Optimizing CO emissions from the 2018 Californian fires using S5P – an inverse modelling study

A clear understanding of carbon monoxide (CO) emissions is important at various scales. On the local scale CO is toxic to living organisms, and on the global scale CO plays in role in the budget of  the hydroxyl radical (OH). OH, in turn, is important for the oxidizing capacity of the atmosphere. Additionally, CO is a precursor of the greenhouse gases ozone and carbon dioxide, hence CO influences also climate on a global scale.

Approximately one quarter of the global atmospheric CO load emanates from wildfires. However, these emissions are sometimes underrepresented in the emission datasets. Among the reasons for this discrepancy are clouds and smoke plumes hampering observations of land cover and active fires and uncertainties in emission factors. These issues are less relevant for top-down approaches like inverse modeling, which allow tracing back an atmospheric signal to its source even if it is only observed days after emission.

In this study, we attempt to improve the emission estimates of an existing inventory by applying an inverse modeling approach to the CO emissions of the California wildfires in 2018, that devastated more than 7500 square kilometers of forested and residential area. More specifically, we used the Fire Emission Inventory from NCAR (FINN) together with the CO observations from the TROPOMI instrument onboard the Sentinel 5 Precursor (S5P) satellite and the TM5-4dvar inverse model. The high resolution of the TROPOMI observations enables better spatial constraints compared to previous instruments. Preliminary results suggest significant positive emission increments compared to FINN.