Spectral characteristics of alpine grassland with different degrees of desertification in Maqu County, the source region of the Yellow River

Alpine grasslands are the ecological barrier in the upper reaches of the Yellow River. The increase in areas of bare sand following decreasing vegetation cover has become severe in grasslands, which is difficult to monitor and control because of natural and human influences. Analyzing the spectral characteristics of alpine grassland with different degrees of desertification can provide sufficient ground verification information for remote sensing monitoring and accurate identification of degrees of desertification in the alpine region, which is of considerable importance. Based on hyperspectral information from alpine grassland with different degrees of desertification in Maqu County field, the study analyzed the spectral characteristics by examining various mathematical variations. In terms of the results, the spectral curves were relatively complex and inefficient and showed representative characteristics of desertified soil. Meanwhile, the parameters of red valve ley, green peak, and the “three edges” of other sanded grasslands were consistent with the spectral characteristics of green vegetation. The red valley and green peaks in the 510－560 nm and 650－690 nm were not well differentiated. The first-order differential and envelope removal mathematical transformations were used to process the reflectance curves of the grassland spectra. It was found that the characteristic bands and parameters that could be used to differentiate between different levels of sandy grassland were 510－560 nm, 650－700 nm, the red-edge amplitude, and the absorption depth at 600－700 nm after continuum removal. After the band calculation, the relative vegetation index (RVI) and green and red vegetation index (GRVI) indices were apparent in distinguishing different degrees of sandy grassland. Meanwhile, the vegetation indices were thresholded, which could be used as a basis for identifying grasslands with different degrees of desertification. The results provide data support and a key reference for monitoring loss of vegetation cover and effective analysis of the degree of desertification in alpine areas.