A New Method for Plant Flammability Estimation Using SMAP Soil Moisture

Live fuel moisture (LFM) is defined as the percentage ratio between the mass of water content in the fresh biomass and the mass of dry biomass. It indicates the level of flammability of plants, therefore serves as an effective indicator of fire risk. Satellite data have been used to derive models to extrapolate, estimate, and forecast the LFM to better address the need for a gap-free LFM product to improve the current fire danger forecasting model. Although empirical models based on vegetation indices (VIs) and process-based method using radiative transfer models (RTMs) both yielded satisfactory results in the fire-prone Mediterranean ecosystems, they cannot provide physically meaningful outlooks of LFM. We proposed a new method for LFM forecasting using SMAP L-band radiometer soil moisture (SM) and leveraging the lagged relationship between SMAP SM and LFM. In the study area of Southern California, SMAP SM changes about 60 days ahead of LFM, varying by locations and years. Moisture accumulates during this period to initiate and sustain the growth of plants. To describe the time-constrained accumulative effect in the contribution from SM to LFM, we summed up the SMAP SM with a moving window ahead of each LFM measurement date to obtain the accumulative SMAP SM. The accumulative SMAP SM was combined with cumulative growing degree days (CGDDs) to build a regression model of LFM estimation. Unlike the previous empirical models using VIs, this model can calculate the outlook of LFM based on the earlier moisture and heat conditions. In addition, we separated the growing cycle into the green up, brown down, and photosynthetically inactive period based on the key phenological metrics and built separate models for each, as the contributing factors to LFM differed between periods. SMAP SM was the major contributing factor during the green up period, while the CGDDs determined the dynamics of LFM during the brown down and the photosynthetically inactive period. This new method outperformed the reference model using MODIS Visible Atmospheric Resistance Index (VARI) for 12 LFM sites in Southern California, with an adjusted R² of 0.53 and RMSE of 19.88%. This work demonstrated the potential of SMAP SM to estimate the plant flammability, which can lead to future improvement in the current fire danger alarming system.