Assessing Fire Severity in Semiarid Environments with the DNBR and RDNBR Indices

Abstract

Available remote sensing historical Landsat TM images allow identifying of first order effects of wildfires also in huge and inaccessible regions. In this paper the usefulness of the best known satellitederived severity indices was tested on a large wildfire occurred in January 1999 in a steppe of Northwestern Patagonia. The main objective of the work was to analyze and compare the behavior of dNBR and RdNBR in their ability to discriminate the degrees of fire severity in semiarid ecosystems principally dominated by herbaceous vegetation. For this purpose the values of the two indexes were compared in all vegetation communities (shrubl and, meadow, grassland and forestation). To interpret the results, we considered the variability of the principal factors that influence the fire severity, as fire intensity, fire duration and vegetation susceptibility to fire. The analysis showed that the interaction between fire and vegetation changes the fire effects because the vegetation parameter as fuel load, moisture content, species composition, horizontal continuity and the topography affect the fire behavior and then the fire severity. Furthermore the results suggest that dNBR and RdNBR provide substantially different information respectively related to the effects on soil and vegetation. This work is an important contribution to the utilization of fire severity indexes in ecosystems dominated by herbaceous species that change more subtly the post-fire biomass than ecosystems dominated by woody species.