Authors: Timothy Assal*, US Geological Survey, Collin Haffey, The Nature Conservancy, Ellis Margolis, US Geological Survey, Craig Allen, US Geological Survey
Topics: Biogeography, Spatial Analysis & Modeling, Mountain Environments
Keywords: Fire, vegetation change, remote sensing, Landsat, randomForests, New Mexico
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Regency Ballroom, Omni, West
Presentation File: No File Uploaded
Numerous fires over the last several decades have eliminated conifer forest in mid to upper elevations of the Jemez Mountains in northern New Mexico. Over a century of fire suppression led to an increase in tree density and forest structure, where ladder fuels favored crown fire development resulting in high mortality. There is concern these fires may have triggered a change from forest to deciduous shrubland where the majority of seed source has been removed. Resprouting woody species may become more prevalent on the landscape as this life-history trait may be an advantage in areas that experience frequent disturbance. The goal of our study is to quantify the amount of vegetation change that has taken place on the landscape at key points in time with respect to major disturbance events. We developed continuous cover models of coniferous and woody deciduous vegetation using a combination of topographic and phenological spectral predictor variables derived from Landsat data. We used aerial photos to scale up vegetation observations across the landscape in a randomForest modeling procedure, then backcast the models to prior time periods. The results indicate a cumulative loss of conifer cover over the last three decades; whereas deciduous shrub cover increased throughout the region, most notably in areas that were burned at least once. We will highlight drivers of post-fire vegetation recovery with respect to burn severity, landscape position and refugia. We will also address the potential application to export this open data framework to other ecosystems to quantify recent vegetation change.