Construction and evaluation of a vegetation coverage retrieval model for alpine prairie in Bayanbulak grassland

We constructed an inversion model of the vegetation coverage of the alpine prairie in Bayanbulak, screening for the best model and checking its accuracy, which will provide a scientific basis for carrying out grassland monitoring and scientific management in the alpine prairie of Bayanbulak in future. An empirical regression model and pixel dichotomy were used to calculate the vegetation coverage and vegetation indices and to construct the inversion model. The results showed that there were significant positive correlations between the vegetation indices, except for the Modified Soil Adjusted Vegetation Index (MSAVI), and vegetation coverage (P<0.01). The best inversion model of alpine steppe vegetation coverage in Bayanbulak is the quadratic polynomial regression model of NDVI:y=-0.894x2+1.467x+0.099 (R2=0.723); its inversion accuracy is higher (R2=0.837), and can be used to estimate the vegetation coverage of Bayanbulak alpine steppes.