Estimating the air pollution impact of the response to the COVID-19 pandemic in California using the surface monitoring network

The COVID-19 pandemic and ensuing response has led to a significant reduction in activity in California, especially for several weeks after the stay-at-home executive order on March 19^th, 2020. We explore the change in atmospheric pollutants (NO₂, CO, particulate matter, and O₃) during the shutdown, relative to previous years, in different air basins using California's extensive network of surface monitoring sites. Statewide, NO₂, CO, and particulate matter are all approximately 25% below the historical average between March 19^th and July 1^st. We find that comparisons to previous years can be problematic when isolating the impacts of the pandemic response because of significant interannual variability in meteorology, especially in Spring. A statistical model trained on meteorological variables in the South Coast basin of California predicts daily average NO₂ concentrations surprisingly well (r² > 0.7 for January - July) when tested on different years between 2015-2020. However, in 2020 alone there is a persistent high bias in the modeled NO₂ after March 19^th relative to the observations. Using the historical relationship between vehicle miles traveled (VMT) and NO₂ concentrations as an additional input to the model reduces the bias in 2020 and improves the correlation with observations. We use this model to estimate the impact of the COVID-19 pandemic response on NO₂ concentrations in the South Coast basin.