Spectral Unmixing vegetation cover of a desert ecotone with High Spatial Resolution Satellite Imagery

Abstract

Vegetation in arid and semiarid ecosystems is sensitive to climate change and anthropogenic effects. Accurate estimation of vegetation cover in drylands is critical for their conservation and management. Spectral unmixing of spaceborne remotely sensed data have provided a promising tool to estimate sparse vegetation in dryland environment. In this study, we explored estimation of fractional vegetation cover using multiple endmembers spectral mixture analysis (MESMA) applied to high spatial resolution (HSR) satellite images. Two HSR datasets were acquired by WorldView-2 and Pleiades-1A on June 4, 2013 and July 13, 1019 respectively. The study area lies on the Manas River watershed on north hillslope of Tian Shan, Xinjiang, China. The ecotone includes three plant communities: desert steppe, dry steppe, and meadow steppe. A field study was conducted in July 2013. A photo-based grid point intercept sampling method were used to quantifying vegetation cover over this desert ecotone. Spectra of different ground covers were collected using ASD FieldSpec spectrometer in the field study Both endmember spectra from USGS spectral library and the field study will be compared with endmembers identified in imagery scenes for unmixing analyses. Vertex component analysis (VCA) algorithm was used to retrieve endmember automatedly from image scene. This study will provide insight in accuracy of fractional vegetation cover retrieved from HSR data and how phenological effects impact vegetation cover of this ecotone. The fractional vegetation cover estimated from HSR data can serve as reference data for assessing unmixing results with moderate spatial resolution data such as Landsat imagery data.