Quantifying Fuel Consumption for Two Western U.S. Fires using Repeat LiDAR
Abstract
Accurate estimation of the type, loading, and combustion completeness of fuels consumed by wildfire is the largest source of uncertainty in estimating the quality and quantity of atmospheric emissions of greenhouse gases. Maps of fuel consumption are also valuable for constraining estimates of carbon fluxes and for better understanding burn severity patterns across the post-fire landscape. Climate change-driven increases in wildfire size and duration of the wildfire season, and the issue of hazardous fuel accumulations due to historic fire exclusion policies, provide further impetus to improve methods for characterizing fuel consumption. Airborne Light Detection and Ranging (LiDAR) is a currently available technology that has proven useful for estimating aboveground biomass (AGB) in conjunction with field plot measurements, although places with LiDAR and field plot data available both pre- and post-fire are still relatively rare. This study exploited two opportunities to estimate fuel consumption (Mg ha-1) using differences in AGB estimated from pre- and post-fire LiDAR. We acquired repeat LiDAR and field data from two fires, the 2011 Las Conchas fire in New Mexico and the 2012 Pole Creek fire in Oregon. We predicted landscape-scale AGB with Random Forest (RF) regression using field-based estimates of AGB and various height and density metrics from the LiDAR data. The models had a pseudo-r-squared as high as 0.86 with a root square mean distance of 108.7 Mg ha-1. Changes in AGB resulting from each fire were quantified and evaluated through comparison with Fire Radiative Energy (FRE) estimations, obtained through the temporal and spatial integration of Fire Radiative Power (FRP) observed from the MODIS spaceborne sensor. FRE has been demonstrated to be linearly related to biomass consumption across a variety of spatial scales and vegetation combustion conditions. Additionally, the models developed for each fire were compared to explore the potential of a generalized model, capable of rapid response assessment of AGB from future fires that intersect existing LiDAR acquisitions and are viable candidates for post-fire LiDAR acquisitions to facilitate rapid assessment of fuel consumption.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.B24B..07M
- Keywords:
-
- 0414 Biogeochemical cycles;
- processes;
- and modeling;
- BIOGEOSCIENCESDE: 0439 Ecosystems;
- structure and dynamics;
- BIOGEOSCIENCESDE: 1640 Remote sensing;
- GLOBAL CHANGEDE: 1817 Extreme events;
- HYDROLOGY