Improved Forest Fire Danger Index estimate using SMAP soil moisture

Fire has a profound influence on altering an ecosystem and impacting the economy of a region. Fire-danger indicators are useful for describing burning conditions at a particular location and time. For example, the McArthur Forest Fire Danger Index (FFDI) is commonly used in Australia. In this study we demonstrate a modified FFDI by using remote sensing-based surface soil moisture from Soil Moisture Active-Passive (SMAP) Level-4 product, as an alternative to the drought factor. Fire probability is derived based on logistic regression between FFDI and the burned dates indicated by the Moderate Resolution Imaging Spectroradiometer (MODIS) based MCD64A1 Version 6 Burned Area data. We analyzed the Australian 2019-20 wildfire based on IGBP landcover classes.

Preliminary results show a very good agreement between McArthur FFDI and SMAP FFDI in Evergreen broadleaf and evergreen needleleaf forest. However, based on the comparisons made between MCD64A1 burned dates and FFDI estimates, for the drier regions like savannas and shrublands SMAP FFDI performs better than McArthur FFDI in terms of indicating a fire danger. Finally, we will highlight the best threshold to identify high probability of fire, defined by landcover type. Landcover based threshold probability can act as a better fire danger indicator than the traditional categorization simply based on FFDI range.