Monitoring changes of Rangeland Vegetation using Remote Sensing

Keywords: MESMA, Sentinel, Spectroscopy, Rangeland, Encroachment
Abstract Type: Paper Abstract

Abstract

Encroachment of woody plants into grasslands occurs in drylands worldwide and often negatively impacts key ecosystem services such as grass production and erosion control. To control or reverse the process, land managers use different strategies, such as herbicide applications. However, the efficacy of treatments across space and through time is typically unclear, because post-treatment vegetation changes are not monitored in a spatially and temporally explicit manner. To address this issue, we tested the utility of Multiple Endmember Spectral Mixture Analysis (MESMA) and image differencing for mapping fractional cover changes of photosynthetic vegetation (woody plants), non-photosynthetic vegetation, and soil in response to herbicide treatment. Our case study area included 20 paired control/treatment plots of honey mesquite (Prosopis glandulosa) encroached rangeland on the Jornada Experimental Range in southern New Mexico, USA. Our input data included Sentinel-2 satellite imagery acquired over the study area before and after treatment in August/September 2021; ii) field spectral reflectance data of green and senescing mesquite, other plants, and soil collected using an ASD FieldSpec 4 Hi-Res Spectroradiometer; and iii) fractional cover data estimated along three transects per plots. We applied MESMA using Sentinel data spectral endmembers selected using three different endmember selection strategies. We assessed the accuracy of the fractional cover maps using the transect data. Our results suggest that MESMA and image differencing may provide a timely, cost-effective, and accurate remote sensing approach for monitoring the effectiveness of herbicide treatments on mesquite-encroached rangelands.