Finescale Air Quality Modeling over the Denver Area: Model Evaluation and Sensitivity Simulations

High personal exposure to air pollution and associated health impacts have been one of key issues introduced by the environmental injustice. Fine-scale monitoring or air quality model data are required for an accurate assessment of personal exposure using exposure models. The main objective of this work is to conduct fine-scale air quality simulations using the state-of-science online-coupled Weather Research and Forecast Model with chemistry (WRF/Chem) over an environmental justice community in the Denver area during the summer of 2017, where the sensor-based ambient monitoring data have been collected over multiple sites in the same area by RTI international. The predicted O₃ and PM_2.5 concentrations will be used to estimate personal exposure to ambient air pollution. A one-month triple-nested simulation using WRF/Chemv3.9 has been previously performed for August, 2017, with the 36-km domain covering the continental U.S., 6-km covering the state of Colorado, and 1-km covering the Denver Metropolitan area. The evaluation of this initial application identified some model biases such as relatively large underprediction of 8-hr O₃ against the AQS data and large overprediction of PM_2.5 against the RTI sensor-based monitoring data at the 1-km grid resolution, due in part to inaccurate model inputs. This study is aiming to improve the model performance for O₃ and PM_2.5, especially over the 1-km domain by adjusting boundary conditions for O₃, improving wildfire emissions, and reducing anthropogenic emissions for gas and aerosol precursors based on comparisons between National Emissions Inventory 2011 and 2014. Multiple one-month sensitivity simulations have been conducted to assess the impacts of the improved model inputs. The preliminary evaluation shows noticeable model improvements for 8-hr O₃ and PM_2.5 in terms of domain-mean biases. The optimal model configurations will be determined based on additional sensitivity simulations and more comprehensive evaluation for improving O₃ and PM_2.5 predictions at 1-km resolution for personal exposure assessment.