Orienting rocky desertification towards sustainable land use: An advanced remote sensing tool to guide the conservation policy

Abstract

Due to non-sustainable land management, desertification has been occurring widely across the world and continues to be a global land use problem. In this context, appropriate methodological tools, which can provide a biased estimation of desertification, are critical for learning from past failures and local successes in orienting desertification towards sustainable land use. This paper proposes a locally adaptive multiple endmember spectral mixture analysis (MESMA) algorithm to extract the rocky desertification information from medium resolution images at subpixel level and applies it to the case of Danjiangkou reservoir region (DRR), China. Quantitative comparisons show that the locally adaptive MESMA has achieved more accurate and reliable estimations of rocky desertification information in DRR than the traditional MESMA. An inversed U-shaped trend is observed for desertified land with different severity levels from 1987 and 2013 in DRR. In particular, the inflection point roughly emerged in period 2000–2005. Casual mechanism-based regressions demonstrate that such dynamics of rocky desertification are closely coupled with socioeconomic, biophysical, and policy factors. More specifically, we identify a significantly positive role of land conservation policy in combating and relieving rocky desertification in DRR. Positive effects are observed particularly through afforestation, investment, and professionals input. Based on the conclusions and lessons of DRR, I finally make relevant recommendations for formulating policies and strategies that attempt to orient desertification towards sustainable land use. The proposed locally adaptive MESMA can act as an advanced remote sensing tool to guide the conservation policy.