Timely, Gridded Estimates of Fuel Moisture Content Based on Machine Learning and Satellite Data

A primary driver of wildfire behavior is the moisture content of dead and live fuels. Fuels with high fuel moisture content (FMC) do not ignite as readily nor burn as quickly and as energetically. FMC also regulates the amount and composition of smoke. Unfortunately, current industry-standard datasets of FMC are based on sparse, automated observations of dead FMC, and on infrequent, labor-intensive, manual sampling of live FMC. The NSF National Center for Atmospheric Research, in collaboration with NOAA, has nearly finished developing a national, machine-learning-based, gridded dataset of estimated dead and live FMC with pixel granularities of orders 0.1–1.0 km and 1–10 h, depending on the constituent satellite data. Presenters will review a public online dashboard that displays live and 10-h dead FMC nearly in real time, and will describe the research and development on which those data are based.