Assimilation of the Global Fire Emissions Database (GFED) burned area dataset into Community Land Model version 5: Its impacts on the carbon and water fluxes in high latitudes

In the high latitudes such as boreal forest and tundra, the wildfire intensity and occurrence have been increased over the past decades. In particular, unprecedented large fires (more than 1.5 Mha of burned area) in interior Alaska were reported in 2004 and 2015. These fires not only caused carbon emission from vegetation but also increase the soil temperature in summer, which could induce the permafrost thaw. However, the most of Earth System Models (ESMs) and Land Surface Models (LSMs) are still limited in representing the fire processes and thus simulating the arctic fire occurrences and their impacts on the land surface processes. In this study, we assimilated the 0.5 degree daily burned area derived from Global Fire Emissions Database (GFED) for twelve years (20012012) over the arctic region into the NCAR Community Land Model version 5 (CLM5) with a biogeochemistry module (CLM5-BGC), one of widely used LSMs, by employing the direct insertion method. In CLM5-BGC, the burned area is predicted based on the empirical relationships among lightning frequency, human population density, and vegetation composition, which is limited in capturing the observed burned area from GFED over several areas including the high latitudes. In this study, we therefore conducted the data assimilation and analyzed the results with focusing on Alaska and Siberia where there were large uncertainties of fire prediction (i.e., prediction of burned area), which would further affect carbon emission as well as carbon and water cycles. Results demonstrated that the land data assimilation with the burned area data remarkably improved carbon emission estimation, while showing that the opposite trends of the net ecosystem exchange were simulated between the open loop (i.e., free run) and data assimilation runs in Alaska. Furthermore, we showed that the evapotranspiration components such as ground evaporation and canopy transpiration significantly changed after assimilation. In the future, innovative methods for simulating the burned area such as using machine learning techniques should be developed for better understanding of future carbon cycles in high northern latitudes. Acknowledgements This work was supported by the Basic Science Research Program through the National Research Foundation of Korea, which was funded by the Ministry of Science, ICT & Future Planning (2020R1A2C2007670) and by the Technology Advancement Research Program through the Korea Agency for Infrastructure Technology Advancement (KAIA) grant funded by the Ministry of Land, Infrastructure and Transport (21CTAP-C163541-01).