Analysis of hyperspectral reflectance characteristics of three main grassland types in Xinjiang

The arid desert meadow,Artemisia desert meadow,and lowland mountain meadow are three main grassland types of Xinjiang Uygur Autonomous Region of China.To assist the extraction of remote sensing data and dynamic monitoring of grassland,hyperspectral reflectance spectra were measured with a portable ground object spectrometer.The reflectance spectral characteristics were analyzed for the three main grassland types that are located on the northern side of the Tianshan Mountains in Fukang city.The canopy spectral reflectance of arid desert meadow was smaller than those of Artemisia desert meadow and lowland mountain meadow in the visible wavelengths(except for Ceratocarpus arenarius) while in the near infrared wavelengths,the spectral reflectance of C.arenarius,Peganum harmala and Haloxylon ammodendron was considerably greater than those of Artemisia desert meadow and lowland mountain meadow.Because of the difference between vegetation type and internal structure of leaves,the differences of spectral reflectance between different vegetations that belong to the same type were significant in the visible and the near infrared wavelengths.The value of the red edge position of H.ammodendron of arid desert meadow was greater than those of Artemisia desert meadow and lowland mountain meadow.The value of Dλred and Sred of P.harmala in the Artemisia desert meadow was greater than those of arid desert meadow and lowland mountain meadow,while those of Carex liparocarpos was the smallest.PRI,OSAVI and MCARI were the greatest for arid desert meadow vegetation and the smallest for Artemisia desert meadow vegetation.NDVI was the greatest for lowland mountain meadow and the lowest for arid desert meadow.In addition,GNDVI was the greatest for lowland mountain meadow and the lowest for Artemisia desert meadow.In conclusion,hyperspectral remote sensing played a vital and significant role in monitoring grassland vegetation classification and remote sensing inversion.