Just Accepted

This article has been peer reviewed and accepted for publication. It is in production and has not been edited, so may differ from the final published form.

Generating fuel consumption maps on prescribed fire experiments from airborne laser scanning

Ryan McCarley , Andrew Hudak, Benjamin Bright, James Cronan, Paige Eagle, Roger Ottmar, Adam Watts

Abstract

Background: Characterization of fuel consumption provides critical insights into fire behavior, effects, and emissions. Stand-replacing prescribed fire experiments in central Utah offered an opportunity to generate consumption estimates in coordination with other research efforts. Aims: We sought to generate fuel consumption maps using pre- and post-fire airborne laser scanning (ALS) and ground measurements and to test the spatial transferability of the ALS-derived fuel models. Methods: Using random forest (RF), we empirically modelled fuel load and estimated consumption from pre- and post-fire differences. We used cross-validation to assess RF model performance and test spatial transferability. Key Results: Consumption estimates for overstory fuels were more precise and accurate than for subcanopy fuels. Transferring RF models to provide consumption estimates in areas without ground training data resulted in loss of precision and accuracy. Conclusions: Fuel consumption maps were produced and are available for researchers who collected coincident fire behavior, effects, and emissions data. The precision and accuracy of these data vary by fuel type. Transferability of the models to novel areas depends on the user’s tolerance for error. Implications: This study fills a critical need in the broader set of research efforts linking fire behavior, effects, and emissions.

WF23160  Accepted 21 June 2024

© CSIRO 2024