those incurred as consequences of poor air quality or changing climate to costs associated with reducing emissions.

Imposing fees on criteria pollutants spurs installation of more stringent emissions control devices, whereas greenhouse gas fees are more likely to induce fuel switching. With GHG fees, use of biomass based fuels increases, while with all fees use of coal decreases compared to the base case. With high criteria pollutant fees coal is used 19% less than the base case in 2040 while with mid-range GHG fees, coal is used 46% less. Natural gas use increases for all fee cases except the very highest greenhouse gas fee modeled. Electricity from wind power increases for all fees. Total fuel use is reduced by 2% from the base case in 2040 with high criteria pollutant fees, and 5% with mid-range GHG fees.

Wenxian Zhang, Erin E. Tullos, Yongtao Hu, Athanasios Nenes, Armistead G. Russell

D. H. Loughlin¹, L. Ran², D. Yang², Z. Adelman², B. H. Baek² and C. G. Nolte¹

¹US Environmental Protection Agency, Office of Research and Development, 109 T.W. Alexander Drive, Research Triangle Park, NC 27711, USA

²University of North Carolina at Chapel Hill, Institute for the Environment, 137 E. Franklin St., Chapel Hill, NC 27514, USA

The Emission Scenario Projection (ESP) method is used to develop multi-decadal projections of U.S. Greenhouse Gas (GHG) and criteria pollutant emissions. The resulting future-year emissions can then be translated into an emissions inventory and applied in climate and air quality modeling. ESP projections accommodate different assumptions about a wide range of emission drivers, including population growth and migration, economic growth and transformation, technological change, land use change, and current and potential policies (e.g., air quality, energy, and climate). At the heart of the ESP method is the MARKet ALlocation (MARKAL) energy system optimization model. Optimization allows the option of prescriptive application, identifying least cost technological and fuel choices that meet air quality and climate management goals simultaneously. The method (ESP 1.0) was first described in a 2011 publication. Since then, it has continued to undergo improvements, including expanded pollutant coverage and a new MARKAL database that reflects recent energy system developments (e.g., increased natural gas supplies and successful commercial introduction of electric vehicles). In addition, ESP 2.0 includes the ability to update the spatial distribution of emission projections to account for population and land use changes, as predicted by the Integrated Climate and Land Use Scenarios (ICLUS) model. In this presentation, we provide an overview of ESP 2.0 and present an illustrative emission projection for 2050, with a particular focus on evaluating the effects of spatially redistributing emissions.

Matthew Woody, Saravanan Arunachalam, J. Jason West, Hsi-Wu Wong

Aviation activities represent an important and unique mode of transportation, but also impact air quality due to gaseous and particulate emissions. In this study, we aim to quantify the impact of aircraft on air quality, focusing on aircraft-attributable PM_2.5 at scales ranging from local (aircraft engine exhaust) to continental (spanning hundreds of kilometers). Aircraft emissions from the Aviation Environmental Design Tool (AEDT) are modeled using the CMAQ-Advanced Plume Treatment (CMAQ-APT) model. In our CMAQ-APT simulations, aircraft emissions from 99 major U.S. airports are modeled over the continental U.S. in January and July using the CB05 chemical mechanism. CMAQ-APT includes a plume-scale treatment of source emissions that are modeled (in this case, aircraft emissions) and in addition to this treatment, we include a number of additional modeling enhancements for predictions of aircraft-attributable PM_2.5. These enhancements, which are based on recent measurements and smog chamber data, include: 1) replacing non-volatile POA aircraft emissions with semi- and intermediate volatile organic compounds (S/IVOCs) mapped to the volatility basis set (VBS), 2) accounting for the formation of non-traditional secondary organic aerosols (NTSOA) from the oxidation of aircraft S/IVOC emissions, and 3) utilizing alternative emission estimates from the Aerosol Dynamics Simulation Code (ADSC), a 1-D plume scale model that estimates engine specific PM and S/IVOC emissions at ambient conditions (accounting for relative humidity and temperature). We performed over 2,300 ADSC simulations - that spanned 6 engines by varying engine thrust (6 levels) and ambient relative humidity and temperature (8 bins each) to create a look-up table of engine-specific inputs, that were then mapped to the remaining engines in the AEDT database. Additional simulations were performed for aircraft emissions from the Hartsfield-Jackson Atlanta international airport in the stand-alone version of the SCICHEM model for the July period, to evaluate additional features in the latest version of the model not available in CMAQ-APT, such as the ability to represent moving emitters. The goal of this work is to provide an enhanced and multiscale modeling framework which builds upon prior work (using CMAQ-APT and using VBS) to reduce uncertainty in modeled predictions of aircraft-attributable PM_2.5 associated with spatial scales, the traditional treatment of aircraft organic PM, and emission estimates.

Shannon L. Capps, Rob W. Pinder, Dan Loughlin, Jesse O. Bash, Matthew D. Turner, Daven K. Henze, Peter B. Percell, Shunliu Zhao, Matthew G. Russell, Amir Hakami

Energy production processes emit gases that contribute to the formation of tropospheric ozone (O₃), which acts as a short-lived climate forcer and degrades human and ecosystem health. The formation of O₃ is dependent upon the meteorological and chemical state of the atmosphere and can occur hundreds of miles from emission sources. Furthermore, the harmful effects of ozone on human health and public welfare differ spatially and mechanistically. Therefore, optimal benefits from emissions control strategies for the energy system, even for one criteria pollutant, can be difficult to evaluate. The adjoint of a chemical transport model uniquely addresses this problem by efficiently evaluating the relative influence of different ozone precursor emissions on concentration-based metrics, such as human health or ecosystem disbenefits from ozone exposure, without perturbing the modeled ozone concentrations.

We present an approach to determining optimal emissions control strategies for the U.S. energy system with a focus on achievable human health and ecosystem benefits from reductions in O₃ precursor emissions. Specifically, we use the adjoint of CMAQ to elucidate the relative influence of spatially- and sectorally-refined emissions on O₃ exposure in the continental U.S. in the summer of 2007. We separately evaluate the relative influences of emissions on the expected human mortalities due to chronic exposure to O₃ and on species-specific crop and timber biomass yield losses due to cumulative exposure over a period of months. These relative influence parameters are used to inform the trade-offs between technologies in the energy system optimization MARKet ALlocation (MARKAL) model.

James Boylan and Byeong-Uk Kim

Georgia Department of Natural Resources (GA DNR)

Facilities applying for PSD air permits are required to model the impact of direct PM2.5 emissions using AERMOD and must account for the impacts of secondary PM2.5 formation from NOx and SO2. Since AERMOD does not contain chemistry or aerosol formation modules, the secondary formation of PM2.5 cannot be modeled directly in AERMOD. However, PM2.5 interpollutant trading ratios (also called PM2.5 offset ratios) can be used to account for secondary PM2.5 formation in AERMOD and other dispersion model. Previously, we presented a detailed technical approach for developing seasonal PM2.5 interpollutant trading ratios in Georgia using the CAMx photochemical grid model with flexi-nesting (12 km/4 km/1.333 km). This paper investigates the spatial variability in seasonal PM2.5 interpollutant trading ratios by looking at the results at eight different locations in Georgia. These results can be used to quantify the impacts of secondary PM2.5 formation from NOx and SO2 for any new PSD project in the state.

Rebecca K. Saari

Noelle E. Selin

Tammy M. Thompson

We examine the influence of proposed ozone reductions on exposure, economic welfare, and social equity using an integrated modeling framework of the US economy, energy system, and atmosphere. We link CAMx and BenMAP to an economic model designed to evaluate climate and energy policy, the US Regional Energy Policy Model (USREP). This coupled modeling framework has previously been used to estimate the fine particulate matter co-benefits of US climate change and clean energy policy. Here, we extend this framework to represent the health-related impacts of tropospheric ozone. We employ this framework to quantify the effect of proposed ozone reductions under the Cross-State Air Pollution Rule and other proposals, and the impact of their delay, on ozone exposure, human health, welfare, and income inequality. We find that planned reductions tend to reduce ozone exposures for low income households more than high income households. We quantify the effect of ozone reductions and the impact of delay as a fraction of per capita welfare, and we find that the proposed reductions could benefit low income households by as much as 0.5% of per capita welfare. Future extensions of this framework could examine specific energy and climate policies and their effects on greenhouse gas emissions, ozone, and fine particulate matter in search of holistic management strategies.

Shih Ying Chang^1,2, Saravanan Arunachalam¹, Brian Naess¹, Alejandro Valencia¹, Vlad Isakov³, Ted Palma^4, Laura Jackson^3, Michael Breen³

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

²Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

³Office of Research and Development, U.S. Environmental Protection Agency

⁴Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency

A typical urban system has high density of roadway and vehicles, and is thus vulnerable to traffic-related air pollution. Therefore, it is desirable to quantify the impact of on-road transportation to urban air quality at high spatial resolution to gain insight into the interaction between urban ecosystem and anthropogenic air pollution. In this study, we demonstrate an approach that can rapidly quantify multiple traffic-related pollutants at census block level. We used a "bottom-up" strategy, under which detailed emission from each road link is calculated based on traffic activity and ambient temperature, combining a line source dispersion model (R-LINE) to model several criteria pollutants and mobile source air toxics (MSATs). To reduce the computational burden, we developed the Annual Stability WInd Clustering (ASWIC) approach, which clustered meteorological data based on dispersion-related parameters to reduce the required run times, by running for select hours of the year, and then using weights to compute the annual average concentrations. Also, traffic count on each road link was used as emission input to R-LINE to reduce the computational burden. We used emission factors from U.S. EPA's Motor Vehicle Emission Simulator 2010b (MOVES), activity and roadway link definition from Federal Highway Administration's Freight Analysis Framework 3 (FAF3), and surface meteorological data from National Weather Service (NWS) stations processed by AERMET. The approach was evaluated at a mid-size urban region (Portland, Maine) and showed a normalized mean error of 30% for all pollutants comparing to an explicit model simulation for each hour of the year. The normalized mean bias for most of the MSATs and PM_2.5 is within 10% but is relatively high (20% to 33%) for NO_X, CO, and Formaldehyde. The method was implemented at two large metropolitan areas - Portland, Oregon (~21,000 census blocks) and Tacoma, Washington (~45,000 census blocks), one mid-sized urban area - Portland, Maine (~6,000 census blocks), and a larger region with both rural and urban area - North Carolina Piedmont (~104,000 census blocks) as illustrative examples. We show how this approach could be connected to the EnviroAtlas national web-based mapping tool with examples in Portland, Maine and Durham, North Carolina, providing information for community decision making with respect to sustainable development planning that benefits public health. EnviroAtlas includes high-resolution data on urban natural infrastructure including percent tree cover along busy roadways. For featured communities, these data and block-group summaries can suggest where near-road populations may benefit most from existing or planned tree buffers, and indicate the distribution of multiple natural benefits across the local demographic spectrum. Additional maps include the heat mitigation, ambient air and water filtration, carbon storage, and recreational opportunities provided by urban tree cover and open space.

Stephen Reid¹, Kenneth Craig¹, Garnet Erdakos¹, Charles Driscoll², Habibollah Fakhraei², Kathy Fallon Lambert³, Jonathan Buonocore⁴

¹Sonoma Technology, Inc., Petaluma, CA

²Syracuse University, Syracuse, NY

³Harvard Forest, Harvard University, Cambridge, MA

⁴Harvard School of Public Health, Boston, MA

The U.S. Environmental Protection Agency (EPA) released the Clean Power Plan on June 2, 2014, a proposed rule for decreasing carbon emissions from existing power plants. Fossil-fueled power plants are the single largest source of anthropogenic carbon dioxide (CO₂) emissions in the United States, accounting for about 40% of total CO₂ emissions nationwide. These power plants are also significant sources of additional pollutants, including sulfur dioxide (SO₂), nitrogen oxides (NO_x), and mercury (Hg). SO₂ and NO_x contribute to the formation of fine particulate matter (PM_2.5) and acid rain, and NO_x is a precursor to ground-level ozone. Policies intended to address climate change by decreasing CO₂ emissions can also reduce emissions of these co-pollutants, thereby providing important co-benefits to human and environmental health.

A project team led by Syracuse and Harvard Universities is conducting an integrated, spatially explicit study of the benefits to human and ecosystem health associated with different approaches to carbon pollution standards for existing power plants. As part of this study, Sonoma Technology, Inc. (STI) used the Community Multiscale Air Quality (CMAQ) model to quantify changes to air quality (ozone and PM_2.5) and pollutant deposition (sulfur, nitrogen, and mercury) associated with a 2020 reference case and three 2020 carbon policy scenarios. The reference case assumes full implementation of current "on the books" air quality programs, and the carbon policy scenarios represent different stringencies (i.e., emissions rates in tons of CO₂/MWh) and flexibility levels (i.e., options available for compliance). STI used 2020 power sector emissions estimates from existing runs of the Integrated Planning Model (IPM) as inputs to CMAQ, and model results are being provided to researchers at Syracuse and Harvard for further analysis of the human health and ecosystem benefits associated with each of the three policy scenarios.

To date, the results of our analysis show that the co-benefits associated with reducing carbon emissions from existing power plants depend on the specific design of the carbon standard. For example, the policy scenario evaluted that is most similar to EPA's proposed rule allows additional renewable energy and energy efficiency and emission trading within a state to count toward compliance. This scenario results in greater reductions of co-pollutant emissions and more widespread air quality benefits than a standard that focuses strictly on "within the fenceline" power plant retrofits.

Maria Teresa Pay¹, Victor Valverde¹, Jose Maria Baldasano^1,2, Roger Kwok³, Sergey Napelenok³, Kirk Baker⁴

¹Earth Science Department, Barcelona Supercomputing Center, Jordi Girona 29, Edificio Nexus II, 08034 Barcelona, Spain

²Environmental Modeling Laboratory, Technical University of Catalonia, Barcelona, Spain

³ORD/NERL/AMAD, U.S. EPA, Research Triangle Park, NC

⁴OAQPS, U.S. EPA, Research Triangle Park, NC

Despite the European effort to reduce anthropogenic emissions of ground-level ozone (O₃) precursors (reduction by 55% for NMVOC, and 44 % for NO_x in the period 1990-2009), the number of exceedances of European ground-level O₃ concentration standards for protecting human health remains a serious problem, especially in summer (EEA, 2013). Meteorological conditions and biogenic NMVOC emissions are also important drivers on O₃ dynamic. Spain is a Mediterranean country under the influence of prevailing stagnant conditions in summer where emissions of O₃ precursors are slowly dispersed into the atmosphere, and chemical reactions leading to O₃ formation are enhanced. The present work attributes emission sources and source region to the O₃ concentration in Spain. On the one hand, it estimates the contribution to the Spanish O₃ exceedances from other European countries. On the other hand, it calculates the contribution of the main sources of O₃ precursors such as road transport, power generation, and industrial and non-industrial combustions. For that purpose, the present work uses the recently released Integrated Source Apportionment Method (ISAM) (Kwok et al., 2013) within the CMAQv5.0.2 model which attributes O₃ and its precursors to sectors/regions of user's interest. First, the ISAM tool, implemented within the CALIOPE air quality forecasting system (CALIOPE AQFS, Baldasano et al., 2011), runs over Europe at 12-km resolution to estimate European contribution to Spanish O₃ concentration; and second it runs over Spain at 4 km to quantify the contribution of the main sources. The meteorological driver is the WRF-ARWv3.5 model which runs accordingly with the same horizontal resolution, and it is initialized by the GFS/FNL global data (0.5 x 0.5). Chemical boundary conditions come from the Monitoring Atmospheric Composition and Climate project (MACC) which provides global forecast for aerosol and reactive gases (1.125 x 1.125). The emissions are estimated by the High-Elective Resolution Modeling Emission System (HERMES version 2.0; Guevara et al., 2013) using a top-down approach in the European domain, and a bottom-up approach in Spain. In this study, ISAM within the CALIOPE AQFS simulates the period 15-31 July 2012, when a high pressure system lingered over western and central Europe, and temperatures exceeded 30C. The episode is the largest in terms of area affected and it accounted the ~30% of the total number of exceedances of the O₃ information (180 μgm^-3, 1h) and alert (240 μgm^-3, 1h) thresholds. The attribution assessments of O₃ concentration in Spain will quantify the contribution by the aforementioned emission sectors and regions in terms of the exceedances of the European standards (Directive 2008/50/EC).

Baldasano, J.M., Pay, M.T., Jorba, O., Gasso, S., Jimenez-Guerrero, P., 2011. An annual assessment of air quality with the CALIOPE modeling system over Spain. Sci. Total Environ., 409, 2163-2178.

EEA, 2013. Air pollution by ozone across Europe during summer 2012. Overview of exceedances of EC ozone threshold values for April-September 2012. European Environmental Agency, EEA Technical Report 3/2013. ISSN 1725-2237. 52 pp.

Guevara, M., Martinez, F., Aravalo, G., Gasso, S., Baldasano, J.M., 2013. An improved system for modelling Spanish emissions: HERMESv2.0. Atmos. Environ., 81, 209-221.

Kwok, R., Napelenok, S., Baker, K., 2013. Evaluation of an ozone attribution diagnostic analysis tool implemented in CMAQ. . In: 12th Annual CMAS Conference. October 28-30, 2013 Chapel Hill, NC (USA).

Daniel Cohan

Recent studies by our group (Pegues et al., 2012; Cohan and Chen, 2014) have shown that state implementation plans (SIPs) developed to attain previous national ambient air quality standards (NAAQS) for ozone and particulate matter (PM) have been largely successful in achieving air quality improvements beyond national trends, and in achieving the air quality improvements predicted by CMAQ or CAMx photochemical modeling. However, the SIPs showed little indication of integrating across ozone and PM strategies, or of quantifying costs and health benefits.

EPA's proposed Clean Power Plan could soon require States to develop an entirely new form of SIP, this time to attain state-by-state caps on power plant emissions of greenhouse gases. Preliminary conversations with state agency officials indicate that plans for attaining these power plant caps are being seen as separate from the SIP development likely to be needed for tighter ozone and PM NAAQS, if they are being considered at all before a final rule is issued. The lack of a direct impact of local greenhouse gas emissions on local ozone and PM, and the different timetables and spatial scope of statewide power plant caps and non-attainment region SIPs, may be contributing to the lack of integration and advanced planning. However, different paths to attaining the carbon caps could have very different impacts on air quality. For example, efficiency and renewables reduce all emissions nearly proportionally, whereas carbon capture can cause some air pollutant emissions to increase on a per kWh basis if the retrofit technologies impair efficiency. Substitution of natural gas for coal has intermediate impacts, based on life cycle analyses conducted by our group and others.

Here, I will present a framework by which States could jointly consider air quality attainment and greenhouse gas mitigation in integrated SIPs. The approach will build upon the framework introduced by Cohan et al. (2007) and Chestnut et al. (2006) for jointly considering control costs, photochemical sensitivities, and health concentration-response functions in the integrated multi-pollutant control of ozone and PM. Opportunities for integrating greenhouse gas mitigation into that framework, both under the EPA-proposed Clean Power Plan and potentially for other sectors, will be discussed. The potential roles of various state-of-the-science options for energy systems modeling (including EPA-MARKAL, NREL's ReEDS, and investment-oriented models developed at Rice) and photochemical sensitivity analysis (including HDDM and adjoint methods) for informing these integrated air quality and climate planning efforts will also be discussed.