Session Presentations

Assessment and Development of National Emission Inventory of Major Air Pollutants from Open Municipal Solid Waste Burning in India (2018)

Swetaleena Dash¹, Saroj Kumar Sahu^1*, Poonam Mangaraj¹, G Beig²

¹Utkal University, Bhubaneswar, India

²Indian Institute of Tropical Meteorology, Pune, India

(*Email: saroj.bot@utkaluniversity.ac.in)

Municipal Solid Waste (MSW) management is one of the major environmental issues of Indian cities, also one of the most important challenges of developing countries. Unmanaged MSW becomes a crucial element for the transmission of numerous ailments as well as provides the ground for open burning of the waste. Open burning can happen at the source of the waste production or at a collection site. In India, the waste management problems are aggravated by poor waste segregation at the source leading to lesser treatment and higher open burning. The extremely low calorific value accompanied with high moisture content of the MSW combustion makes it more vulnerable to public health. In the present work, the per capita waste generation in urban and rural as well as major cities have been taken into account to estimate a gridded fine resolution emission inventory of MSW based air pollutants load in 2018 for Indian geographic region. It is found that nearly 180 million tons (MTs) of MSW is being produced every year where the major fraction is being burnt in open. The India level emission for the pollutants i.e. PM₁₀, PM_2.5, CO, NO_X, SO₂, BC, and OC have been estimated to 746.98 Gg/yr, 693.62 Gg/yr, 3574.82 Gg/yr, 199.55 Gg/yr, 26.68 Gg/yr, 53.36 Gg/yr and 693.2 Gg/yr respectively is illustrated in this work. The findings will be an important tool for policymakers as well as the regional air quality modeling studies.

Redistribution of the EDGAR global emissions inventory using local spatial proxies: Assessment of environmental equity metrics.

Espitia Cano, Sebastian Orlando; Montejo-Barato, Alejandra; Morales, Ricardo

Air quality models are essential tools for proper air quality management. However, the quality of atmospheric chemical transport modeling and its use as a tool to inform public policy decisions, strongly depends on the quality of emission inventories. In Colombia, the use of these models has been limited, and only a handful of cities systematically build and revise spatially distributed local atmospheric emissions inventories. Therefore, chemical transport modelling uses detailed emissions from local sources within city limits but has to rely heavily on the use of global emission inventories everywhere outside city limits. Global emission inventories have a resolution of 11 km, and use limited local information to spatially allocate emissions. This dependence on global inventories hinders the applicability of atmospheric modeling for the design of air quality management plans. Therefore, it is necessary to have an improved spatial distribution of the emissions beyond the city limits of Bogota. In this study we use the chemical transport model WRF-Chem v3.9.1. to evaluate the impact of redistributing global emissions according to local spatial proxies in combination with detailed local emission information. Anthropogenic emissions from the global emissions inventory EDGARv4.3.1 were merged with a local emissions inventory for the city of Bogota. Two simulation scenarios were evaluated and compared with the observed concentrations from the air quality monitoring network for Bogota. The base-case (Scenario 1) used the default EDGAR spatial distribution of emissions. In the second scenario (Scenario 2), the emissions from industrial, commercial and mobile sources of the EDGARv4.3.1 inventory were redistributed according to the spatial distribution of the population without considered Bogota (population density data files were obtained the most recent census in Colombia (DANE 2020)). The modeling period for this assessment was February 2018. Three nested domains were used with resolutions of 27 km, 9 km, and 3 km respectively and 41 vertical levels. Gas phase chemistry is described by the RACM mechanism, and aerosol physics by the MADE modal scheme. SOA formation is described with the Volatility Basis Set. Biogenic emissions and biomass burning are also included using MEGAN and FINN-emission-inventory respectively. Our simulations suggest that even the limited use of spatial redistribution applied in the sensitivity Scenario 2 can have a positive impact on improving modeling metrics when compared to the observations. Most of the improvements are achieved in the surroundings of Bogota due to the redistribution of emissions from the EDGARv4.3.1 inventory. Those areas are also the most densely populated, and those with worse air quality. The improved allocation of emissions were then used to compute environmental inequality metrics in the city of Bogota, and were compared with similar metrics calculated using observations. We believe this approach is a step forward to extend the use of air quality modelling in the country, and are a temporary fix while better local atmospheric emissions are available.

Development of a 2017 Emissions Modeling Platform

A. Eyth, M. Strum, J. Vukovich, C. Farkas, J. Godfrey, R. Mason, S. Roberts, C. Allen, J. Beidler

The EPA has developed an emissions modeling platform for the year 2017 based on the 2017 National Emissions Inventory (NEI). Some of the sectors were developed in a consistent way with the 2016v1 Collaborative Emissions Modeling Platform released in the fall of 2019. The 2017 modeling platform has been used for some preliminary air quality modeling studies of the year 2017. The magnitude of emissions in the platform will be reviewed along with some of recent improvements regarding the modeling of the emissions.

Development of 2002-2017 meteorology, emissions, and air quality modeling for the Northern Hemisphere and the Conterminous United States

Kristen Foley, George Pouliot, Alison Eyth, Norm Possiel, Michael Aldridge, Chris Allen, Wyat Appel, Jesse Bash, Megan Beardsley, James Beidler, David Choi, Brian Eder, Caroline Farkas, Rob Gilliam, Janice Godfrey, Barron Henderson, Christian Hogrefe, Shannon Koplitz, Rich Mason, Rohit Mathur, Chris Misenis, Havala Pye, Lara Reynolds, Matthew Roark, Sarah Roberts, Donna Schwede, Karl Seltzer, Darrell Sonntag, Kevin Talgo, Claudia Toro, Jeff Vukovich

EQUATES: EPA's Air QUAlity TimE Series Project

The US EPA is developing a set of modeled meteorology, emissions, air quality and pollutant deposition from 2002 through 2017. Modeled datasets cover the Conterminous US at a 12km horizontal grid spacing and the Northern Hemisphere at a 108km using WRFv4.1.1 for meteorology and CMAQv5.3.2 for air quality modeling. New hemispheric and North American emissions inventories were developed using, to the extent possible, consistent input data and methods across all years.  The new emissions datasets and CMAQ output will be made publicly available to support a wide variety of human health and ecological applications. Model estimated trends will be used for dynamic and diagnostic evaluation of the CMAQ system to inform model development and build confidence in the use of the model for quantifying the impact of meteorological and emissions changes on air quality.  This presentation will describe the development of the emissions inventories and model simulations and will provide initial evaluation results.

Tehran emission inventory development using a new bottom-up emission inventory calculation and reporting system

Hossein Shahbazi^1,2, Amin Hassani³, Vahid Hosseini⁴

^{hshahbazi@mech.sharif.ir}

1- PhD candidate, Mechanical engineering department, Sharif university of technology

2- Tehran Air Quality Control Company

3- PhD candidate, Energy engineering department, Sharif university of technology

4- Associate professor, Mechanical engineering department, Sharif university of technology

The Emission Inventory system is a web-based, user-friendly platform that has been developed to calculate and update Tehran emission inventory. Using this system, it is possible to import source specific input data (including traffic volume, fleet composition, fuel consumption, population distribution ect,) and calculate emission from following sources with high spatio-temporal resolution:

1. Mobile Sources (exhaust, non-exhaust, evaporative and cold start emission)
2. Power plants
3. Bus Terminals
4. Airports
5. Household, general and commercial Sources
6. Refineries
7. Railways
8. Petrol Stations
9. Industries

Here are different applications of the system:

• Tracking emission rates for regulatory purposes and police making
• Running if-then scenarios
• Permitting and compliance
• Providing data for national programs
• Tracking GHGs and GHG mitigation activities
• Easy to use system with minimum technical background
• Capabilities of generating various types of reports
• Custom-made to regional and local needs

In the presentation we will introduce the system and will discuss how the system works, input data, emission calculation methodology and different applications of the system. Also, The results for Tehran city will be presented.

Using CrIS Ammonia Observations to Improve Decision Making on PM2.5 Control Policies

Nicholas Heath, Matthew Alvarado, Amy McVey, Karen Cady-Pereira, Jeana Mascio, Mark Shephard

Air quality managers and forecasters need accurate emissions estimates of PM2.5 precursors, such as ammonia (NH3), to analyze and forecast how these emissions impact human health and air quality. However, current emission inventories are too uncertain to provide reliable estimates of the health effects of NH3. Observations from the Cross-track Infrared Spectrometer (CrIS) provide an opportunity to address this problem and improve NH3 emissions estimates using inversion-based modeling techniques. Moreover, as new CrIS instruments are expected to be launched over the next two decades as part of the JPSS series, designing an infrastructure and methodology to use these observations in operational air quality policymaking and forecasting will provide benefits extended through 2030 and possibly beyond. In the current study, CrIS total-column NH3 observations are used in a finite-difference mass-balance approach to constrain NH3 emissions in the Community Multiscale Air Quality (CMAQ) model. CMAQ is run over the continental United States using 12 km grid spacing for June 2015. A baseline simulation is made with unperturbed NH3 emissions. Then, a second simulation is performed with NH3 emissions perturbed by 20%. The resulting total column concentrations of NH3 are compared to CrIS observations to derive a monthly-mean scaling factor for the a priori NH3 emissions. This scaling factor accounts for the relationship of NH3 concentrations to NH3 emissions in the baseline model run and is used to derive updated emissions, which are utilized in a final CMAQ simulation. This finite-difference inversion method has been incorporated into Amazon Web Services, and the data will be made publicly available. It will ultimately allow air quality managers and other stakeholders to obtain more accurate NH3 emissions estimates that can be implemented directly into their air quality modeling.

Sensitivity Analysis of PM2.5 Concentrations to the Emission Sources in the Megacity of Tehran

1-Farzaneh Taksibi, 2-Hossein Khajehpour, 3-Yadollah Saboohi

1-Master of Energy Systems Engineering, Tarasht Energy Science and Technology co., taksibi@energy.sharif.edu

2-Postdoctoral researcher, Energy Engineering Department, Sharif University of Technology, khajehpour@energy.sharif.edu

3- Professor of Energy Systems Engineering, Sharif University of Technology, saboohi@sharif.edu

There are various ways to estimates the share of different emission sources in air pollution. Emission inventory from different sources shows the share of sources in total primary emissions of pollutants during a specific time and in a limited geographical area. However, the observed concentration at specific point is a result of the dispersion of pollutants released from different sources of emissions and formation of secondary pollutants by chemical reactions in the atmosphere. Therefore, identifying the complete contribution of different sources to the concentration is possible either through source apportionment or through sensitivity analysis of validated dispersion models to emission sources. In this study, based on the sensitivity analysis of a validated PM2.5 dispersion model in Tehran using ADMS-Urabn, the contribution of different sources in air pollution concentration at receptor point in Tehran has been estimated. According to the results of the emission inventory, the share of various sources in primary PM2.5 emission rate in the Tehran megacity is 49%, 10%, 4% and 37%, for industries and power plants, domestic and commercial, agriculture, and mobile sources, respectively. However, based on sensitivity analysis study, the contribution of these sources to the concentration at the Sharif University Station, a concentrated residential area in the middle-western of Tehran, is estimated to be 19%, 34% and 47% from the emissions from industry and power plant, domestic and commercial, and mobile sources, respectively. The observed difference illustrates the functionality and necessity of the sensitivity analysis approach of emission sources in planning the environmental management of air pollution as a complement to emission inventory and source apportionment studies.

Enhanced Representation of Agricultural Emissions of Reactive Nitrogen in Fertilizer Emissions Scenario Tool for CMAQ (FEST-C)

Lina Luo and Daniel S. Cohan, Department of Civil and Environmental Engineering, Rice University, Houston, TX

Fertilizer-intensive agriculture has become the largest source of reactive nitrogen emissions in the United States. Almost half of added nitrogen on croplands is lost to the environment in multiple reactive forms, including the air pollutants ammonia (NH₃) and nitric oxide (NO), and a potent greenhouse gas, nitrous oxide (N₂O). Furthermore, NH₃ and NO are important precursors of secondary air pollutants, fine particulate matter (PM_2.5) and ozone (O₃). Integrated assessments show that agriculture is now the leading contributor to PM_2.5 pollution-related health impacts. Managing nitrogen emissions are thus essential to mitigate their negative effects on air quality and climate. Some farming practices such as fertilizer management could reduce emissions of all forms of reactive nitrogen species (Nr, including NH₃, NO, HONO, and N₂O), while others such as no-tillage could generate trade-offs among different species. Therefore, there is a need to not only predict Nr emissions but also illuminate how these emissions vary with different farming practices.

Currently, the process-based mechanistic nitrogen (N) scheme in the Community Multiscale Air Quality (CMAQ) model computes emissions as functions of meteorological conditions and soil properties (chemical and physical), and part of these soil properties are derived from agroecosystem modeling with the Fertilizer Emissions Scenario Tool for CMAQ (FEST-C). FEST-C is a regional-scale integrated ecosystem assessment model adapted from the field-scale Environmental Policy Integrated Climate (EPIC) model. EPIC simulates how soil properties and N-cycling processes are affected by farming practices and weather conditions. Obtaining soil properties from FEST-C enables Nr emissions to be modeled with spatio-temporal detail during the growing season. However, since only some of the required soil properties are extracted from FEST-C, CMAQ cannot dynamically reflect how soil properties and also Nr emissions change with farming practices. In this study, instead of linking agroecosystem (FEST-C) and air quality (CMAQ) models, we take an alternative method that directly uses FEST-C with modifications to its N-cycling schemes to generate Nr emissions. Because FEST-C neglects NO and nitrous acid (HONO) emissions from N-cycling and N₂O emissions from nitrification, we update its nitrification and denitrification schemes to predict NH₃, NO, HONO, and N₂O emissions adapted from mechanisms in the DayCENT model. To evaluate the performance of enhanced FEST-C modeling, we compare its estimates of agricultural Nr with those from a mechanistic soil N scheme and other existing schemes. Our enhanced FEST-C could not only consistently generate Nr emissions but also enables us to estimate how they vary with farming practices to identify practices which could yield co-benefits for air quality and climate.

Evaluating a Framework for Refining Ammonia Emissions Estimates with Satellite-based Observations with Regional Air Quality Modeling

Congmeng Lyu, Drexel University, Civil, Architectural, and Environmental Engineering, Philadelphia, Pennsylvania, USA

Shannon Capps, Drexel University, Civil, Architectural, and Environmental Engineering, Philadelphia, PA, USA

Matthew Lombardo, Johns Hopkins University, Baltimore, Maryland, USA

Mark Shephard, Environment and Climate Change Canada, Toronto, Ontario, Canada

Amir Hakami, Carleton University, Civil and Environmental Engineering, Ottawa, Ontario, Canada

Daven Henze, University of Colorado, Mechanical Engineering, Boulder, Colorado, USA

Steven Thomas, University of Melbourne, School of Earth Science, Melbourne, Victoria, Australia

Peter Rayner, University of Melbourne, School of Earth Science, Melbourne, Victoria, Australia

The Community Multiscale Air Quality (CMAQ) model calculates the impact of emission on atmospheric composition, including inorganic aerosols, while considering the transport and reactions of chemical constituents. Adjusting emissions by comparing modeled concentrations with observations is possible when the science processes are well understood as is the case for inorganic species such as ammonia (NH3). Four-dimensional variational data assimilation leverages differences in simulated and actual observations to revise estimates of emissions with spatial specificity. In this study, we evaluate the capacity of a CMAQ-based data assimilation system to improve NH3 emissions, which are relatively uncertain given the diversity of emissions processes in the agricultural sector. To do so, a Python-based four-dimensional variational framework (py4dvar) is integrated with CMAQ and its adjoint to constrain NH3 emissions with observations from the satellite-based Cross-track Infrared Sounder (CrIS). Pseudo-observation tests are conducted with the CrIS observation operator to evaluate the extent to which emissions are expected to be recovered with the assimilation. Then, the framework is ported to a 2017 modeling platform for assimilation of CrIS NH3 observations. Three suitable periods are selected from April through October 2017 for assimilation.

Inventory of multi-pollutant emissions from the predominant sectors in India: Road Transport and Thermal Power Plant

Poonam Mangaraj¹, Saroj Kumar Sahu¹, Swetaleena Dash¹, G Beig²

¹Utkal University, Bhubaneswar, India

²Indian Institute of Tropical Meteorology, Pune, India

(*Email: saroj.bot@utkaluniversity.ac.in)

The urge to understand the sources of emissions, particularly those from the traditionally dominant sectors is the initial gesture to meet the major requirements of regional air quality management. The up-surging demand for personal transport equipment has put forth the transport sector of India to witness a rapid transitional phase about several policy interventions related to emission norms. Along a similar line, the power sector too is of a high priority in the national planning processes in India where nearly 70 % of electric energy demand is met by the fossil fuel based thermal power plants. Emissions from road transport as well as thermal power stations have significant impacts on the regional as well as global climate change. Having said that, this study attempts to develop a very high resolution gridded ~ (10km X 10km) Emission Inventory (EI) to assess the load of PM_2.5, PM₁₀, CO, NO_x, VOC, SO₂, BC and OC emission from the above-mentioned two sectors over Indian sub-continent for the base year 2018. Emissions of 1522.1 Gg y^-1 PM_2.5, 1550.3 Gg y^-1 PM₁₀, 14864.5 Gg y^-1 CO, 11523.2 Gg y^-1 NO_x, 8802.6 Gg y^-1 VOC, 1896.7 Gg y^-1 SO₂, 673.1 Gg y^-1 BC and 1009.7 Gg y^-1 OC were estimated for on-road transport sector. Followed by the emissions of 378.8 Gg y^-1 PM_2.5, 1452.1 Gg y^-1 PM₁₀, 37.9 Gg y^-1 CO, 2045.5 Gg y^-1 NO_x, 0.7 Gg y^-1 VOC, 7386.5 Gg y^-1 SO₂, 18.9 Gg y^-1 BC and 0.6 Gg y^-1 OC were estimated for thermal power plants. This assessment would give an overview of emissions from road transport and power plants which can be critical sensitive input to atmospheric chemical transport models. Ultimately, this inventory could be used to ascertain the impacts on the atmospheric composition and air quality, on human health and environment, and on options for mitigation.

Emissions of condensable particulate matter from stationary combustion sources

Yu Morino,¹ Satoru Chatani,¹ Kiyoshi Tanabe,¹ Yuji Fujitani,¹ Katsuyuki Takahashi,² Kei Sato,¹

¹ National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba, Ibaraki, 305-8506, Japan

² Japan Environmental Sanitation Center, 10-6 Yotsuyakami-Cho, Kawasaki, Kanagawa, 210-0828, Japan

Emission factors of particulate matters (PM) from stationary combustion sources have been measured without dilution or cooling in Japan and other Asian countries, thus condensable PM were not included in the PM emission inventory. In our previous studies (Morino et al., ES&T, 2018), emissions of organic aerosol (OA) in condensable PM were estimated without any consideration of dependence of condensable PM emissions to temperature and total OA concentrations. In this study, we modified the emission data by considering the effect of temperature and total OA concentrations on the condensable PM emissions.

A Multi-Year Reanalysis of EPA's Fire Emissions Inventory

George Pouliot,Kristen Foley,James Beidler,Jeff Vukovich, Kirk Baker

The EPA estimates area burned and emissions from wildland fires, prescribed fires, grassland fires, and crop residue burning every three years for the National Emissions Inventory (NEI). For NEI and non-NEI years, different methods and datasets have been applied. To fully characterize the air quality (AQ) impacts of different fire types, a consistent method for inventories was developed and retrospectively applied. A multi-year reanalysis of fires from 2002 - 2017 was developed using the best available data and methods to estimate emission factors and fire activity. The new long term fire emissions estimates will be used in 2002-2017 CMAQ simulations for long-term AQ trend analysis and human health and ecological applications. A sensitivity analysis will be used to compare national approaches vs. the use of state and local data. Comparisons with the 2016v1 platform and the 2017 NEI will be presented.

Future scenarios in vehicular emissions and the impact of new vehicle technologies, fuel and mobility trends on air quality

Ribeiro, Flavia Noronha Dutra

Umezaki, Arissa Sary

Chiquetto, Julio Barboza

The present work estimates future trends of active fleet, emission factors, and vehicular ativity in the Metropolitan Region of Sao Paulo, Brazil. Then, several scenarios are developed to investigate the impact of changes in fuel and in modes of transportation and electrification. The emissions are calculated for 2028 and 2038, including criteria pollutants, greenhouse gases, and heat. Additionally, simulations using SMOKE and WRF-CHEM are performed to test some of the changes in emissions totals. Current trends point to an increase in greenhouse gases and heat emissions in Sao Paulo. Despite a decreasing tendency in most criteria pollutants, secondary pollutants may not follow the same trend. Additionally, the increased contribution to climate change is disturbing, particularly joined by an increase in vehicular anthropogenic heat, raising concerns of synergistic effects between global warming and urban heat island. Among the different scenarios tested, the measures that show better results in emission totals are increase in the use of biofuels, such as sugarcane ethanol, and decrease in private passenger cars activity. Simulations show the influence of atmospheric conditions on the pollutants concentrations.

Reactive Organic Carbon Emissions from Volatile Chemical Products

Karl Seltzer, Ben Murphy, Venkatesh Rao, Elyse Pennington, Madeleine Strum, Kristin Isaacs, Havala Pye

Volatile chemical products (VCPs) are an increasingly important source of anthropogenic reactive organic carbon (ROC) emissions. Among these sources are everyday items, such as personal care products, general cleaners, architectural coatings, pesticides, adhesives, and printing inks. These emissions have long been accounted for in the US EPA's National Emission Inventory (NEI) as the solvent sector, but new inventory methods have suggested the NEI could be biased low by factors of 2-3. As the influence of VCPs on secondarily formed pollutants grows in relevance, the need to resolve these differences becomes increasingly important. Here, we develop VCPy, a new framework to model ROC emissions from VCPs throughout the United States, with additional applications to spatially allocate these emissions to regional and local scales. In this framework, fate and transport assumptions are a function of the use timescale for product-use categories and the evaporation timescales of individual compounds within these categories, which are a function of their physiochemical properties. Since ingredients in these product categories are considered individually, determination of speciated emission profiles is explicit. This approach also enables quantification of emission volatility distributions and the abundance of different compound classes. Overall, we find National-level emissions of ROC from VCPy to be consistent with the NEI, but State and County-level estimates can differ substantially. In addition, we test the sensitivity of predicted emission factors to uncertain parameters, such as use and evaporation timescales, through Monte Carlo analysis, evaluate the inventory using published emission ratios, and map emissions to common chemical mechanisms for ease of research use in the chemical transport modelling community.

Urban Scale Inverse Framework Using Carbon Monoxide Total Column Retrievals from TROPOMI for Tehran: Preliminary Results

Nasimeh Shahrokhishahraki^{1, 2}, Peter J. Rayner^{1, 2}, Jeremy D. Silver¹, Steven Thomas¹, and Robyn Schofield^{1, 2}

1 School of Earth Sciences, The University of Melbourne

2 The Centre of Excellence for Climate Extremes (CLEX)

Improved air quality estimations rely on decreasing uncertainties in the modeling system and, principally, emission inventories (EIs), which contain spatiotemporal data about the emission sources and released ratio of pollutants. Improving EIs will improve air quality forecasts and provide a more qualified basis for policy. An inverse framework is applied in this study to derive a posteriori EIs using the comparison of observed and modeled concentration with a priori EIs in urban scale for Tehran. This study uses global datasets to prepare fine-resolution inventory data for an urban area. Emissions Database for Global Atmospheric Research (EDGAR), Gridded Population of the World, Night-time Lights Composite datasets, and Fossil Fuel Data Assimilation System are used to downscale the spatiotemporal resolution of global EIs to the finer scale. The resultant high-resolution inventory is applied to run the forward WRF-CMAQ Modeling System to simulate the concentrations of air pollutants. The TROPOspheric Monitoring Instrument (TROPOMI) data product and surface measurements are used to compare with the modeled concentration of CO. After comparison, an adjoint model is applied to generate gradients to provide directions for gradient-based optimization in four-dimensional variational data assimilation (4D-Var). The main goal of the inverse modeling framework is refining knowledge of the CO-EIs in the target urban area. Independent satellite and in-situ measurements are utilized to assess the inverse framework capability in improving CO-EIs.

Non-linear Joint Inversion of NOx and NMVOC Emissions Using Satellite Observations over East Asia

Amir H. Souri, Caroline R. Nowlan, Gonzalo Gonzalez Abad, Lei Zhu, Donald R. Blake, Alan Fried, Andrew J. Weinheimer, Jung-Hun Woo, Qiang Zhang, Christopher E. Chan Miller, Xiong Liu, and Kelly Chance

Ozone is a secondary pollutant that adversely affects both human health and crop yields. Concern over ozone and its precursors over East Asia has globally gained more attention, since this pollutant can spread hemispherically through the atmosphere, affecting the background levels in various places. Promisingly, Chinese governmental regulatory agencies have started taking action to reduce the magnitude of several ozone precursors since 2011-2012 by implementing selective catalytic reduction for thermal power plants under the clean air act. On the other hand, countries such as South Korea and Japan have undergone a hiatus in the reduction of NOx emission primarily due to increases in the number of diesel vehicles and thermal power plants. Unraveling the origin of ozone is complicated by a number of factors including the nonlinearity of ozone formation to sources, primarily from nitrogen oxides (NOx) and non-methane volatile organic compounds (NMVOC). Therefore, to be able to quantify the impact of recent emission policies in Asia on tropospheric ozone and oxidation capacity of the atmosphere we are required to provide a top-down estimate of emission inventories using well-characterized observations. For the first time, we provide a joint non-linear analytical estimate of NOx and NMVOC during the KORUS-AQ campaign by simultaneously incorporating both SAO's new product of HCHO columns from Ozone Mapping and Profiles Suite Nadir Mapper (OMPS-NM) and NASA's Ozone Monitoring Instrument (OMI) tropospheric NO2 columns into a regional chemical transport model (here CMAQ). Results demonstrate a promising improvement in the performance of the model in terms of HCHO and NO2 columns, which in turn, it enabled us to quantify the effect of the emission changes on different pathways of ozone formation and HOx chemistry.

Improving EPAs SPECIATE Program

Madeleine Strum, EPA, OAQPS; Marc Menetrez EPA ORD; George Pouliot, EPA ORD;  Art Diem, EPA OAQPS; Casey Bray, EPA OAQPS; Venkatesh Rao, EPA OAQPS; Julia Black, EPA OAQPS; Souad Benromdhane, EPA OAQPS; Heather Simon,  EPA OAQPS; Ben Murphy, EPA ORD; George Pouliot, EPA ORD; Havala Pye, EPA ORD;  Amara Holder, EPA ORD; Mike Hays, EPA ORD; Ingrid George,EPA, ORD;  Alison Eyth, EPA OAQPS; Libby Nessley,  EPA ORD; Justine Geidosch, EPA OTAQ;  Ying Hsu and Frank Divita (Abt Associates), B.H. Baek (UNC School of the Environment), Tejas Shah (Ramboll)

EPA continues to update SPECIATE, the U.S. Environmental Protection Agency's (EPA) repository of chemical speciation profiles of many types of air pollution sources. This presentation discusses the most recent improvements to the SPECIATE program, culminating in one of the quickest-ever releases of the updated version, SPECIATE 5.1, on July 20, 2020. It also reviews the profiles of importance for use emissions platforms used for air quality modeling and the EPA's National Emissions Inventory (NEI). The updated version, SPECIATE 5.1, completed about a year after SPECIATE 5.0, added 92 total profiles (16 organic gas, 18 particulate matter and 58 mercury profiles). These will assist the user community with PM, VOC, and mercury species characterization related, primarily, to oil and gas (VOC), fires (PM) and geothermal power (Hg).  The release also includes improvements to the database structure, species properties information and interface with the Speciation Tool.  Finally, we have completed updates to the Web Browser and added a browser's user's guide as well as a data developer's guide, providing the structure for how submitters can submit/prepare data for SPECIATE use. For the first time, our Data Developer's Guide/Template was used for community-developed data submission to SPECIATE. The SPECIATE Workgroup continues seek new data for future versions of SPECIATE. Priorities for profile sources have been presented in previous CMAS conferences and include region-specific oil and gas profiles for VOC, fires profiles for VOC and PM2.5, and residential wood combustion profiles for VOC. We will review these needs and identify other source categories for which updated speciation would improve our estimates of air toxics in the NEI.

Evaluating methane emissions using extended Kalman filter and Iterative Kalman smoother inversions with CMAQ: Application of OSSEs

Sina Voshtani, Richard Menard , Thomas Walker , Amir Hakami

Inverse modelling capabilities based on the extended Kalman filter (EKF) and the iterative Kalman smoother (IKS) approaches are developed with the Community Multiscale Air Quality (CMAQ) regional model at 36 km resolution. The analyses use observations from five satellite instruments (GOSAT, SCIAMACHY, IASI, AIRS, TES) to estimate methane emission corrections in three main source categories: agriculture, energy plus industry, and waste. First, we process methane anthropogenic emissions of EDGARv5 and map it into SMOKE, and second, we modify CMAQ to account for methane. Our study addresses the spatial distribution of prior and posterior methane emissions, treated as a tracer in CMAQ, and the impact of different types of observations. To show that the methods can work properly, we conduct an observing system simulation experiments (OSSEs) on a realistic configuration of atmospheric inversion that maintains the average kernel of the corresponding satellite instruments. OSSEs are carried out with different methane prior emissions and flux distributions. Both approaches show a degree of robustness to account for the estimation of the emissions and their bias and uncertainties, as opposed to the classical variational method with a frozen error statistic. Still, the method requires aggregation of source regions and some expert knowledge to define such aggregations.

The results and performance of both methods are assessed. They both prove to reproduce the "truth" in most of the cases. IKS provides a faster rate of error variance reduction compared to EKF, partly due to a larger amount of measurements assimilated through a forward-backward integration of the inverse model. EKF, on the other hand, promises less computational cost. We further discuss the objective of using real observations and deducing realistic statistics for the emission, transport, and satellite observation errors.