PhD Defense by Xinxin Zhai

Primary tabs

School of Civil and Environmental Engineering   Ph.D. Thesis Defense Announcement Spatial-temporal Air Quality Modeling Using Multiple Modeling Techniques at High Resolutions   By Xinxin Zhai   Advisors: Dr. James A. Mulholland and Dr. Armistead G. Russell   Committee Members: Dr. Rodney Weber (EAS) , Dr. Yongtao Hu (CEE) , Dr. Heather Holmes (UNR)   Date & Time: Monday, October 30th, 2017, 11:00am-12:00pm Location: Mason Conference Room 2119 Ambient air quality is found to be associated with human mortality and morbidity. To better estimate the human exposure to air pollution, multiple air quality modeling techniques are developed and employed to increase the accuracy and resolutions of exposure information including source impacts and pollutant concentrations. First, Chemical Mass Balance (CMB) receptor model based on observational data is applied to investigate source impacts in the State of Georgia. The results show that over 2002 to 2013, the secondary sulfate and nitrate species decreased by 58% and 44%, respectively; total mobile source impacts decreased more at urban sites (39%) than rural sites (23%); biomass burning impacts decreased more at rural sites (34%) than urban sites (27%). Second, to understand mobile source impacts on PM2.5 at finer spatially and temporally scales, we developed an approach using EC, CO, and NOx measurements as indicators of mobile source impacts based on an integrated mobile source indicator (IMSI) method. The generated total mobile and separate vehicle source impacts agree well with daily CMB results for 2002 to 2010 in Georgia, with high temporal correlations and low biases. Third, to estimate city-level source and pollutant exposure information, a procedure is developed that generates observation-calibrated hourly concentrations of PM2.5, CO, and NOx from mobile sources using RLINE at 250m resolution in the 20-county Atlanta area. The results show that RLINE overestimated the annual averages of CO and NOx daily 1-hour maximum concentrations by factors of 1.3 and 4.2 on average, respectively, and PM2.5 mobile source impacts by a factor of 1.8 compared with estimates by CMB with gas constraints. Based on observational data, we calibrated the RLINE estimates of CO, NOx, and PM2.5 emitted by mobile sources from 2002 to 2011 at multiple sites. The calibration largely reduced modeling biases. Finally, to further estimate the near-road exposure variation in the trafficked and populated areas of cities, near-road spatial gradients were characterized for NOx, CO, PM2.5, and eight volatile organic compounds (VOCs) using passive sampler detectors (PSD) and dispersion model RLINE in Atlanta, GA. We found that the spatial gradients show a decrease of up to 3.1 fold from highway adjacent areas (<100m away from highway) to remote areas (>1500m away from highway) for all pollutants in the PSD measurements and up to 4.2 fold in RLINE estimates. Comparison of NOx by the two methods show good correlations, indicating RLINE captures the spatial variations well but overestimate concentrations. After calibrating RLINE estimation using the PSD measurements, the near-road spatial gradients of NOx were well captured. Overall, this research developed and evaluated multiple modeling techniques that simulate concentrations and source impacts at different temporal resolutions and spatial scales with a focus on the mobile sources. These approaches help improve the estimation of fine-scale concentration fields by calibrating modeling results to observational levels with better spatial and temporal coverage. The methods are being applied in studies in different areas and years, and the generated concentration fields have helped evaluating the health impacts of air pollution on the cohort in Georgia.  


  • Workflow Status: Published
  • Created By: Tatianna Richardson
  • Created: 10/18/2017
  • Modified By: Tatianna Richardson
  • Modified: 10/18/2017