Session Presentations

Marco Rodriguez (Ramboll US Consulting Inc)

Yuge Shi (Ramboll US Consulting Inc)

Courtney Taylor (Ramboll US Consulting Inc)

Ralph Morris (Ramboll US Consulting Inc)

Photochemical Grid Models (PGMs) are capable to assess the chemical formation and transport of secondarily formed pollutants, and this makes them suitable for use in a wide variety of regulatory and research purposes. PGM can produce vast amounts of data in the form of gridded air quality concentrations for multiple pollutants over a defined three-dimensional modeling domain. An advantage of data generated with PGMs is that enables scientists and regulators to have a more complete understanding of pollutant spatial variability than what is possible through monitoring data alone. However, the wealth of detailed information generated by PGMs is not commonly available to the broader public due to logistical constraints associated with transmitting, hosting, and visualizing large datasets. As the public and elected officials increasingly emphasize the importance of government transparency and equal access to clean air and water, so too our tools need to be upgraded and re-envisioned to serve these broader needs. PGMs data can aid regulatory agencies and communities make informed decisions for areas most in need of enhanced air monitoring funding to further Environmental Justice initiatives recently announced by the United States Environmental Protection Agency. In this presentation, we detail a method to develop an air quality screening tool based on local PGM results in combination with equity indicators to assist local and state government agencies identify areas most likely to have disproportionate air quality burdens that are a high priority for further monitoring efforts. This approach presents distinct advantages relative to federal screening tools, such as EJSCREEN, because is based on best available local data, rather than national defaults, and it exclusively uses air quality indicators, rather than have results confounded by other environmental variables. Further, the screening tool can be readily shared with the public to engage community members in meaningful and well-informed dialogue regarding potential air pollutant burdens as an important step to increase transparency, communication, and trust with historically underserved communities. Since many states and regions across the U.S. have access to PGM results from past analyses, leveraging local PGM data to assess environmental justice could provide a more robust method to prioritize future efforts and support informed community engagement.

Michael Breen, Vlad Isakov, Catherine Seppanen, Sarav Arunachalam, Miyuki Breen, Steven Prince, Thomas Long, David Heist, Parikshit Deshmukh, Wyat Appel, Christian Hogrefe, Benjamin Murphy, Chris Nolte, George Pouliot, Havala Pye, Jacky Rosati

Epidemiologic studies of ambient fine particulate matter (PM_2.5) and ozone (O₃) often use outdoor concentrations as exposure surrogates, which can induce measurement error. We developed an exposure model called TracMyAir, which is an iPhone application that determines eight tiers of individual-level exposure metrics in real-time for ambient PM_2.5 and O₃ using outdoor concentrations, home building characteristics, weather, time-locations, and time-activities. In this study, we extended TracMyAir by including (1) outdoor concentrations from an air quality model (CMAQ) and air monitoring network (OpenAQ), (2) indoor and outdoor PM_2.5 concentrations from low-cost air sensors (PurpleAir), (3) microenvironment model (MicroTrac) based on time-resolved smartphone geolocations, and (4) inhaled ventilation models based on physical activity data from smartphone and smartwatch accelerometers and heart rate sensors. The exposure metrics for the extended TracMyAir include: residential air exchange rates (AER, Tier 1), infiltration factors (F_{inf_home}, Tier 2), indoor concentrations (C_{in_home}, Tier 3); and personal outdoor concentrations (C_out, Tier 4), time spent in ME (T_ME, Tier 5), exposure factors (F_PEX, Tier 6), exposures (E, Tier 7), and inhaled doses (D, Tier 8). We applied TracMyAir to determine hourly PM_2.5 and O3 exposure metrics for two panel studies with nine participants living in central North Carolina: one study with 216 participant hours during November 2019, and another study with 648 participant-hours during September-October 2020. The TracMyAir predictions showed considerable temporal and house-to-house variability of AER, F_{inf_home}, and C_{in_home} (Tiers 1-3), and person-to-person variability of C_out, T_ME, F_PEX, E, and D (Tiers 4-8). Our study demonstrates the capability of extending TracMyAir with air quality and time-activity sensors and models to determine individual-level ambient PM_2.5 and O₃ exposure metrics, in support of epidemiologic studies and public health strategies to help individuals reduce their exposures to ambient air pollutants.

Shih Ying “Changsy” Chang¹, Jiaoyan Huang¹, Marcus Alexander², Doug Eisinger¹, Nathan Pavlovic¹, and Eladio Knipping²

¹ Sonoma Technology, Petaluma, CA

² Electric Power Research Institute, Palo Alto, CA

As substantial progress is made on improving regional air quality, there is increased attention on addressing the ongoing challenge of reducing pollution exposure for individuals in disadvantaged communities living near heavily traveled roads such as major freeways and truck routes. Accelerated penetration of electric vehicles (EVs) into the vehicle fleet offers potential air quality benefits for near-road communities, but these benefits, to our knowledge, have not been previously quantified at a community scale. This case study aims to quantify the air quality benefit of EV penetration in environmental justice (EJ) communities in the Southern California Interstate 710 Corridor using a fine resolution modeling framework. Four future-year EV penetration scenarios in 2040 will be evaluated. These scenarios include Scenario A, business as usual (BAU); Scenario B, high EV penetration; Scenario C, medium EV penetration; and Scenario D, EV penetration for medium- and heavy-duty vehicle. The penetration of light-duty EV for each scenario is modeled with the Market Acceptance of Advanced Automotive Technologies (MA³T) model. The penetration of medium- and heavy-duty EV is based on policy analysis by the California Air Resource Board. The estimated EV penetration is used to adjust the emission factors modeled with the EMission FACtor mobile source emission model (EMFAC2017) to generate emission factors for each scenario. These emission factors will be combined with detailed travel activity data to estimate roadway emissions within a modeling domain covering Long Beach, California, and surrounding communities. These emissions will then be used by the Research LINE source dispersion model (RLINE) to estimate NO_x and PM_2.5 concentrations at a census block group level for both EJ and non-EJ communities in the modeling domain. The comparison between the four scenarios will provide information on how accelerated EV fleet penetration can close the gap in air quality between EJ and non-EJ communities.

Gavendra Pandey1, Chowdhury Moniruzzaman1, Saravanan Arunachalam1, and Akula Venkatram2

1Institute for the Environment, University of North Carolina at Chapel Hill 2University of California at Riverside

Aircraft emissions from an airport are a significant source of total emissions that have an impact on air quality in the airport vicinity. These sources are unique due to the transient nature of the emissions from each source, as well as the buoyant exhaust. These sources emit the pollutants in short bursts especially during landing and takeoff operations (LTO). It is difficult to quantify these short bursts emissions and model the governing processes. An added complexity occurs when the wind speed is low and variable. In this condition, winds can go in several directions during an hour, resulting in multiple concentration “lobes”, and large plume spread. The classical steady-state model such as AERMOD does not account for the meandering effect and short bursts of aircraft emissions due to the hourly nature of inputs and outputs in typical applications. To account these features that characterize the dispersion of aircraft emissions, a sensitivity analysis based on the sub-hourly approach is being described in this paper. We used the SO2 concentration measurements from the LAX Air Quality Source Apportionment Study (AQSAS) conducted at the Los Angeles International Airport (LAX) on four core locations named AQ (Air Quality), CN (Community North), CS (Community South), and CE (Community East) for 42 days period during February and March 2012. This paper focuses on the effect of the sub-hourly approach on dispersion. We used AERMINUTEplus, the enhanced version of the AERMET preprocessor AERMINUTE that outputs sub-hourly wind averages. The output from AERMINUTEplus is used in AERMET, and then AERMOD is run multiple times for each portion of an hour, and output concentration files are simply averaged on that hour using a previously developed approach called SHARP (Sub-Hourly AERMOD Run Procedure). Hourly AERMOD simulations produced concentrations that were significantly greater than those at AQ and CS and significantly lower than those at CN and CE. The fractional bias based on top 26 robust highest concentrations improved from -1.50 and -1.37 to -0.91 and -0.79 respectively at sites AQ and CN sites, and showed little change at sites CS and CE. The ratio of medians of the observed to the modeled concentrations is improved from 3.6 to 2.2 at the CS site. In addition, the model performance improved substantially by replacing meteorological parameters such as friction velocity and Monin-Obukhov (M-O) length during stable and unstable conditions with those corresponding to neutral conditions. The use of the SHARP approach magnifies the mid to lower range concentrations and lower concentrations nearly match the observations. The fraction of model estimates within a factor of two of the observations improved from 32 to 55% at the CN site and at the CS site, by up to 46%. Overall, our sub-hourly modeling results using SHARP are relatively closer to observations and demonstrate that this alternate approach should be seriously explored in circumstances when low-wind meander conditions predominate and for treating aircraft sources.

Abdullah Mahmud, Pingkuan Di, and Jeremy Avise

The Community Multi-scale Air Quality Model version 5.3.2 (CMAQv5.3.2) was utilized to simulate toxic Volatile Organic Compounds (VOCs) in the atmosphere for 2017 within major air basins in Southern California including the South Coast, San Diego, and Imperial. Formaldehyde, acetaldehyde, benzene, 1,3-butadiene, perchloroethylene, and p-dichloroethylebezene are some of the VOCs that have been identified as human carcinogens and toxic air contaminants (TACs) in California, which pose significant health risks through inhalation exposure. In this study, the annual average and population weighted concentrations of the toxic VOCs were calculated at a 2-km spatial resolution and at aggregated to population census tract levels within the model domain. The annual average concentrations were then applied to calculate cancer risks and burden with an objective to understand the distributions of toxic VOCs within southern California and to identify disadvantaged communities that are likely to experience higher pollution burden as required by the California community air protection Assembly Bill 617 (AB 617). In addition, the modeling results will be used to quantify the contribution of each toxic VOC to overall impact in Southern California. For formaldehyde and acetaldehyde, their relative contributions from primary (direct emissions) and secondary formed from atmospheric chemistry will be specified. The results of source attribution can be used to evaluate the benefits of emission reduction efforts the state has taken and to direct the efforts to further reducing toxic VOCs. The 2017 simulations were driven by the meteorological fields generated by the Weather Research Forecast version 3.9.1 (WRFv3.9.1) model. The gridded and speciated sector specific emissions were combined by the SMOKE emissions processing system. The air quality simulations were carried out separately using two emissions scenarios: 1) emissions that include sources within the US border only and 2) emissions within the US plus the sources adjacent to the southern border in Mexican side within the modeling domain. This was done to portray air quality as accurately as possible, particularly, within the communities that are near to the US-Mexican border. The boundary conditions for the 2-km model runs were generated from the state-wide 12-km model output. The model performance analysis was carried out using available data including satellite measurements and ground based monitored data for toxic VOCs, fine particulate matter (PM2.5), and ozone.

Shuming Du, Sue Chen, Karry Liu, Zhen Liu, Abdullah Mahmud, Wenli Yang, Melissa Venecek, Roger Kwok, Pingkuan Di, and Jeremy Avise

Diesel particulate matter (DPM) has been identified as a human carcinogen and toxic air contaminant (TAC) by many jurisdictions. In the past two decades, tremendous efforts have been made to reduce exposure of DPM to the public in California. However, since DPM cannot be directly measured, DPM concentrations and associated health risks need to be assessed using air quality models. Some existing studies unitized grid-based photochemical models such as CMAQ or CAMx to calculate DPM concentrations. Since most DPM emissions are emitted from on-road and off-road mobile diesel engines (such as trucks, locomotives, etc.), aggregated gridded emissions treated by CMAQ or CAMx may not fully reflect/represent actual emission source characterization. As such, accurate source-receptor relationships cannot be established, especially at a local/community scale. In this study, we utilize an air dispersion model - CALPUFF to simulate DPM concentrations for each emission source and category in 2017. Specifically, on-road mobile sources and some off-road mobile sources (such as OGV, locomotives, airport takeoff and landing, etc.) are characterized as individual line sources, major stationary sources are treated as individual point sources, and the remaining DPM emissions are simulated as area sources with fine resolution. DPM concentrations are calculated at every centroid of the census blocks and hundreds of explicit locations of interests, such as schools, daycare centers, elder centers, etc. As a comparison, CMAQ modeling is also conducted using the same emission inventory and meteorology but with different emission treatments (2km x 2km gridded). The similarities and differences of the modeling results between CALPUFF and CMAQ will be discussed. Both models’ performance will be conducted by comparing to available observations. Elemental carbon (EC) is chosen as a surrogate of DPM because DPM cannot be measured directly, and it is understood that EC is likely to have a high correlation with DPM.

Dianna M. Francisco¹, David K. Heist², Akula Venkatram³, Lydia H. Brouwer⁴, and Steven G. Perry²

¹ORAU ORISE Research Participation Program hosted at the U.S. EPA, RTP, NC, USA

²U.S. EPA, ORD, CEMM, RTP, NC, USA

³U.C. Riverside, Riverside, CA, USA

⁴Jacobs Technology Inc., RTP, NC, USA

Noise barriers have been proposed as a mitigation measure to reduce human exposure to traffic-related air pollutants. This paper presents new results from wind tunnel studies examining the effect of solid roadside barriers of varying heights (from 3 m to 9 m) on concentrations downwind of a roadway. These wind tunnel results are used to further develop a mixed-wake algorithm that accounts for barrier effects within a steady-state air dispersion model. The mixed-wake algorithm combines two flow regimes, one for flow above the barrier and one for the recirculation region that develops downwind of the barrier. The regime representing the recirculation region exhibits a uniform vertical concentration profile while the concentrations in the upper regime rise to a peak above the barrier before decreasing with height. Observations of the barrier-induced flow field and concentration patterns, and the mixed-wake algorithm implemented into the air dispersion model will be described. The algorithm was implemented in the U.S. EPA regulatory air dispersion model, AERMOD, using the non-regulatory line-source option, RLINEXT. Both the wind tunnel and mixed-wake algorithm show that relative to a roadway without a barrier, a solid barrier substantially reduces concentrations immediately downwind of the barrier with the greatest reductions occurring with the tallest barriers. Model performance will be evaluated using geometric mean (MG), geometric standard deviation (SG), and the fraction of model results that are within a factor of two of the observations (FAC2).

Chowdhury Moniruzzaman, Gavendra Pandey, and Saravanan Arunachalam

Aircraft’s landing and take-off (LTO) emissions contribute to surface air quality in and around an airport. Dispersion modeling helps to quantify the effects of aircraft emissions on surface air quality. We present results of two tasks: 1) effects of airborne aircraft emissions on surface air quality and 2) effects of varying number of sources in dispersion modeling. Although there are many studies on the effects of aircraft’s surface emission on surface air quality, effects of airborne emissions on surface air quality in dispersion modeling remains poorly understood. In the first task, AERMOD modeling was performed for the Los Angeles International Airport (LAX) for SO2 for 1 month in February, 2012 for 2 cases: emissions with and without air-sources. The difference between the 2 cases gave the air emission's contribution to surface concentration which were found to be 3 to 6% at 3 monitors (AQ, CN and CS) and 13% at another monitor (CE) where all 4 monitors were within 2 km of the runway. The second task was motivated by a research need to reduce the number of sources in the Federal Aviation Administration’s (FAA’s) Aviation Environmental Design Tool (AEDT) produced AERMOD hourly emission file. Both the surface (idle, taxi and takeoff) and airborne sources were reduced by accumulating the emissions in an hour in a smaller number of sources based on their locations by developing a python-based emission processor which uses raw AEDT flight segment data as input. AEDT-segment surface emission data were extracted for 2 source characterizations: 1) 31-sources having 16 runway sources for landing and take-off in both direction with highly simplified 15 non-runway sources and 2) 138-sources having same 16 runway sources of 31-source case but having spatially detailed 122 non-runway sources having separate detailed taxiway, taxi-ramp, terminal and gate sources. AERMOD modeling was performed for the LAX airport for SO2 for Feb 1, 2012 for 3 surface source number sensitivity cases: 1) 1440, 2) 138 and 3) 31 surface sources. Based on 1 day’s results for the 138-source run, the simulation found that 90% surface source reduction (from 1440 to 138) decreased computation time by 90% and changed the mean absolute error (MAE) by +8%, -10%, +1% at 3 sites (AQ, CN, CE) and +98% at 1 site (CS). For the 31-source run for 1 day’s results, the simulation found that 98% surface source reduction (from 1440 to 31) decreased computation time by 95% and changed the MAE by -5%, -0.06%, -0.8% at 3 sites (AQ, CN, CE) and +85% at 1 site (CS). As airborne source’s contribution is found to be 3-6% at 3 of the 4 sites, the number of airborne sources can be reduced significantly in AEDT for AERMOD modeling. The treatment of surface sources can also be reduced significantly to gain computational efficiency without affecting model performance in AERMOD predictions.

Chi-Tsan Wang, William Vizuete, Marc Serre, Richard Strott, Jaime Green, Bok Haeng Baek*

These ambient toxicants, Styrene, Benzene, Toluene, Ethylbenzene, and Xylenes (SBTEX) , are emitted from the petrochemical industry, combustion process, transport emission, and solvent usage sources. There have been several SBTEX toxic assessment studies, but they mainly relied on the ambient measurements using GC-MS (Gas Chromatography-Mass Spectrometer)/FID (Flame Ionization Detector) combined with the statistical models or dispersion models to understand their behaviors in the atmosphere. However, the measurements are limited to specific locations only during the short monitoring period. To overcome the spatial and temporal coverage limitation, we have developed the long-term air quality modeling system over the U.S. gulf region using the state-of-science chemical transport model, Comprehensive Air Quality Model with Extensions (CAMx), the Sparse Matrix Operator Kerner Emissions (SMOKE), and the Weather Research and Forecasting (WRF) modeling systems. Our study will simulate the spatially and temporally enhanced SBTEX ambient concentrations over the U.S. gulf region over six years from 2011 to 2016. Because these individual SBTEX toxicants are not considered as the chemical model species with the complete chemical reactions in the atmosphere, we utilized the unique post-analysis feature called “Reactive Tracer” in the CAMx modeling system to simulate the ambient SBTEX concentrations. The Reactive Tracer applies the major oxidants concentration result (O₃, OH, and NO₃) to calculate the chemical reduction rate of SBTEX. Therefore, we can compare our SBTEX modeling outcomes against the Texas Commission on Environmental Quality (TCEQ) State Implementation Plan (SIP) air quality model results, and Ambient Monitoring Technology Information Center (AMTIC) data, to evaluate the model performance. The result of this study will be used in the human exposure assessment and epidemiological cohort study. In our preliminary result of the 2011 total annual SBTEX emission, the largest contribution section is the on-road transportation emission (32.5%), and the second largest is point fire emission (29.8%).

Margret Velizarova¹, Reneta Dimitrova ^1,2 and Angel Burov³

¹Department of Meteorology and Geophysics, Faculty of Physics, Sofia University “St. K. Ohridski”,5 James Bourchier Blvd., 1164 Sofia, Bulgaria; margretv@phys.uni-sofia.bg

²National Institute of Geophysics, Geodesy and Geography—Bulgarian Academy of Sciences, Acad. Georgi Bonchev Str., Bl. 3, 1113 Sofia, Bulgaria;

³University of Architecture Civil Engineering and Geodesy, Faculty of Architecture, Urban planning department, Hristo Smirnenski Bld., 1164 Sofia, Bulgaria

The urban area is a unique system, which includes all aspects of the natural environment, which was subsequently modified by the introduction of a build environment (so-called physical environment). Despite of all air quality studies conducted in Bulgaria during the last decades, the effect caused by the physical environment on the urban pollution hasn’t been studied till this day.

     The numerical experiments were conducted with the Air Quality Management & Assessment System: ADMS-Urban, which is a comprehensive system for air quality modeling in urbanized areas, and it is developed and supported by the private organization Cambridge Environmental Research Consultants (CERC). The main goal of this study is to present the spatial and temporal concentration distributions of atmospheric pollutants in street canyons, typical for Sofia city. In addition, the impact of the street canyon parameters on the pollutant concentrations from transport sources, was evaluated.

     The utilization of detailed spatial representation of the urban infrastructure (as 3D maps) can considerably affect the air quality modelling. It can describe the pollution field structure and help with locating the ‘hot spots’ of pollution, and by this way to secure valuable tools in favor of urban planning, design and management, in terms of creating regulations and guidance for future urban development projects, public works and activities in the urban environment.

Clarissa May, Robson Will, Vinicius Fin, Ana Paula Stein Santos, Otávio dos Santos, Leonardo Hoinaski

This study evaluated the influence of temperature and humidity data on the health risk associated with particulate matter (PM₁₀) in two cities, Camboriú and Florianópolis, in the state of Santa Catarina, Brazil. We use data from hospital admissions for cardiorespiratory diseases and PM₁₀ concentration data collected with a high volume sampler (AGV-MP₁₀). We calculate the relative risk of cardiorespiratory diseases related to air PM₁₀ in children under five years of age and adults over forty, due to their greater vulnerability. The health costs caused by these diseases associated with particulate matter are estimated by applying the Disability Adjusted Life Years (DALY) methodology. The highest calculated risks are found in the city of Florianópolis. In both cities, the annual risk variation is not significant, as well as the costs. The air pollution cost related to PM10 was larger in Florianópolis, whose population is about seven times larger than that of Camboriú. Analyzing the seasonally, there are some cases of significant variations in both cities, however, following different patterns. Children are the most affected as they have higher rates of variation. Comparing meteorological data from the respective cities, it was possible to observe that the humidity data had little relationship with the risk data. Temperature and humidity are inversely correlated to the risks.