Accurate assessment of fire severity is invaluable for monitoring fire effects and recovery trajectories. This is particularly true for fire severity ratings derived using remote sensing data, which can be used to assess fire severity across burn units, and also retroactively. Several indices exist that use remote sensing data to calculate burn severity. Change in the Normalized Burn Ratio (dNBR), which is a pre-fire to post-fire change in the ratio of middle-infrared to near-infrared reflected light, has emerged as the leading effective index of burn severity when compared to field-based measures. dNBR as a measure of burn severity has not been as successfully utilized in central hardwood forests of the southern Appalachians as it has in conifer-dominated forests of the western United States, particularly for low and moderate fire severities. The objective of this study was to evaluate the ability of Normalized Burn Ratios (i.e., NBRpre, NBRpost, dNBR, RdNBR) to predict fire severity using pre- and post-burn monitoring data from 13 prescribed burn units located in eastern TN, western NC, northwestern SC, and northern GA. We compared results to Normalized Difference Vegetation Index (NDVI) values, which is an index of plant “greenness” or photosynthetic activity, calculated for the same areas.
Results/Conclusions
Characterization of the post-burn landscape aids examinations of pre-fire conditions and resulting burn severity patterns as they affect post-fire trajectories. To track ecosystem recovery we must have accurate information about pre- and post-fire forest structure, topography, and burn severity. Initial results indicate dNBR is not effective for identifying areas of low burn severity in the southern Appalachians. We believe it is because low severity fires generally do not impact the upper canopy, which causes issues when analyzing burn severity using dNBR. However, heat generated from a fire can change the physical and/or chemical composition of canopy leaves, which NDVI may be able to detect. dNBR appears to be effective in identifying areas of high burn severity where canopy was burned during a fire, while NDVI appears to be better at exposing subtle changes to vegetation post-fire, or increased sensitivity of ecosystem response. These findings are consistent with those from studies in other areas. We are currently assessing the effectiveness of other remote sensing-derived burn indices. Knowing how vegetation influences fire behavior, and thus burn severity, is key to understanding how vegetation recovery trajectories might impact future ecosystem structure, function, and services, which are particularly critical in areas of extraordinary biodiversity.