Presentation 3
LiDAR Combined with Landsat Increases the Accuracy and Resolution of Landscape-scale Estimates of Fire Effects (286)
Research
Landsat-based fire severity maps have limited ecological resolution that can hinder assessments of change to specific resources. Therefore, we evaluated the use of pre- and post-fire LiDAR, and combined LiDAR with Landsat-based (RdNBR) estimates, to increase the accuracy and resolution of basal area mortality. We vertically segmented point clouds and performed model selection on spectral and spatial pre- and post-fire LiDAR metrics and their absolute differences. Our top multi-temporal LiDAR model included change in mean intensity values 2-10 m aboveground, the proportion of canopy reflection sum above 10 m, and differences in maximum height. This model reduced root mean squared error (RMSE) and root mean squared prediction error (RMSPE) by 39% and 37%, and bias by 58%. Our top combined model integrated RdNBR with LiDAR return proportions < 2 m aboveground, pre-fire 95% 31 heights and pre-fire return proportions < 2 m aboveground. This model reduced RMSE and RMSPE by 38% and 34%, and bias by 75%. Our results confirm that 3-dimensional spectral and spatial information from multi-temporal LiDAR can isolate disturbance effects on specific ecological resources with higher accuracy and ecological resolution than Landsat-based estimates, offering a new frontier in landscape scale estimates of fire effects.
Author
Michael Hoe
SCS Global Services