Publication

Abstract : This study addresses the question of how skillful regional air quality modeling can be when using downscaled globally-available emission inventories. This paper applies global datasets to prepare fine-resolution priori emission inventories for an urban area. Emissions Database for Global Atmospheric Research (EDGAR), Gridded Population of the World, and output from the Fossil Fuel Data Assimilation System are used to downscale the spatiotemporal resolution of global emission inventories to the finer scale. The resultant high-resolution inventory is taken as input for an off-line run of the Community Multi-Scale Air Quality (CMAQ) modeling system to simulate the concentrations of air pollutants in Tehran during August 2018, November 2018, February 2019, and May 2019. These runs are forced with meteorology from the Weather Research and Forecasting (WRF) model. Retrievals of atmospheric composition from the TROPOspheric Monitoring Instrument (TROPOMI) and surface measurements (split into ‘road’ or ‘city’ type stations) are used to compare with the modeled concentrations of NO2, CO, and O3 to assess the capability of the applied modeling framework and emission inventories in concentration estimations. Comparison of modeled NO2 concentration with NO2 tropospheric column shows that the model captures the spatial and temporal distribution with Pearson correlation above 0.7 and 0.6, respectively and absolute bias under 0.08 ppb. The offline WRF-CMAQ simulations overestimate surface measurements of NO2 and CO and underestimate O3. The model captures the diurnal variations for NO2, CO, and O3 with a correlation between 0.7 and 0.91, 0.6–0.8, and 0.5–0.98 and absolute bias less than 69 ppb, 1.15 ppm, and 12.5 ppb, respectively. The overall performance of the observing and modeling system is sufficient for a credible inversion of surface emissions, which is the intended purpose of this modeling set-up.