Simulation and evaluation of forage height in alpine meadow grassland in headwaters of the Yellow River

Canopy height in grassland is an important factor for reflecting growth status, and is closely related to monitoring of aboveground biomass and grazing capacity evaluation as well. Presently, accurate remote sensing methods for monitoring canopy height in natural grassland communities are still lacking. In this study, height estimation models based on MODIS vegetation indices, canopy height model, and grassland hyperspectral reflectance were established, and assessed in the source region of the Yellow River, individually. The results showed that: 1) NDVI was most sensitive to grassland canopy height in 12 MODIS vegetation indices. However, the R2 values of four regression models derived from NDVI were low, which were 0.203~0.241. The NDVI exponential model (y=0.789e3.186x,R2=0.241) had the best fitting effect; however, inversion error was greater (RMSE=4.2 cm, CVRMSE = 45.7%); 2) The accuracy of the canopy height model in alpine meadow was low (RMSE=5.8 cm, CVRMSE=62.1%); 3) There was a significant negative correlation between height of alpine meadow community and canopy spectral reflectance ranging within 519.4~583.17 nm band (P0.05). The linear model y=-156.375x+20.384 (R2=0.489, RMSE=3.5 cm) based on R'510.59 was suitable for estimating the grassland canopy height in alpine meadows.