Characterizing the Response of Piñon-Juniper Woodlands to Mechanical Restoration Using High-Resolution Satellite Imagery

Abstract

In northern New Mexico, complex interactions among climate, land use, and the associated reduction of surface fire in forest and savanna communities facilitated the expansion of piñon-juniper woodlands. Because increasing tree cover can outcompete herbaceous vegetation (e.g., forbs and grasses), woodland expansion resulted in reduced herbaceous cover and increased soil exposure, leading to increased runoff and erosion. To improve hydrologic function in these degraded woodlands, an overstory thinning and slash-mulch treatment was applied to more than 1 000 hectares within Bandelier National Monument (New Mexico, United States) between 2007 and 2010. Our objective was to develop a remote sensing strategy to quantify land-cover changes following mechanical treatment of piñon-juniper woodlands. In this study, we 1) established a simple and repeatable method for assessing treatment effectiveness using high-resolution satellite imagery, 2) quantified vegetation response at six times since treatment (from 0.5 to 3.5 years), and 3) delineated areas of degraded woodland before and after treatment. We classified a 2006 QuickBird satellite image (before treatment) and a 2011 WorldView image (after treatment) to map tree cover, herbaceous cover, bare soil, and shadow with average overall accuracies of 95.0% and 94.7% for the 2006 and 2011 images, respectively. Following treatment, average tree cover, bare soil, and bare soil patch size decreased 9%, 4%, and 18%, respectively, while herbaceous cover increased 14%. Overall, the total amount of woodland area classified as degraded decreased by 34% post treatment. By comparison, control areas remained relatively unchanged across the same time period. Our results demonstrate the utility and promise of high-resolution satellite imagery to inform rapid and objective assessments of landscape-scale restoration treatments.