Correctly representing the optical properties of black carbon in the integrated WRF-CMAQ system

Francis S. Binkowski

Institute for the Environment

The University of North Carolina At Chapel Hill

The WRF-CMAQ integrated system is designed to include the feedback effect from aerosols on the short and long wave radiation budgets. Contemporary measurements of ambient aerosol particles, especially PM2.5, indicate that these particles often consist of a Black Carbon (BC) core surrounded by a coating or shell of species such as sulfates, organics and water. Jacobson (2000) has argued that the best way of representing the absorption and scattering properties of such particles is to model them directly as a coated BC sphere. Thus, it seems very appropriate to include this coated-sphere (core-shell) approach in a new aerosol-optics package for WRF_CMAQ.

Because CMAQ produces log-normal size distributions for the aerosols, the scattering and absorption codes must be integrated over these distributions. Our computer codes use the very efficient Gauss-Hermite (GH) numerical integration method to accomplish this integration.

The integration algorithms internally call the widely accepted routines BHMIE and BHCOAT (Bohren & Huffman, 1983) to calculate the extinction and scattering efficiency coefficients as well as the asymmetry parameter. When Black Carbon is present as an aerosol species, the code automatically invokes the coated-sphere (core-shell) approach of BHCOAT. Otherwise, invoking BHMIE is the path of choice for homogeneous particles. The code also chooses an optimal set of abscissas and weights depending upon the Mie size-parameter ( pi times the geometric-mean diameter / wavelength) and the refractive index. These calculations are done for both short and long wave radiant fields.

Thus, the radiative transfer calculations in WRF-CMAQ will be more aligned with current aerosol particle observation.

Reference:

Bohren, C.F. and D.R. Huffman, 1983, Absorption and Scattering of Light by Small Particles, Wiley-Interscience , New York, copyright 1983. (Paperback published 1998).

Jacobson, M.Z., 2000, A physically-based treatment of elemental carbon optics: Implications for global direct forcing of aerosols, Geophysical Research Letters, Vol. 27, No. 2, pp 217-220, January 15, 2000

Application and Evaluation of 2006 NLCD and MODIS in the WRF/CMAQ System

Limei Ran¹, Robert Gilliam², Jonathan Pleim², William Benjey²_, Adel Hanna¹

¹Center for Environmental Modeling for Policy Development

Institute for the Environment

University of North Carolina at Chapel Hill, NC USA

²Atmospheric Modeling and Analysis Division

ORD NERL/USEPA, Research Triangle Park, NC

Consistent and accurate land cover data plays an important role in meteorology and air quality modeling systems. To improve the simulation results, land cover data types have to be correctly defined (e.g. albedo, LAI, and maximum stomatal conductance) in the land surface model (LSM). We have enhanced the land use processing tool previously developed in the Spatial Allocator Raster Tools to generate land cover data set for WRF/CMAQ modeling from 2001 NLCD and MODIS land cover data to 2006. The newly generated 2006 NLCD and MODIS land cover data has a total of 40 land cover classes (20 from NLCD and 20 from MODIS) in comparison with the 2001 data with 50 classes (30 from NLCD and 20 from MODIS). The PX LSM has been modified to use the new 2006 gridded data for the WRF/CMAQ simulation. The objective of this presentation is to apply 2006 NLCD and MODIS data in the modified WRF/CMAQ system and evaluate the improvement from using the most current and high resolution land cover data. The presentation will address the degree of sensitivity of the WRF/CMAQ system to four different land cover data sets with different classification and resolution (2006 NLCD/MODIS, 2006 MODIS, WPS 24 category USGS, and 1992 NLCD land cover data sets) for three nested domains (12km, 4km, 1km). Simulated meteorology (e.g. temperature, moisture, wind speed and direction, surface radiation budgets, and PBL height) and air quality (e.g. ozone, PM2.5, NO_x, CO, and N and S deposition) will be compared among different simulations and analyzed with measurement data.

Cloud Assimilation in WRF Model and Impact on Sub-domains

Yun-Hee Park1, Arastoo Pour Biazar1, Richard T. McNider1, Bright Dornblaser3,Maudood Khan2, Kevin Doty1

1. University of Alabama in Huntsville
2. University Space Research Association (USRA)
3. Texas Commission on Environmental Quality (TCEQ)

Geostationary satellites offer observations at a temporal and spatial scale comparable to numerical meteorological models and therefore can be used to improve the performance of these models. One area of significance to air quality studies that needs to be improved is the simulation of clouds at the right time and place. Clouds have a significant impact on tropospheric chemistry as they affect photolysis rates. Solar radiation is reduced below a cloud, resulting in decreased photochemical activity in the shaded region from the cloud base to the surface. The purpose of this study is to improve the cloud simulations in WRF model by using geostationary satellite data to adjust key model parameters relevant to cloud simulation. We present a technique to estimate the maximum vertical velocity and its location in a column that is conducive to forming/dissipating a cloud based on Geostationary Operational Environmental Satellite (GOES) Imager observations. Using a variational technique a balanced three-dimensional wind field will be constructed to sustain the target vertical velocities and the model is nudged accordingly. The technique was implemented in WRF and used for a case study during the summer of 2006. The modeling study covers the continental US at 36-km grid spacing with nests of 12- and 4-km grid spacing that are centered over Texas. The presentation here will show the improvements in model cloud simulation and the propagation of the improvements to the sub-domains.

Developing chemical mechanisms that are more robust to changes in atmospheric composition

Gookyoung Heo,^a William P.L. Carter,^a Greg Yarwood^b

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

^bENVIRON International Corporation, 773 San Marin Drive, Suite 2115, Novato, CA 94998, USA

Chemical mechanisms commonly used for 3-dimensional air quality modeling represent the atmospheric reactions of many different volatile organic compounds (VOCs) with highly condensed reactions of a limited number of model species by using assumptions on the average air compositions of major VOC classes (e.g., alkenes) or major carbon bond types (e.g., terminal C=C bonds). These condensed chemical mechanisms are designed to model O₃ formation from typical urban ambient VOC mixtures while reducing the computational burden caused by gas-phase chemistry calculations and allowing available resources to be used for improving other components of the air quality modeling (e.g., better representing emissions and meteorology). However, the atmospheric composition is not constant and changes temporally and spatially. For example, in southeast Texas, the atmospheric composition is often markedly different from that of the typical composition in most urban areas in the U.S. due to industrial emissions; the VOC composition of mobile emissions also changes as fuels and technologies change. As a result, certain compounds may warrant more attention for some regions than for other regions, and certain compounds may need more attention due to changes in emissions over time. However, using detailed mechanisms with many additional reaction parameters beyond some point where basic laboratory studies and environmental chamber studies cannot reasonably support could give us an illusion that we know better than we actually know. Furthermore, using more detailed chemical mechanisms does not guarantee significantly better model predictions (e.g., predicted O₃ concentrations) than condensed chemical mechanisms while demanding more resources such as more computer memory, disk storage, and computational time. In this presentation, we will present a practical and scientifically reasonable approach to developing chemical mechanisms that are more robust to temporal and spatial changes in atmospheric composition.

Key words: chemical mechanism, condensed chemical mechanism, atmospheric composition, air quality modeling

Constraining Regional CO2 Budgets using CMAQ-WRF

Zhen Liu¹, Cosmin Safta¹, Khachik Sargsyan¹, Bart G. van Bloemen Waanders ², Habib Najm¹, Bert Debusschere¹, Garrett Barter¹, John Roskovensky², Mark Ivey², Paul Schrader¹, Fred Helsel¹, Ray P. Bambha¹, Hope A. Michelsen¹

¹Sandia National Laboratories, Livermore, CA 94551

²Sandia National Laboratories, Albuquerque, NM 87122

In response to a need to understand and reduce greenhouse gas emissions, substantial modeling efforts have been made to constrain CO₂ budgets based on satellite and in situ observations. Most of these efforts have either focused on global and continental scales using global chemical transport models (CTMs) or have relied on inversion of CO₂ alone. Despite these efforts, inferences of regional CO₂ budgets and anthropogenic contributions are plagued by uncertainties, and considerable work is required to understand and reduce these uncertainties.

We develop the capability of CMAQ to simulate CO2 and use CMAQ driven by WRF to analyze observations of CO₂ and other species measured from a mobile laboratory deployed in Oklahoma during the fall of 2010. We investigate the impacts of various factors, including biosphere-atmosphere exchange, anthropogenic emissions, boundary layer heights, transport and boundary conditions to understand diurnal and daily variation patterns of CO₂. We identify the key aspects of uncertainties in both the model and observations and the associated implications for regional inverse modeling of CO₂.

A Forward and Adjoint Neighborhood Scale Air Quality Model

Eduardo P. Olaguer

The HARC air quality model is a fine (200 m) horizontal resolution 3D chemical transport model with a customized chemical mechanism designed for near-source applications. The performance of the model has been evaluated against both routine and research grade monitoring observations of ozone and its precursors, including radical species. An adjoint version of the full model has been developed for applications such as Computer Aided Tomography based on DOAS remote sensing measurements and inverse modeling of industrial emissions. Examples of forward and adjoint applications of the HARC model will be provided, including the simulation of historical flare emission events in the Houston region.

Modeling the impacts of alternative burning strategies on local and regional air quality

Aika Yano, Yongtao Hu, M. Talat Odman

One of the main requirements land managers must follow while performing prescribed burnings is to limit smoke from reaching populated areas. We surveyed land managers and prescribed burning experts in the Southeast to identify the most relevant burning scenarios they would like us to simulate with our smoke impact prediction system. In response to the survey results, we focused on the impacts of changing: 1) frequency of burn, 2) season of burn, 3) size of burn, 4) ignition type, and 5) time of burn. The objective of this study was to find the benefits and harms associated with changing these five factors, giving guidance to land managers on alternative burn strategies that may reduce the impacts on air quality downwind.

Our smoke impact prediction system incorporates new elements into CMAQ and is called Adaptive Grid Daysmoke-CMAQ (AGD-CMAQ). WRF provides the meteorological inputs and CONSUME (Version 3.0) and Fire Emission Production Simulator (FEPS) are used to generate the emissions inputs. AGD-CMAQ was built to strengthen the ability to predict air quality impacts from biomass burnings. Adaptive Grid CMAQ has the capability to increase grid resolution for tracking biomass burning plumes at the regional scale. Daysmoke, a prescribed burning plume dispersion model, is used to track the plume at subgrid scales. While CONSUME estimates total emissions, FEPS provides time-varying emissions and important plume parameters such as fire diameter and heat flux. These parameters that can be controlled by adopting certain burn strategies, along with meteorology, play important roles in shaping the dispersion and long-range transport of plumes and affect the ground-level concentrations of pollutants downwind.

A burn conducted at Fort Benning Army Base on April 9, 2008 (Compartment F5) under southeasterly winds was used as our study case. The modeling system is evaluated with ground-level smoke measurements 2-5 km downwind of the burn and PM2.5 observations at Columbus Airport. Results of different burning strategies are analyzed based on the amounts of pollutants emitted along with how smoke plumes are dispersed and pollutants are transported to impact downwind air quality.

Development and evaluation of plume-in-grid and volatility basis set modules in CMAQ 5.01
Prakash Karamchandani, Bonyoung Koo, Greg Yarwood, Jeremiah Johnson
ENVIRON International Corporation

Eladio Knipping
EPRI

The Community Multiscale Air Quality (CMAQ) modeling system Version 5.0 (CMAQv5.0) was released by the U.S. EPA in February 2012, and an interim release (v5.01) is currently undergoing testing before being publicly released. Because CMAQ is a community model, EPA encourages the development of proven alternative science treatments by external scientists and developers that can be incorporated as part of an official CMAQ release. This paper describes the implementation and evaluation of two new science modules for CMAQ 5.01: (1) a plume-in-grid (PinG) module, based on a reactive puff model, SCICHEM; and (2) an alternative framework for organic particulate matter (PM) formation using a volatility basis set (VBS) approach. The PinG module, also referred to as Advanced Plume Treatment (APT), provides the capability of resolving sub-grid scale processes, such as the transport and chemistry of point source plumes, in a grid model. The VBS approach provides a unified framework for gas-aerosol partitioning of both primary organic aerosols (POA) and secondary organic aerosols (SOA), including chemical aging. The VBS implementation in CMAQ 5.01 uses four separate basis sets to differentiate anthropogenic POA and SOA, biogenic SOA, and OA from biomass burning. Each basis set consists of five volatility bins including a zero-volatility bin for non-volatile compounds. Molecular weight increases as volatility decreases to account for mass gain from chemical aging. The new PinG and VBS modules in CMAQ 5.01 are applied and evaluated separately for two 15-day summer and winter periods in 2005 to the eastern U.S., and their results compared with those from the base CMAQ 5.01. Twenty large point sources of NOx in the eastern U.S. are selected for explicit plume treatment with APT in the PinG simulation. We present results from these applications of the two new modules in CMAQ 5.01.

Aircraft Emissions Contribution to Organic Aerosols using the Volatility Basis Set

Matthew Woody, Saravanan Arunachalam, J. Jason West, Shantanu Jathar, Allen Robinson

Aviation, while an integral part of daily global activities, is a source of various air pollutants, including PM_2.5 and organic aerosols. In this study, aircraft-specific organic aerosol contributions comprised of primary organic aerosols (POA), secondary organic aerosols (SOA) formed from traditional precursors (e.g. aromatics and long-chain alkanes), and non-typical SOA formed from unidentified precursors previously unaccounted for in air quality models are modeled using the volatility basis set approach in CMAQ v4.7.1. This approach treats oxidation reactions of traditional SOA (both biogenic and anthropogenic) and non-typical SOA (specific to aircraft emissions) with OH to produce products of lower volatility (represented by 4 volatility bins with C* values ranging from 10¹ to 10⁴ for traditional SOA and by 10 volatility bins with C* values ranging from 10^-2 to 10⁷ for non-typical SOA). Non-typical SOA yields and precursor emission estimates for idle and non-idle aircraft activities are based on recent sampling and smog chamber experiments using a CFM56-2B gas turbine engine. Primary organic aerosols from all sources in the model, including aircraft, are treated as semi-volatile and are aged through reactions with OH. This model is applied to quantify organic aerosols and total PM_2.5 formed from aircraft emissions due to landing and takeoff activities at the Hartsfield-Jackson International Airport in Atlanta during January and July, 2002. Overall model results are compared against various monitoring networks to determine the impacts on model performance when using VBS within CMAQ.

Potential role of isoprene epoxydiols in organic aerosol formation over the United States

Havala O. T. Pye (1), Deborah Luecken (1), Ying Xie (1), Bill Hutzell (1), Rob Pinder (1), Jason Surratt (2)

(1) US Environmental Protection Agency

(2) University of North Carolina at Chapel Hill

Isoprene is the most abundant non-methane hydrocarbon emitted into earth's atmosphere and contributes significantly to ambient organic aerosol in locations such as the southeast United States. CMAQ parameterizes isoprene SOA based on initial reaction with OH leading to two semivolatile (Odum type) products followed by oligomerization and enhancement under conditions of strong acidity. Laboratory and ambient work indicates secondary organic aerosol (SOA) from isoprene involves multigenerational gas-phase chemistry subject to biogenic and anthropogenic influences, and isoprene epoxydiols have been identified as precursors to SOA under low-NO_x conditions. In this work, we simulate summer 2006 conditions over the United States with an updated version of CMAQ v5.0 to examine the potential role of SOA from isoprene epoxydiols and compare to the current parameterization in CMAQ.

Modeling ozone depletion in the marine boundary layer caused by natural iodine emissions

Greg Yarwood, Jeremiah Johnson, Jaegun Jung

ENVIRON International Corporation

Gary Z. Whitten

SmogReyes

Mark Estes, Jim Smith, Jocelyn Mellberg

Texas Commission on Environmental Quality

Abstract

Iodine chemistry can destroy ozone (O₃) in the marine boundary layer and influence background O₃ in coastal regions such as Houston. In this study, we develop iodine reaction mechanisms and emission inventories and use the Comprehensive Air Quality Model with extensions (CAMx) to investigate how iodine chemistry influences background ozone along the Texas Gulf Coast. Oceans emit methyl iodide (CH₃I) and other halo-methanes that are formed by biological and/or photochemical processes in sea water. Methyl iodide is destroyed by photolysis and reaction with hydroxyl radical (OH) and has a tropospheric lifetime of several days which provides a source of iodine atoms (I-atoms) throughout the troposphere. Other iodine containing halo-methanes (e.g., CH₂I₂, CH₂IBr, CH₂ICl) and molecular iodine (I₂) photolyze very rapidly and release iodine atoms predominantly within the marine boundary layer. I-atoms can destroy ozone catalytically (i.e., one iodine atom can destroy many ozone molecules) because I-atoms and simple iodine oxides (IO, OIO) can interconvert. The catalytic cycles can proceed efficiently in the troposphere because, in contrast to chlorine atoms, I-atoms are unreactive toward organic compounds. The catalytic cycles are interrupted by reactions with nitrogen oxides and terminated when higher iodine oxides form and condense to particles. We describe the preparation of gridded, hourly emission inventories of halo-methanes derived from satellite measurements of seawater chlorophyll content. We discuss the reactions that should be included in condensed chemical mechanisms for iodine intended for use in regional photochemical models such as CAMx. We present the results of CAMx sensitivity tests that show the potential changes in background O₃ along the Texas Gulf Coast caused by natural iodine emissions.

Effects of Implementing Subgrid-Scale Cloud-Radiation Interactions in WRF

Kiran Alapaty¹, Jerold Herwehe¹, Chris Nolte¹, Russ Bullock¹, Tanya Otte¹, Megan Mallard¹, Jimy Dudhia², and Jack Kain³

^{1Atmospheric Modeling and Analysis Division, U.S. Environmental Protection Agency, Research Triangle Park, NC}

^{2National Center for Atmospheric Research, Boulder, CO}

^{3National Severe Storms Laboratory, National Oceanic & Atmospheric Administration, Norman, OK}

Abstract: Interactions between atmospheric radiation, clouds, and aerosols are the most important processes that determine the climate and its variability, affecting air quality and environmental systems. The Weather Research & Forecasting model (WRF) is being used as a regional climate model (RCM) by many groups, including the RCM group at the EPA. One of the findings from our RCM studies is that the summertime convective systems simulated by the WRF model are highly energetic, leading to excessive surface precipitation. We also found that the WRF model does not consider the interactions between convective clouds and radiation, thereby omitting an important process that drives the climate. Thus, the subgrid-scale cloudiness associated with convective clouds (from shallow cumuli to thunderstorms) does not exist and radiation passes through the atmosphere nearly unimpeded, potentially leading to overly energetic convection. The Atmospheric Modeling and Analysis Division (AMAD) at the EPA is advancing the climate science of the cloud-aerosol-radiation (CAR) interactions by continued development of integrated modeling systems for meteorology/climate (such as WRF) and air quality (the Community Multiscale Air Quality model, CMAQ). In these models, when used at coarse spatial resolutions (e.g., larger than 1 km), convective cumulus clouds need to be modeled as subgrid-scale clouds. Thus, in order to achieve full CAR interactions of cumulus clouds with aerosols and radiation, cumulus cloud parameterizations need to be linked first with radiation processes. To this end, our research group has implemented into WRF a cloudiness formulation that is widely used in global earth system models (e.g., CESM/CAM5) to account for the effects of the cumulus clouds on radiation. Estimated grid-scale cloudiness and associated condensate are adjusted to account for the subgrid clouds and then passed to the WRF radiation schemes to affect the shortwave and longwave radiative processes. To test the implementation, two sets of WRF simulations were conducted: a one-week case study (July 24-30, 2010) to represent running WRF in numerical weather prediction (NWP) mode; and simulation of a three-year period (1988-1990) to study the effects on WRF in a regional climate mode. Results will be presented to show the effects of introducing the subgrid-scale cloud-radiation interactions on quantities such as precipitation and temperature, and the potential implications for regional air quality.

A modeling study to investigate the meteorological factors for inducing high ozone concentration in southern Taiwan

Ka-Wa Chan, Chin-Fang Lin, Ming-Tung Chuang and Fang-Yi Cheng

In order to understand the causes for inducing the high ozone concentrations in Taiwan, the Weather Research Forecasting (WRF) meteorological model and CMAQ air quality modeling were performed to investigate the relationship between the meteorology and high ozone concentrations. In this study, the WRF meteorological modeling is improved significantly with the update of the land use type, aerodynamic roughness length and the terrain height. The emission inventory is generated using the Taiwan Emission Data System (TEDS) data.

Through out the whole study episode, the local weather in Taiwan was mostly dominated by a strong synoptic northeasterly monsoonal flow which was induced by the Asian continental high-pressure system. Some of the days the land-sea breeze circulation also dominated the local scale flow patterns. The preliminary CMAQ simulation results show that the high ozone concentration is induced by the accumulation of the ozone precursors as well as the land sea breeze recirculation flows.

Single Column Model: A useful tool for evaluating and improving the aerosol component of CMAQ model.

Chuen-Meei Gan, Francis Binkowski, Jonathan Pleim and Rohit Mathur

The main goal of this project is to evaluate the aerosol component of Community Multiscale Air Quality (CMAQ) model, particularly the physical and optical aspects by developing a single column model (SCM). This model has 54 layers in the vertical extending form the surface to 50 hPA. The SCM is designed to read in a vertical profile of observed aerosol properties. The SCM then calculates a vertical profile of aerosol optical depth (AOD) from values of extinction calculated from the CMAQ algorithms. Analogous profiles of single scattering albedo (SSALB) and asymmetry factor are also calculated within the SCM.  The YYTMG codes for short wave (SW) and long wave (LW) are then invoked to produce vertical profiles of the upward and downward fluxes of radiant energy.  Thus, calculated values of the AOD and SSALB along with the flux profiles are available for comparison with field observations.

The observations are taken during the Aerosol Intensive Operating Period (AIOP) on May 1-31, 2003 at the Department of Energy’s Atmospheric Radiation Southern Great Plains (SGP) site. Observations were made at the ground at the Climate Research Facility and aboard the Twin Otter aircraft.  The observations provided size distributions and some species information. Each instrumental system has unique protocols for processing. The vertical profiles extended only to 5 km above the ground level. In addition, important gas-absorbers were not considered. In order to provide values above 5 km, a set of profile calculations using the widely accepted Mid-Latitude-Summer (MLS) case supplemented the observations. The temperature, pressure and relative humidity from radiosonde observations were used to represent the environment between surface to 5 km. The observed aerosol information was mapped into the five model species used in the radiative transfer codes in WRF-CMAQ. These are water-soluble, insoluble, sea-salt, elemental-carbon, and water.  The new codes developed for WRF-CMAQ were used to calculate the optical properties. The innovation in these new codes is the inclusion of a coated-sphere (core-shell) option. 

Our poster will provide direct comparison of outputs from the SCM with observed field data.  Information obtained from the evaluation using the SCM will lead to further improvements in the new WRF-CMAQ codes.

Issues of Air Pollution and Air Quality: Effects on the Egyptian Public

Dana M. Greene, Ph.D., M.A.

This presentation represents a case study of air pollution as an increasingly important environmental and social issue in Egypt. As has been documented by Mohammad al-Raey, high levels of suspended particulates have become common issues in many areas, thereby rivaling emissions of sulfur dioxide into the air (as a result of increasing industrialization). While it is clear that air pollution emanating from local industry and sources (lack of catalytic converters on cars, use of diesel fuel, etc.) is, indeed, an important starting point from which to launch a discussion of the effects of significant air pollution on human health outcomes, it is clear that greater awareness of how to manage air pollution must begin with community-based decision-makers who are willing to implement quality control markers for products produced within Egypt. Here, Egypt faces a unique fate: while levels of production are high, so, too, are levels of air pollution and noxious chemicals emanating from plants, factories, and other insults to the ozone layer. Thus, as a means of studying pollution levels in the country, meteorological dispersion models of pollutants in the country will be examined and discussed. This type of high-level meteorological modeling is extremely useful for monitoring different forms and content of air pollution by providing satellite models that illustrate, visually, the types of pollution that affect Egypt, by placing Egypt in direct contrast with other countries located in close proximity in the Middle East as well as with more developed nations. In this way, emphasis will also be placed on how Egypt (as a nation with documented high levels of pollutants) compare with other nations with high levels of pollutants (but who have greater resources with which to manage emerging environmental issues and mitigate any potential health effects prior to their development. As a developing country, however, Egypt does not enjoy such luxury, and is, instead, facing "intensified environmental problems as a result of population growth, economic development, and rapid urbanization" (see al-Raey). Further, because Egypt is a desert nation whose depletion of natural resources has also been documented in the literature. To this end, this presentation will focus on issues of air quality control in Egypt and on the impact of such control on overall health outcomes for the Egyptian citizenry. In this way, perhaps inroads can be made toward increasing the overall health for Egyptian citizens who, presently, breathe in air that is equivalent (in terms of pollutants) to smoking a pack of cigarettes per day.

Recent inter-annual trend of spherical aerosol in East Asia based on integrated analysis of remote sensing and a chemical transport model

Hara, Y¹., Uno, I.¹, Shimizu, A.², Sugimoto, N.², Matsui, I.², Itahashi, S.³ and Ohara, T.²

^{1 Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka, Japan}

^{2 National Institute for Environmental Studies, Tsukuba, Ibaraki, Japan.}

^{3 Department of Earth System Science and Tecnology, Kyushu University, Fukuoka, Japan}

Recent rapid economic growth of Eastern Asian countries has caused a marked increase of anthropogenic emission (Ohara et al., 2007) since 2000. However, Lu et al. (2011) reported that SO₂ emission in China decreased by 9.2% from 2006 to 2010 due to the wide application of flue-gas desulfurization (FGD) devices in power plants. On the other hand, Lamsal et al. (2011) showed that NOx emission from East Asia increased by 18.8 % during 2006-2009. Anthropogenic emission in East Asia has changed dramatically by the balance between economic development and political emission control.

National Institute for Environmental Studies (NIES) has been constructing a ground-based network of automated dual-wavelength (532, 1064 nm), polarization-sensitive (532 nm) Mie-lidar systems to examine air quality continuously in eastern Asia since 2001 (Shimizu et al., 2004). In this paper, recent inter-annual trend of anthropogenic aerosols was investigated using ground-based lidar data, space-borne lidar, the Moderate resolution Imaging Spectroradiometer (MODIS) data and the Community Multi-scale Air Quality Modeling System (CMAQ) chemical transport model simulation during 2004-2011.

The increase trend of spherical aerosol optical depth (AOD) was observed over the wide region of East Asia by aerosol retrievals of lidar and MODIS from 2004 to 2008. After 2008, the decrease trend of AOD was seen in downwind region (around Japan), while increase trend of AOD was still observed in northeastern China by remote sensing. CMAQ showed that the main composition of spherical aerosol around Japan is sulfate, therefore recent AOD decrease trend in downwind region may be basically caused by the SO₂ emission reduction due to the FGD devices in China. The AOD increase trend in northeastern China might be attributed by increase of secondary aerosol formed from precursor species including NOx after 2008.

Modeling wildfire impacts on air quality during the ARCTAS-CARB campaign: evaluation using surface, airborne and satellite measurements

Yongtao Hu¹, M. Talat Odman¹, Armistead G. Russell¹, Xiaoyang Zhang^2,3, Shobha Kondragunta³, Hongbin Yu^2,4, Huisheng Bian^5,4, Lorraine Remer⁵,

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

² Earth System Science Interdisciplinary Center, University of Maryland, College Park, MD 20740

³ NOAA/NESDIS/STAR, College Park, College Park, MD 20740

⁴ Earth Science Directorate, NASA Goddard Space Flight Center, Greenbelt, MD 20771

⁵Joint Center for Earth Systems Technology, University of Maryland at Baltimore County, Baltimore, MD 21228

Increasingly frequent wildfires in the US have led to imposed adverse impacts on rural and urban air quality. During large wildfire episodes, exceedances of NAAQS for ozone and PM_2.5 are encountered. The US EPA allows these "exceptional events" to be exempted from being designated as exceeding NAAQS. However, how and how much the wildfires really contribute to elevated ozone and PM_2.5 observations are poorly understood. The 2008 summer northern California wildfires are suspected of causing severe air pollution in the urban areas in California. With ample airborne data collected in flights chasing the fire plumes, in addition to the regular surface network measurements and satellite observations, the June-July 2008 ARCTAS-CARB campaign provides a good opportunity to evaluate the capability of regional air quality models in capturing the wildfire impacts on air quality.

We employed a state of art air quality model, CMAQ, equipped with a new SOA module including the multi-generational oxidation process to simulate the air quality impacts of the 2008 northern California wildfires. The simulation covers the period of June 15 through July 14, 2008. Three nesting grids are used with the 36-km grid covering the CONUS, the 12-km grid covering California and the 4-km grid covering most metro areas in California. All the three grids have 34 vertical layers extending to ~16km above the ground with the first layer ~18m thick. We evaluate model performance by examining ozone and PM_2.5 as well as other gaseous and PM components against measurements at multi-platforms: surface, airborne and space. Further, the sensitivity analysis tool DDM-3D is used to assess the air quality impacts of fire emissions and to investigate the sources of model deficiencies regarding modeling of the fire impacts. Model bias due to uncertainties in emissions estimates, fire plume rise estimations, as well as meteorology parameters will be addressed. An inverse modeling technique will be used to quantify the uncertainties in estimates of fire emissions and in layering of those emissions vertically in the CMAQ model.

High resolution CMAQ simulation of air pollution over the Chesapeake Bay during DISCOVER-AQ

Christopher P. Loughner, Daniel Goldberg, Chinmay Satam, Maria Tzortziou, Melanie Follete-Cook, Kenneth E. Pickering, Andrew Weinheimer, David J. Knapp, Denise D. Montzka, Laura Landry, James H. Crawford, Antonio Mannino, Lackson Tambaoga Marufu, Jeffrey W. Stehr, and Russell R. Dickerson

Regional models have difficulty representing the large daytime surface temperature gradient present along coastlines between relatively cool surface waters and the warm ground during the summertime. This temperature gradient is required to initiate a sea or bay breeze circulation. High resolution simulations (horizontal resolution <5 km) capture this local scale circulation and reveal large spatial and temporal gradients in air pollution associated with a sea or bay breeze. Two concurrent field experiments in July 2011, DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) and GeoCAPE-CBODAQ (Geostationary Coastal and Air Pollution Events-Chesapeake Bay Oceanographic Campaign with DISCOVER-AQ), made air quality observations in the Washington, DC and Baltimore, MD metropolitan areas and over the Chesapeake Bay. These observations along with a high resolution CMAQ simulation (1.33 km horizontal resolution) are used to investigate the role of the Chesapeake Bay on air pollution in the region. The horizontal and vertical distribution of air pollution over the Chesapeake Bay is contrasted with those over land using ground-, ship-, and aircraft-based observations obtained during the field campaigns and CMAQ model output. In-situ aircraft profiles and ship-based in-situ O₃, NO, and NOy mixing ratios show a strong vertical gradient of air pollution near the surface of the water which reveals that there is little mixing taking place in the stable marine boundary layer. In addition, observed ozone shows a peak aloft indicating that a local scale circulation influenced by the temperature gradient between the cool surface waters and warm land surface is lofting pollutants upward near the coastline. High resolution WRF and CMAQ simulations are used to demonstrate how the Chesapeake Bay impacts local scale circulation and pollutant transport.

SATMAQ: bringing satellite data to the CMAQ community

Daniel Tong^1,2, Steven Kempler¹, and Wenli Yang^1,2

1 NASA GES/DISC, NASA Goddard Space Flight Center, Greenbelt, MD

2. CSISS, George Mason University, Fairfax, VA

Satellite remote sensing data are being increasingly used in chemical transport models, including CMAQ. This work introduces a new effort at NASA Goddard Flight Science Center Data Information Service Center (GSFC DISC) to make these data more accessible to the CMAQ community. Existing data tools, such as the Giovanni system, can facilitate data download and visualization for scientists unfamiliar with remote sensing datasets. This new project, called SATellite for CMAQ (SATMAQ), advances previous data services by providing the following new features: 1) casting original satellite swaths into a user-defined CMAQ domain; 2) converting netCDF or HDF into IOAPI format; and 3) retaining high resolution. Gridded data (Level 3 and above) downloaded from existing online platforms are usually of coarser resolution (half or one degree), designed for use of global, not regional, models. Here we demonstrate how the new service can be used to access irregular satellite swaths (of high resolution), and convert the data into the IOAPI format in a Lambert Conformal CMAQ domain. Three case studies are presented, including (1) using the MODIS Deep Blue AOD data for grid-to-grid comparison with the CMAQ dust prediction; (2) using the NDVI data to generate dynamic dust emission source region; (3) using the satellite soil moisture data to evaluate the land surface output. Finally, we briefly discuss a list of satellite products that can be potentially used in air quality applications.

A Novel Formation Mechanism of Atmospheric Low-Molecular Weight Carbonyls over Marine Regions

Jonathan Trueblood, Nicholas Meskhidze

Formaldehyde (HCHO) is a ubiquitous oxidation product that exists in polluted rural and urban areas, as well as remote background sites where it is an important photochemical intermediate. HCHO levels of up to six times above what is typically predicted by photochemical models have been reported over the marine boundary layer (MBL). As proposed mechanisms for HCHO production remain to be insufficient to explain such large discrepancies between model predictions and measured values, the role of marine regions in the creation of HCHO continues to be one of the largest sources of uncertainty in current global chemistry-transport models. Furthermore, measurements of glyoxal (CHOCHO) precursors (isoprene and acetylene) above biologically active oceanic regions in the remote tropical Pacific Ocean are too low to account for the amount of glyoxal found. The poorly understood formation mechanism in conjunction with its significance as a secondary organic aerosol (SOA) precursor over the MBL makes the improved understanding of glyoxal an important area of research in the remote marine regions. Here we examine the viability of a proposed mechanism for the photochemical production of formaldehyde, glyoxal, and several other low-molecular weight (LMW) carbonyls involving aerosols enriched with biologically produced organic matter.

In this study, the phytoplankton Emiliania Huxleyi was incubated in autoclaved seawater contained within a 9 liter Pyrex glass bottle. Quantitative analysis of the enrichment of transparent exopolymer particles (TEP) and other biologically produced organic matter (dissolved and particulate) in the surface microlayer was carried out by employing Alldredge’s alcian blue staining technique. To produce organic aerosols, enriched seawater was bubbled with hydrocarbon free air using a sintered glass filter placed 5 cm below the surface. Utilizing a mixed flow reaction scheme, produced aerosols were then pushed through stainless steel flow tubes into a separate 9-liter Pyrex glass container acting as a residence chamber. The container was surrounded with six 40W Sylvania 350 BL black lights to allow for the irradiation of aerosols at 350 nm. A flow rate of approximately 0.1 l/min allowed for an average aerosol residence time of 90 minutes inside the residence chamber. All air from the chamber was then passed through a 5” long Pyrex desorber tube packed with 60/80 Tenax that had been soaked in the derivatizing agent pentafluorophenyl hydrazine (PFPH). Subsequent thermal desorption of the sampling tube using a CDS-8000 Preconcentrator and analysis with a Varian 450 GC/220 MS was carried out to detect several LMW compounds, including HCHO, CHOCHO, acetaldehyde, and acetone. The presentation will discuss the approximate production rates for LMW carbonyls for given TEP enriched seawater.

Investigation of Multi-decadal Trends in Aerosol Direct Radiative Effect from Anthropogenic Emission Changes over North America by Using a Two-way Coupled Meteorology-chemistry Model

Chao Wei, Jonathan Pleim, Rohit Mathur, David Wong , Jia Xing, Chuen Meei Gan, ST Rao (USEPA, RTP, NC)

and Francis S. Binkowski (UNC)

The complex feedback mechanisms among chemistry-aerosol-cloud-radiation-climate exist ubiquitously in the Earth system. These feedbacks have been observed in numerous field experiments or through analyses of long-term historic surface and satellite observational data. While aerosol radiative effects have been recognized as some of the largest sources of uncertainty among the forcers of climate change, there has been little effort devoted to verification of the spatial and temporal variability of the magnitude and directionality of aerosol radiative forcing. A comprehensive investigation of the processes regulating aerosol distributions, their optical properties, and their radiative effects from past and current human activities will help us to build more confidence in the estimates of the projected impacts from changes in anthropogenic forcing and climate change. This study addresses this issue through a systematic investigation of changes in anthropogenic emissions of short-lived aerosol-precursors over the past two decades (1990–2010) in the United States, their impacts on aerosol loading in North America, and subsequent impacts on regional radiation budgets. During this period, U.S. emissions of SO2 and NOx have been reduced by about 66% and 50%, respectively, mainly due to Title IV of the U.S. Clean Air Act Amendments that aimed to reduce emissions that contribute to acid deposition. Significant decadal brightening of downwelling shortwave is observed in the continental United States. In this study we test the hypothesis that changes in surface solar radiation over time are caused by the changing patterns of anthropogenic emissions of aerosols and aerosol precursors.

A newly developed two-way coupled meteorology and air quality model composed of the Weather Research and Forecasting (WRF) model and the Community Multiscale Air Quality (CMAQ) model is being run for 20 years (1990–2010) on a 12-km resolution grid that covers most of North America. A newly developed 20-years emission inventory is used in order to accurately reflect the emission trends resulting from progressively more stringent air quality regulations as well as population trends, economic conditions, and technology changes in motor vehicles and electric power generation. The WRF/CMAQ model includes direct effects of aerosols on SW radiation and the direct effects of tropospheric ozone on LW. New algorithms on the calculation of aerosol optical properties and radiation have been developed for considering of both computational efficiency and more realistic aerosol states. By simulating this period we can assess model performance for reproducing observed trends in air pollutants, such as sulfate and nitrate aerosols, and their consequences on trends in surface radiation. Preliminary model simulations for 1990 and 2006 are being evaluated both for their performance in comparison to observed concentrations and simulation of observed trends in concentrations and surface radiation.

Impact of Vegetation Variability on Biogenic Emissions

Erin Chavez-Figueroa

Dan Cohan

Adetutu Aghedo

Ben Lash

Ranga Myneni

Jian Bi

Biogenic volatile organic carbons (BVOCs) interact with manmade emissions to form ozone and particulate matter in the troposphere. Accurately accounting for these emissions is important when making regulatory decisions, particularly in transition regimes where ozone may be jointly sensitive to NOx and VOCs. . Recent studies have yielded conflicting conclusions regarding the extent to which interannual variation in vegetation, as quantified by the leaf area index (LAI), may influence biogenic emissions estimates. Current modeling often ignores this variability, which may lead to less accurate model output.

In this study, we first assess the interannual variability of leaf area index (LAI) derived from MODIS and its correlation with meteorological phenomena such as drought or extreme temperatures. Preliminary results indicate that summer droughts can lead to significant decreases in vegetation, while wet conditions lead to increases in vegetation. We then assess the impact of the interannual variability in LAI on isoprene and other BVOC emissions estimates from the biogenic emission model MEGAN. MODIS is the most widely used source for LAI data, but has some issues with noise and may underestimate variability in heavily forested areas. MODIS data therefore requires significant processing before use, which can include temporal or spatial smoothing as well as application of a climatology mask. Some studies also consider satellite measurements together with phenology modeling methods to generate alternate estimates of LAI. Consensus has not yet been reached on the best approach to use. In order to capture the uncertainties in LAI, we will use LAI from multiple sources. By running these data sets through MEGAN, we will assess the impact of interannual variability in LAI on the final model output. Future work will run a photochemical model coupled with MEGAN to estimate the impact on air quality from variations in LAI.

Addressing Science and Policy Needs with Community Emissions Efforts

Gregory Frost (1), Leonor Tarrasón (2), Claire Granier (1,3), Paulette Middleton (4)
(1) University of Colorado/CIRES, NOAA/ESRL, Boulder, Colorado, USA
(2) Norwegian Institute for Air Research, Kjeller, Norway
(3) CNRS/INSU, LATMOS-IPSL, UPMC, Paris, France
(4) Panorama Pathways, Boulder, CO, USA
Contact: gregory.j.frost@noaa.gov

We present community-driven emissions efforts within the Global Emissions InitiAtive
(GEIA, http://www.geiacenter.org/), a joint IGAC/iLEAPS/AIMES initiative of the
International Geosphere-Biosphere Programme. Since 1990, GEIA has served as a
forum for the exchange of expertise and information on emissions. GEIA’s mission is to
quantify anthropogenic emissions and natural exchanges of trace gases and aerosols,
and to facilitate the use of this information by the research, assessment, and policy
communities. GEIA supports a worldwide network of about 1300 developers and users
in international scientific projects, providing a solid scientific foundation for atmospheric
chemistry research. Moving forward, GEIA is broadening its role to help serve the
emissions needs of the research, assessment, regulatory, operational, and policy
communities. GEIA intends to demonstrate the potential for improving emission
information by promoting the interoperability of datasets and tools and by making use of
near-real-time observations. As a step toward these goals, GEIA is being linked with
ECCAD (Emissions of chemical Compounds & Compilation of Ancillary Data,
http://eccad.sedoo.fr/) and CIERA (Community Initiative for Emissions Research &
Applications, http://ciera-air.org/). ECCAD is GEIA’s new interactive emissions data
portal, providing consistent access to emission inventories and ancillary data with easyto-
use tools for analysis and visualization. CIERA is a GEIA community project to
develop interoperability in emissions datasets and tools and to support evaluations of
inventories. GEIA is also implementing new approaches to communicate emissions
information and to connect scientific and regulatory emissions efforts. We invite the
CMAS community to join the GEIA network and build partnerships with GEIA to advance
emissions knowledge for the future.

Preparation of Mexico National Emissions Inventory 2005 for Air Quality modeling using WRF-chem in Mexico

Agustin R. Garcia and Bertha Mar

Centro de Ciencias de la Atmosfera, Universidad Nacional Autonoma de Mexico, Mexico

The National Emissions Inventory was updated for base year 2005 in order to be useful and comparable among US and Canada inventories. It considers seven different pollutants (PM₁₀, PM_2.5, CO, NOx, SO₂, VOC and NH₃), there are four different source types, point, area, mobile and natural sources. Emissions are allocated spatially in the 2,454 municipalities.

Several categories are used for different sources 17 for point sources, 34 for area, 28 for mobile.

Spatial allocation was made using population density at AGEB level; forest and agricultural land use type were used to allocate the fire forest and agricultural emissions. Highways and streets were used for mobile emissions.

Temporal distribution was based on the source classification codes (SCC) as well the chemical speciation, for the aggregation to photochemical categories the carter's work was used. The EI for modeling contains for RADM2 mechanism 37 different species, 25 gas phase and 12 for particle and aerosol emissions. For SAPRC99 mechanism 50 different species, 38 gas phase and 12 for particle and aerosol emissions. The intermediate files are in ascii and the final file is in netcdf format.

This work presents the methodology used in order to convert MNEI 2005 for air quality modeling using WRF-chem.

Preparation of Oil and Gas Emissions Inventories for Use in Photochemical Grid Modeling

Erin K. Pollard and Stephen B. Reid, Sonoma Technology, Inc.

Jason Reed and Courtney Taylor, AECOM

Brian Bohlmann, Wyoming Department of Environmental Quality

In recent years, elevated 8-hour ozone concentrations have been observed during winter months in the Upper Green River Basin (UGRB) in southwest Wyoming, where significant oil and gas development activities are occurring. To support air quality management in the region, AECOM and Sonoma Technology, Inc. are conducting photochemical grid modeling with the Community Multiscale Air Quality model (CMAQ) and the Comprehensive Air Quality Model with extensions (CAMx) to determine the model that best replicates winter ozone formation processes in the UGRB.

To support this effort, the project team is converting detailed oil and gas emissions inventories for the winter of 2008 to air quality model-ready formats. These inventories were developed by the Air Quality Division (AQD) of the Wyoming Department of Environmental Quality (WDEQ), which has instituted an extensive minor source permitting program that covers all oil and gas production wells in the state. This program has led to the collection of detailed emissions data for all permitted wells, including emissions estimates for criteria pollutants, nitrous acid (HONO), and formaldehyde for a variety of sources, such as drill rigs, tanks and pressurized vessels, dehydration units, pneumatic pumps, and process heaters. The inventory also contains speciated volatile organic compound (VOC) emissions for some processes, as well as process-specific stack parameters and temporal information.

Emissions data for the winter of 2008 are being converted to formats compatible with the Sparse Matrix Operator Kernel Emissions (SMOKE) model, with individual oil and gas wells being treated as discrete point sources with multiple emissions processes. In addition, temporal information and speciated VOC data are being incorporated into SMOKE-ready temporal and speciation profiles. This paper will describe the processes used to prepare the detailed oil and gas inventories for use in air quality modeling applications.

Simulation of N2O Production and Transport in the US Cornbelt Compared to Tower Measurements

TL Roche (UNC ENVR), EJ Cooter (EPA AMAD)

N₂O (nitrous oxide) is, among anthropogenic emissions, currently the most important ozone depleter and the third most important radiative forcer. Its atmospheric chemistry is relatively simple, but the chemistry of its production is not. Attempts to spatially and temporally quantify its fluxes have achieved mixed results; however, many researchers agree that agriculture is the dominant source sector, with fertilizer application the management practice most closely related to N₂O emission.

We present the results of modeling N₂O concentrations with CMAQ-5 over the continental US. We input emissions from EPIC, a process model for agricultural soils (et al), and other inventories (e.g., EDGAR). We compare results with observations and recent investigations. We address the relation between N₂O concentrations and agricultural processes, and seek to provide direction to future research.
 

Inverse modeling of Texas NOx emissions using space-based NO2 observations

Wei Tang^a,*, Daniel Cohan^a, Lok Lamsal^b

^aDepartment of Civil and Environmental Engineering, Rice University, 6100 Main Street MS-519, Houston, TX 77005, USA; Email: Wei.Tang@rice.edu

^bNASA Goddard Space Flight Center, Greenbelt, MD 20771, USA

This work demonstrates the application of combining an inverse modeling technique and satellite NO₂ observations to develop a top-down NO_x (NO + NO₂) emission inventory for Texas air quality modeling. Current NO_x emission inventories were developed by a bottom-up approach based on estimated emission activities and emission factors, which have been classified as one of the top uncertainties in air quality modeling. Uncertain NO_x emission inventories may have a significant impact on the model performance and mislead control strategies in air quality management. Applying inverse modeling techniques combined with satellite NO₂ observations to constrain NO_x emissions offers a novel approach to indicate possible biases in air quality modeling. Several inverse modeling studies using satellite observations have been conducted recently in both global and regional scale models. This study applies the Discrete Kalman filter (DKF) inversion technique combined with Ozone Monitoring Instrument (OMI) satellite NO₂ measurements to the regional Comprehensive Air Quality Model with extensions (CAMx) model, to create a top-down NO_x emission inventory for Texas ozone abatement modeling. The Decoupled Direct Method (DDM) in CAMx provides relationships between NO_x emissions and NO₂ concentrations in simulations. Averaging kernels associated with the OMI retrieval are used to compute the NO₂ column density that OMI would have observed for the CAMx-simulated NO₂ concentrations. The impacts of applying inversed NO_x emissions on model estimates of ozone concentrations and their sensitivities to control strategies are being tested.

Historical gaseous and primary aerosol emissions in the United States from 1990-2010

Jia Xing¹, Jonathan Pleim¹, Rohit Mathur¹, George Pouliot¹, Christian Hogrefe¹, Chuen-Meei Gan¹, Chao Wei¹

¹ The U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA

It is believed that the observed evolution of decadal "dimming" to "brightening" during the 1990s in the continental United States was strongly influenced by the reductions in anthropogenic emissions of aerosol precursors. Regional chemistry or climate models are good tools for improving our understanding of the role of aerosols in the decadal changes of solar radiation. However, to perform such simulations in order to reproduce and interpret the observed phenomena, an accurate description of emission changes over such extended time periods is crucial but challenging.

In this study, we used an approach based on activity data to develop a consistent series of spatially resolved emissions in the United States from 1990 to 2010. The state-level anthropogenic emissions of SO₂, NO_x, CO, NMVOC, NH₃, PM₁₀ and PM_2.5 for three major sectors (incl. 48 sub-sectors) were estimated based on several long-term databases containing information about changes in activity data and emission controls. Activity data for energy-related stationary sources were derived from the State Energy Data System. Corresponding emission factors reflecting implemented emission controls were calculated back from the national emission inventory (NEI) for seven years (i.e., 1990, 1995, 1996, 1999, 2001, 2002 and 2005), and constrained by the AP-42 (US EPA's Compilation of Air Pollutant Emissions Factors) dataset. Activity data for mobile sources including all types of highway vehicles and non-highway equipments was obtained from the highway statistics reported by Federal Highway Administration. The trends in emission factors for highway mobile source were informed by the 2011 National Transportation Statistics. Emissions for all non-energy related source were either scaled by the growth ratio of activity indicator or adjusted based on the NEI trends report and EDGAR (Emissions Database for Global Atmospheric Research) dataset.

Because of the strengthened control standards, particularly for the power sector and mobile sources, emissions of all pollutant except NH₃ were reduced by half over the last two decades. The emission trends developed in this study are comparable with the NEI trend report and EDGAR data, but better constrained by trends in activity data. Reduction ratios in SO₂ and NO_x emissions agree well with the observed changes in ambient SO₂ and NO₂ concentrations, suggesting that the various control on SO₂ and NO_x emissions implemented over the last two decades are well represented in the emission inventories developed in this study. These inventories were processed by SMOKE and are now ready to be used for regional chemistry transport model simulations over the 1990-2010 period.

Multiyear boundary conditions for CMAQ 5.0 from GEOS-Chem with Secondary Organic Aerosol Extensions

 

 

 

 

Recent work has demonstrated the importance of hemispheric and global transport on estimating the concentrations of ozone and other pollutants in regional air quality models. Dynamic boundary conditions derived from global simulations capture the variability of global circulation patterns and international emissions on the long-range transport of air pollution. In this study, a new tool to dynamically convert hourly GEOS-Chem (with SOA extensions) output into boundary and initial conditions for CMAQ version 5.0 is described. This species mapping tool is compatible with several chemical mechanisms, including SAPRAC07 and CB05.  The new mapping tool has been applied to 10 years (2001 – 2010) of retrospective GEOS-Chem simulations to create boundary conditions for air quality modeling applications over North America.  This dataset of boundary conditions is available to the CMAQ modeling community and can provide boundary conditions for any North American domain. A comparison of the GEOS-Chem model output against available surface and upper-air (e.g. ozonesonde and aircraft) measurements will be presented. 

The Impact of Climate Change and Projected Water Scarcity on the Electric Power Sector in the US

 

C. Cameron, W. Yelverton, R. Dodder, J.J. West

The electric power sector is currently one of the largest water withdrawers and fastest growing water consumers in the U.S.  Water supply in the United States is becoming increasingly stressed due to growth in population, per capita energy use and industrial water consumption.  At the same time, climate change is expected to decrease water availability and increase water temperature.  These changes have the potential to decrease cooling capacity and, as a result, to compromise U.S. electric power production.  Water availability will thus be a significant factor in determining how we meet demand for electricity in the future.  Here we analyze the impact of multiple water scarcity scenarios on the electric power sector through 2055 using simulations of the U.S. 9-region (EPAUS9r) MARKAL (MARKet ALlocation) integrated energy systems model.  MARKAL is a least-cost optimization model that represents the entire U.S. energy system.  Water withdrawal and consumption factors for electricity generating technologies were assigned in the EPAUS9r database using data from the National Renewable Energy Laboratory (NREL)¹.  Water availability scenarios were constructed for each of the nine model regions based on literature projections to capture a range of possible future conditions.  Water availability is represented through explicit total water withdrawal and consumption constraints.  Results will be analyzed for the market share of different sources of electricity generation, in particular to see whether renewable energy sources become more attractive. In addition, model outputs will be evaluated for the total system greenhouse gas emissions compared to a base case scenario with no constraint on water availability.  Results will indicate whether water limitations may stimulate increased development of renewable energy resources due to their lower water usage.

The impact of biomass burning aerosols from Southeast Asia on the West Pacific

Ming Tung Chuang¹, Joshua S. Fu², Fang-Yi Cheng³, Neng-Huei Lin³, Chung-Te Lee¹

1. Graduate Institute of Energy Engineering & Sustainable Environmental Technology Research Center, National Central University, Jhung-Li, Taoyuan 32001, Taiwan

2. Department of Civil and Environmental Engineering, University of Tennessee, Knoxville, Tennessee, USA

3. Department of Atmospheric Sciences, National Central University, Jhung-Li, Taoyuan 32001, Taiwan

As for the influence of biomass burning aerosols from Indochina on Taiwan and neighborhood area during long-range transport, certain understanding of aerosol properties has been obtained through years of observation experiments. However, all these experiments can only tell us the information of aerosols at Mt. Lulin Mountains in Central Taiwan when air masses were from Indochina or southern China. The characteristics of aerosol components analyzed at the receptor cannot tell us the evolution of chemical species during long-range transport and the radiative effects over Taiwan and neighborhood.

This study will apply CMAQ model to simulate aerosol chemistry along the path of long-range transport from Southeast Asia to Taiwan. The meteorology input was from WRF model. The emissions of anthropogenic sources, biogenic sources, and biomass burning emissions were processed into hourly resolution, respectively. The simulation results were validated by observations from other sub-projects which analyzed chemical components of aerosols collected at Mt. Lulin and regional observation in South China Sea. It is found that even the simulation uncertainty has reduced the precision of comparison between the simulated and the measured concentration, this preliminary study still reveals the chemical evolution of biomass burning aerosols and precursors in the biomass burning plume from Indochina, and influence of aerosols on radiative effects around Taiwan.

A Prescribed Burn Decision Support System for the Kansas Flint Hills Region

Kenneth Craig¹, Clinton MacDonald¹, Neil Wheeler¹, Alan Healy¹, Patrick Zahn¹, Tom Gross², Doug Watson²

¹Sonoma Technology, Inc.
²Kansas Department of Health and Environment

Each spring, ranchers and farmers in the Flint Hills region burn approximately two million acres of grasslands during a 6-8 week period. These annual burns have been a longstanding land management practice in the Flint Hills. However, air quality impacts on the public from smoke generated by these fires have gained more attention in recent years. In addition, as the National Ambient Air Quality Standards (NAAQS) become more stringent, there is added concern that the smoke impacts could contribute to violations of the NAAQS in urban areas.

To help mitigate these impacts, a prescribed burn decision support system (http://ksfire.sonomatechdata.com) has been developed for the Kansas Department of Health and Environment (KDHE). An intuitive web interface provides access to smoke predictions and localized guidance to support current and next day burn/no-burn decisions and help reduce downwind air quality impacts from potential fires in the Flint Hills. This guidance consists of model-based products, supplemented by a human forecast discussion of regional weather conditions and a 5-day outlook. Model products are based on a matrix of NOAA HYSPLIT dispersion model simulations of smoke emitted from hypothetical fires of varying location, size, and fuel load, and driven by operational 40 km North American Model weather forecasts. Simulations are executed through the BlueSky Framework, which also provides fire emissions and plume rise algorithms. This decision support system has operated for each of the last two Flint Hills burn seasons. An evaluation of system performance for the 2011 burn season shows that daily forecasts were accurate, and provided stakeholders with useful information on when and where to burn to avoid adverse air quality impacts.

Coupled Meteorology-Chemistry Model Application for Evaluation of Climate Change Influence on Air Quality: Central Mexico Case Study.

Agustín R. García¹,  Aidee Vega-Rodriguez¹, Ernesto Caetano², Daniela Cruz-Pastrana¹, Manuel Suarez-Lastra²

Universidad Nacional Autónoma de México

¹Centro de Ciencias de la Atmósfera

²Instituto de Geografía

Meteorology can impact the ambient concentrations of pollutants; warmer weather can enhance atmospheric reactivity but also can influence atmospheric dynamic processes. With a coupled model both effects can be taken into account. In order to evaluate the impact of climate change on air quality, a coupled meteorology chemistry model is applied for central Mexico for two different climate scenarios, (base case and A1B for 2070).

Meteorological data using a dynamic downscaling from climate scenario is used in order to obtain the initial and boundary conditions for base case 1980 and 2070 decades. The radiation module used considers modifications in gas trace compounds. In order to identify the influence of climate two different scenarios were considered a) using base case emissions and 1980 meteorological conditions, b) using base case emissions and 2070 meteorological conditions. Emissions projection was based only in the urban sprawl projection. Case a) and b) were used for identify only the climate influence in the air quality in the region.

 The WRF-chem model incorporates chemical transformations and complex gas-phase chemistry, photolysis and aerosols interactions, and is applied using different meteorological senarios.

To evaluate the effect of ozone concentrations in the population, three metrics from Georgopoulos et al. 1997 were used: severity, pervasiveness and potential integrated exposure. Severity sums the times that a concentration exceeds a threshold level (110ppb for ozone). Pervasiveness is the sum of grid cells that registered concentrations above the threshold level during the evaluation period, and the integrated potential exposure measures the exposure in time and space, considering the number of population potentially exposed to outdoor hazardous levels.  Results from the base scenario 1980 and 2070 climate scenario.

 Results show an increase in temperature in the region; in some periods the mixing height layer increased inducing a decrement of pollutants, like ozone and aerosols, concentrations.

Integrated process and reaction rate analysis for wintertime nitrate formation and chemical transport in the Midwestern U.S using CMAQ

Yoo Jung Kim¹, Scott Spak¹, Charles Stanier¹, Gregory Carmichael¹, Nicole Riemer², Jaemeen Baek¹, Abigail Fontaine³, Mark Janssen³, Donna Kenski³, Stephanie Shaw⁴

¹University of Iowa, ²University of Illinois, ³LADCO, ⁴EPRI

Analysis of air quality and related measurements and air quality modeling during the LADCO (Lake Michigan Air Directors Consortium) Winter Nitrate Study was performed in order to better understand wintertime episodes of elevated fine particle (PM2.5) concentrations in the U.S. Midwest. The episodes are found to be regional and are characterized by low wind speeds, near-freezing temperatures, and elevated levels of ammonium nitrate. Integrated process rate (IPR) analysis and integrated reaction rate analysis (IYY) in the CMAQ process analysis tool is employed to understand relative contributions to nitrate formation from all simulated processes, most importantly local chemical production and transport. Regional simulations for the study period are performed using CMAQ with 12 km resolution.

Results indicate that nitric acid is produced most efficiently at night, 50-200 m above the surface. Surface concentrations of nitrate are maintained by downward transport via vertical diffusion. Aerosol nitrate is predominantly removed by dry deposition. The aerosol production rate is significantly higher during episodes, and both daytime and nighttime (N2O5) chemical pathways are important contributors to elevated concentrations. Near the surface at Milwaukee, the daytime formation pathway for nitric acid is larger than the nighttime pathways, while in rural areas, the relative magnitude of the nighttime pathways is larger than that of the daytime pathway. Across the region, the importance of the nighttime pathway increases with altitude. The spatial distribution of aerosol nitrate formation is found to vary significantly spatially and on synoptic time scales.

Impact of biomass burning aerosols on regional climate over Southeast USA

Peng Liu, Yongtao Hu, Alexandra Tsimpidi, Athanasios Nenes, Armistead Russell

Prescribed burning, as an important forest management practice, is used for increasing productivity, reducing wildfire risk and sustaining wildlife habitat. Aerosol from biomass burning, which is rich in black carbon and organic compounds, is a major contributor to the particulate matters over Southeast USA, and may have significant feedbacks to regional climate through the direct and indirect effects.

In this study, the cloud droplet activation parameterization of Kumar et al.(2009), which considers the competition between soluble and in soluble aerosols for water vapor during cloud droplet formation in ascending air parcels, is implemented in WRF-Chem. The water uptake properties of the biomass burning aerosol (required for predicting optical depth for direct radiative forcing, and CCN activity for indirect effects) are constrained using observations of fresh and aged biomass burning aerosol sampled during the 2008 ARctas campaign (Jacob et al., 2010). In order to isolate the direct and indirect effects, we first couple the aerosol module only with radiation module to estimate direct effect, and then with microphysics module for the indirect effect.

Use of a coupled lake model with WRF

Megan Mallard, Russ Bullock, Jonathan Gula, Chris Nolte, Kiran Alapaty, Tanya Otte

Large freshwater lakes, such as the Great Lakes, can play a significant role in regional climate, modulating temperatures inland and modifying passing air masses with fluxes of heat and moisture. As increasingly finer horizontal resolution is used, inland bodies of water represent a more significant challenge to regional climate modeling. Currently available methods of simulating lake surface temperatures (LSTs) in the Weather Research and Forecasting (WRF) model can result in large errors which also impact simulated ice coverage. Although for retrospective runs these errors can be mitigated by including observed LSTs, this approach is not viable for regional climate downscaling applications, because global climate models typically lack sufficient resolution to resolve lake temperatures.

FLake, a 1D bulk freshwater lake model, has recently been dynamically coupled with WRF, providing updated LSTs and ice coverage throughout the integration (Gula and Peltier 2012). FLake is a two-layer model, with a temperature-depth profile including a homogeneous mixed layer at the surface and a thermocline below. The shape of the thermocline is assumed, based on past theoretical and observational studies. Therefore, additional variables required for FLake to run are minimal, and it does not require tuning for individual lakes. These characteristics are advantageous for a downscaling approach, where little information is available about future changes in specific lakes. In the current work, retrospective year-long simulations are performed at 12 km grid spacing. The performance of the FLake model in the Great Lakes is evaluated by comparison with a 1/12^th degree global sea-surface temperature product. Also, resulting effects inland are evaluated by comparison with 2-m temperature observations and a high-resolution gridded precipitation product. Recommendations are made for the initialization and spin-up of WRF-FLake for regional climate modeling applications.

The impact of projected future emissions on global human mortality due to ozone and PM2.5 outdoor air pollution

Raquel A. Silva, Susan C. Anenberg, J. Jason West, Jean-François Lamarque, Drew T. Shindell, Daniel Bergmann, T. K. Berntsen, Philip Cameron-Smith, William J. Collins, Steven J. Ghan, Beatrice Josse, Tatsuya Nagashima, Vaishali Naik, David Plummer, Jose. M. Rodriguez, Sophie Szopa, Guang Zeng

Exposure to ground-level ozone and fine particulate matter (PM_2.5) has been associated with premature human mortality. The IPCC AR5 Representative Concentration Pathways (RCPs) project a range of plausible future global greenhouse gas and air pollutant emissions to 2100. The main objectives of this study are to quantify global mortality impacts of: 1) future ozone and PM_2.5 as projected in RCP scenarios; 2) future effects of climate change on air quality isolated from emission changes. We will obtain O₃ and PM_2.5 concentrations resulting from simulations of the RCP scenarios from five or more global models of atmospheric dynamics and chemistry, for a base year (present-day) and future control years, with present or future climate. All model outputs will be regridded to the same resolution to estimate multi-model medians per grid cell, and to calculate the differences between base year and each control year. Resulting premature deaths will be calculated using these medians along with epidemiologically-derived concentration-response functions, and future projections of population and baseline mortality rates, considering aging and transitioning disease rates over time. The spatial distributions of future global premature mortalities due to ozone and PM_2.5 outdoor air pollution will be presented separately. These results will strengthen our understanding of future air pollution health impacts as simulated with widely used emission projection scenarios, and of the contributions of direct emissions changes relative to climate change in affecting future air quality.

Mercury Modeling for North Carolina using the MATS modeling platform

Nicholas Witcraft

The NC Division of Air Quality used the MATS version of CMAQ to estimate mercury deposition within North Carolina. The primarily goal was to quantify the amount/percentage of mercury deposition in North Carolina that is from North Carolina sources.  A secondary goal was to quantify the percentage that is from international sources.  To accomplish these goals, a series of 'brute-force' zero-out runs were run.  The EPA modeling results for the MATS rule suggests that mercury deposition in North Carolina should decrease by 10% between 2005 and 2016.  The NC DAQ sensitivity modeling indicates that in 2005, approximately 15%  of the atmospheric mercury deposition in North Carolina comes from air emission sources located in North Carolina and by 2016 that fraction is expected to drop to 3.5%.  Finally, based on the sensitivity run that removed mercury emissions from the boundary conditions from the 2005 modeling, approximately 70% of the atmospheric mercury deposition in North Carolina originates from outside the central and eastern United States.  For 2016, the boundary contribution rises to 90%.

Multi-scale Air Quality and Health Impacts of U.S. Aviation Emissions

S.H.L. Yim¹, G.L. Lee¹, I.H. Lee¹, A. Ashok¹, F. Caiazzo¹, S.R.H. Barrett¹

¹ Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, United States

Recent has research found that the aircraft-attributable air quality impact could cause ~160 premature mortalities in the US each year. This estimation has only included aviation emissions within landing-takeoff cycle (<3000 ft above field elevation), and the long-term localized particulate matter exposure has not yet been fully captured. Since aviation activity is forecast to grow every year, the air quality and health impacts due to aviation become increasingly important to quantify and mitigate. To study these impacts of aviation emissions in the U.S., we use a multi-scale modelling approach to simulate the aircraft emissions-attributable PM_2.5 concentration on both regional and local scales. With boundary conditions provided from GEOS-Chem, CMAQ is used to simulate the regional air quality impact due to aviation emissions. This study investigates the air quality impact due to the aircraft emissions within the landing-takeoff cycle as well as those associated with full flight activity, and incorporates the effects of near-airport pollutants by using a hybrid dispersion approach. The increased premature mortality risk arising from aviation-attributable PM_2.5 exposure is estimated. To examine the potential health benefits of a jet fuel desulfurization scenario, we assess the regional and local air quality changes associated with reducing jet fuel average fuel sulfur content from 600 to 15 ppm.

A Simulation Study of the Climate Change Effects on Air Quality in Taiwan

Hsin-Chih Lai¹*, Jiun-Horng Tsai², Yu-Han Cheng³, Der-Ming Tsai⁴,

¹Department of Engineering & management of Advanced Tech., Chang Jung Christian University, Tainan 711, Taiwan

²Department of Environmental Engineering, National Cheng Kung University, Tainan 701, Taiwan

³Department of Information Management, Chang Jung Christian University, Tainan 711, Taiwan

⁴Department of information Communication, Kun Shan University, Tainan 710, Taiwan

Air pollution problems in Taiwan are very difficult to solve with particularity of its geographical location, seasonal monsoon and island terrain with high mountain make the transportation and diffusion of air pollutants more complex than the other Asian areas. Recently, climate change has strong and rapid effects on the regional atmospheric environment then coupled with the changing of air quality, it is very important to study the climate change effect on air quality in Taiwan. In this study, MM5 mesoscale meteorological model and Models-3/CMAQ air quality model are using to simulate the air quality variation under climate change. Scenarios for fixed emission sources (2003) are designed for the analysis of climate change impacts of seasonal and regional air quality over the past decade during different meteorological conditions. The results showed that the trend of temperature variation in the warm season decayed about 1 ~ 2 and rise up to 2 ° C in the cold season. Air quality simulation results show a good correlation of ozone and atmospheric temperature trends. The evolution of PM10 is closely related to changes in the wind field, wind direction changed from east to southeast during the past decade, also the average wind speed rise up to 1 m/s, that making stronger transportation of PM10 from the north to the south then accumulated in southern more 6ppm than before.

Effects of changes in emissions and climate change on global air quality: a study of the air quality co-benefits of GHGs mitigation

 

Yuqiang Zhang¹, J. Jason West¹, Meridith Fry¹, Zac Adelman¹, Raquel Silva¹, Steve Smith², Vaishali Naik³, Susan  Anenberg⁴,  Larry W. Horowitz⁵, Jean-Francois Lamarque⁶

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

²Joint Global Change Research Institute, Pacific Northwest National Laboratory, 5825 University Research Court,  Suite 3500, College Park, MD 20740, USA

³Atmospheric Physics, Chemistry, and Climate Group, UCAR GFDL, Princeton, New Jersey, USA

⁴U.S. Environmental Protection Agency, Washington, DC, USA

⁵Geophysical Fluid Dynamics Laboratory, National Oceanic and Atmospheric Administration, 201 Forrestal Road, Princeton, NJ 08540, USA

⁶Atmospheric Chemistry and Climate and Global Dynamics Divisions, National Center for Atmospheric Research, Boulder, CO 80307, USA

 

Air quality can be affected by both pollutant emissions and meteorological conditions which influence the transport and chemical transformation processes. Actions to mitigate Greenhouse Gases (GHGs) emissions therefore will yield air quality co-benefits from two mechanisms in the future: one directly through the reductions in the emissions of co-emitted air pollutants (short-term and local); another indirectly from slowing the influence of climate change on air pollution (long-term and global). We have completed simulations using the MOZART-4 global chemical transport model that are designed to quantify the co-benefits of GHG mitigation by these two mechanisms in future years.  Here we aim to explain the changes in ozone and fine particular mater (PM_2.5) that we modeled, in terms of changes in emissions and meteorological variables due to climate change. We emphasize the different factors controlling the air quality co-benefits by these two mechanisms, including changes in co-emitted air pollutants, biogenic emissions variations, soil NO_x emissions, and other meteorological factors such as water vapor, temperature, and surface solar radiation, and analyze the influence of these factors in contributing to the change of ozone and PM_2.5. We will present results for the changes in global air quality due to GHG mitigation in 2030, 2050, and 2100.  For this study, future emission scenarios were developed by the GCAM global energy-economics model as part of the Representative Concentration Pathways (RCP) process, where we focus on the RCP4.5 scenario, and the GCAM reference emission scenario. Future meteorological conditions are from general circulation model simulations of RCP4.5 and RCP8.5 from the Geophysical Fluid Dynamics Laboratory (GFDL). By analyzing combinations of emissions and meteorology from different scenarios, we isolate the co-benefits of GHG mitigation from reductions in co-emitted species and from those due to slowing down climate change. 

High-Resolution Three-Dimensional Plume Modeling with CMAQ

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

The goal of this research project is to develop a three-dimensional modeling technique that allows simulations of atmospheric plumes under unprecedented levels of grid resolution. The model developed overcomes traditional limitations to resolution while benefiting from the state of the science representations of complex chemical and physical processes in gridded air quality modeling systems. These objectives are achieved by applying an adaptive grid modeling technique to the Community Multiscale Air Quality modeling system (CMAQ). Adaptive gridding is a method to increase accuracy by dynamically restructuring the grid on which solution fields are estimated and providing refinement at the regions where accuracy is most dependent on resolution. Adaptive grids in air quality modeling have been previously described and an adaptive grid version of CMAQ has been developed as precursor to this work. However, adaptation in comprehensive air quality models has been limited to horizontal resolution. Accurate plume modeling also necessitates fine grid resolution along the vertical plane. In this work, the importance of vertical emissions distribution is demonstrated and shows that plume modeling calls for increased vertical grid resolution. This research has led to a fully-adaptive three-dimensional air quality model that can accurately represent the phenomenon and provide insight into plume dynamics unattainable with static grid models. Model evaluation is accomplished by simulating biomass burning plumes from selected wildland fire episodes in the Southeastern U.S.

Alaska adapted CMAQ model

Ketsiri Leelasakultum and Nicole Molders

The Community Multiscale Air Quality (CMAQ) model version 4.7.1 was adapted to represent Alaska conditions to be able to simulate the PM_2.5-concentrations in Fairbanks, the interior of Alaska during episodes in January-February and November, 2008 accurately. In doing so, Alaska specific initial and background conditions were developed. Further modifications include reducing the eddy diffusivity coefficients by half and scaling them according to the fraction of land-use, updating the minimum stomatal resistances for Alaska typical vegetation and reducing minimum PBL height in accord with observations. Since according to studies for other regions CMAQ model is likely to underestimate sulfate during winter, the background values of iron and manganese, which are catalyst for SO₂ oxidation reactions, were updated by data from observations made in Fairbanks. The liquid-water threshold for resolvable scale clouds was decreased by 50%. The parameterization for the sulfuric acid-water nucleation rate was changed to be more representative for low temperature conditions as they occur in Fairbanks. The model evaluation with the observational data from the official monitoring site at the State Office Building shows that the Alaska adapted CMAQ model captures the PM_2.5-concentrations well on days that have concentrations above the standard. The correlation coefficients are 0.52 and 0.31 for the January-February and November episode, respectively. Simulated sulfate are still underestimated with mean biases of -4.7 (-3.1) μg/m³ and normalized mean biases and errors of 70% (60%) for the January (November) episode. The evaluation suggests a reassessment of the partitioning of PM_2.5 in the emission inventory as sulfate is extremely underestimated as compared with other studies.

A comparison of model predictions with the RACM2 and CB05 mechanisms

Golam Sarwar, Barron Henderson, Daiwen Kang, George Pouliot, William Hutzell, Wendy Goliff, William Stockwell

We incorporated the Regional Atmospheric Chemistry Mechanism, version 2 (RACM2) into the Community Multiscale Air Quality (CMAQv5.0) modeling system. Model simulations are conducted using the RACM2 and the 2005 Carbon Bond (CB05) mechanism for September, 2006. The RACM2 has been developed for regional model applications. The Carbon Bond chemical mechanism was originally developed for summertime urban applications but the CB05 mechanism expands to more accurately simulate wintertime, pristine and high altitude situations. The CB05 mechanism, used in this study, is further revised with the updated toluene chemistry and chlorine chemistry. The modeling domain covers Canada, United States, and Mexico using 12-km horizontal grid-resolution and contains 35 vertical layers. Meteorological inputs are obtained from the Weather Research and Forecasting (version 3.3) model. Emissions are derived from the 2005 National Emissions Inventory. Emissions from soil and vegetation are derived from the Biogenic Emissions Inventory System (version 3.14). On a domain-wide average basis, the RACM2 simulation produces 8% additional ozone than the CB05 simulation due to differences in inorganic as well as organic chemistry. Compared to the CB05, the RACM2 enhances most oxidation products (nitric acid by 25%, hydrogen peroxide by 3%, aerosol sulfate by 10%, aerosol nitrate by 6%, ammonium by 10%, anthropogenic secondary organic aerosols by 42%, and biogenic secondary organic aerosols by 5%) but reduces organic nitrate products (peroxy acyl nitrate by 34% and organic nitrate by 51%). The accompanying paper describes the mechanistic reasons of the differences in model predictions and the temporal as well as spatial variation of the predictions of the two mechanisms. Model performances are calculated by comparing model predictions with observed data from the Air Quality System network, the Clean Air Status and Trends Network, the Chemical Speciation Network and the Interagency Monitoring of Protected Visual Environments network. The accompanying paper describes the model performance statistics for both mechanisms.

Effect of Land Cover on Atmospheric Processes and Air Quality

 

Zhining Tao^1,2, Joseph A. Santanello², Qian Tan^1,2, Mian Chin², and Christa D. Peters-Lidard²

1. Universities Space Research Association, 10211 Wincopin Circle, Columbia, MD 21044
2. NASA Goddard Space Flight Center, Greenbelt, MD 20771

 

Land surface properties play a key role in regulating land-atmosphere (L-A) processes. Soil water content shows drastic difference among various land coverage, which is of major relevance to the water and energy cycles that in turn impacts air temperature, boundary layer stability, precipitation, and air quality. This presentation shows the results of a modeling study of effect of land cover and its associated properties on some key L-A processes with focus on air quality. The newly developed NASA Unified Weather Research and Forecast (NU-WRF) modeling system couples the Land Information System (LIS) with WRF that allows the users to explore the L-A processes and feedbacks in an internal consistent way.  Three commonly used land cover datasets, i.e., USGS, UMD, and MODIS, have been explored in NU-WRF to investigate the land cover effects in the Continental United States (CONUS) domain. The weeklong simulation demonstrates the large differences in latent/sensible heat flux, soil temperature/moisture, PBL height, ozone and PM2.5 air quality.

Recent updates in the CMAQ windblown dust emission module

Daniel Tong^1,2, George Bowker³, Fantine Ngan⁴, George Pouliot⁵, Pius Lee⁴

1 NASA GES/DISC, NASA Goddard Space Flight Center, Greenbelt, MD

2 CSISS, George Mason University, Fairfax, VA

3 US Environmental Protection Agency, Clean Air Markets Division, Washington, DC

4 NOAA Air Resources Laboratory, Silver Spring, MD

5 US EPA Atmospheric Modeling and Analysis Division/NERL/ORD, RTP, NC

Dust emitted during windstorms is an important component of the atmosphere, affecting climate as well as air quality. Previously, we have created a windblown dust emission module for use within CMAQ. Here, we present three potential improvements to this module: chemical composition profile, soil moisture effect, and wind bias adjustment. Currently, the chemical profile of CMAQ windblown dust is based on anthropogenic dust composition, which may not be consistent with natural dust particles originating from deserts and semi-arid regions. We have developed a new approach to identify local dust events from the IMPROVE routine aerosol monitoring network, and reconstructed a database of dust events for the western United States. An updated chemical profile for natural dust particles is compiled through the analysis of the speciated aerosol data from these identified dust events and through the comparison of the US dust profile to previously reported profiles from elsewhere. We have also updated the soil moisture treatment in the dust module, as the previous codes did not consider variations in the units of soil moisture data provided by different meteorology models. Finally, we present the impact of wind bias on dust emission estimates, and propose an adjustment measure to improve dust modeling.

Evaluation of the ozone concentrations over the Metropolitan Area of Vitoria, Esperito Santo State, Brazil, using the CMAQ model

Taciana T. de A. Albuquerque1, Renato S. Marinho2, Ayres Loriatto1, Brigida G. Maioli1, Rita Y. Ynoue3, Erick G. Sperandio4, Alexandre S. Magalhaes1, Neyval C. Reis Jr.1, Jane M. Santos1.
albuquerque.taciana@gmail.com

1 Federal University of Esperito Santo - Environmental Engineering Department
2 Federal University of Esperito Santo - Department of Geography
3 University of Sao Paulo - Atmospheric Science Department
4 Federal University of Esperito Santo - Department of Computer Science

Ozone (O3) is one of the most important pollutants in the atmosphere, causing harmful health effects and agricultural damages. Growing levels of urbanization in developing countries have generally resulted in increasing air pollution due to higher activity in the transportation, energy, and industrial sectors, injuring the air pollution control programs. This study aims to evaluate the ozone levels in the Metropolitan Area of Vitoria (MAV), Esperito Santo State, Brazil, in comparison to the Models-3 Community Multiscale Air Quality Modeling System (CMAQ) simulations. It was used ozone measured data from a local monitoring dataset network (Laranjeiras, Cariacica, Enseada and Vila Velha-Ibes Stations) during May 2011 to compare with modeling results. Meteorological fields were modeled using the Weather Research and Forecasting model WRFv3.1, for the 31-day period (01 - 31 May, 2011) and after the SMOKE emissions model was applied to build a spatially and temporally resolved emissions inventory for a high resolution domain of 1-km (73 x 64 cells). The air quality simulation used local measured concentrations as initial and boundary conditions. Aerosol processes and aqueous chemistry in CMAQ (AERO4) were used, as well as the Carbon Bond V gas phase mechanism. The numerical results showed a good agreement with measurement data, however the concentrations were overestimated by the model. It was verified for Vila Velha Ibes Station a mean error of 1%, showing the best agreement with CMAQ model.

Keywords: Ozone concentrations, CMAQ model, urban air pollution.

Examining the performance of trace metals estimates in CMAQv5.0
K. Wyat Appel, Heather Simon, George Pouliot, Havala Pye, Sergey Napelenok, Farhan Akhtar and Shawn J. Roselle

The latest version of the Community Multiscale Air Quality Model (version 5.0) includes explicit tracking and treatment of trace metals (i.e. Fe, Ca, Mg, Mn, K, Si, Al, Ti), which in previous versions of the model were treated as a single soil species. The explicit treatment of trace metals is important, as they can make up a sizeable fraction of the total fine particulate matter (dependent on location and time of year) and can also affect the model chemistry. There are both natural (e.g. dust) and anthropogenic (e.g. coal combustion) sources of trace metals, and as such the model performance for these elements is particularly sensitive to the specification of both the emissions inputs and boundary conditions. The latest version of the CMAQ model contains an updated treatment for wind-blown dust and updates to the anthropogenic fugitive dust estimates, both of which have a large impact on the model estimates of trace metals. The CMAQv5.0 model has been used to simulate the entire year of 2006, and the model estimates of the trace metals has been compared to observations from the CSN and IMPROVE networks. The results of this evaluation will be presented, along with results from several model sensitivity simulations highlighting the impact that the wind-blown dust treatment and anthropogenic fugitive dust updates have on the model estimates.

A New Parallel Sparse Chemistry Solver for CMAQ

George Delic, HiPERiSM Consulting, LLC, P.O. Box 569, Chapel Hill, NC
27514

Historically CMAQ Gear and Rosenbrock chemistry solvers have relied on the JSparse [1] procedure to perform sparse LU decomposition followed by forward/backward solve steps. JSparse suffers from the deficiency that parallelism is allowed only at the instruction level in vector loops over cells for each block of the grid domain passed to the solver. The fundamental inhibitors to higher levels of parallelism in JSparse reside in the use of layers of indirect addressing subscripts that cannot be overcome by compiler options alone. A radical approach has been implemented at HiPERiSM in three steps to enable more parallelism. The first step is to replace JSparse by modern sparse solver algorithms (CSparse) developed to reduce arithmetic operation counts [2]. The second step is to translate selected CSparse C procedures into Fortran for integration into CMAQ. The third step is to modify CSparse algorithms to enable nested parallel loops. This last step is required because native CSparse procedures inhibit parallelism. This new formulation removes these inhibitors by a specific choice of indirect addressing of subscripts in critical loops to enable multiple levels of parallelism. The new CMAQ sparse solver (FSparse) has been implemented in the Rosenbrock algorithm and is ported to either multi-core (OpenMP) or many-core GPGPU (OpenACC) paradigms. A side benefit of FSparse is that various norms may easily be computed as the computation progresses to monitor numerical stability. Examples of such norms for Ax=y, include |A|, |x|, and |Ax-y|. Since matrix A in CMAQ is diagonally dominant, and |A| has a large magnitude, scaling has been applied to allow underflow, but avoid overflow exceptions. The FSparse version of CMAQ has been tested with compilers from the Portland Group, Intel Corporation, and the Absoft Corporation for CMAQ 4.7.1 with 64-bit Linux operating systems on Intel and AMD platforms using a 24-hour simulation for a domain of 2.3 million grid cells.
[1] M. Jacobson and R.P. Turco (1994), Atmos. Environ. 28, 273-284
[2] T.A. Davis, Direct Methods for Sparse Linear Systems, SIAM, Philadelphia, 2006.

Bayesian Analysis of a Reduced-form Air Quality Model

Kristen M. Foley, US EPA

Sergey L. Napelenok, US EPA 

Brian J. Reich, NC State University

The US EPA uses numerical air quality models to design emission control strategies for improving ambient ozone concentrations across the US.  A Bayesian hierarchical model is used to combine air quality model output and monitoring data in order to characterize the impact of emissions reductions while accounting for different degrees of uncertainty in the modeled emissions inputs. The probabilistic model predictions are weighted based on population density in order to better quantify the societal benefits/disbenefits of four hypothetical emission reduction scenarios in which domain-wide NOx emissions from various sectors are reduced individually and then simultaneously. Cross validation analysis shows the statistical model performs well compared to observed ozone levels. Accounting for the variability and uncertainty in the emissions and atmospheric systems being modeled is shown to impact how emission reduction scenarios would be ranked in terms of effectiveness in reducing maxium 8-hour ozone levels, compared to standard methodology.

Evaluation of WRF-CHEM Simulations with the Unified Post Processor (UPP) and Model Evaluation Tools (MET)

Agustin R. Garcia and M.A. Mora-Ramirez

Centro de Ciencias de la Atmosfera, Universidad Nacional Autonoma de Mexico, Mexico

An important step in the use of models is its evaluation. A coupled model with air quality and meteorology requires evaluation of both features. The WRF-chem model is an updated version of the Weather Research Forecast model that incorporates chemical transformations and complex gas-phase chemistry, photolysis and aerosols interactions. There are several programs that can handle the output for analysis and visualization like NCAR command Language (NCL), Grid Analysis and Display System (GrADS), and the Unified Post Processor Software Package (UPP) by Developmental Testbed Center (DTC).

The UPP replaced the WRF Pos Processor (WPP) and has enhanced capabilities to post-process output from different models as WRF-NMM, WRF-ARW and Global Forecast System (GS) among others. UPP interpolate the output from native grids to GRIB format, in order to be use for other programs.

For meteorological analysis of the WRF the Model Evaluation Tool (MET) is commonly used. This software is configurable and is a state-of-the-art suite of verification tools.

The WRF-chem evaluation methodology consists in process the output from WRF with UPP to produce input for MET system in order to compute statistical metrics and parameters between model results and observations. Modifications in the codes of UPP and MET were made in order to process the additional chemical variables generated byt e model, as ozone, sulfur dioxide, nitrogen oxides among others.

In this work a detailed description of modifications and a case study in Mexico City using surface stations are presented. The modifications made over different files in UPP and MET packages could be interesting especially for users and developers of the WRF-chem model concern about forecasting or analysis episodes in air quality

Model Evaluation Tools (MET) was developed at the National Center for Atmospheric Research (NCAR) through a grant from the United States Air Force Weather Agency (AFWA). NCAR is sponsored by the United States National Science Foundation

A Modeling Study of Ozone Production Rates during the Summers of 2002 and 2006

James Godowitch, Golam Sarwar, Shuang Chen

Ozone production rates (P(O₃)) have been calculated using Community Multiscale Air Quality (CMAQv5.0) model results for the summers of 2002 and 2006. Model simulations were conducted using the Carbon Bond 2005 (CB05) chemical mechanism. The modeling domain covers the continental United States, Canada, and Mexico using a 12-km grid cell size and contains 35 vertical layers, including a 20 m thick layer 1. Meteorological inputs were generated from the Weather Research and Forecasting (WRFv3.3) model. Anthropogenic emissions were derived from the 2002 and 2005 National Emissions Inventories (NEI), while natural emissions were determined by the Biogenic Emissions Inventory System (BEISv3.14).

Peak P(O₃) values were found in the early afternoon, which coincides with the highest photochemical activity. Values at night are negligible due to a lack of photochemical activity. Calculated daytime P(O₃) values are considerably higher in urban areas than rural locations. Major metropolitan areas exhibiting the highest production rates include: New York City, Atlanta, Houston, St. Louis area, and Los Angeles basin with peak values in these urban areas exceeding 25 ppb/hr. In contrast, P(O₃) values in rural areas are lower than 10 ppb/hr.Our calculated values are in good agreement with those determined by other investigators. A noticeable decrease in peak P(O₃) values of 5-25% occurred from summer 2002 to 2006, especially in the Ohio River Valley and Tennesse River Valley regions which experienced significant point source NO_X emission reductions between these periods, along the northeast corridor, portions of the mid-Atlantic region, and parts of Florida and California. However, peak values also increased by 5-25% in 2006 in some areas of the southern US and California compared to those in 2002. The accompanying paper will describe the method of deriving P(O₃) and will present the temporal evolution, spatial variability and change in P(O₃) from summer 2002 to 2006.

Sensitivity analysis of source regions to PM2.5 concentration at Fukue Island, Japan

Ikeda Kohei¹, Yamaji Kazuyo¹, Kanaya Yugo¹, Taketani Fumikazu¹, Pan Xiaole¹, Komazaki Yuichi¹, Kurokawa Jun-ichi², Ohara Toshimasa³

1 Japan Agency for Marine-Earth Science and Technology

2 Asia Center for Air Pollution Research

3 National Institute for Environmental Studies

We have continuously monitored the PM_2.5 mass concentration at Fukue Island located in the western part of Japan since February 2009. Although the influence of local emission is small at Fukue Island, it is reported that the observed daily averaged value exceeds an environmental standard (short-term standard value), which was newly introduced in September 2009. In order to estimate the contribution of trans-boundary transport from Asian continent, emission sensitivity simulations are performed by WRF/CMAQ for the full year 2010, and the sensitivities of source regions are analyzed. The source regions are divided to six areas: Japan, the Korean Peninsula, and four areas for China. In each sensitivity simulation, the anthropogenic emission is reduced by 20% in each source region. The control simulation in which the emission is not changed generally reproduces the seasonal variation of the PM_2.5 concentration at Fukue Island. However, the model tends to underestimate the PM_2.5 concentration; monthly averaged concentrations are 17-63 % less than those of observations. Throughout the year, the contribution of China is dominant; annual averaged sensitivities of 20 % reduction of anthropogenic emission for Japan, the Korean Peninsula, and China (total of four areas) are 2.5, 2.8, and 12 %, respectively. The sensitivities of each source region have seasonal changes. In summer, the influence of southern China increases, while those of northern and central China and the Korean Peninsula decrease.  In winter, while contribution of Japan decreases, the influence of the Korean Peninsula and central China increases.

Inter-comparison of Photochemical Modeling by U.S. EPA and CARB for the Calnex 2010 Study Period

James T Kelly and Kirk R Baker, Office of Air Quality Planning & Standards, US EPA, RTP, NC 27711

Chenxia Cai, Jeremy Avise, and Ajith Kaduwela, Planning and Technical Support Division, California Air Resources Board, Sacramento, CA 95812

Despite much progress in recent decades, many regions of California suffer from poor air quality due to ozone and particulate matter (PM) pollution. To develop optimal emission control strategies, air quality modeling systems are needed that can accurately predict the response of pollution concentrations to emission reductions. However, simulating air quality in California is challenging due to the complex terrain, diverse emissions, and other complicating factors. In May-June 2010, the Calnex field study was conducted to answer important scientific questions about emissions, chemical transformations, climate processes, transport, and meteorology in California. The study provides a rich dataset of observations from aircrafts, ships, and supersites, which are supplemented by California's routine monitoring networks. This dataset is being used by the California Air Resources Board (CARB) and US EPA in diagnostic evaluations of their photochemical modeling systems to improve understanding of their modeling capabilities and to identify areas for model development. In this study, predictions of the CARB and U.S. EPA modeling platforms for fine-scale (4-km) simulations of the Calnex 2010 study period are inter-compared and evaluated against observations. An emphasis is placed on isolating the causes of differences in predictions for PM components, ozone, and precursor species. Future work will involve improving both modeling systems based on the findings of this inter-comparison and evaluation.

Meteorological Evaluation in Support of a 2007-Based Air Quality Modeling Platform

Chris Misenis and Kirk Baker

As part of EPA's Office of Air Quality Planning and Standards development of a 2007-based air quality modeling platform, an evaluation of meteorological performance is necessary. This meteorological data will be used to support upcoming regulatory and policy assessments for a broad spectrum of pollutants. The Weather Research and Forecasting model (WRF) v3.3 has been applied for the entire year of 2007 on both 36 & 12 km continental United States domains. A vertical resolution of 35 layers was used up to 50 mb, with a 20-m layer spacing at the surface. WRF version 3.3 with the Advanced Research WRF (ARW) core was implemented, with several notable physics options: Pleim-Xiu land surface model, Asymmetic Convective Model version 2 planetary boundary layer scheme, Kain-Fritsch cumulus parameterization, Morrison double moment microphysics, and the YYTMG longwave and shortwave radiation schemes. The WRF model was initialized using the 12 km NAM analysis provided by NCDC. Analysis nudging for temperature and moisture (above the PBL only) and wind (above and below the PBL) was also used. The model simulations will be compared to observations of temperature, wind speed and direction, mixing ratio, precipitation, and solar radiation. A spatial and statistical evaluation of the model performance is presented here.

An application of CAMx process analysis tools: Exploring process contributions to extreme O3, NOX and SO2 conditions over NYC

David-anthony Murray, James Schwab, Kenneth Demerjian, Mark Beauharnois

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

The University at Albany Air Quality Forecasting Modeling System (AQFMS) is a state-of-the-art model that generates reliable daily and “day-ahead” air quality forecasts for the Northeastern United States.  The three major categories of processes which dictate regional air quality are horizontal and vertical transport within the prevailing meteorology, production from emission sources and chemical transformations.  The Advanced Research WRF (ARW) produces meteorological fields.  The Sparse Matrix Operator for Kernel Emissions (SMOKE) creates an inventory of emissions.  The Comprehensive Air Quality Model with extension (CAMx) handles both chemical processes and the integration of ARW-WRF and SMOKE in devising separate quantitative contributions to pollutant concentrations from process categories.  An AQFMS forecast represents the sum of these process contributions.  This forecast, though indicative of the temporal and spatial changes in the ambient condition, does not tell us exactly how and why those changes occurred.  High concentrations of criteria pollutants during “extreme” conditions could come about in many ways.  Process analysis takes a step back in numerical procedures to showcase the partial contribution of 18 different processes to the predicted concentration.  Production from emissions is divided into area and point sources. Horizontal and vertical transport in the form of advection and diffusion through the west, east, south, north, bottom and top boundary accounts for 12 of these processes.  Chemical transformations are broken up into contributions from gas phase and heterogenous reactions.  Dominant modes in which processes contribute to extreme predictions in the first model layer were explored in a 12km grid spacing over the New York Botanical Gardens.  Hours of the July 2010 model run were grouped into intraday periods of 5 to 6 hour lengths to reflect the diurnal pattern of human activity depicted in national time use surveys (Klepeis et al., 2001 and Phipps et al., 2009).  These intraday periods were classified here as the Ante Meridiem Commute (amc), Mid-day (mid), the Post Meridiem Commute (pmc) and Evening (eve).  Intraday periods that forecasted one or more hourly O₃, NO_x, or SO₂ concentrations equal to or greater than the monthly 90^th percentile for that particular pollutant were considered extreme conditions.  Process contributions are explored in the vertical layers up to the first 4km of the model atmosphere for these cases.  Individual events of consecutive intraday periods that spanned the length of more than a day were singled out to showcase process features.  Visualization of the dominant modes in which processes contribute to cases of extreme modeled pollutant concentrations in a selected grid cell over New York City offers enhanced explanatory value to the pursuits of operational forecasting, exposure surveying and regulatory planning.

How does the concentration of surface ozone change in CONUS due to a new paradigm for emission upgrade for the national forecasting system

Li Pan^{1, 2}, Daniel Tong¹, Pius Lee¹, Hyun Kim^{1, 2}, Tianfeng Chai^{1, 2} and Charles Ding¹

Affiliations:

¹: NOAA/OAR/ARL, Silver Spring, MD 20910

²: Earth Resources Technology, Inc.  Laurel, MD 20707

Accurate forecasting of surface ozone concentration is important for public health. To simulate ozone concentration, CMAQ model strongly relies on model inputs especially on its emission input. However the emission inventory used in CMAQ in forecasting application is often out of date. In forecasting model emission project factors are applied to the model emission processes, which represent emission inventory changes by upgrading the old inventory to the current inventory. In this analysis, we compare surface ozone concentration change due to two emission scenarios in July of 2011. Model results are verified using ground station observations and measurements collected during the NASA DISCOVER-AQ period. Whether CMAQ performance in ozone prediction is improved under a new inventory and how emission change influences ozone precursors and how the change is finally reflected on surface ozone concentration have been investigated through process analyses. This would shed insight on realistic configuration of the emission projection processes for surface Ozone concentration forecasting.

Process analysis of PM concentration over Pearl River Delta region, China using MM5-CMAQ model

Qi Fan¹, Wei Yu¹, Shaojia Fan¹, Jing Lan¹, Yerong Feng²

1 Department of Atmospheric Sciences, Sun yat-sen University, Guangzhou, 510275, China

2 Guangzhou Central Meteorological Observatory, Guangzhou, 510080, China

As one of the three large economic regions of China, the regional air pollution, in which the primary is aerosol pollution, is serious over the Pearl River Delta (PRD) region. The main synoptic systems influencing the air quality include the Stationary Front, the Warm High, the Tropical Cyclone and so on. This study focuses on the influences of the warm high on the aerosol pollutions, with the simulations by Models-3/CMAQ system and the observations from China “973” national projects in 2004.

The results show the spatial distributions of air pollutants are circular around Guangzhou and Foshan cities with high emissions. The primary pollutant is particulate matter over the PRD. The pollution range of NO_X is bigger than that of SO₂, though with lower concentrations. Both MM5 and CMAQ show reasonable performance for major meteorological variables (i.e., temperature, relative humidity, wind direction, planetary boundary layer height) with normalized mean biases (NMBs) of 4.5–38.8%. The temporal variations of surface concentrations SO2, NO2, O3 and PM2.5 were captured well by CMAQ model. Relatively poor performance was found in the simulated maximum concentrations of all pollutants, the CMAQ systematically underpredicted the mass concentrations.

The process analysis (PA) results show that the emission, dry deposition, horizontal transport and vertical transport are four main processes to air pollutants. The contributions of horizontal and vertical transport processes were different during the period, but in all, these two processes contributed to the removal of air pollutants. Besides, the contributions of the same physical process were different for various pollutants, the dry deposition was vital to the removal of PM10. NOx was affected by the transport process obviously. For this high pressure case, the contributions of various processes show higher correlations in the cities with similar geographical environment. According to the statistic results, the cities in PRD region can be divided into four types with different features.

Simulation of PM2.5 Exposures in Metropolitan Statistical Areas (MSA) Using CMAQ and Intercomparison to The American Cancer Society Study Methods, Evaluation And Application

JAEMEEN BAEK (1), Michelle C. Turner (2), Sinan Sousan (1), Jacob Oleson (1), Daniel Krewski (2), Gregory Carmichael (1), Susan M Gapstur (3), Michael J Thun (4), Charles Stanier (1)

(1) University of Iowa

(2) University of Ottawa

(3) American Cancer Society

Epidemiological studies have shown that elevated concentrations of fine particulate matter with aerodynamic diameter less than 2.5 micrometer (PM2.5) are related with cardiovascular and respiratory related hospital admissions or mortality rates. PM2.5 measurements are crucial to understanding exposures, but they are limited by time (one measurement for every three to six days) and space (discrete point locations). Adverse impacts of PM2.5 vary with PM2.5 constituents but there are fewer measurements available for detailed PM2.5 constituents than total PM2.5 mass. Chemical transport models, such as the Community Multiscale Air Quality (CMAQ) model, provide another way to obtain exposure estimates, but this approach has seen limited use in epidemiological studies due to uncertainties in the resulting exposures. The goal of this study is to identify pros and cons in using CMAQ simulated PM2.5 concentrations vs. PM2.5 measurements in U.S. Metropolitan Statistical Areas (MSAs) containing health outcome data that were used in the American Cancer Society (ACS) study of particulate air pollution and mortality. 3D model skill in the areas of focus for the epidemiological study will be discussed, both with and without data assimilation of surface PM2.5 measurement. The effect of model resolution (36, 12 and 4km domains) on model error will also be discussed.

Characterization of Air Quality Model Performance Using Land Use Regression Model

Yadong Xu, William Vizuete, Marc Serre

In the past few years, Bayesian Maximum Entropy (BME) integration of air monitoring observations and numerical model predictions has been used to improve spatial predictions of ozone concentrations. In our study, this approach is used to provide ozone exposure assessment for epidemiology studies. To obtain optimal estimates of exposure, the model predictions weighed according to model performance are integrated with the observations treated as an error free proxy. Therefore, how we characterize the performance of the air quality model is critical to the accuracy and the precision of ozone estimations.

Rather than assuming the model performance is homogeneous across the domain as in previous studies, we developed a stochastic land use regression model to characterize the prediction errors and also the distribution of the prediction errors that change across space/time.

The development of this model performance land use regression model uses predictor variables obtained through geographic information systems (GIS), such as traffic representations and population density, and also the gridded fractional land use coverage data in the chemical transport models.

We apply this methodology in the continental U.S domain, using the U.S Environmental Protection Agency (EPA) Air Quality System (AQS) ozone monitoring network in combination with outputs from Community Multi-scale Air Quality (CMAQ) model. We will show the improvements of the accuracy and the precision of ozone estimations across the U.S.

Application of WRF/Chem over East Asia: Evaluation, Seasonality, and Aerosol Feedbacks

Xin Zhang and Yang Zhang

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

Weather Research and Forecasting (WRF) model with chemistry (WRF/Chem) that includes online-coupled meteorology and chemistry is under rapid development since its public release. Application and evaluation of WRF/Chem over various testbeds will provide useful information for further model improvement. East Asia provides an ideal testbed for testing and improving model performance of WRF/Chem, because of high emissions of primary air pollutants as a result of dense population and complex mixtures of anthropogenic and natural sources, a complex weather system that favors the formation of secondary air pollutants and the export of air pollutants to downwind countries through long range transport, potentially strong feedbacks of greenhouse gases and aerosols into meteorology and climate systems, and a need to develop effective emission control strategies to reduce air pollution and associated human exposure and mitigate adverse climate changes. In this work, WRF/Chem version 3.3.1 is applied over East Asia for January, April, July and October 2005 to simulate the fates of air pollutants and their interactions with meteorology and associated seasonalities. Simulated surface meteorological variables are evaluated using hourly global surface observational data from the National Climatic Data Center (NCDC). Simulated chemical and radiative properties are evaluated using surface measurements in China including Hong Kong and Taiwan and Japan, as well as satellite data from Terra-MODIS, MOPITT, SCIAMACHY and OMI.

Model evaluation shows that WRF/Chem simulates temperature and water vapor mixing ratios at 2 meters reasonably well, but significantly overpredicts wind speeds at 10 meters for the four months with the largest deviation in January (with a normalized mean bias (NMB) of 113%). It well simulates precipitations in October but overprediction occurs in the other three months with the largest NMB of 32% in January. The surface mixing ratios of CO, NO_x and SO₂ are significantly underpredicted and that of O₃ is overpredicted except for Japan, with NMBs of -75% to -44%, -94% to -64%, -83% to -38%, and 3.1% to 160% respectively. Surface PM₁₀ concentration is underpredicted with the largest underprediction (NMB of -49%) in July, except for Japan in April when an overprediction (NMB of 37%) occurs due likely to high dust concentrations from the Gobi Desert. Column CO mass abundance is well predicted in all months except for April (with an NMB of -25%). Observed seasonalities of column NO₂ and SO₂ mass abundances are reproduced reasonably well, with the highest concentration in January and the lowest in July. Simulated column O₃ mass abundance is well simulated in October; it is overpredicted in January and April and underpredicted in July. Column AOD is well predicted in October but underpredicted in other months, especially April with an NMB of -39%. Sensitivity simulations are being conducted to address uncertainties in anthropogenic and dust emissions and initial and boundary conditions used in the simulations. The comparison of the simulations with and without emissions of anthropogenic aerosol using improved inputs will be conducted to examine aerosol direct and indirect effects over East Asia.

Contributions of major industrial emissions sources in the United Kingdom to projected reduction of population exposure to PM2.5

Charles Chemel (1), Xavier V. Francis (1), Nicholas Good (1), Xin Kong (1), Ranjeet S. Sokhi (1), Bill Collins (2), Gerd Folberth (2), and Bernard E. A. Fisher (3)

(1) Centre for Atmospheric & Instrumentation Research, University of Hertfordshire, Hatfield, UK; (2) Met Office Hadley Centre, Exeter, UK; (3) Risk and Forecasting Science, Environment Agency, Reading, UK

The management of regulated industrial emissions sources is essential for the development of a cost effective strategy to meet pollutant specific limit and target values set by legislative bodies, aimed, in particular, to reduce health impacts. We assess, using results from an application of the Community Multiscale Air Quality modelling system, the impact of major industrial emissions sources in the United Kingdom to projected reduction of population exposure to PM2.5 from 2006 to 2020. This work indicates that those emissions sources would make up 5 to 15% of the annual mean PM2.5 concentrations across most of the United Kingdom in 2020, a reduction of about 5% in comparison with 2006. The reduction in PM2.5 concentrations over the UK attributed to climate change is found to be much smaller that that associated with emission reductions (by a factor of about five). Results are discussed in relation to source-specific projected emission reductions. The contribution of individual PM2.5 species to the projected reduction of population exposure to PM2.5 is quantified in terms of risk of early death and implications for policy are discussed.

A Ensemble-Trained Bayesian Method for Source Apportionment of PM2.5

Sivaraman Balachandran

Howar H. Chang

Jorge E. Pachon

James A. Mulholland

Armistead G. Russell

A source apportionment (SA) technique is developed that uses an ensemble of SA methods and a Bayesian technique to estimate ensemble averages and uncertainties. This method allows for the development of new source profiles that can be used in a chemical mass balance (CMB) model to apportion sources of fine particulate matter (PM2.5). Source impacts from a short term application of three receptor-based models and one chemical transport model, the community multiscale air quality (CMAQ) model are used as inputs into the ensemble. The method can also be used to evaluate differences between various SA methods and to estimate uncertainties in source impact estimates from chemical transport models

The method uses a two step process to estimate uncertainties in SA results. These uncertainties are used as weights in calculating the ensemble average. In the first step, all four methods are treated equally to calculate an un-weighted initial ensemble average. Next the root mean square errors (RMSEs) between each method's source impact and the initial ensemble average is calculated. In the Bayesian framework, a non-informative prior is assumed for the four SA methods and the RMSEs are treated as the updated data. Next, a posterior distribution of SA method uncertainties is calculated. Monte Carlo sampling of the posterior distribution is used to estimate weighted ensemble averages for each day of the short term application (30 days for summer and winter each) of the four SA methods. For each day in the ensemble, 30 realizations of source impacts are estimated, which represent distributions of source impacts.

For each realization of source impacts, source profiles are calculated, resulting in 30 estimates of source profiles for each day of the ensemble. This process is done for winter and summer ensemble separately resulting in 900 summer and 900 winter source profiles. Source apportionment is conducted for along term PM2.5 data set using a CMB application. For each day in the long term data set, 10 sets of source impacts are calculated using 10 source profiles that are sampled from these distributions of 900 source profiles. The 10 SA results for each day can be viewed as a distribution of possible daily source impacts rather than a single value with an estimated uncertainty.

The BlueSky Western Canada Wildfire Smoke Forecasting System

Kenneth Craig¹, Sean Raffuse¹, Steve Sakiyama², David Lyder³, George Hicks II⁴

¹Sonoma Technology, Inc.
²British Columbia Ministry of Environment
³Alberta Environment and Sustainable Resource Development
⁴Department of Earth & Ocean Sciences, University of British Columbia

Through an inter-agency partnership, the British Columbia Ministry of Environment has developed an integrated system for forecasting the spread of smoke from wildfires within British Columbia (BC) and Alberta. This forecast system, a first of its kind in Canada, is based on the USDA Forest Service BlueSky Framework, which was customized for Canadian wildfire applications. The forecast system couples satellite-based forest-fire hot spot information and forest fuel consumption estimates from the Canadian Wildland Fire Information System (CWFIS), MM5 meteorological forecasts at 4-km horizontal resolution produced by the University of British Columbia (UBC), and the NOAA HYSPLIT dispersion model. The Western Canada BlueSky system began producing daily smoke forecasts on a 4-km resolution receptor grid covering BC and Alberta during the 2010 wildfire season. A qualitative evaluation of the forecasts indicates that the system does a good job at predicting regional smoke coverage in BC during large wildfire outbreaks. The system also captures long range smoke transport from BC into Alberta and other provinces.

In 2011, UBC's 12 km resolution MM5 grid was added to facilitate smoke forecasts over all of western Canada. For 2012, the SmartFire2 fire processing system is being added to reduce redundant fire detections and to facilitate the use of additional fire information data sources as they become available. Future developments will include expanding modeling domains, inclusion of carryover smoke, and optimization to relieve computational burdens. Forecast products are made available via (http://www.bcairquality.ca/bluesky), to inform the public, decision makers, and other interests that impacted by wildfire smoke.

Assessment of air quality response to anthropogenic emission changes over Taiwan using response surface modeling method

Xinyi Dong1, Joshua S. Fu1, Carey Jang2, and Hsin-Chih Lai3, and Ciao-Kai Liang4

1Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, TN
2Office of Air Quality Planning and Standards, USEPA, Research Triangle Park, NC
3Chang Jung Christian University, Environmental Research and Information Center, Taipei, Taiwan
4Department of Air Pollution Control, Taiwan EPA, Taipei, Taiwan

Air quality is usually strong correlated with local anthropogenic emissions. A number of studies have been conducted with air quality models for Europe and United States: Tsimpidi et al. (2007) applied a three-dimensional chemical transport model (PMCAMx) for in January and July 2002 to examine the NH3 emission impact on inorganic aerosol; Pinder et al. (2007) applied the same model for July 2001 and January 2002 and reported that NH3 control in winter is more cost-effective to reduce PM2.5 than controlling NOx and SO2; Derwent et al. (2009) reported the nonlinear mass concentrations of PM with precursor emission in UK based on air parcel trajectory. While these studies were targeted to examine impacts of NH3 on total PM2.5 over populated area, knowledge of impacts of anthropogenic emission on ozone and aerosol formation in Taiwan, which is one of the most populated area, is still very limited. In addition, the remaining uncertainties, such as the nonlinearity and seasonality of ozone and aerosol responses, from different emission sectors remain unknown from the published studies.

Therefore, in this study the integrated MM5/CMAQ modeling system is applied over Taiwan to examine the O3 and PM2.5 response to ammonia emission changes using Response Surface Modeling (RSM) method. RSM has been successfully applied in China to predict O3 and PM2.5 response under sector-dependent emission control (Xing et al., 2011; Wang et al., 2011). In this study, seasonality of inorganic PM2.5 response is also investigated with four different months (January, April, July, and October) as representatives for different seasons in 2007. Impacts of ammonia emission from four different sub-areas in Taiwan is also identified and analyzed to quantify the inter-state transport impacts on inorganic PM2.5 formation over Taiwan.

Reference:

Pinder, R. W.; Adams, P. J.; Pandis, S. N. Ammonia Emission Controls as a Cost-Effective Strategy for Reducing Atmospheric Particulate Matter in the Eastern United States. Environ. Sci. Technol. 2007, 41, 380-386.

Tsimpidi, A. P.; Karydis, V. A.; Pandis, S. N. Response of Inorganic Fine Particulate Matter to Emission Changes of Sulfur Dioxide and Ammonia: The Eastern United States as a Case Study. J. Air Waste Manage. Assoc. 2007, 57, 1489-1498, DOI: 10.3155/1047-3289.57.12.1489.

Wang, S.X.; Xinag, J.; Zhu, Y.; Fu, J.S.; Hao, J.M. Impact assessment of ammonia emissions on inorganic aerosols in east China using response surface modeling technique. Environ. Sci. Technol. 2011, 45, 9293-9300.

Xing, J.; Wang, S. X.; Chatani, S.; Zhang, C. Y.; Wei, W.;Klimont, Z.; Cofala, J.; Amann, M.; Hao, J. M. Projections of Air Pollutant Emissions and its Impacts on Regional Air Quality in China in 2020. Atmos. Chem. Phys. 2011, 11, 3119-3136, DOI: 10.5194/acp-11-3119-2011.

Evaluation of PM2.5 control policies through application of Source Apportionment Technology in the Pearl River Delta region

Dongwei Wu¹, Jimmy Fung^1,2, Alexis Lau¹ and Teng Yao¹

^{1Division of Environment, Hong Kong University of Science & Technology, Kowloon, Hong Kong}

^{2Department of Mathematics, Hong Kong University of Science & Technology, Kowloon, Hong Kong}

Hong Kong and the enclosed Pearl River Delta Region (PRD) situated on China's south coast have been experiencing severe air pollution problems with increasing population and booming economics. To solve the complex air pollution problem in this region, one of the main challenges is to identify the source categories / regions of the pollutants (i.e. ozone and PM_2.5) which can cause serious health and environmental damage. In this study, we apply the Particulate Source Apportionment Technology (PSAT) which has been implemented in the Comprehensive Air quality Model with extension (CAMx) to identify source contribution in the Hong Kong and Pearl River Delta region to PM_2.5 levels of all cities for different time.

The results indicate that average contributions from emission sources within HK/PRD region to ambient PM_2.5 level is various amount different cities and the ranges are 18.3%-49.6% in December and 44.6%-71.8% in April. PM_2.5 level in downwind regions is significantly affected by the emission sources in upwind areas. Mobile, point and area source always are top three source categories in this region except for Hong Kong. The dominant source categories contributing to Hong Kong's PM_2.5 are marine and mobile emission, while area and point emission contributions are similar. The result shows that by reducing mobile and point source emissions in the region, as well as the area emission could potentially lower the PM_2.5 level in the whole PRD region.

Using CAMx/HDDM to estimate ozone levels across the USat various levels of anthropogenic NOx and VOC

Chris Emery, Greg Yarwood, Jaegun Jung

ENVIRON International Corporation

Nicole Downey

Earth System Sciences, LLC

Abstract

The US EPA is conducting a Risk and Exposure Assessment (REA) for the current review of the ozone National Ambient Air Quality Standard (NAAQS). In the past, as well as for the first draft of the current REA, rollback techniques have been applied in 12 cities to estimate distributions of 1-hour ozone after monitored Design Values (DVs) are reduced to the proposed standard. The rollback technique does not change the distribution of ozone as precursor emissions are reduced. NOx reductions can both increase or decrease ozone, changing the ozone distribution. We present an application of the CAMx photochemical grid model with the High-Order Decoupled Direct Method (HDDM) to estimate second-order ozone sensitivity to US anthropogenic NOx and VOC emissions at 23 cities and all CASTNET sites across the US. HDDM accounts for non-linear ozone response to precursor emission changes efficiently in one model run instead of running a matrix of NOx/VOC emission levels (as in a "brute force" method). CAMx with HDDM was run for the entire year of 2006 on a nested grid system that resolves the US at 12 km grid spacing. The accuracy of HDDM-predicted ozone response to large emission changes was evaluated by comparing against brute force simulations with 100% and 0% US anthropogenic emissions. By post-processing the CAMx/HDDM results we can efficiently generate annual time series of hourly ozone at each location for any NOx and/or VOC emission level, as well as ozone DV response surface diagrams. We are using these results to explore specific policy-relevant questions, such as how the distribution of ozone concentrations "just meeting" a proposed standard derived from the quadratic rollback technique compares to the distribution predicted by a photochemical grid model.

Development and Evaluation of Two Reduced Form Versions of the CMAQ Modeling System

Kristen Foley, US EPA

Sergey Napelenok, US EPA

Sharon Phillips, US EPA

Cary Jang, US EPA

Due to the computational cost of running regional-scale air quality models, reduced form models (RFM) have been proposed as computationally efficient simulation tools for characterizing the pollutant response to many different types of emissions reductions. EPA has developed two types of reduced form models based upon simulations of the Community Multiscale Air Quality (CMAQ) modeling system. One is a based on statistical response surface modeling (RSM) techniques and the other is based on sensitivity calculations from the Higher-Order Decoupled Direct Method in 3 dimensions (DDM-3D). Both types of models are used to estimate the air quality changes associated with emissions reductions of NOx, SO2 and PM2.5 from EGUs and other sectors in the eastern United States. This study provides a direct comparison of the RSM and DDM RFMs in terms of: computational cost, model performance against brute force runs, and model response. The results are used to identify the different strengths and weaknesses of the RSM and DDM RFM approaches when used for control strategy assessments.

Global net radiative forcing responses to regional CO and NMVOC reductions

Meridith M. Fry¹, M. Daniel Schwarzkopf², Zachariah Adelman¹, J. Jason West¹

¹Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.

² Atmospheric Physics, Chemistry, and Climate Group, NOAA GFDL, Princeton, New Jersey, USA.

We analyze the global annual net radiative forcing (RF) impacts of regional reductions in anthropogenic carbon monoxide (CO) and non-methane volatile organic compound (NMVOC) emissions due to changes in the tropospheric concentrations of ozone (O₃), methane (CH₄), and aerosols.  We present the RF responses to CO and NMVOC emission reductions from 10 regions (North America, South America, Europe, Former Soviet Union, Southern Africa, India, East Asia, Southeast Asia, Australia and New Zealand, and Middle East and Northern Africa).  The global chemical transport model MOZART-4 is used to simulate the tropospheric concentration changes, using the IPCC AR5 Representative Concentration Pathway 8.5 (RCP 8.5) emissions inventory for 2005 and global meteorology from the Goddard Earth Observing System Model, version 5 (GEOS-5) for the years 2004-2005.  We utilize the NOAA Geophysical Fluid Dynamics Laboratory standalone radiative transfer model to calculate the stratospheric-adjusted net RF for each regional CO and NMVOC reduction, relative to the base.  We present estimates of the global annual net RF and the RF by latitude band (90^oS – 28^oS, 28^oS – 28^oN, 28^oN – 60^oN, and 60^oN – 90^oN).  The net RF distributions for the CO and NMVOC reductions show widespread cooling across the northern and southern hemispheres corresponding to the patterns of O₃ and CH₄ decreases, and localized positive and negative net RFs due to increases and decreases in aerosols.  The strongest annual net RF impacts occur within the tropics (28^oS – 28^oN) followed by the northern mid-latitudes (28^oN – 60^oN), for all CO emissions reductions and many of the NMVOC reductions.  These estimates of RF by world region are intended to inform coordinated actions addressing air quality and climate forcing.

Estimation of 3-D VOC Reactivities Through Adjoint Sensitivity Analysis

Saba Hajaghassi, Amir Hakami

The Ozone formation potential of volatile organic compounds (VOCs), referred to as incremental reactivity, varies significantly among species as well as for various locations and times. Reactivity of organic compounds was traditionally calculated in box model simulations that do not account for physical variability in the atmosphere. Previous efforts for three-dimensional simulation of organic reactivities have been limited to forward sensitivity analysis methods where responses of ozone at various receptors to uniform changes in source emissions are calculated. This work is focused on quantification of three-dimensional reactivities using adjoint sensitivity analysis tools. The adjoint of CMAQ with SAPRC-99 chemistry is used to elucidate location-dependent reactivities of 28 (lumped or explicit) species while preserving source specificity over a North American domain for the summer of 2007. Our results indicate the adjoint method is capable of capturing spatial variability. For most of the compounds our 3-D reactivities are comparable to the box model scales but notable (and at times significant) differences exist. Impact of different base mixtures on reactivity calculations is also explored but no significant impact on calculated reactivity scales is observed.

Development of a Mixtures Characterization Toolkit: Estimating air pollution source impacts to investigate air quality and human health associations using time-series epidemiologic analysis

Heather A. Holmes¹, Marissa L. Maier¹, Mariel Friberg¹, Sivaraman Balachandran¹, Cesunica Ivey¹, Yongtao Hu¹, Armistead G. Russell¹, James A. Mulholland¹, Stefanie E. Sarnat², Jeremy A. Sarnat², Matthew J. Strickland², Andrea Winquist², Mitchel Klein², Paige E. Tolbert²  

¹Georgia Institute of Technology, Atlanta, GA, USA ²Emory University, Atlanta, GA, USA

 

Air pollution concentrations measured from regulatory monitoring networks are commonly used as air quality metrics in time-series epidemiologic analysis to investigate air quality and human health associations.  While these data provide useful indicators for air pollution impacts in a region, the data are limited temporally, spatially and chemically.  Additionally, the species concentrations cannot directly identify the emission sources or characterize the pollutant mixtures.  These data in combination with chemical transport models (CTM) and source apportionment (SA) techniques can be used to characterize pollutant mixtures, sources and species impacting both individual locations and wider areas.  Extensive analysis using a combination of these approaches may be beneficial for health studies whose goal is to assess the health impacts of pollutant mixtures.  As part of the EPA-funded Southeastern Center for Air Pollution and Epidemiology (SCAPE), a Mixtures Characterization (MC) Toolkit is being developed to effectively analyze air pollution and air quality modeling data to better understand and quantify how emission sources combine to impact air quality and to provide metrics for use in health assessments. 

Statistical data analysis is done using the species concentrations to calculate a time-series of spatial air pollution metrics based on both area and population weighted averages.  Source impacts are estimated using receptor oriented SA methods (CMB, PMF, integrated indicators) and source oriented modeling from a CTM.  The CTM used in this work is CMAQ, which uses emissions modeling (SMOKE) to generate a spatial and temporal allocation of the source emissions.  The MC Toolkit incorporates two novel SA techniques, the first an ensemble method that generates new source profiles for CMB based on an ensemble-trained approach utilizing SA results from CMB, PMF and CMAQ.  The second is a hybrid source-receptor model approach that adjusts the CMAQ source impacts based on a scaling factor obtained using CMAQ simulations and observations in a CMB-approach.  In addition to source characterization, exposure estimates are a knowledge gap in health studies, thus another objective of the MC Toolkit is to develop spatial averaging and temporal interpolation components to improve the concentration distribution modeling for species concentrations and source impact estimates.

This work will present results using the MC Toolkit for two regions with differing source emissions, St. Louis, Missouri and Atlanta, Georgia.  There is a significant impact of point source emissions, e.g., chemicals manufacturing and metals processing, on the air pollution in St. Louis.  While in Atlanta large biogenic emissions interact with emissions from mobile sources and power plants which lead to a large amount of secondary organic aerosol formation.  In the future, the MC Toolkit will be used with observations from regulatory and specialty monitoring networks to investigate the extent of city-to-city variability in pollutant mixtures and associations with health outcomes. 

*This research was supported, in part, by USEPA grant R834799. Its contents are solely
the responsibility of the grantee and do not necessarily represent the official views of the
USEPA. Further, USEPA does not endorse the purchase of any commercial products or
services mentioned in the publication.  Additional support was made possible by grants from Georgia Power and the Southern Company.

Impacts of Interstate Transport of Pollutants on Ozone Air Quality Attainment in the Mid-Atlantic Region

Kuo-Jen Liao

Xiangting Hou

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

Impacts of Interstate Transport of Pollutants on Ozone Air Quality Attainment in the Mid-Atlantic Region

Xiangting Hou and Kuo-Jen Liao

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

Ozone is form in the atmosphere through photochemical reactions and has been found to be associated with adverse human health effects. One important indicator of regional air quality is the daily maximum 8-hr average ozone (MDA8h O₃) which is regulated by National Ambient Air Quality Standards (NAAQS) under the U.S. Clean Air Act. Summertime ozone has caused nonattainment of NAAQS over the Mid-Atlantic region in the U.S. The goal of this study is to investigate how emissions from neighboring regions affect ozone air quality attainment in the Mid-Atlantic region. We used the U.S. Environmental Protection Agency's (EPA) Community Multiscale Air Quality Model (CMAQ) version 4.7.1 to simulate three-dimensional gridded concentrations of ambient ozone, fine particulate matters and other air pollutants for the episode of August 1-15, 2007. Observation shows high MDA8h O₃ concentrations in the Mid-Atlantic during the episode. The modeling domain covers the whole eastern U.S. and was divided into four regions (CENRAP, LADCO, MANE-VU and SEMAP). A uniform grid of 12 by 12 km horizontal cells with 34 vertical layers was employed in the simulations. Daily maximum 8-h ozone in four nonattainment areas (Pittsburgh-Beaver Valley, Washington DC, Baltimore and Philadelphia-Wilmin-Atlantic City) in the Mid-Atlantic area was specifically investigated. Day-to-day comparison between modeled and observed MDA8hO₃ shows that the CMAQ model underestimates MDA8hO₃ for peaks ozone days. The decoupled direct method 3D (DDM-3D) in the CMAQ version 4.7.1 was used for calculating sensitivities of MDA8h O₃ in the four nonattainment areas to precursor emissions from the four regions. The results of sensitivity analyses show that MDA8h O₃ in the Pittsburgh-Beaver Valley area was attributed to anthropogenic NOx emissions from the MANE-VU, LADCO and SEMAP regions. Overall, emissions of anthropogenic NOx and VOC from the MANE-VU contributed to high ozone levels in the four nonattainment areas in the Mid-Atlantic region. It's expected that reductions in anthropogenic NOx and VOC emissions from the MANE-VU were effective for improving ozone air quality over the nonattainment areas in the Mid-Atlantic area. Detailed investigation of how emissions from each of the states in MANE-VU affect ozone air quality attainment in the Mid-Atlantic region will be required in the future.

Extending the Hybrid Source Apportionment Method by Spatial Interpolation of Source Impact Adjustment Factors

Cesunica Ivey¹, Heather Holmes¹, Yongtao Hu¹, Armistead G. Russell¹, and James A. Mulholland¹

¹Georgia Institute of Technology, Atlanta, GA

Data from air quality studies are increasingly used in health studies to assess associations between ambient concentrations of air pollutants and health. Of interest are the sources of pollutants and the magnitude of the impact that the sources have on ambient concentrations. Source impacts cannot be directly measured, leading to the application of source apportionment methods to measurable air quality indicators (e.g. daily pollutant concentrations, daily emissions from regulated sources, species profiles of emission plumes) to determine source impacts. Current source apportionment methods are derived from receptor modeling (RM) or chemical transport modeling (CTM) techniques. Both techniques present inherent challenges and uncertainties including the absence of physical and chemical processes in RM and the exclusion of monitor data in CTM, in most cases.

A hybrid source apportionment method was developed that utilizes aspects of both RMs and CTMs to integrate physical and chemical processes, observation data, as well as uncertainty estimates for model inputs. Previously, this hybrid method was applied to Chemical Speciation Network (CSN) data to determine source impacts at receptors for reporting days in January 2004. Here, the hybrid method was extended to perform source apportionment for the contiguous U.S. by spatial kriging of source impact adjustment factors, which modify original source impact simulations to reflect observed concentrations. These factors were applied to the original source impact estimates to reconstruct species concentrations for PM_2.5, EC,SO₄^2-, K, Fe, Se, and Sigat Interagency Monitoring for the Protection of Visual Environments (IMPROVE) monitor locations.

Species reconstruction at IMPROVE locations was employed to assess the performance of the hybrid-kriging (HK) method at locations beyond the CSN receptor locations. On average, HK reconstructed concentrations were closer to observations than original CMAQ simulations. The average concentrations of PM_2.5 for reporting days in January 2004 were 6.15 - 1.47, 11.08 - 3.75, and 9.28 - 2.70 μg/m³for observations, CMAQ, and HK simulations, respectively. Trends were similar for all other species, with the exception of Se. Both methods under-simulated Se concentrations, but CMAQ simulations were closer to observations. CMAQ and HK concentrations were biased high for all species except Se, and for all other metal species the simulated concentrations were significantly higher than observations. For nonmetal species and total PM, 60.4 - 11.9% of CMAQ simulations and 68.8 - 14.2% of HK simulations were within a factor of two of observations. For metal species, 30.6 - 14.4% of CMAQ simulations and 35.0- 12.5% of HK simulations were within a factor of two of observations. Overall, the HK method is useful for performing source apportionment over large domains with enhanced spatial resolution, taking into account observed data, emissions, meteorology, and terrain.

Simulation of source-receptor relationships for ozone in South Korea

Jong-Jae, Lee, Cheol-Hee, Kim

The rapid growth in many Asian economies in recent years has resulted in degradation of air quality in the region. In particular, ozone and aerosol aggravates respiratory illness and may lead to premature mortality [World Health Organization, 2005]. Tropospheric ozone is produced via the photochemical oxidation of volatile organic compounds (VOC) and carbon monoxide (CO) in the presence of nitrogen oxides (NOx). However some precursors (or ozone) are transported from source regions to receptor regions. In this study, source-receptor relationships of ozone are simulated in South Korea, and the impact of long-range transport on ozone concentrations are interpreted through the simulations of source-receptor relationships.

During study periods (April and July, 2009), the weather conditions of East Asia were simulated with ARW-WRF version 3.2 at 60 km horizontal grid resolution with 23 vertical layers. The NCEP FNL data was used as initial and boundary conditions for the simulations. For air quality simulations, the Community Multiscale Air Quality (CMAQ) model version 4.7 was used with a domain targeting the East Asia. Source-receptor relationships are quantified using the Brute Force method as emission changes (25% reduction of NOx and VOC) over source regions (3 parts of China, South Korea and Japan).

We calculate the spatial average surface ozone concentration over each of the five regions by carrying out six simulations including base case and five cases of emission reductions in 3 regions of China (Northern, Central, and Southern China), Japan, and South Korea. In the base simulation, spatial mean surface ozone concentrations are found to be similar: 42.1 ppb for North China, 51.1 ppb Middle China, 45.3 ppb South China, 41.9 ppb for South Korea and 33 ppb for Japan in July. The simulation generally captures the observed daily cycle and is simulated to be close to the observed ozone concentration in Seoul in South Korea.

Reduction of emissions in the three Chinese regions simulated a decreased spatial mean concentration of 3.1 ppb over South Korea especially in April, showing much lower than that of case of the emission reduction in South Korean which is 4.3 ppb. The response of South Korean surface ozone concentrations to Chinese emissions is much larger in April due to relatively lower precipitation than that in other months such as in July. High ozone concentrations in April in South Korea are relatively much more sensitive to foreign emissions than other months, as indicated by the foreign (Chinese) emission reduction cases of 13.8 ppb with the daily maximum 8-h average ozone concentrations above 60 ppb. In July, however, domestic (i.e., over the source region itself) sensitivity is larger (19.1 ppb) than foreign sensitivity (10.7 ppb).

Sensitivity of boundary conditions to source receptor relationships show higher in April than other months. Our study shows, for example in April, boundary condition contributed to Seoul concentrations about 65 % in April but it shows much lower value of about 30% in July over South Korea.

The Adjoint of the CMAQ Aqueous Chemistry Module

JAEMEEN BAEK (1), Stanier, C. O. (1), Saide, P. E. (1), Carmichael, G. R. (1); Henze, D. (2); Turner, M.(2); Zhao, S. (3); Hakami, A.(3); Resler, J.(4); Sandu, A. (5); Russell, A. (6); Jeong, G. (2); Nenes, A. (6); Capps, S.(6); Percell, P. B. (7); Pinder, R. (8); Napelenok, S.(8); Bash, J. (8); Chai, T. (9); Byun, D. (9)

(1) University of Iowa; (2) University of Colorado at Boulder; (3) Carleton University; (4) ICS Prague; (5) Virginia Tech; (6) Georgia Tech; (7) University of Houston; (8) US EPA; (9) NOAA

In the presence of clouds, gas and aerosol phase species can dissolve into water droplets and participate in aqueous phase chemical reactions, changing the concentrations of those species significantly. These important processes oxidize S(IV) to S(VI), convert gaseous species to aerosol phase species, and remove both aerosol and gas phase components via deposition. The Community Multiscale Air Quality (CMAQ) model simulates aqueous chemistry as a part of cloud dynamics. The aqueous chemistry module includes wet deposition, scavenging, gas-aqueous phase partitioning, dissociation, formation of secondary organic aerosol and aqueous phase oxidation reactions that form sulfate and secondary organic aerosol. As part of a broader effort to develop adjoint applications of CMAQ for aerosol (CMAQ-ADJ) such as data assimilation and receptor-based sensitivity calculations, our objective is to create the adjoint model for the aqueous chemistry processes. CMAQ aqueous chemistry is implemented using the Kinetic PreProcessor (KPP), which allows for easier updating of mechanisms, and automated generation of the adjoint model. We discuss the pros and cons of the new KPP-based forward model for aqueous chemistry, and intercompare calculated sensitivities between the forward model (finite differences) and the adjoint sensitivities within both a box model and CMAQ.

Adjoint-Based Source Attribution of PM Health Impacts

 

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

 

Long-term exposure to fine particulate matter has been associated with adverse health effects, including premature mortality. In 2011 the World Health Organization estimated that urban outdoor air pollution is the cause of approximately 1.3 million premature deaths worldwide per year. Studies have suggested that PM mixtures with a high BC percentage may have greater effects on mortality than mixtures low in BC. Quantifying the role of emissions from different sectors and different locations in governing the total health impacts is critical towards developing effective control strategies. To answer such questions, an adjoint model can provide sensitivities of excess mortality (through the use of the concentration response functions) with respect to emissions at a highly resolved spatial and sectoral level of specificity. This tool can be used to determine the sensitivity of mortality in a region with respect to emissions throughout the modeled domain. From a single simulation of the CMAQ Adjoint model, we are able to obtain sensitivities of premature mortality due to regional BC concentrations with respect to BC emissions from any and all sectors of the continental United States. 

Development of a 2007-Based Air Quality Modeling Platform

Sharon Phillips, Heather Simon, and Norm Possiel

EPA's Office of Air Quality Planning and Standards is developing a 2007-based air quality modeling platform to support upcoming regulatory and policy assessments. This platform includes meteorology for 2007 obtained from simulations of WRFv3.3[1] and emissions for the US that are representative of the 2007/2008 time period[2] which can be used with either the CAMx or CMAQ photochemical model . The "national scale" modeling domain covers the lower 48 states plus adjacent portions of Canada and Mexico using a horizontal grid resolution of 12 x 12 km. Pollutant species concentrations along the lateral boundaries of this domain were derived from a simulation of the GEOS-Chem hemispheric photochemical model for 2007. As part of the development of this platform we have performed a series of sensitivity tests to understand the effects on predicted species concentrations of several optional science features that are included in the most recent public release of CMAQ, version 5.0.1. We created a benchmark annual simulation for 2007 using CMAQ version 4.7.1 with AERO5 coupled with meteorology from a previous version of WRF, version 3.1. We then used these same inputs to run CMAQ version 5.0.1 with AERO6 and inline photolysis to examine the effects of using the default science features in CMAQ. This simulation was followed by three emissions-related incremental sensitivity tests that included adding (1) bi-directional flux for fertilizer ammonia emissions, (2) lightning NOx emissions, and (3) emissions of wind-blown dust. Finally, we remodeled the 2007 case using meteorology based on simulations from WRF version 3.3 to examine the impacts of changes in meteorology between WRF version 3.1 and 3.3 along with increasing the vertical resolution of the lowest layer from a thickness of 38 meters with version 3.1 to 20 meters with version 3.3. In this paper, we will present an overview of the components of the 2007 platform and describe the impacts on species concentrations and deposition of simulating 2007 with CMAQ version 5.0.1 compared to version 4.7.1 along with the effects of the incremental changes to emissions and meteorology identified above. We will also discuss the effects on model performance of each simulation by comparing model predictions to the corresponding measured concentrations and deposition.

Evaluation of Urban PM2.5 Emission Inventories across the U.S.

Prakash Bhave, Adam Reff, Alexis Zubrow, Venkatesh Rao

Over the past decade, improving the accuracy of emission inputs has been the most successful route toward enhancing air quality model performance for PM_2.5. Along the source-categorical, spatial, and chemical dimensions, the U.S. has the world's most detailed PM_2.5 inventory. However, comprehensive evaluation of this inventory eludes researchers. In this work, we develop and apply a new methodology for evaluating speciated PM_2.5 inventories directly against ambient air measurements. First, we process the 2005 National Emissions Inventory (NEI) through the SMOKE system and apply source-specific speciation profiles (Reff et al., Environ. Sci Technol., 2009, 43:5790-5796) to create a 12km-gridded database of U.S. emissions disaggregated into 85 source categories and 38 chemical species (OC, EC, ions, and trace elements). Next, we dilute the emissions using grid- and month-specific meteorological data to approximate the ambient concentrations of each non-reactive PM_2.5 species that would be obtained from a chemical transport model like CMAQ. Finally, we compare our diluted emissions against ambient measurements collected at 159 sites across the Chemical Speciation Network (CSN). We perform various statistical and graphical comparisons between the diluted emissions and ambient concentrations to diagnose source-specific errors in the PM_2.5 mass emission rates, the temporal allocation profiles, spatial surrogates, and speciation profiles. Our most significant findings will be summarized in this presentation. We anticipate that these results will inform the prioritization of future enhancements to the NEI for PM_2.5.

Comparing differences in model performance between the coupled and non-coupled versions of WRF-CMAQv5.0
K. Wyat Appel, Shawn J. Roselle and David C. Wong

The U.S EPA recently released version 5.0 of the Community Multiscale Air Quality (CMAQ) model. The latest version of the model includes the ability to run the model in coupled mode, meaning that the meteorological driver (WRF) and chemical transport model (CTM) are run in conjunction with each other, as opposed to the non-coupled version of the model in which the meteorological driver is run first, and then used to drive the CTM. The advantage of the coupled WRF-CMAQ simulation is that it allows for feedback to occur between the meteorological and chemical fields within WRF and CMAQ. These feedbacks can result, for example, in changes in radiation and cloud formation within WRF and CMAQ, which cannot be accounted for in the non-coupled version of the model. This work will examine the difference in model performance between a coupled WRF-CMAQ simulation and non-coupled WRF-CMAQ simulation covering the entire year of 2006. Differences in meteorological and chemical fields will be examined to assess the impact that coupling the meteorological and chemical models have on the model estimates. Both model-to-model and model-to-observation comparisons will be made.

A New Interface to Model Global Commercial Aircraft Emissions from the FAA Aviation Environmental Design Tool (AEDT) in Air Quality Models

B.H. Baek, Saravanan Arunachalam, Matt Woody, Pradeepa Lakshimi, Mohammad Omary, Frank Binkowski
UNC Institute for the Environment

Gregg Fleming
Volpe National Transportation System Center

Not available

Space-Time Analysis of AQMEII Phase 1 Air Quality Simulations

Christian Hogrefe, Shawn Roselle, Rohit Mathur, S.Trivikrama Rao, and Stefano Galmarini

During Phase 1 of the Air Quality Model Evaluation International Initiative (AQMEII), 22 modeling groups performed air quality simulations over North America and Europe for the year 2006. In this study, we compare space-time patterns in observations and model outputs in an effort to elucidate the strengths and weaknesses of current generation air quality modeling systems. To this end, we applied a synoptic typing approach to determine model performance under different meteorological regimes in different regions of the modeling domain and to assess how well the modeling systems capture the impact of changing weather conditions on pollutant concentrations. While all modeling systems exhibited skill in capturing the directional impact of changing synoptic patterns on ozone anomalies, there were significant differences in the magnitude of the anomalies simulated by the various models. Furthermore, we determined how well observed and modeled synoptic-scale pollutant fluctuations were correlated in space since these correlation distances can serve as a metric to assess the modeling system's ability to capture the interplay between local vs. regional effects. Results of this analysis show differences among these models. In addition, we show that spatial correlation analyses can be used to distinguish monitoring sites that are regionally representative from those more influenced by local-scale phenomena, and that model performance varies between these two sets of sites for all modeling systems participating in AQMEII Phase 1.

Coupled Energy Market Trading and Air Quality models for improved simulation of peak AQ episodes

Caroline M. Farkas, Annmarie G. Carlton, Frank A. Felder, Mark Rogers, Tyler Wibbelt

Inadequate emission inventories hinder the ability of photochemical models to aid in the development of effective air quality management strategies. Episodic peak ozone events often occur on the sunniest and hottest days, when electricity demand is highest. On high electricity demand days (HEDD), baseload and intermediate electric generating units (EGUs) are supplemented with "peaking units". As opposed to the baseload and intermediate units (typically newer with cleaner emissions), peaking units are located in highly populated areas and consist of older units with dirtier emissions. Peaking units that produce d 25 MWe are not subject to federal regulations (e.g., CAIR), and as a consequence emissions are uncontrolled and not reported. On the days most favorable for ozone and PM formation, emission data is the least reliable.

In this work we simulate an intense heat wave in July and August 2006 in the U.S. that brought both record high and low temperatures to most of the country and caused more than 200 fatalities. In-line emissions developed from the National Emissions Inventory (NEI) and Continuous Emissions Monitoring (CEM) data from the EGU sector for the Community Multi-scale Air Quality model (CMAQv5) for the state of New Jersey confirm missing peaking units. It is observed that EGU NOx and SO₂ emissions are consistently lower than actual power and pollution generation in New Jersey.

In an effort to remedy this inconsistency, we develop a method to couple the Day-ahead Market Analyzer (DAYZER) which simulates actual power generation, transmission and dispatch with CMAQ. We develop emission factors and temporal profiles for EGU sector pollution based on fuel type and actual use. We add these emissions to the CMAQ in-line emissions. CMAQ simulations of both scenarios, with and without the additional point sources, and differences in predicted air quality are compared and evaluated with O₃ and PM measurement data from EPA's National Air Monitoring Stations (NAMS) and State and Local Air Monitoring Stations (SLAMS).

Sensitivity analysis of influencing factors on PM2.5 nitrate simulation

Hikari Shimadera¹, Hiroshi Hayami¹, Satoru Chatani², Yu Morino³, Yasuaki Mori⁴, Tazuko Morikawa⁵, Kazuyo Yamaji⁶, Toshimasa Ohara³

¹Central Research Institute of Electric Power Industry, Abiko, Japan

²Toyota Central R&D Labs., Inc., Nagakute, Japan

³National Institute for Environmental Studies, Tsukuba, Japan

⁴Japan Weather Association, Tokyo, Japan

⁵Japan Automobile Research Institute, Tsukuba, Japan

⁶Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan

Fine particulate matter (PM_2.5) has been of increasing concern because it may seriously affect human health. The air quality standard for PM_2.5 is not attained in most urban areas in Japan. Although air quality models with reasonable accuracy are essential tools to develop and evaluate measures to attain the standard, current air quality models cannot adequately simulate atmospheric mass concentrations of PM_2.5 and its components. In this study, the Community Multiscale Air Quality modeling system (CMAQ) was applied to the Kanto region of Japan in winter 2010 and summer 2011. Comparisons of the model results with PM_2.5 observations indicated that CMAQ overestimated PM_2.5 NO₃^- concentration. Sensitivity analysis was conducted in order to investigate factors influencing the model performance for PM_2.5 NO₃^- simulation. The sensitivity analysis included following factors: variations of air temperature (-5 to +5 K), variations of NO_X emission (-40 to +40 %), modification of seasonal variation of NH₃ emission, parameterizations of N₂O₅ heterogeneous reaction probability, and variations of gaseous HNO₃ and NH₃ dry deposition velocities (0.2 to 5 times). These analyses indicated considerable sensitivity of air temperature, NH₃ emission and dry deposition velocity to PM_2.5 NO₃^- concentration. While current meteorological models well simulate air temperature, there are large uncertainties in estimates of NH₃ emission and dry deposition velocity. Therefore these two are key factors to improve the model performance for PM_2.5 NO₃^- simulation.

2007/2008 Emissions Modeling Platform Components and New Tools

Rich Mason, Alison Eyth and Alexis Zubrow

U. S. Environmental Protection Agency,

Office of Air Quality Planning and Standards, Air Quality Analysis Division,

Research Triangle Park, NC 27711

Mail Drop C339-02

mason.rich@epa.gov

Zacariah Adelman

Institute for the Environment,

University of North Carolina,

Chapel Hill, NC

The U.S. Environmental Protection Agency (EPA)'s Office of Air Quality Planning and Standards (OAQPS) is currently developing a 2007/2008 emissions modeling platform. The emissions for this platform are based on Version 2 of the 2008 National Emissions Inventory (NEI) along with several updates and non-NEI components in the oil and gas, onroad mobile, nonroad mobile and EGU sectors. We will briefly describe why we are developing a year 2007 modeling platform in addition to a 2008 platform. We will describe several of the analyses performed to improve the emissions in the modeling platform, including reconciliation between state/local, EPA-generated, and regional planning organization (RPO) data. We will also describe newly-incorporated enhancements to the processing for residential wood combustion, agricultural ammonia, fugitive dust and fire emissions. We will discuss steps taken to support CMAQ version5 along with enhancements to ancillary input files for the platform such as new spatial surrogates, temporal profiles, and speciation profiles. We will describe updates made to V2 of the NEI as a result of our analyses, as well as sources that undergo the largest changes between our most recent 2005 platform and the 2007 modeling platform.

Recent performance of the National Air Quality Forecast Capability

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

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

The National Air Quality Forecasting Capability (NAQFC) provides operational predictions of ozone and wildfire smoke nationwide and operational predictions of airborne dust from dust storms over the contiguous 48 states. Predictions are produced beyond midnight of the following day at 12km resolution and 1 hour time intervals and are distributed in numerical format and in graphical format at http://airquality.weather.gov/. Ozone predictions combine the NOAA National Centers for Environmental Prediction (NCEP) operational Nonhydrostatic Mesoscale Model on Arakawa B-grid (NMMB) weather predictions with inventory based emissions estimates from the EPA and chemical processes within the Community Multiscale Air Quality (CMAQ) model. The Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model is used to provide standalone predictions of wildfire smoke and dust storm predictions, both of which have highly variable intermittent sources. Smoke sources are based on NOAA/NESDIS analysis of satellite imagery for detection of smoke source locations that is combined with US Forest Service's BlueSky framework to estimate smoke emissions. Dust source locations are based on satellite climatology of the frequency of dust emissions. Dust is emitted in the model when threshold velocity is exceeded and the emissions are modulated by real time estimates of soil moisture. Routine verification of ozone 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.

In summer of 2012 estimates of emissions used in ozone prediction have been updated to use projections for 2012 for mobile emissions. This was also the first summer since the new meteorological model NMMB became operational. We present examples of performance of ozone predictions. Dust predictions were implemented in 2012. With widespread dust storms the predictions would sometimes take longer than desired and ways of reducing the run time for predictions are being explored. Examples of smoke predictions for recent wildfires will be shown. We will also highlight connections with several related research efforts such as those aiming to improve emissions and examine impacts of fires from Central America, increased model resolution, and assimilation of data collected in an intensive field campaign.

SMOKE-MOVES: Description and Recent Enhancements

Alexis Zubrow

U. S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Air Quality Analysis Division,
Research Triangle Park, NC 27711
Mail Drop C339-02
zubrow.alexis@epa.gov

B.H. Baek
Institute for the Environment, UNC - Chapel Hill

EPA's Office of Transportation and Air Quality (OTAQ) and Office of Air Quality Planning and Standards (OAQPS) have developed an integrated system for mobile emissions, SMOKE-MOVES. The major motivations for this new system are: to closely integrate MOVES (Motor Vehicle Emission Simulator) into the emissions process, the sensitivity of many pollutants to temperature and humidity, and the computational demands for running MOVES. SMOKE-MOVES provides high spatial and temporal resolution of onroad inventories with respect to temperature, which is a key parameter with high temporal and spatial variability in most air quality modeling applications. We will present an overview of the system and will highlight recent improvements to the integrated system. In addition, we will present a brief comparison of SMOKE-MOVES with previous modeling results using MOVES.

Evaluation of the CMAQ5.0 in the framework of the CALIOPE air quality forecasting system over Europe

Maria-Teresa Pay^a, Santiago Gasse^a,b, Jose Maria Baldasano^a,b

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

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

CALIOPE is a high-resolution air quality system which provides forecast for 24 and 48h since October 2006 for Europe (12 km x 12 km, 1h) and Spain (4 km x 4 km, 1h). The meteorological model is the WRF-ARW model (version 3.2.1) initialized by the FNL/NCEP data. The emissions are estimated by means a top-down approach implemented in the High-Elective Resolution Modeling Emission System (HERMES version 2.0) based on the EMEP inventory for the year 2008. The Chemical Transport Model (CTM) is the CMAQ (version 4.5) using the CB-IV chemical mechanism and AERO4 mode for aerosols. Several evaluation studies and near-real time evaluation (NRT) against air quality measurements on an hourly basis support the confidence on the system. The present contribution evaluates the CALIOPE system over Europe using CMAQ CTM version 5.0 (CB05 and AERO5) which is working in forecast mode since 9 April 2012. The comparison between both CMAQ versions is done in terms of gaseous and aerosol concentrations (O₃, NO₂, SO₂, PM10, and PM2.5) at the lowest level. Forecast concentrations are compared against observations on an hourly basis from the European air quality database (Airbase) which classifies stations as background rural/suburban. Results indicate that CMAQv5.0 improves O₃ forecast daily cycle, especially at nighttime over suburban stations, where O₃ biases are reduced between 20 and 40 μgm^-3. The CMAQv5.0 improves the forecast of NO₂ peaks at suburban stations reducing biases ~10-20 μgm^-3. PM10 forecast also improves with the new CMAQ version. Episodes of secondary aerosol formation are now reproduced (i.e. 7-14 may 2012), where bias are reduce in ~10-20 μgm^-3. Furthermore, PM10 hourly peaks in suburban stations are better reproduced reducing hourly biases ~5-10 μgm^-3. The contribution also evaluates the effect of using the Kalman filter post-process to reduce systematic bias in both CMAQ versions. Results show that the bias-adjustment technique is more effective over CMAQv5.0.

High NOx emissions bias of the EPA National Emissions Inventory 2005: two case studies over Los Angeles and Houston for August 2009

Yunsoo Choi, Department of Earth and Atmospheric Sciences, University of Houston

Simulation results from the Community Multiscale Air Quality (CMAQ) model version 4.7.1 over Los Angeles and Houston for August 2009 are analyzed to investigate the uncertainty of the EPA National Emissions Inventory 2005 (NEI2005). The NO_x emissions from NEI2005 are adjusted by comparing the Global Monitoring Experiment 2 (GOME-2) and CMAQ NO₂ columns. The CMAQ simulations using the GOME-2-derived NO_x emissions adjustment ("adjusted CMAQ", decreasing from 462 Gg N to 426 Gg N over the US for August 2009) show large reductions of simulated NO_x concentrations and mitigate the large discrepancies between the absolute amount of NO_x concentrations of the EPA Air Quality System (AQS) and those of baseline of CMAQ (particularly over urban regions). In particular, we found that NO_x emissions reduction decreased the surface NO_x concentrations significantly over Los Angeles (e20 ppbv) and Houston (e10 ppbv) where the simulated NO_x concentrations from adjusted CMAQ generally compared well to those of in-situ AQS observations. Over Los Angeles, the significant reduction of NO_x concentrations results in comparable increase of surface O₃ concentrations over Los Angeles (representing a typical characteristic of NO_x-saturated regime region) and the resulting simulated O₃ concentrations are compared well with in-situ surface O₃ measurements. Similarly, over Houston, the large reductions of NO_x emissions in adjusted CMAQ reduce the large amounts of simulated NO_x concentrations, which are also well compared with those of EPA AQS observations. However, high simulated O₃ bias still remains over Houston in the adjusted simulation, even with some reduction of the high bias in the outflow regions from central Houston. Both results suggest that anthropogenic NO_x emissions over Los Angeles and Houston reported in NEI2005 are too high and remote sensing could be used to estimate the uncertainty of the emissions inventory.

Utilization of MOZAIC profiles in the evaluation of CMAQ as part of the Air Quality Model Evaluation International Initiative (AQMEII)

Morgan L. Silverman, James Szykman, Christian Hogrefe, Brian Eder, James H. Crawford, Thomas Pierce, Todd Plessel, Matt Freeman, S.T. Rao, Jean Pierre Cammas, and Andreas Volz-Thomas

CMAQ is a key component of EPA's decision support system in managing air quality domestically and also in meeting our international obligations. As new health science results guide revisions to the National Ambient Air Quality Standards for ground-level O₃ and PM_2.5, there is an increased emphasis in CMAQ's ability to accurately simulate the contribution of O₃, PM_2.5, and associated precursors aloft to concentrations at the surface. Rao [2009] and Scheffe et al., [2009] discuss the critical need to assess air chemistry at multiple levels in the atmosphere with accurate and high-fidelity chemical and physical observations to better develop and evaluate a key decision support tool like CMAQ. While recent efforts within EPA have started to evaluate the model performance in the free troposphere using data collected from ozonesondes (Eder et al., 2010), understanding and improving air quality predictions is often hampered by the scarcity of routine and systematic atmospheric observations that exist above the surface. Accordingly, this research focuses on the use of the European-based aircraft program Measurement of OZone, water vapor, carbon monoxide and nitrogen oxides by in-Service AIrbus airCraft (MOZAIC - currently part of the European program In-service Aircraft for a Global Observing System (IAGOS) http://www.iagos.fr/web/) (Cammas and Volz-Thomas 2007; Volz-Thomas et al., 2009) as a key data source of profile measurements for CMAQ evaluation.This analysis will characterize the MOZAIC measurements in the mid to lower troposphere (< 6km) and evaluate the 2006 CMAQ AQMEII run using key chemical and physical parameters (O₃, CO, NOy, H₂O, T, and P) from the MOZAIC flights. Appropriate data for comparison with the MOZAIC profiles has also become available under the NASA Earth Venture -1 (EV-1) DISCOVER-AQ (Deriving Information on Surface conditions from COlumn and VERtically resolved observations relevant to Air Quality) mission. As part of the first DISCOVER-AQ field mission, the NASA P-3B aircraft performed over 250 profiles in the Baltimore-DC region, including one landing at the Baltimore Washington International Airport. These profiles offer comparative observations of O₃, CO, and reactive nitrogen in the vicinity of major air traffic patterns, local urban sources, and typical background environments. Sensitivity to these different environments will be evaluated. In addition, the presentation will include a prototype application of using both MOZAIC and CMAQ data within the US EPA's Remote Sensing Information Gateway (RSIG) -- www.epa.gov/rsig.

Comparison of Link-Based and SMOKE processed Motor Vehicle Emissions over the Greater Toronto Area

Motor vehicle emissions contribute significantly to air pollution, especially in cities where emissions from motor vehicles are the main precursors of smog. Mobile sources also emit air toxics which have serious health effects. It is very important to characterize the spatial and temporal distribution of mobile source emissions in modelling local air quality and its effects on public health at a fine grid scale. Typically, emissions from the inventory are disaggregated spatially and temporally using spatial surrogates and predefined temporal profiles by emission processing software, such as SMOKE (Sparse Matrix Operator Kernel Emissions). However, the spatial surrogates and temporal profiles may not represent the mobile source activities well, especially at high grid resolutions. On the other hand, link-based emissions are estimated from traffic flow characteristics by each roadway and are usually deemed to be better representative of on-road mobile emissions. In this presentation, we will show the comparisons of the motor vehicle emissions over the Toronto Census Metropolitan Area (CMA) estimated from the two methods: 1) SMOKE processed emissions based on a set of road type specific surrogates and temporal profiles; 2) Link-based emissions calculated from traffic flows and congested travel speeds on the road network within the study area estimated by a traffic flow simulation software (TRAFFIC), which was developed by the Centre for Spatial Analysis, McMaster University. Based on the comparison results, possible improvements to SMOKE surrogates and temporal profiles will be discussed.

Evaluating NOx Emission Inventories for Air Quality Modeling Using Satellite, Model and SEARCH NO2 Data

Greg Yarwood, Sue Kemball-Cook, Jeremiah Johnson and Gary Wilson

ENVIRON International Corporation

Bright Dornblaser and Mark Estes

Texas Commission on Environmental Quality

Past studies of ozone transport into Texas indicate that the southeastern U.S. sometimes is a source of ozone transport into Texas and that systematic errors in the modeling of southeastern U.S. ozone can confound efforts to quantify the effects of ozone transport. Model overestimates of southeastern U.S. ozone may result from biased NOx emissions, and the purpose of this study is to assess the accuracy of the NOx emissions data in the Texas Commission on Environmental Quality's State Implementation Plan (SIP) modeling inventories which are based on the 2005 National Emission Inventory in the southeastern U.S. We used satellite NO₂ column data from the Ozone Monitoring Instrument (OMI) and near-surface NO₂ data from the SouthEastern Aerosol Research and Characterization and Air Quality System (SEARCH) together with modeled NO₂ columns from the Comprehensive Air quality Model with extensions (CAMx) to make top-down NOx emissions estimates for the southeastern U.S. Lightning and aircraft NOx emissions were added to the CAMx emission inventory to allow comparison between modeled and satellite-derived NO₂ columns. Research-quality photolytic NO₂ measurements from SEARCH provide unambiguous NO₂ observations at a few rural sites in the southeastern U.S. locations and were used to evaluate modeled surface layer NO₂. In contrast, the OMI satellite data provide broad spatial coverage but measure column rather than surface NO₂. CAMx and OMI tropospheric vertical NO₂ columns were compared across the southeast, and were then used to estimate top-down NOx emissions. The resulting top-down NOx emission estimates were compared with TCEQ's NOx emission inventory and the potential for bias in the NOx emission inventory in the southeastern U.S. was evaluated.

Estimating Emissions for the 2008 Evans Road Fire in North Carolina

George A Pouliot

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

pouliot.george@epa.gov

Ana G Rappold, Wayne E Cascio

Environmental Public Health Division, National Health and Environmental Effects Research Laboratory, US Environmental Protection Agency, Research Triangle Park, NC 27711

Amber Soja

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

The Evans Road Fire in 2008 was started from a series of lightning strikes on June 1, 2008. When the fire was contained by the end of August, it had burned nearly 41,000 acres. This fire also burned a depth of two to three feet deep into organic soil, smoldering for many days. Smoke from the fire was observed as far west at the Raleigh Durham area and air quality was severely degraded by the fire. This fire was unique in that the majority of the emissions came from the below ground smoldering of peat. We will summarize our method for estimating emissions from this fire using field estimates of the Carbon above and below the ground. In addition, we will compare our revised emission estimates with those obtained from the National Emissions Inventory. Finally, we will show some chemical transport modeling simulations of the fire using CMAQ version 5.

Diagnostic Evaluation of a Modeling System for Predicting the Air Quality Impacts of Prescribed Burns

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

Gary L. Achtemeier, Yongqiang Liu and Scott L. Goodrick
Center for Forest Disturbance Science, USDA Forest Service, Athens, GA, 30602-2044, USA

Roby Greenwald
Rollins School of Public Health, Emory University, Atlanta, GA, 30322, USA

Brian Gullett, Johanna Aurell, and William R. Stevens
Office of Research and Development, US Environmental Protection Agency, Research Triangle Park, NC, 27711, USA

Roger D. Ottmar
Pacific Northwest Research Station, Pacific Wildland Fire Sciences Laboratory, USDA Forest Service, Seattle, WA, 98103, USA

Robert J. Yokelson and Sheryl K. Akagi
Department of Chemistry, University of Montana, Missoula, MT, 59812, USA

A modeling system has been developed with the purpose of predicting the downwind air quality impacts of prescribed burns. The system incorporated several new elements into emission, dispersion and transport process components, all with significant potential for improving the accuracy of the predictions. Components of the modeling system were evaluated individually and together using field data from monitored burns and observations from regional networks and satellites. Field data include measurements of: 1) fuels before and after the burns, 2) emissions with an aerostat, 3) winds with ground-based and airborne instruments, 4) plume height with a lidar ceilometer, 5) short-range smoke with ground-based stationary and mobile platforms, and 6) long-range smoke with aircraft. Modeling results were compared to other models. Uncertainties in the estimation of emissions and prediction of winds were identified and quantified. Sensitivities of predicted downwind smoke concentrations to emission strength and timing, plume injection height versus PBL height, and wind speed and direction were calculated. The objective of this presentation is to synthesize the results of component and system evaluations, uncertainty estimations and sensitivity analyses to determine where future research would be most beneficial for increasing the accuracy of the modeling system. Lessons learned will be pieced together towards a general approach for diagnostic evaluation of complex modeling systems.

Integrating source and receptor models for the purpose of emissions inventory improvement application to residential wood combustion in the Southeast.

S. Napelenok, R. Vedantham, G. Pouliot, P. Bhave

Although residential wood combustion (RWC) is a relatively small part of the NEI, it is a major contributor of non-attainment of the PM_2.5 NAAQS in some areas. Biomass combustion constitutes 51% of the particulate carbon concentrations in the Southeast, and more accurately resolving these emissions sources would lead to improved results from regulatory analysis in the region.

The newly developed 2008 version of the NEI is used as the base inventory. CMAQ is used to processes pollutant precursor emissions as specified by the current NEI. The resulting modeled concentrations are compared to pollutant measurements taken at national network sites and specialized measurement campaigns. Receptor-based source apportionment models are then used to attribute measured pollutant concentrations to sources based on source characterization and signature. The coupling of the two modeling systems is used as the basis for inferring improvements to the NEI.

Measurements for this study are provided from the extensive field campaign that was conducted in the southeastern U.S. in 2007 (Zhang et al., 2010). This data set is applicable, because it can be used both to evaluate the current NEI estimates for these sectors through physical and chemical transformation described by the forward source-based model, and at the same time, it can be used for discern source strengths described by the receptor-based source apportionment models.

References

Zhang, X., Hecobian, A., Zheng, M., Frank, N. H., Weber, R. J. (2010). "Biomass burning impact on PM(2.5) over the southeastern US during 2007: integrating chemically speciated FRM filter measurements, MODIS fire counts and PMF analysis." Atmos Chem Phys 10(14): 6839-6853.

Modeling impact of biomass burning on air quality in Southeast and East Asia

Kan Huang ^a, Joshua S. Fu ^a,b,*, N. Christina Hsu ^c, Yang Gao ^a, Xinyi Dong ^a, Si-Chee Tsay ^c, Yun Fat Lam ^a

^a Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, Tennessee, USA

^b UTK-ORNL Center for Interdisciplinary Research and Graduate Education, Knoxville, Tennessee, USA

^c Goddard Space Flight Center, NASA, Greenbelt, Maryland, USA

A synergy of numerical simulation, ground-based measurement and satellite observation was applied to evaluate the impact of biomass burning originating from Southeast Asia (SE Asia) within the framework of NASA's 2006 Biomass-burning Aerosols in South-East Asia: Smoke Impact Assessment (BASE-ASIA). Biomass burning emissions in the spring of 2006 peaked in March - April when most intense biomass burning occurred in Myanmar, northern Thailand, Laos, and parts of Vietnam and Cambodia. CMAQ could reasonably well simulate the spatial pattern and temporal variations of CO in comparison to both satellite and ground measurement. Overestimation or underestimation occurred in different regions due to great uncertainties of the biomass burning emission inventory. The largest discrepancy between modeled and observed CO occurred in northern Thailand, which overestimated the peak episodes by a factor of 2 - 3. In the other regions of Thailand where less fires occurred, the simulation performed much better. Scenario simulation not only modeled significant impact of biomass burning on AOD (0.4 - 0.6) in SE Asia source regions, but also in the downwind regions. The Asian spring monsoon facilitated the impact of biomass burning extending from peninsular SE Asia to the South China Sea, the Taiwan Strait, and some provinces in southern China. Different spatial distribution patterns of biomass burning AOD between March and April were caused by the different wind fields. Evaluation of several datasets, i.e, EANET, AERONET, and Taiwan supersites' data, illustrated distinct regional differences of aerosol chemical and optical properties. Local biomass burning, anthropogenic sources and long-range/regional transport were the main factors controlling the aerosol chemistry. Highest concentrations of particulate K⁺, SO₄^2-, NO₃^-, and NH₄⁺ were observed at Hanoi in Vietnam. Correspondingly, highest AOD and AAOD were modeled at Bac Giang, also in Vietnam. The contribution of biomass burning to AOD was estimated to be over 56%, indicating the significant influence of biomass burning over this region. In Thailand, the magnitudes of major aerosol chemical components, AOD, and AAOD were much lower than in Vietnam. The contribution of biomass burning to AOD was about 18 - 50 %, indicating that Thailand was moderately impacted. In the downwind regions, the contribution of biomass burning to AOD at Hong Kong and Taiwan was significant within the range of 26% - 62%. The observed monthly mean SSA was around 0.90 during intense biomass burning periods, suggesting the great production of strongly absorbing substances (i.e, EC) due to biomass burning. Modeled concentrations of aerosol chemical components were biased low at most circumstances, and the modeled AOD values were biased low about a factor of 2, probably due to the underestimation of biomass burning emission.

Estimating Secondary Pollutant Impacts from Single Sources

Kirk Baker

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

Sources with proposed modifications or new sources may be required to estimate the impact due to these emissions as part of the New Source Review (NSR) and Prevention of Significant Deterioration (PSD) programs. Sources are required to estimate the impacts of primarily emitted, resulting secondarily formed pollution, and acid deposition (sulfur and nitrogen) at Class I and sensitive Class II areas. Steady-state gaussian plume dispersion models such as AERMOD are used to estimate the impacts of chemically inert compounds up to 50 km from the source. A variety of Lagrangian and Eulerian modeling systems have been used to estimate single source impacts on secondarily formed pollution such as ozone and PM2.5 with varying levels of complexity in the treatment of plume chemistry and the chemical and physical environment which the plume exists.

Eulerian photochemical transport models such as the Comprehensive Air Quality Model with Extensions (CAMx) and the Community Multiscale Air Quality (CMAQ) model treat emissions, chemical transformation, transport, and deposition using time and space variant meteorology. These modeling systems treat primarily emitted species and secondarily formed pollutants such as ozone and PM2.5. Some photochemical grid models have been instrumented with extensions that may allow for the identification of impacts from specific sources to important receptor locations. These extensions generally fall in the categories of sub-grid plume treatment and sampling, source apportionment, and source sensitivity.

Some photochemical models have been instrumented with source apportionment, which tracks emissions from specific sources through chemical transformation, transport, and deposition processes to estimate a contribution to predicted air quality at downwind receptors. Another approach to differentiate the impacts of single sources on changes in model predicted air quality is the direct decoupled method (DDM), which tracks the sensitivity of an emissions source through all chemical and physical processes in the modeling system. Sensitivity coefficients relating source emissions to air quality are estimated during the model simulation and output at the resolution of the host model.

An illustrative example of modeling the impacts of a single source on primary and secondarily formed pollutants using a variety of photochemical grid model source contribution and sensitivity approaches will be presented. The approaches include brute force emissions adjustments, DDM sensitivity, and source apportionment. Single source contributions are compared with in-plume measurements taken by helicopter at multiple downwind distances from the source on multiple days. This type of field measurement data provides a unique opportunity to evaluate how well these modeling systems place plumes, replicate the chemical environment of the source, and simulate the chemical evolution of a plume as source emissions are transported away from the facility.

Ensemble Air Quality Modeling using the Coupled WRF-CMAQ Model

Robert Gilliam, Jonathan Pleim, Christian Hogrefe, Jim Godowitch, ST Rao

Uncertainty in air quality modeling is largely impacted by the uncertain inputs of meteorology and emissions. Transport of urban pollutants, for example, can be altered in CMAQ if the wind speed or direction is misrepresented by a level of instrument error. Inaccurately modeled temperature or moisture can have an adverse impact on boundary layer stability, clouds and radiation, all of which impact mixing, chemistry and photochemistry in air quality models.

Over the last decade ensemble modeling has received attention because of its ability to better estimate the uncertainty contained in weather forecasts. Ensembles are initialized with slightly different, or perturbed initial conditions, which consider measurement uncertainty. One such ensemble modeling system is the Short-Range Ensemble Forecast system (SREF) developed and managed by the National Center for Environmental Prediction (NCEP). Several models are used by the SREF. Each model has a control run with multiple perturbation members. These ensemble members have slightly different initial conditions that represent a statistically probable state of the atmosphere. We leverage these varied initially conditions in retrospective modeling using four-dimensional data assimilation (grid nudging) in the coupled WRF-CMAQ model system. Sixteen WRF-CMAQ simulations were nudged towards SREF member's initial conditions every six hours over a 4 day high-ozone case study in June of 2011. The variability in ozone and some meteorological variables are explored as well as an evaluation of the both the air quality and meteorological model ensemble.

Development of PM2.5 Interpollutant Trading Ratios in Georgia

Jim Boylan¹ and Byeong-Uk Kim¹

¹Georgia Department of Natural Resources (GA DNR)

Sources applying for permits in areas designated nonattainment for PM2.5 can offset emissions increases of direct PM2.5 emissions or PM2.5 precursors with reductions of either direct PM2.5 emissions or PM2.5 precursors in accordance with offset ratios contained in the approved SIP for the applicable nonattainment area. In addition, PM2.5 offset ratios can be used to account for secondary formation of PM2.5 in a dispersion model. EPA developed "preferred" PM2.5 offset ratios in 2008 (e.g., 40:1 SO2 tons for PM2.5 tons). However, in 2011 EPA released a memo stating that those preferred ratios would no longer be presumptively approvable. Instead, any ratio involving PM2.5 precursors adopted by the state must be accompanied by a technical demonstration. This paper describes a technical approach for developing PM2.5 interpollutant trading ratios in Georgia.

Our approach involves the use of the CAMx photochemical grid model with flexi-nesting (12 km/4 km/1.333 km). A series of emissions sensitivity runs are conducted by zeroing out the stack emissions of SO2, NOx, EC from a single stack in three different model runs to evaluate the resulting changes in modeled PM2.5 concentrations. The absolute sensitivities (ug/m3) are normalized by the emissions (tpy) and then PM2.5 offset ratios are calculated by taking the ratios of normalized sensitivities (EC/SO2 and EC/NOx). We will evaluate how the PM2.5 offset rations vary as a function of: (1) distance from the source, (2) season of the year, (3) grid resolution, (4) stack height, (5) emission rates, and (6) region within the state.

Observation-constrained probabilistic evaluation of modeled concentrations and sensitivities

Daniel Cohan and Antara Digar

Photochemical models are typically applied in a deterministic manner to generate the best possible estimates of pollutant concentrations and their sensitivities to emissions. Recent development of techniques such as reduced form models have provided efficient methods for Monte Carlo characterization of the uncertainties in model outputs arising from parametric and structural uncertainties in model inputs and formulations. However, applications of these techniques have typically considered each of the Monte Carlo cases to be equally likely.

Here, we show how probabilistic evaluations of photochemical model results can be constrained by observations of pollutant concentrations. The methods are applied to a historical attainment modeling episode for ozone in Texas, with numerous structural and parametric uncertainties considered in the CAMx model. Three observational metrics including a Bayesian likelihood approach are introduced to weight or screen thousands of alternate simulations generated by a reduced form model. The metrics evaluate model performance against ground-based observations of ozone and nitrogen oxides, but could readily be extended to consider other pollutants or observations aloft. The approach yields observation-constrained probability distributions not only for pollutant concentrations, but also for the responsiveness of those concentrations to perturbations in emission rates. We also generate posterior probability distributions for the true values of uncertain model inputs such as emission inventories and reaction rate constants. Results vary with the observational metric used, highlighting the importance of metric selection. Implications for probabilistic evaluation of regulatory and scientific modeling of historical air pollution episodes are discussed.

Demonstration on the use of a Photochemical Grid Model to AssessSingle-Source Impacts and Comparison with CALPUFF

Chris Emery, Ralph E. Morris, Tanarit Sakulyanontvittaya, Darren Wilton and Lynsey Parker

ENVIRON International Corporation

Abstract

New sources or modified existing sources may be required to assess their impacts to air quality and air quality related values (AQRVs) at distant (50-300 km) Class I areas. Long Range Transport (LRT) models are needed to fill this regulatory niche. Since 2003, CALPUFF has been the recommended LRT model in EPA's Modeling Guidance for simulating chemically inert pollutants. However, AQRVs include visibility and deposition, where secondary particulates (e.g., sulfate and nitrate) are important and ozone impacts are of increasing concern. CALPUFF does not treat ozone and has a highly simplified representation of PM chemistry. Photochemical grid models (PGMs) include detailed non-linear chemistry, but tend to be more costly, are computationally intensive to apply, and are limited by grid resolution. This paper presents the application of CALPUFF and the CAMx PGM for several test sources in the western U.S., and compares their far-field air quality and AQRV impacts at several Class I areas. CALPUFF was applied using meteorological inputs based on the CALMET diagnostic wind model as well as the Mesoscale Model Interface (MMIF) tool that performs a direct "pass through" of the prognostic meteorological output data directly to CALPUFF. CALPUFF/CALMET tended to estimate slightly higher concentrations than CALMET/MMIF. For annual SO₂ and NO₂, there was very good agreement between CALPUFF/CALMET and CAMx for maximum concentrations at Class I areas. CALPUFF tended to estimate higher visibility impacts compared to CAMx, mainly because CALPUFF overstated particulate nitrate. CALPUFF and CAMx estimated similar sulfur deposition at Class I areas, but CAMx estimated much higher nitrogen deposition. CAMx generated more nitric acid than CALPUFF, which deposits much faster than particulate nitrate. The study developed post-processing software and demonstrated how PGMs can be used for single-source air quality and AQRV assessments at far-field distances.

On Model's Ability to Capture Key Measures Relevant to Air Quality Policies through Analysis of Multi-Year O3 and PM2.5 Observations and CMAQ Simulations

Daiwen Kang, Shawn Roselle, Christian Hogrefe, Rohit Mathur, and S. Trivikrama Rao

Photochemical air quality models are being used to assess impacts of control strategies on ambient pollutant concentrations as part of attainment demonstrations. One critical requirement for these models is their performance stability and consistency over different meteorological conditions and varying emissions inputs. In this study, we investigate the interannual variability and trends in observed daily maximum 8-hr O₃ mixing ratios during 2002 - 2008 and the corresponding Community Multiscale Air Quality (CMAQ) model simulations over the eastern United States. Results indicate that the year-to-year variations in the mean and standard deviation of summertime daily maximum 8-hr O₃ mixing ratios were reproduced by the model simulations. Both observed and simulated summertime daily maximum 8-hr O₃ mixing ratios can be characterized by a normal distribution, thereby facilitating probabilistic model evaluations in regulatory applications. Based on the parameters of normal distributions, the model's ability to capture the regulatory metrics such as the 4^th highest daily maximum 8-hr O₃ mixing ratios are assessed against the observed values from the AQS monitoring network. The spatial and temporal characteristics of various metrics are also analyzed. Analysis on PM_2.5 mass and PM_2.5 major species concentrations will also be presented.

Summary of updates to SCICHEM-2012 model and comparison of results with observations and previous version results

Biswanath Chowdhury¹, Ian Sykes¹, Doug Henn¹, Eladio Knipping², Prakash Karamchandani³

1. Sage Management, 15 Roszel Rd., Suite 102, Princeton NJ 08540

2. EPRI, 3420 Hillview Avenue, Palo Alto, California

3. Environ, 773 San Marin Drive, Suite 2115, Novato, CA 94998

The SCICHEM model was developed for EPRI to include complete gas phase, aqueous and aerosol phase chemistry in a Lagrangian puff model SCIPUFF (Second-order Closure Integrated Puff). The SCICHEM model was branched from the SCIPUFF model more than a decade ago. There have been significant advances in the dispersion model in SCIPUFF and these have been incorporated in the SCICHEM-2012 model. Some of the main updates have been modeling of skew turbulence for convective boundary layer, flexible interface for linking aqueous aerosol modules such as AE5, reading AERMOD input files, modeling area and volume sources, multiple PRIME sources and single code for running in serial or in parallel with MPI.

The results from the SCICHEM-2012 model are compared with the previous version of SCICHEM and with observations for a test case. The results also compare the difference in results due to change in chemistry mechanism from CBM-IV to CB05. Comparison of the tracer concentrations for some AERMOD test database will also be presented.

Influence of Air Quality Model Resolution on Uncertainty Associated with Health Impacts, Part II.

Tammy M. Thompson and Noelle E. Selin

Joint Program on the Science and Policy of Global Change, Massachusetts Institute of Technology, 77 Massachusetts, Ave., Bldg E19-411, Cambridge, MA 02139. Phone: 617-452-3192. Fax: 617-253-9845 TammyT@mit.edu

We examine at national scale how air quality model resolution influences the uncertainty associated with the estimated change in human health impacts resulting from policy scenarios proposed to reduce ozone and particulate matter. We use results from a nation-wide, year-long air quality modeling episode designed by the U.S. EPA in support of the recently finalized Cross State Air Pollution Rule, to evaluate how air quality modeling resolution influences the uncertainty associated with human health impacts in multiple regions of the U.S., with each region representing unique pollutant loads and meteorological trends. We evaluate large cities, rural areas, and areas both in attainment, and not in attainment of the current air quality standards set by the U.S.EPA. Our model inputs include emissions inventories representing 2005 and 2014, and meteorological inputs representing 2005. We evaluate model predicted ozone and particulate matter at 36, 12 and 4 km resolution and how concentrations change between 2005 and 2014 due to emissions reductions proposed in the 2014 inventory. Using simulated changes in pollution we calculate the change in human health impacts at each resolution with uncertainty related to health response. Finally, we use our results to assess the influence of differing chemistry, emissions and pollutant behavior on the resolution necessary for air quality modeling to inform policies to address health impacts.

Evaluation of Single-Source Plume Chemistry Simulations with the SCICHEM Reactive Plume Model

James T Kelly and Kirk R Baker

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

A need exists for regulatory modeling approaches that can accurately account for secondary PM_2.5 and ozone formation due to emissions from single sources. The Second-order Closure Integrated puff model with CHEMistry (SCICHEM) is a reactive plume model that could potentially play a role in single-source regulatory studies where secondary pollutants are of concern. However, tools must be developed for processing SCICHEM model inputs and outputs and SCICHEM predictions must be evaluated under conditions relevant to regulatory applications before recommendations can be made on its use. In this presentation, a preliminary evaluation of SCICHEM simulations using in-plume aircraft observations is described. Future work will consider simulations of larger domains and longer time periods that are more relevant to common regulatory applications.

Influence of model grid resolution on NO2 vertical column densities over east Asia.

Kazuyo Yamaji (Japan Agency for Marine-Earth Science and Technology), Hitoshi Irie (Chiba university), Jun-ichi Kurokawa (Asia Center for Air Pollution Research), Toshimasa Ohara (National Institute for Environmental Studies)

Simulated NO₂ in chemical transport models depends on the magnitude and spatial extent of the NO₂ source. Especially for source region, therefore, there are concerns about that affects on differences between simulated and observed NO₂ vertical column densities (NO₂ VCDs). In this study, we examine the sensitivities of NO₂, the other nitrogen oxides, and their related species by running East Asian scale CMAQ simulations at four different resolutions between 80km and 10km. We find higher NO₂ VCDs in the troposphere at higher resolutions. NO₂ VCDs in the North China Plain at 10km scales are 17 % higher than those at 80km scales at 10 am (local time) in June and December. In the afternoon, 2 am (local time), the NO₂ VCDs at 10km scales are 38% higher in June and 18% higher in December than those at 80km scales. Meanwhile, surface O₃ concentrations tend to be lower at higher resolutions, and that contributes to reduce overestimation of O₃ in this region.

Application of the reactive plume model, SCICHEM-2012, to simulate near-source 1-hour NO2 concentrations
Prakash Karamchandani, Ralph Morris, Greg Yarwood, Bart Brashers
ENVIRON International Corporation

Eladio Knipping
EPRI

Biswanath Chowdhury, Ian Sykes
Sage Management

SCICHEM is a state-of-the-science non-steady-state puff model with complete chemistry treatment, including gas-phase chemistry, aerosol chemistry, and cloud chemistry. It can be used for both local-scale and long-range transport applications. It accounts for the effects of building downwash on plume dispersion using the Plume Rise Model Enhancements (PRIME) algorithm. Several improvements have recently been made to SCICHEM, including enhancements to the dispersion component of the model, extensions to treat area sources and volume sources, an updated user-friendly interface (UFI) for setting up model applications, and updates to the chemistry modules. The model has also been adapted to read input files for AERMOD, the U.S. EPA regulatory model for short-range impacts. The new version of SCICHEM, referred to as SCICHEM-2012, will be available to the user community as a public domain model after it has been tested and evaluated. Here, we present the application of SCICHEM-2012 in a regulatory setting: the calculation of 1-hour NO2 concentrations to address the January 2010 short-term National Ambient Air Quality Standard (NAAQS) for NO2. We implemented gas-phase kinetic reactions for NO, NO2 and O3 in SCICHEM-2012 that are both scientifically rigorous, and suitable for such near-source, long-term applications. We use an AERMOD modeling database for a source that has difficulties demonstrating compliance with the 1-hour NO2 standard to conduct annual simulations with AERMOD and SCICHEM-2012. We compare the NO2/NOx ratios and maximum 1-hour NO2 concentrations predicted by the two models.

Springtime Transport of Central American Fire Emissions to the United States

Hyun Cheol Kim ^1,2, Fantine Ngan ^1,2, Pius Lee ¹, and Rick Saylor ³

¹ NOAA/ARL, Silver Spring, MD

² ERT, Inc, Laurel, MD

³ NOAA/ARL/ATDD, Oak Ridge, TN

The impact of springtime fire emissions from Central America on air quality in the southern U. S. is analyzed. Understanding primary emission sources for Particulate Matter (PM) simulation is important for each U. S. state to meet the National Ambient Air Quality Standards (NAAQS). Many states have made enormous efforts to control in-state sources of pollutants, but it is difficult to simulate out-of-state sources from long range transport, especially from sources outside of the simulation domain. Air Quality Forecast (AQF) systems have exhibited considerable underestimations of PM during the springtime in many southern coastal states such as Texas and Louisiana, especially when southerly winds prevail. This period coincides with high fire activity in Central America, near southern Mexico and the Yucatan peninsula. In this work we have used multiple AQF outputs combined with observational data to show the impact of fire emission transport from Central America over the Gulf of Mexico to the southern U. S. Surface PM and speciation observations from the U.S. Environmental Protection Agency (EPA) AIRNow and Air Quality System (AQS) and Interagency Monitoring of Protected Visual Environments (IMPROVE) networks are utilized to compare with AQF outputs from the University of Houston AQF (UH-AQF) system and the NOAA/NWS National Air Quality Forecast Capability (NAQFC). Satellite-based fire detection information (e.g. Hazard Mapping System (HMS), HMS fire smoke detection from visible channel, Moderate Resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Depth (AOD), MODIS Fire Radiative Power (FRP), and MODIS area burned data) are also utilized and compared. Multi-year trends are compared with climate indices, such and Multivariate ENSO Index (MEI) and Palmer Drought Severity Index (PDSI). Analyses suggest that the frequency of Central American fire events do not strongly correlate with small variations in the climatology indices since a large fraction of springtime Central American fires are due to agriculture-related human activity. However, extreme shifts in climate may alter fire management practices and significantly change the impact of these fires on the southern U. S.

Applications and Evaluations of Response Surface Model (RSM) Using CMAQ in the United States and Asia

Carey Jang, Ph.D.,

Office of Air Quality Planning and Standards, USEPA,

MC: C439-01, 109 T.W. Alexander Drive, RTP, NC 27711,

Tel: 919-541-5638; Fax: 919-541-0044, E-mail: jang.carey@epa.gov

abstract: Using computationally demanding air quality models as regulatory tools to meet time-pressing policy analysis and support always presents a challenge to the regulatory as well as the scientific community. To address this issue and build a better bridge between scientists and policy makers, an innovative policy analysis tool, the Response Surface Model (RSM), has been developed by U.S.EPA to provide a real-time assessment of efficacy of various emissions control strategies. RSM is a reduced-form modeling tool which utilizes advanced mathematical and statistical techniques to characterize the relationship between model input parameters (i.e., pollutant emissions) and outputs (i.e., pollutant concentrations) and in a highly economical manner. The RSM methodology has recently been successfully applied in the USA for as series of studies and the RSM implementations have been extended to China and Taiwan. At the same time, the RSM approach has been tested and stringently validated (through "out-of-sample validation" and "cross validation") for PM 2.5, ozone, and deposition, respectively. An overview of the development of multi-pollutant RSM using the Community Multi-Scale Air Quality (CMAQ) model and its ongoing applications in the USA, China and Taiwan will be discussed and presented.

Implementation of GEM-MACH10, A New Higher-Resolution Version of the Canadian Operational Air Quality Forecast Model

Mike Moran, Sylvain Manard, Radenko Pavlovic, Sylvie Gravel, Samuel Gilbert, Hugo Landry,Wanmin Gong, Craig Stroud, Sunling Gong, and Qiong Zheng

GEM-MACH15 has been Environment Canada's operational regional air quality forecast model since November 2009. GEM-MACH15 is a limited-area configuration of GEM-MACH, an on-line chemical transport model that is embedded within GEM, Environment Canada's multi-scale operational weather forecast model. It is run twice daily to produce 48 hour forecasts of hourly O₃, PM_2.5, and NO₂ fields over a North American grid with 15 km horizontal grid spacing, 58 vertical levels from the surface to 0.1 hPa, and a 450 s time step. A new model version, called GEM-MACH10, is now being tested for operational implementation. It uses 10 km horizontal grid spacing, 80 vertical levels, a 300 s time step, updated model source code, and an improved treatment of emissions.

The computational cost of GEM-MACH10 is roughly a factor of four larger than that of GEM-MACH15 due to the increased spatial resolution. The improved forecast performance resulting from these changes will be described by means of a number of evaluation metrics and analysis techniques. Some of the challenges encountered in developing this new model version will also be discussed, including apparent scale dependencies in some of the spatial surrogate fields used to allocate inventory emissions to the 10 km and 15 km model grids.

Using CAMx Source Apportionment Modeling to Characterize the PM2.5 and Ozone-Related Health Burden Attributable to 23 Emission Sectors in the Continental U.S.

Neal L. Fann

Kirk R. Baker

Charles M. Fulcher

Using source apportionment photochemical modeling, we seek to attribute PM_2.5 and ozone air quality levels and health burden among industrial point, area, mobile and international emission sectors in the Continental U.S. in 2005 and 2016. We also investigate the burden posed by sectors to populations most vulnerable to PM_2.5-related premature mortality. We find that the estimated PM_2.5 and ozone-related health burden declines between 2005 and 2016. The number of PM_2.5 and ozone-related deaths among EGUs and mobile sources falls from about 69,000 in 2005 to about 36,000 in 2016. Populations with less than a grade 12 education, whom we identify as most vulnerable, experience the greatest reduction in mortality risk. Mortality risk from area sources grows slightly between 2005 and 2016, due largely to population growth. While the overall burden on public health of PM_2.5 and ozone air pollution will decline, many sectors continue to pose a significant risk. The estimated magnitude and spatial and temporal distribution of mortality and morbidity risk by sector may be useful to policymakers as they design effective air quality management policies.

Uncertainty analysis of regional air quality models using high-order sensitivities

Wenxian Zhang, Marcus Trail, Alexandra Tsimpidi, Yongtao Hu, Athanasios Nenes, Armistead Russell

An efficient approach is presented to quantify the uncertainty of pollutant concentrations simulated by an air quality model. Given that air quality models are widely used to evaluate control strategy effectiveness, information to help understand the accuracy of the model simulations can be useful in decision making. This study focuses on characterization of model uncertainty due to uncertain emission estimates used in the modeling process. The uncertain emission inventories studied are NO_x, SO₂, NH₃, and primary PM emissions, both domain-wide and from specific sectors (e.g., point, area, on-raod and non-road mobile). A reduced form model (RFM) of the Community Multi-scale Air Quality (CMAQ) model is employed to propagate the uncertainty in Monte Carlo simulation. The RFM is constructed using first- and second-order sensitivities obtained from a single CMAQ-HDDM-3D simulation. It well represents the pollutant-precursor response and largely reduces the computational effort. This efficient approach is applied to an episode in Houston area. For domain-wide emissions with an uncertainty factor of 2, the uncertainties of 24-hour averaged PM_2.5 and 8-hour averaged ozone concentrations are up to 10% and 30%, respectively. The response of pollutant concentrations to different emission control strategies in the presence of uncertain emissions rates is investigated. Comparison with observation data will also be conducted to analyze the bias of the model simulation.

Modeling Regional Ozone: A Comparison between a global and a regional model

Barron Henderson, Lin Zhang, Chris Emery and Joseph Pinto

Exposure to ozone is associated with a variety of health outcomes ranging from mild breathing discomfort to mortality. Ozone exposure fields are produced from precursors that can be broadly categorized as natural, intra-continental anthropogenic, and inter-continental anthropogenic. Understanding inter-continental contributions require global scale simulations. Understanding ozone exposure field features, however, requires finer spatial detail than can be supplied by most current global simulations. We compare the results from a nested version of the GEOS-Chem model (~ 50 km X 50 km resolution; Zhang et al. 2011) with that from CAMx (12 km X 12 km resolution; Emery et al. 2012) at CASTNET, AQS and NOAA monitoring sites for 2006. Both models are nested within a coarse resolution version of GEOS-Chem (~ 2 X 2.5). Both models share similarities in simulated MDA8 O3. For example, both models indicate that intercontinental transport and stratospheric intrusions are the major sources of ozone in the intermountain West in spring; and in general when mean surface ozone is compared on a regional basis for averaging periods of a month or more, reasonable agreement is found between them and with measurements.

Most apparent differences between the models and with measurements occur at the ends of the concentration distribution (e.g., for 1^st and 99^th percentiles) on a day specific basis, e.g., both models show limited ability to capture the timing and strength of episodic sources such as stratospheric intrusions. They however, show better agreement with observations if the requirement of day specificity is removed, i.e., if episodic events are treated as stochastic. Even when day specificity is removed, differences between the simulations and between the simulations and observations persist. Some persistent differences are attributed to spatial resolution as expected. Others, however, can be attributed to the models' representation of chemistry, their emissions from wildfires, and their production of NO_x from lightning.

These results indicate that spatial resolution is not the only factor to be considered when attempting to simulate or attribute regional air quality. Rather, difference in models' treatments of atmospheric chemistry and physics must be considered. These results are also in accord with earlier findings showing that agreement between models and measurements is improved as the averaging time of the simulation and measurements are increased. It is also apparent that in analyzing time series over long time periods (e.g., months), special care should be taken to examine temporal trends in bias as this will improve understanding of the processes in the model.

Emery, C.; Jung, J.; Downey, N.; Johnson, J.; Jimenez, M.; Yarwood, G.; Morris, R. Regional and global modeling estimates of policy relevant background ozone over the United States. Atmospheric Environment 2012, 47, 206-217.

Zhang, L.; Jacob, D. J.; Downey, N. V.; Wood, D. A.; Blewitt, D.; Carouge, C. C.; van Donkelaar, A.; Jones, D. B. A.; Murray, L. T.; Wang, Y. Improved estimate of the policy-relevant background ozone in the United States using the GEOS-Chem global model with 1/2 x 2/3 horizontal resolution over North America. Atmospheric Environment 2011, 45, 6769-6776.

Application of DDM for Predicting Ozone Responses within an Urban Area

Heather Simon, Kirk Baker, Norm Possiel, Farhan Akhtar, Sergey Napelenok, Brian Timin, Benjamin Wells

In the United States, ozone monitoring data are compared to regulatory standards based on a 3-year average of the annual 4^th highest 8-hour daily maximum (MDA8) concentration. Since exposure and epidemiology studies often use other ozone metrics to estimate the health benefits associated with lowered ozone levels within a particular urban area, researchers must estimate how lowering the 4^th highest MDA8 ozone value will affect the entire distribution of hourly ozone concentrations in that area. In the past, mathematical techniques such as quadratic rollback have been used to accomplish this task. However, quadratic rollback requires that all monitors in an urban area exhibit the same response to emissions changes and this technique assumes that ozone concentrations never increase in response to emission reductions. Here we present a new methodology utilizing a series of CMAQ/Higher-order Direct Decoupled Method (HDDM) runs over different emissions scenarios to estimate the spatially and temporally varying ozone responses over the entire range of precursor emissions reductions. This new method generalizes the CMAQ/HDDM sensitivities so that they can be applied to non-modeled time periods and allows for bias correction in cases where the modeled and observed ozone concentrations disagree. This new methodology is applied to case studies in Detroit and Atlanta.

Using Hemispheric-CMAQ to Provide Initial and Boundary Conditions for Regional Model Application

Joshua S. Fu¹, Xinyi Dong¹, Kan Huang¹, Yang Gao¹, and Carey Jang²

¹Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, TN

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

Initial and boundary condition (IC/BC) inputs for regional model are usually provided by global model through downscaling, for example, IC/BC derived from GEOS-Chem has been successfully employed in CMAQ applications for regional modeling (Lam and Fu, 2009). However, some limitations remain in the global model downscaling process, such as the stationary assumption within statistically downscaling limit the relationship between observation and models, and dynamic downscaling is usually computationally demanding and requires considerable implementation effort. In addition, the inconsistency between global and regional model in terms of both meteorology and chemical mechanisms may also introduce unrealistic chemical transports (Bullock et al., 2008; 2009). Therefore, it is necessary to extend the ability of regional model to conduct global scale of simulation and generate the nest-down IC/BC for regional applications. In this study, we are using CMAQ with polar stereographic projection for the whole North Hemisphere targeted to provide IC/BC for regional model application (HCMAQ).

The hemispheric modeling domain has horizontal resolution of 108km x 108km, and 34 vertical sigma-pressure coordinates with top height at 50 hPa. Meteorology filed is provided by WRF. Although emission inventories are available for regional application with CMAQ such as the National Emission Inventory (NEI) for US, the Intercontinental Chemical Transport Experiment-Phase B (INTEX-B) emission for China (Zhang et al., 2009), they are not directly used in this application because we want to examine the difference between HCMAQ and global model GEOS-Chem in terms of providing IC/BC, so emission dataset used in GEOS-Chem is extracted and re-projected for HCMAQ in order to keep the consistency between two different models. GEOS-Chem model simulation has been conducted for whole year 2008, and HCMAQ model simulation will be conducted for January and July 2008 for the first phase test run. Two sets of IC/BC will be derived for 36km US domain, one from GEOS-Chem results and the other from HCMAQ. These two sets will be compared to examine the difference caused by different chemical mechanisms, map projections, meteorological fields while generating IC/BC for regional application.

Reference:

Bullock, O.R., Jr., D. Atkinson, T. Braverman, K. Civerolo, A. Dastoor, D. Davignon, J-.Y. Ku, K. Lohman, T.C. Myers, R.J. Park, C. Seigneur, N.E. Selin, G. Sistla, K. Vijayaraghavan (2008), The North American Mercury Model Intercomparison Study (NAMMIS): Study description and model-to-model comparisons, J. of Geophys. Res., 113, D17310, doi:10.1029/2008JD009803

Bullock, O.R., Jr., D. Atkinson, T. Braverman, K. Civerolo, A. Dastoor, D. Davignon, J-.Y. Ku, K. Lohman, T.C. Myers, R.J. Park, C. Seigneur, N.E. Selin, G. Sistla, K. Vijayaraghavan (2009), An Analysis of Simulated Wet Deposition of Mercury from the North American Mercury Model Intercomparison Study (NAMMIS), J. of Geophys. Res., 114, D08301, doi:10.1029/2008JD011224

Lam Y.F. and J.S. Fu A novel downscaling technique for the linkage of global and regional air quality modeling, Atmos. Che. Phys., 9, 9169-9185, 2009

Zhang, Q., Streets, D. G., Carmichael, G. R., He, K. B., Huo, H., Kannari, A., Klimont, Z., Park, I. S., Reddy, S., Fu, J. S., Chen, D., Duan, L., Lei, Y., Wang, L. T., and Yao, Z. L.:Asian emissions in 2006 for the NASA INTEX-B mission,Atmos. Chem. Phys.,9, 5131-5153, 2009

Quantifying the sensitivity of U.S. ozone concentrations to domestic vs international emissions through coupled GEOS-Chem Adjoint and CMAQ DDM source-receptor modeling

Farhan Akhtar, Barron Henderson,Sergey Napelenok, Daven Henze, Susan Anenberg, John Langstaff, Rob Pinder

Previous studies have estimated international contributions to U.S. ozone concentrations by modeling the concentration response to large changes in international or local emissions. A major limitation of these studies is that broad perturbations in anthropogenic emissions may dramatically alter the chemical environment in which ozone is created or destroyed, limiting the applicability of the results to evaluations of alternative emission control scenarios used in air quality management policy development.

Instrumented modeling techniques such as GEOS-Chem adjoint and CMAQ Direct Decoupled Method (CMAQ/DDM) overcome these limitations by directly calculating the U.S. local concentration response to local and global emission changes. Since all emissions are unchanged when these tools are used, the simulations more closely represent the sensitivity of present concentrations to small changes in emissions similar to those resulting from air pollution management policies.

We run both the GEOS-Chem adjoint and CMAQ/DDM for a global pollution transport episode in April 2008. With the GEOS-Chem adjoint model, we find the sensitivity of ozone and ozone precursor concentrations at the North American CMAQ domain boundary to emissions of ozone precursors throughout the globe at a 2x2.5 resolution. Using CMAQ/DDM, we find the sensitivity of local concentrations at 12km to ozone and ozone precursors transported through the domain boundary. We also include ozone sensitivity calculations to local emissions of NOx and VOCs as a part of the CMAQ/DDM runs, allowing for comparison of ozone sensitivity to domestic and international emissions.

In themselves, both model applications allow for insight into the effects of international transport on ozone in the U.S. By using a consistent set of boundary conditions as an intermediary, we combine these studies for a high resolution analysis of how changes in global emissions at 2x2.5 affect U.S. ozone concentrations at 12km. Linking the global and regional models allows us to exploit the benefits of both techniques: addressing a large number of global sources with GEOS-Chem/Adjoint while having high-resolution, local concentration responses from CMAQ/DDM.

Assessing the Impact of Changes in Climate and Emission on Global Air Quality

Timothy Glotfelty, Yang Zhang, and Shuai Zhu

Air Quality Forecasting Lab, North Carolina State University, Raleigh, NC

Prakash Karamchandani

ENVIRON International Corporation, Novato, CA

David G. Streets

Decision and Information Sciences Division, Argonne National Laboratory, Argonne, IL

removed at author's request

Refining Ammonia Emissions Estimates with Observations during CalNEX

Shannon Capps¹, Daven Henze², Armistead Russell¹, Athanasios Nenes¹

¹Georgia Institute of Technology, Atlanta, GA

²University of Colorado, Boulder, CO

Contributions of reactive nitrogen emissions from dense agricultural sources to regional air quality, especially particulates, can be challenging to represent accurately in chemical transport models (CTMs). Detailed process refinement necessary to represent emissions from seasonal fertilization and soil or vegetative bi-directional flux suggests that an alternate initial step toward accurate representation would be beneficial. Receptor-oriented sensitivity analysis and assimilation of observations provide a means of assessing the extent of the role of ammonia in shaping the air quality of the west coast of the US.

Here, the adjoint of GEOS-Chem is employed to conduct receptor-oriented sensitivity analysis and to assimilate satellite and ground-based observations of atmospheric composition. CalNEX, a 2010 field campaign centered over California, provides a relatively dense observational dataset for optimization of emissions parameters through inverse modeling. Additionally, receptor-oriented sensitivity analysis reveals the relative importance of various sources at specific locations to air quality along the west coast.

Cost-effective dynamical downscaling:
An illustration of downscaling CESM with the WRF Model

Jared H. Bowden
Saravanan Arunachalam

Institute for the Environment, UNC - Chapel Hill

Dynamical downscaling of climate change projections from Global Climate Models (GCMs) requires large computational resources (high cost) because dynamical downscaling requires simulating multiple years to decades to provide statistics of the weather. With cost-effective dynamical downscaling, we only simulate a year each in the contemporary and future periods (low cost) to project the climate change signal. Our strategy is to understand and acknowledge potential limitations (effectiveness) of using a select few years while selecting years of interest using a quantitative measure. Such a strategy is important for applications wanting to use dynamical downscaling information for select years. For instance, we have an interest in the impact of climate change on future year air quality as emissions change, where emissions estimates are available only for 2005 and 2025.

For illustrative purposes of cost-effective dynamical downscaling, we will downscale Climate Earth System Model (CESM) using the Weather Research and Forecasting (WRF) model for a year each in the contemporary and future climate. A comparison will be made between the dynamical downscaling projection to the GCM projection for both a 30-year mean climate and for the same select years. We will describe the quantitative method used to arrive at the years selected, discuss the potential limitations of those years selected, and summarize the effectiveness of downscaling select years.

Source Attribution of Health-based and Attainment-based Ozone Metrics in North America

Amanda Pappin, Morteza Mesbah, Amir Hakami

Air quality improvements achieved in North America over the past few decades are largely attributable to an attainment-based regulatory framework. Such an approach to controlling air pollution incorporates chemical transport modeling to establish relationships between sources and non-attainment of air quality standards across receptors. A tandem approach to designing emission control strategies is the use of exposure-based metrics (e.g. mortality or morbidity) as a means of quantifying the health (and other) benefits of individual source controls. Adjoint sensitivity analysis is a tool that allows for calculation of spatiotemporal influences of emission source reductions to such attainment or health metrics. In adjoint sensitivity analysis, source specificity is conserved, making it a viable option for evaluating the relative importance of emission sources in a straightforward manner. In this work, we attribute short-term mortality (valuated as an overall "health benefit") in Canada and the U.S. to anthropogenic NO_xand VOC emissions across North America. We then perform similar source attribution of national average standard non-attainment to those same emissions across the domain. Our results for both Canada and the U.S. show significant spatial and temporal variability in health benefit influences.Furthermore, our work suggests that potential health benefits are vastly under-represented in the current benefit-cost analysis framework that lacks source specificity.We also find a significant influence of epidemiological averaging period on the magnitude of health benefit influences seen from emission reductions across the domain.For example, U.S. health benefits from NO_xsource reductions appear to be substantially larger (a factor of about 2 times) when calculated for daily 1-hr maximum O₃exposure than for a 24-hr O₃averaging period.We examine the potential policy applications of this work using a comparison between source attribution of exposure-based health metrics and attainment-based metrics.

Development of Methodology to Downscale Global Climate Fields to 12km Resolution

Russ Bullock, Kiran Alapaty, Jerry Herwehe, Megan Mallard, Tanya Otte

Previous dynamical downscaling results developed using WRF at 108- and 36-km resolution are applied for further downscaling to 12-km resolution over the eastern United States and southeastern Canada. The previous work investigated both analysis nudging and spectral nudging to constrain the WRF simulation. This work investigates adjustments to the nudging parameters and various physics options required for an optimum application at 12-km resolution. Lake surface temperatures for the previous work were determined from 2.5-degree latitude/longitude resolved sea surface temperature data and that technique was also applied here with results suggesting that a physics-based simulation of lake surface temperatures is needed for 12-km downscaling. Simulated surface-level temperature, water vapor mixing ratio, and wind speed are compared to hourly observations collected by the Meteorological Assimilation Data Ingest System (MADIS) to demonstrate the accuracy of the 12-km downscaling results. Simulated precipitation is compared to Multisensor Precipitation Estimator (MPE) data showing a general excess of precipitation, especially in the southeast during summer.

Implementation and evaluation of PM2.5 source contribution analysis in a photochemical model

Roger Kwok, Sergey Napelenok, Kirk Baker

Emission-based source apportionment algorithm is implemented in CMAQv4.7.1 to track sector/region contributions to particulate matter concentration as well as dry and wet. Flexibility is enhanced by allowing users to choose any number of sectors to be tracked, with or without specifying any source regions at runtime. Results are compared with several zero-out scenarios. The developed methodology compares well for more cases where a more linear relationship between precursor and modeled pollutant exist, as expected. A series of process-level tests are performed for more nonlinear cases in order to highlight the strengths and weakness of the methodology. The algorithm is also applied to an annual 2005 CONUS 36-km simulation, where several emissions sectors including mobile, electricity generation, and agriculture are tracked as a demonstration of the method capability. Seasonal variability in resulting sector contributions as well as in the bulk concentrations is discussed.

Improvement of extreme climate predictions from dynamical climate downscaling

Yang Gao¹, Joshua S. Fu¹, John B. Drake¹, Yang Liu², Jean-Francois Lamarque³, Kan Huang¹, Xinyi Dong¹ and David Wong⁴

¹Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, TN

²Rollins School of Public Health, Emory University, Atlanta, Georgia

³Atmospheric Chemistry and Climate and Global Dynamics Divisions, National Center for Atmospheric Research, Boulder, CO

⁴Atmospheric Modeling and Analysis Division, NERL, USEPA, Research Triangle Park, NC

Due to the limitations of spatial resolution in global climate models, the mean climate studies are more preferable on a global scale. However, the extreme climate conditions play much more significant roles on human health. Thus, dynamical downscaling technique is used to link global and regional climate models. In this study, regional climate model Weather Research and Forecasting Model (WRF) was used to downscale three hourly global climate outputs from Community Earth System Model (CESM) version 1.0. The spatial resolution of CESM is 0.9 by 1.25 degree, while the resolutions of downscaled WRF domains are 36 km by 36 km and nested to 12 km and 4 km.

In this study, the present climate (2001-2004) and fossil fuel intensive scenario Coupled Model Intercomparison Project Phase 5 (CMIP 5) Representative Community Pathways (RCP) RCP 8.5 (2057-2059) was downscaled in order to determine future climate change. We first evaluated the predictions of extreme climate, including heat waves and flood between CESM and WRF. Compared with the observational data from National Climatic Data Center (NCDC), significant improvements in extreme events predictions were achieved after downscaling, and large bias was reduced in WRF in comparison to CESM. Compared to present climate, more intense heat waves and extreme rainfall will occur across the Eastern US.

Modeled Nitrogen Deposition Source Apportionment at Rocky Mountain National Park for RoMANS2

Michael Barna¹, Marco Rodriguez², Kristi Gebhart¹, Bret Schichtel¹, William Malm³

1National Park Service, Fort Collins, CO
2AECOM, Fort Collins, CO
3CIRA, Fort Collins, CO

The CAMx regional air quality model was used to estimate a source apportionment for the wet and dry deposition of nitrogen-containing compounds as part of the second Rocky Mountain Atmospheric Nitrogen and Sulfur study (RoMANS2). In particular, the impacts of oxidized and reduced nitrogen emission sources on deposition at Rocky Mountain National Park (RMNP), which is located in northern Colorado, were considered. Recent research has indicated that RMNP is above a "critical load" with regard to deposited nitrogen, and that sensitive alpine ecosystems are threatened under current conditions. RoMANS2 is a follow-up study to RoMANS, and includes an extended field campaign that was conducted in 2009. Nitrogen species included in the evaluation were particulate nitrate and ammonium, nitric acid and ammonia. Several different emission source categories contribute to deposited nitrogen at RMNP, including nitrogen oxide emissions from mobile sources along the Denver I-25 corridor and ammonia emissions from agricultural operations in northeastern Colorado.

Influences of Regional Climate Change on Air Quality across the Continental U.S. Projected from Downscaling IPCC AR5 Simulations

Chris Nolte, Tanya Otte, Rob Pinder, Jared Bowden, Greg Faluvegi, and Drew Shindell

Projecting climate change scenarios to local scales is important for understanding, mitigating, and adapting to the effects of climate change on society and the environment. Many of the global climate models (GCMs) that are participating in the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) do not fully resolve regional-scale processes and therefore cannot capture regional-scale changes in temperatures and precipitation. We have used the Weather Research and Forecasting (WRF) model as a regional climate model (RCM) to dynamically downscale the NASA GISS ModelE2 GCM over North America to 36 km for two 11-year periods that represent current climate (circa 2000) and climate at 2030 following the RCP6.0 scenario. In this work, these RCM scenarios are used as meteorological inputs for the Community Multiscale Air Quality modeling system to explore influences of regional climate change on air quality. We will examine the representativeness of the downscaled meteorology for current climate and the interannual variability in simulated air quality induced by the downscaled meteorology.

Airborne remote sensing of air quality over London and Western Europe: Validation and modelling of novel airborne FTIR-derived trace gas concentrations during the ClearFLo and MAMM campaigns.

Grant Allen, University of Manchester, UK

D. P. Moore, University of Leicester, UK

S. Newman, UK Met Office

A. Vance, UK Met Office

C. Percival, University of Manchester, UK

M. Gallagher, University of Manchester, UK

H. Coe, University of Manchester, UK

J. Lee, University of York, UK

J. McQuaid, University of Leeds, UK

Poor air quality in urban environments has been linked to premature death. Moreover, the ventilation of the urban boundary layer and impacts of advected urban pollution on the regional scale downwind may also contribute to adverse health effects in remote rural environments.

Airborne remote sensing of trace gases pertinent to air quality monitoring provide a measurement strategy that informs air quality models at both the urban and regional scales. In this study we present novel trace retrievals from an airborne FTIR - the Met Office Airborne Research Interferometer Evaluation System (ARIES). The new data provided by the ARIES instrument permits simultaneous nadir total column (and resolved profile) measurement of a wide range of trace gases (e.g. O_3, CO₂, H₂O, N₂O, CH₄, HNO₃) at ~1km spatial scale. Such data are well-suited to assimilation into air quality models and also serve as a useful validation of model output as well as facilitating inverse flux analyses.

Trace gas retrievals from ARIES will be presented in the context of simultaneous airborne and ground-based (in situ and remotely-sensed) data collected during the Clear Air For London (ClearFLo, July 2012) campaign and the Arctic Methane measurement and modelling (MAMM) campaign (July 2012). Both campaigns sample polluted airmasses in the urban environment and in the regional plume downwind, allowing excellent validation of the new retrieval strategy and a comprehensive dataset from which to conduct air quality modelling.

We will conclude by discussing planned applications of these measurements in flux determination and source-receptor analysis.

Development of a Regional-Scale Pollen Emission and Transport Modeling Framework for Investigating the Impact of Climate Change on Allergic Airway Disease

Rui Zhang¹, Tiffany Duhl², Muhammad T. Salam³, James M. House⁴, Richard C. Flagan⁴, Edward L. Avol³, Frank D. Gilliland³, Alex Guenther², Serena H. Chung¹, Brian K. Lamb¹, and Timothy M. VanReken¹

¹Washington State University, Pullman, WA

²National Center for Atmospheric Research, Boulder, CO

³University of Southern California, Los Angeles, CA

⁴California Institute of Technology, Pasadena, CA

Exposure to bioaerosol allergens such as pollen can trigger allergic airway disease (AAD) for sensitive populations and thus can cause serious public health problems. Assessing these health impacts by linking the airborne pollen levels, concentrations of respirable allergenic material, and human allergenic response under current and future climate conditions is a key step toward developing effective preventive actions. To that end, a regional-scale pollen release and transport modeling framework has been developed that treats allergenic pollens as non-reactive tracers within the WRF/CMAQ air-quality modeling system. The pollen potential model, a module within the Model of Emissions of Gases and Aerosols from Nature (MEGAN), was developed to generate daily pool of pollen available for release. It is driven by species-specific meteorological threshold conditions (temperature and/or precipitation) and is flexible with respect to the representation of vegetation species and plant functional types. The hourly pollen emission flux was parameterized by considering available pollen grains, friction velocity and wind threshold values. The dry deposition velocity of each species of pollen was estimated based on pollen grain size and density.

An evaluation of this new pollen modeling framework was conducted over southern California for March to June 2010. This period coincided with observations by the University of Southern California's Children's Health Study (CHS), which included O₃, PM_2.5, and pollen count, as well as sensitization data at nine sites. Two nesting domains with horizontal resolution of 12 km and 4 km were constructed and six representative allergenic pollen genera were included: birch tree, walnut tree, mulberry tree, olive tree, oak tree, and brome grasses. Under the current parameterization scheme, the modeling framework tends to underestimate peaks in oak pollen concentration and overestimate grass pollen concentration, but it shows reasonable agreement with observed birch, olive, and mulberry tree pollen concentrations. Sensitivity studies suggested that the estimation of pollen available for release is a major source of uncertainty for simulated pollen concentrations. Achieving agreement between emission modeling and observed pattern of pollen releases is the key for successful pollen concentration simulations. The framework is being used to evaluate the impact of climate change on pollen release and concentration by comparing the simulation results with current decade (1995-2004) and future decade (2045-2054) WRF meteorological fields downscaled from the ECHAM5 global climate model results.

Analysis of ultra-fine particle measurements of urban and rural sites in Delaware

Mohammed A Majeed[1], Golam Sarwar[2], Michael McDowell1, Betsy Frey1, Ali Mirzakhalili1

The particle number size distributions in urban areas vary in time and space in a complex fashion as a result of interaction between local- and long-range transport, and the meteorological conditions. Air quality studies have indicated that particle number size distributions are unevenly spread in urban areas and vary significantly from urban to rural areas. These distributions also vary from winter to summer and from day to night time. Ultra-fine particles of different sizes are typically associated with different aerosol particle modes. High concentrations of particles are found closer to the sources, and the concentrations decrease with distance from the source. Particle number concentrations are high closer to the roadways. Urban aerosols mass distributions are characterized by three modes - nuclei model (particles < 0.1 μm), accumulation mode (0.1 - 1 μm), and coarse mode (> 1 μm). Sources and chemical composition of the fine and coarse particles are different. The number distribution of rural continental aerosols is characterized by two modes while the mass distribution is dominated by the coarse mode, which is not influenced by local sources.

Delaware Department of Natural Resources and Environment Control's Division of Air Quality measured size-resolved ultra-fine particle number concentrations in an urban area (Wilmington, New Castle County) for nearly a year, and a rural area (Sussex County) for six months by using (TSI Model 3031 Ultrafine Particle Monitor). Particle numbers were measured in six different size bins (all in nm): 20-30, 30-50, 50-70, 70-100, 100-200, and 200-500 with a sampling interval of 15 seconds. In this study, we compare urban particle number concentrations to the rural concentrations. Mean rural particle number concentrations are substantially lower than the urban concentrations. We compare diurnal, weekly, and seasonal pattern of the particle number, surface, and volume concentrations at rural site to those of urban site. We investigate the particle formation and growth rates, and identify particle nucleation events. Our goal is to improve the understanding of the influence of local and regional nucleation events on ultra-fine particle concentrations in urban and rural areas.

Modeling Emission and Transport of Allergenic Tree Pollens under Climate Change

Yong Zhang¹, Sastry Isukapalli¹, Leonard Bielory², Lai-yung Ruby Leung³,Panos G. Georgopoulos¹

1 Environmental and Occupational Health Sciences Institute (EOHSI), A Joint Institute of UMDNJ-RW Johnson Medical School \& Rutgers University, 170 Frelinghuysen Road, Piscataway, NJ 08854, USA

2 Center for Environmental Prediction, Rutgers University, New Brunswick, NJ 08901, USA

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

Abstract: Climate change is expected to alter the emission and transport of allergenic pollens, and potentially increase occurrence of allergic airway disease (AAD). A modeling system is presented here for studying emission and transport of two representative allergenic tree pollens (birch and oak) under climatic change conditions. The emission module considers major physical processes such as direct emission and re-suspension of pollen particles, and accounts for meteorological parameters such as ground surface temperature, friction velocity, humidity, precipitation, etc. This module also incorporates statistical modeling based on historical data for estimating effects of climate change on annual flux and start date of pollen emissions. The transport of pollen was simulated via the combined application of the Weather Research and Forecasting (WRF) model, the Sparse Matrix Operator Kernel Emissions (SMOKE) model and an adapted version of Community Multiscale and Air Quality (CMAQ) model. Spatiotemporal profiles of pollen emissions and ambient airborne distributions in 2004, 2040 and 2050 were estimated through the combined application of (a) analysis results of historical data, (b) meteorological data from WRF under representative climate change scenarios, (c) land use and land cover data from Biogenic Emissions Land use Database, version 3 (BELD3). Simulation results of ambient distributions of pollen in 2004 were evaluated using observed pollen data from multiple pollen stations of the American Academy of Allergy, Asthma & Immunology (AAAAI). It is demonstrated that simulation results from the SMOKE-WRF-CMAQ modeling system could characterize reasonably well the spatiotemporal distribution of birch and oak pollen; and that the simulation estimates were comparable with those from observed climatologic means. It is also shown that responses of pollen timing and quantity to future climatic conditions will be different for different allergenic genus and different regions. Simulation results improve our understanding of climatic change effects on timing and quantity of emission and transport of allergenic pollens, and provide information useful in managing public health problems associated with expected increases in cases of AAD.

Keyword: Climate change, CMAQ, Pollen, Birch, Oak

AERLINE: Air Exposure Research model for LINE sources

Michelle G. Snyder, David K. Heist, Steven G. Perry, and Vlad Isakov (U.S. EPA, RTP, NC), Akula Venkatram (UCR), Sarav Arunchalam (UNC)

WHAT IS AERLINE
AERLINE (under development) is a research-level, line-source dispersion model being developed by EPA's Office of Research and Development as a part of the ongoing effort to further develop tools for a comprehensive evaluation of air quality impacts in the near-road environment. AERLINE is currently formulated for near-surface releases, contains new (field study and wind tunnel based) formulations for the vertical and lateral dispersion rates, simulates low wind meander conditions, includes Monin-Obukhov similarity profiling of winds near the surface and selects plume-weighted winds for transport and dispersion calculations. The model utilizes the surface meteorology provided by the AERMET model and includes user-friendly input requirements such as simplified road-link specifications. Model simulation with integrated point sources has been formulated for computational efficiency and careful attention to appropriately simulate dispersion for receptors very near the source line. The current version of the model is designed for flat roadways (no surrounding complexities). The model framework is designed to accommodate future algorithms for simulating the near-source effects of complex roadway configurations (noise barriers, depressed roadways, etc).

WHY WAS AERLINE DEVELOPED
In 2007, the Science Advisory Board and the National Research Council recommended a paradigm shift from a pollutant-by-pollutant approach to a multi-pollutant approach where the emphasis for air research would be on understanding how specific source types contribute to health effects of concern. The initial focus was to understand the relationship of traffic (roadway emissions) on exposures and adverse health effects in populations living, attending school and working near major roadways. While there are some existing algorithms designed for estimating concentrations in the vicinity of roadways (i.e. CALINE), these approaches are currently not supported nor have they given sufficient attention to near-road complexities and very near road concentrations (i.e. within a few meters of the road). A research effort was initiated to design and conduct wind tunnel and field studies to evaluate pollutant transport and dispersion which provided new and expanded databases for development and evaluation of improved line source algorithms. AERLINE 1.0 Beta is the initial modeling product of this development program. It is a research tool that is designed primarily to support risk assessments and health studies related to near-road pollutants.

WHAT AERLINE IS NOT.
AERLINE 1.0 is not a modeling tool for regulatory applications (e.g. NAAQS enforcement, New Source Review, PM Hot Spot Conformity, SIP analysis, etc). It is contained in a research platform and has not gone through the rigorous review and public comment required for inclusion in the list of recommended regulatory models.

Sensitivities of Spectral Nudging Toward Moisture for Regional Climate Modeling

Tanya L. Otte¹, Martin J. Otte¹, Jared H. Bowden² and Christopher G. Nolte¹

¹U.S. EPA, Research Triangle Park, North Carolina, USA

²University of North Carolina, Chapel Hill, North Carolina, USA

Several studies with WRF and other models have shown that spectral nudging can improve predictions of 2-m temperature and large-scale fields for regional climate modeling. However, compared to analysis nudging, precipitation predictions with spectral nudging are inferior for both means and extremes (Bowden et al. 2012; Otte et al. 2012). In this work, we implemented spectral nudging toward moisture and performed several multi-decadal sensitivity runs with WRF. We show that careful and conservative use of spectral nudging toward moisture (specific humidity) improves the prediction of precipitation without compromising other thermodynamic fields.

Evaluation of CMAQ and WRF-Chem Simulations of Air Quality over the Baltimore-Washington Region during the July 2011 DISCOVER-AQ Field Campaign

Kenneth Pickering, Melanie Follette-Cook, Christopher Loughner, James Crawford, and the DISCOVER-AQ Observation Team

The July 2011 NASA Earth Venture DISCOVER-AQ (Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality) field program period was simulated with the CMAQ v5.0 and WRF-Chem v3.3.1 regional chemical models and evaluated using in-situ data collected by the NASA P-3B aircraft. CMAQ was run with the CB05 chemical mechanism and the AE5 aerosol module, while WRF-Chem was run with the CBMZ mechanism and the MOSAIC 8-bin aerosol scheme. The evaluation was conducted using output from a 4-km horizontal resolution domain over the mid-Atlantic states. During the campaign, over 250 lower tropospheric in-situ profiles of trace gases (e.g. O3, NO, NO2, HCHO, CO, HNO3, peroxy nitrates, alkyl nitrates, and selected VOCs) and aerosols (sulfate, nitrate, black carbon, water-soluble organic carbon, etc.) were observed over a set of selected Maryland Department of the Environment air quality monitoring stations. Other observational components included remote sensing of aerosols and trace gases by airborne and ground-based instruments, as well as ozonesonde launches. The initial analysis was conducted by constructing curtain plots of each species using profiles extracted from the model at the aircraft locations as a function of time. Observed mixing ratios are overplotted on these curtains. This technique allows easy identification of under or overprediction by the model for individual species. In addition, scatter plots of model versus observed data were constructed and correlation analysis was conducted. We also evaluate the model boundary layer depths against those derived from the aircraft profiles and assess the errors in model-predicted species concentrations due to boundary layer depth errors. We compare performance between the two models in terms of the photochemistry and aerosol generation.

Simple urban parameterization for WRF-CMAQ

Jonathan Pleim and Robert Gilliam

Atmospheric Modeling and Analysis Division, USEPA

RTP, NC

Recent evaluation studies of the WRF-CMAQ meteorology and air quality modeling system have shown a persistent tendency to overpredict ground-level concentrations of locally emitted pollutants such as CO and NO_x in developed areas. Emissions of these chemical species are predominantly from mobile sources and typically peak during the morning and evening rush hours which correspond to transition times for PBL development. We have hypothesized that model overpredictions, especially during the evening transition time, are due in part to inadequate representation of urban heat island (UHI) effects in the Pleim-Xiu land surface model (PX LSM). Therefore, we made several modifications to the PX LSM, such as scaling the surface heat capacity according to the fractional coverage of impervious surfaces from the National Land Cover Data (NLCD) and increasing the roughness length and decreasing albedo for the 4 developed land-use categories to better represent the effects of building canopies. Also, the deep soil temperature nudging was reduced to account for poor representation of UHI effects in the temperature analyses used for the nudging.

The WRF-CMAQ was run for August and half of September 2006 on three domains: 12 km CONUS, 4km Texas, and 1 km over the Houston area. Comparisons to tethersonde observations made during the evening and overnight at the University of Houston for several nights in September 2006 show reduced stability in the lowest 100 - 200 m in the run with the urban modifications compared to a base run, which is in much better agreement with the measured profiles. These modifications also reduce the overpredictions of surface CO and NO_x in the Houston area.

Chemical data assimilation with CMAQ and DISCOVER-AQ field measurements

Tianfeng Chai ^1,2, Hyun-Cheol Kim^1,2, Pius Lee¹, and Li Pan^1,2

1. NOAA Air Resource Laboratory, Silver Spring, MD

2. ERT, Inc.

In previous studies, assimilation of MODIS (MOderate Resolution Imaging Spectroradiometer) aerosol optical depth (AOD) shows marginal effect in improving surface PM_2.5 predictions. This is partially due to the complex association of the column quantities of AOD with surface particular matter. An effort has been made to assimilate AIRNow PM_2.5 in-situ observations to complement the satellite observations. However, the vertical profiles are important to simulate aerosol transport and transformation but are often missing. A rich data set of speciated aerosol were obtained during the Deriving Information on Surface conditions from Column and Vertically Resolved Observations Relevant to Air Quality (DISCOVER-AQ) field experiment taken place in June and July 2011 along the Interstate 95/Baltimore-Washington area. It provides multiple vertical aerosol profiles under different meteorological conditions in both urban and rural areas. In this study, the DISCOVER-AQ observations are first used to evaluate the CMAQ model predictions. Then different assimilation strategies are investigated utilizing the DISCOVER-AQ flight measurements. At last, the assimilation tests are validated and the information contents of observations from MODIS, AIRNow, and DISCOVER-AQ flights are studied using a leave-one-out method, where assimilations are carried out with measurements from one platform removed from assimilation data sets.

Does Temperature Nudging Overwhelm Aerosol Radiative Forcing in Regional Integrated Climate Models

Kiran Alapaty, Shaocai Yu, Tanya Otte, and Chris Nolte

Regional climate models are often constrained to a reference atmospheric state by using data assimilation (nudging) to develop credible climate information. In integrated regional atmospheric modeling systems, meteorology and air quality interactions are built in such that changes in both systems will impact the other throughout the model integration. For example, predicted aerosols in an air quality modeling system can modulate radiative fluxes shaping climate simulated by a meteorological model. Temperature is often continuously nudged in meteorological model simulations and such a required constraint may counteract the benefit of using integrated modeling systems, especially in simulating regional climate. To investigate this matter, we developed a simple methodology to estimate the pseudo radiative forcing that arises due to nudging. This pseudo radiative forcing then can be compared with the estimates of aerosol radiative forcing from observations and model simulations. In the context of regional climate modeling, the results have important implications on whether there is an added benefit of using integrated climate modeling systems when continuous nudging is applied.

Fine resolution Air Quality forecasting capability for limited domains over Eastern Texas in support of validation attempt of wild fire emission retrievals by geostationary and polar-orbiting satellites

Pius Lee ¹, Shobha Kondragunta², Hyuncheol Kim³, Xiaoyang Zhang³, Li Pan³, Rick Saylor⁴, Tianfeng Chai³, Daniel Tong³, Ariel Stein⁵

1. Air Resource Laboratory (ARL), NOAA, College Park, MD

2. NESDIS Center for Satellite Applications and Research, NOAA, College Park, MD

3. University of Maryland, College Park stationed at ARL, College Park, MD

4. Air Resource Laboratory (ARL), NOAA, Oak Ridge, TN

5. Earth Resources & Technology (ERT), Annapolis, MD

A novel approach to parameterize heat flux and buoyancy of wild fire plumes is being attempted by NESDIS. The satellite-derived "Fire Radiative Power (FRP)" provides the basic input to the parameterization concerning initial dispersion and plume rise of the plumes. The FRP measurements are retrieved from instrument constellations on board of both geostationary satellites such as GOES and Meteosat and polar orbiting satellites such as Aqua and Terra. This parameterization will be tested in an air quality forecasting experiment. It is decided that a 2 week intensive measurement in September 2012 over Eastern Texas will be conducted involving a few local state or local air quality forecasters to understand the impact of wild fire and the possible leverage of such parameterization for wild fire emission to improving air pollution forecast. A limited fine resolution domain over Eastern Texas nested-in a 12km Conterminous U.S. domain is configured to provide forecasting guidance for the forecasters. NCEP's NAM provides the meteorological input to both of the parent and the nested-in domains. CMAQ version 4.7.1 is coupled to the meteorological model offline at hourly intervals. Real time wild fire emission is ingested into the domains with Blue-sky HSPLIT Hazardous Mapping System and off-line plume-rise calculation approach. The forecast results are evaluated with both AIRNow surface monitoring network and column AOD obtained by MODIS.

Indirect radiative forcing of climate due to aerosols in the two-way coupled WRF-CMAQ: model description, development, evaluation and regional analysis

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

Rob Gilliam, Steve Howard, and S.T. Rao

Atmospheric Modeling and Analysis Division,

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

Atmospheric emissions resulting from consumption of fossil fuels by human activities contribute to global warming and degrade air quality. The IPCC (2007) concludes that the radiative forcing due to the cloud albedo effect (also referred to as first indirect), is estimated to be -0.7 [-1.1, +0.4] W m^-2, with a low level of scientific understanding. For a given cloud liquid water content, an increase in the cloud droplet number concentration implies a decrease in the effective radius, thus increasing the cloud reflectivity; this is know as the first indirect aerosol effect. The second indirect aerosol effect is based on the idea that decreasing the mean droplet size in the presence of enhanced aerosols decreases the cloud precipitation efficiency, producing clouds with a larger liquid water content and longer lifetime. This study implemented indirect aerosol radiative forcing in the two-way coupled WRF3.3-CMAQ5.0 modeling system by including parameterizations for both cloud drop and ice number concentrations. The cloud droplet number concentrations were calculated from the CMAQ-predicted aerosol particles using a parameterization based on a maximum supersaturation determined from a Gaussian spectrum of updraft velocities and the internally mixed aerosol properties within each mode. For each aerosol mode, both the number and mass fractions of aerosol particles activated each time step are determined. The cloud condensation nuclei (CCN) concentrations at six supersaturations (0.02%, 0.05%, 0.1%, 0.2%, 0.5%, 1.0%) are also estimated. On the other hand, the cloud ice number concentrations for the CMAQ-predicted sulfate, black carbon and dust were estimated with an ice nucleation scheme used in the NCAR Community Atmospheric Model (CAM) which considers homogeneous nucleation (-60 C2.5 and regional analysis are carried out over the Houston and continental U.S.

Evaluation of lower and middle tropospheric nighttime ozone from air quality models using Aura/TES and ozonesonde data

Greg Osterman, Jessica Neu, Annmarie Eldering

Jet Propulsion Laboratory/California Institute of Technology

At night, ozone can be transported long distances above the surface inversion layer without chemical destruction or deposition. As the boundary layer breaks up in the morning, this nocturnal ozone can be mixed down to the surface and rapidly increase ozone concentrations at a rate that can rival chemical ozone production. Most regional scale models that are used for air quality forecasts and ozone source attribution do not adequately capture nighttime ozone concentrations and transport. We combine ozone profile data from the NASA Earth Observing System (EOS) Tropospheric Emission Spectrometer (TES) and other sensors, ozonesonde data collected during the INTEX Ozonesonde Network Study (IONS), EPA AirNow ground station ozone data, the Community Multi-Scale Air Quality (CMAQ) model, and the National Air Quality Forecast Capability (NAQFC) model to examine air quality events during August 2006. We present both aggregated statistics and case-study analyses that assess the relationship between the models' ability to reproduce surface air quality events and their ability to capture the vertical distribution of ozone at night. We find that the models lack the mid-tropospheric ozone variability seen in TES and the ozonesonde data, and discuss the conditions under which this variability appears to be important for surface air quality.

Coupled Meteorology and Chemistry Within GEM-MACH: Investigating Feedbacks with Canadas Air Quality Forecast Model

P.A. Makar, W. Gong and J. Milbrandt

Environment Canada is a participant in 2^nd phase of the Air-Quality Model Evaluation International Initiative (AQMEII). In this phase of the multi-model comparison, coupled models are the focus, and specifically, fully coupled models, in which the chemical and meteorological components are allowed to interact. Environment Canada's operational air-quality model GEM-MACH is on-line, but only partially coupled, in that the chemistry portions of the model are influenced by meteorology, but feedbacks from chemistry to meteorology are not allowed.

Progress towards incorporating feedbacks between chemistry and weather in GEM-MACH will be described in this talk. These feedbacks are considered in two process pathways: the radiative transfer and the cloud microphysics. In the current GEM radiative transfer scheme, aerosols are assumed to be uniformly distributed over a 1500m thick layer at the surface, with a latitude-dependant variation in total aerosol optical depth, with standard radiative parameters for continental versus maritime aerosols. O₃ is taken from a latitude-height 2D monthly climatology up to 0.3mb, with the UV-Visible portion of the short-wave spectrum divided into 9 bands. The impact of aerosol on cloud droplet nucleation is taken into account in the Milbrandt-Yau 2-moment bulk cloud microphysics scheme, using two fixed cloud condensation nucleii supersaturation spectra, one for continental and one for marine aerosols (Cohard et al (1998).

In contrast, the air-quality components of GEM-MACH include prognostically varying ozone and aerosols, the latter being spatially and chemically resolved. A suite of different aerosol activation/cloud droplet nucleation schemes have been examined. The process by which these more detailed chemical components are being fed backwards into the meteorology will be described in this talk, along with initial tests of the revised modelling system.

Assimilation of Satellite Observed Clouds in CAMx Modeling System

Arastoo Pour Biazar, Richard McNider, Yun-Hee Park
University of Alabama in Huntsville

Daniel Cohan
Rice University