Evaluation and Application of Big Data Derived Datasets for High Resolution Updates to Transport Emissions in the Wake of COVID-19

Abstract:

The COVID-19 pandemic and ensuing shutdown of many US States has resulted in rapid changes to motor vehicle traffic and the emissions associated with this sector. This presents a significant challenge for atmospheric chemistry modelling and air quality forecasting during this time period, in that transportation emission inventories need to be updated more rapidly. Additionally, some of the traditionally well parameterized aspects of transportation such as diurnal and day-of-week trends may no longer be applicable during the pandemic versus business-as-usual. Most transportation emission inventories take years to develop for chemical transport models and are only updated every few years. Here, we update the previously developed Fuel-based Inventory of Vehicle Emissions (FIVE) to account for changes in traffic emissions due to COVID-19. We first construct a 2018 baseline inventory and adjust the emissions for 2020 based on weekly fuel sales information, and we also assess the utility and potential for integrating big data products (e.g., Streetlight Data, Apple Mobility Index) with embedded traffic monitoring sites in a variety of cities located throughout the US. The relationship between these mobility datasets and embedded traffic counters is used to determine scaling factors for adjusting traffic emissions during COVID-19. Changes in day-of-week and diurnal patterns of light- and heavy-duty vehicular traffic are estimated quantitatively. Overall, there was ~50% drop in on-road NO_x emissions in April 2020 and ~25% drop in June 2020 versus business-as-usual. Beyond the applicability to chemical transport modelling and forecasting for 2020, this work also represents a methodology for quickly updating transportation inventories over the continental US.