Häusler, M., Nunes, J.P., Silva, J.M., Keizer, J.J., Warneke, T. & Pereira, J.M. (2019) A promising new approach to estimate drought indices for fire danger assessment using remotely sensed data.Agricultural and Forest Meteorology, 274, 195-209. DOI:10.1016/j.agrformet.2019.04.015 (IF2018 4,189; Q1 Forestry)
One of the crucial input variables in fire danger rating systems is the water content of the soil, as well as of the living and dead fuels. This study concentrates on the Keetch-Byram Drought Index (KBDI) and the Drought Code (DC), which are both used to estimate the soil moisture deficit (SMD) and are inherent parts of well-established fire danger rating systems.
These indices assume a simple water balance equation where effective rainfall is the only input and evapotranspiration is the only loss term, retrieved from empirical equations. We propose a method to increase spatial resolution of estimated SMD from KBDI and DC by replacing the empirical estimation of evapotranspiration with the actual evapotranspiration (ETa) estimated by the remote-sensing based Simplified Two-Source Energy Balance model (STSEB). For the days with no satellite observations, the basal crop coefficient (Kcb) and the adjusted crop coefficient (Kc,adj) approach for water stress conditions were applied to obtain continuous daily estimates for fire danger rating. A new approach was suggested to find a threshold dividing areas into fire-prone and fire-safe, regardless of the status of the drought indices.
Both drought indices (KBDI and DC) benefited from the estimated ETa of the existing STSEB model obtaining SMD maps with higher spatial resolution (compared to conventional methods and products using interpolation techniques). Areas with higher proneness to drought and therefore to occurrence to fire were identified in the north-central part of Portugal, where e.g. on 16 July 2015, in the case of KBDI 47% of a specific area was classified as fire-prone and about 67% of the fires occurred within the identified area of this year.
This new approach generated information on the water status of the soil at a spatial resolution of 500 m, which can be useful for forest management to identify endangered areas, as the danger of fire increases with increasing drought conditions.