Sub-pixel mapping of karst rocky desertification using multiple endmember spectral mixture analysis

This study tests the performance of multiple endmember spectral mixture analysis (MESMA) for estimation of karst rocky desertification (KRD) in karst studies. We present a methodology to build a regionally specific spectral library of karst materials including: vegetation, carbonate rock, soil, and water. Using this library, we applied MESMA to generate a total of 1000 two-, three-, and four-endmember models for each Landsat 8 OLI pixel, and the mapped fractions were validated using WorldView-2 plots. The results showed that a majority of the image could be modeled successfully with two- or three-endmember SMA models. The validation results indicated the robustness of MESMA for deriving spatially continuous variables quantified at the sub-pixel level. Quantitative analysis of the resulting percent KRD and comparative analyses with traditional Liner Spectral Unmixing (LSU) indicate that MESMA improved the estimation accuracy of KRD.