Monitoring models of the grass yield for grassland in Gansu Province

Based on MODIS data and field data, the optimal prediction models of remote sensing were established for different grassland distribution regions, and these established models were also used to estimate the grass yields during the period of 2005 to 2009 in Gansu, and discuss the feature spatial distribution of grass yield. The results showed that optimal vegetation index and inversion model was different for different grassland distribution regions. The vegetation index RVI with logarithm model was suitable for the desert grassland areas of high plain in Hexi Corridor, and vegetation index EVI with exponent model was suitable for meadow grassland areas in Gannan High-land, and vegetation index NDVI with quadratic polynomial model was suitable for temperate steppe grassland areas in the Loess Plateau. The total grass yield of grassland in Gansu was 3.718 30×107 t in 2009, with an increasing trend from northeast regions to southwest regions. There were 3 high yield regions which were in Gannan High-land, the middle-east part of North Slope of Qilan Mountains and south region of Longnan. Compared with that of year 1985, the grass yield of grassland in Ganau province obviously decreased with percent of 19.34%.