Assessing responses of grasslands to grazing management using remote sensing approaches.

摘要

Grazing caused grassland degradation has occurred worldwide in recent decades. In spite of numerous efforts that have been invested to explore the mechanism of grassland responses to grazing management, the major challenge remains monitoring the responses over large areas. This research evaluates the synthetic use of remote sensing data and the Milchunas-Sala-Lauenroth (MSL) model for grazing impact assessment, aiming to explore the potential of remotely sensed data to investigate the responses of grasslands to various grazing intensities across different grassland types.;By combining field collected biophysical parameters, ground hyperspectral data and satellite imagery with different resolutions, this research concluded that 1) sampling scale played an important role in vegetation condition assessment. Adjusted transformed soil-adjusted vegetation index (ATSAVI) derived from remote sensing imagery with 10m or 20m spatial resolution was suitable for measuring leaf area index (LAI) changes in post-grazing treatment in the grazing experimental site; 2) canopy height and the ratio of photosynthetically to non-photosynthetically active vegetation cover were identified as the most sensitive biophysical parameters to reflect vegetation changes in mixed grasslands under light to moderate grazing intensities; 3) OSAVI (Optimised soil adjusted vegetation index) derived from Landsat Thematic Mapper (TM) image can be used for grassland production estimation under various grazing intensities in three types of grasslands in Inner Mongolia, China, with an accuracy of 76%; and 4) Grassland production predicted by NCI (Normalized canopy index) showed significant differences between grazed and ungrazed sites in years with above average and average growing season precipitation, but not in dry years, and 75% of the variation in production was explained by growing season precipitation (April-August) for both grazed and ungrazed sites.