Research On The Remote Sensing Means Of Monitoring Grassland Vegetation Change

In this study a quantitative method of directly estimating parameters indicative of grassland vegetation change from existent Landsat Thematic Mapper (TM) image data and unified models for the estimation in a study area around Lake Qinghai of Qinghai Province, China were developed using modern satellite remote sensing technology, in conjunction with in situ grassland samples concurrently collected with satellite data recording, ground remote sensing, as well as Global Positioning System (GPS) technology. Invariant ground targets were used to radiometrically calibrate historical image data and to standardize multitemporal satellite data. These calibrated data were input into a GIS to detect the spatiotemporal changes of grassland vegetation in the study area. Such detected changes were comprehensively analyzed in association with a wide variety of natural and socioeconomic factors that might affect grassland vegetation changes in an attempt to determine its mechanism.This study has led to the following findings and conclusions:(1) Density of grass cover and the proportion of palatable species of grass in the study area are most closely correlated with normalized difference vegetation index (NDVI) and soil adjusted vegetation index (SAVI), respectively. Their regression models take the following form:Density of grass cover = 210.57xNDVI - 47.915 (R2=0.74) Proportion of palatable grass species = 318.11XSAVI - 52.13 (R2=0.73) Field tests suggest that the overall accuracy of estimating these parameters was as high as 89% and 87%, respectively.(2) In case that grass cover density is not spatially homogeneous on the ground and that in situ sampling plots are much smaller than those from space, it is impossible to establish an accurate regression relationship between in situ sampled parameters and the value of their corresponding pixels on the satellite image as is the case in the conventional way in the past. In this study I undertook the calibration of the TM image first with the results acquired from ground remote sensing before itwas used to establish the relationship between grassland vegetation change and satellite imagery so that the space-borne data could be used to estimate features on the ground. This processing has led to satisfactory results.(3) Inside the study area both grassland degradation and grassland vegetation regeneration co-exist next to each other. At some locations grassland vegetation has regenerated while at some others it has degraded at the same time. These two processes have resulted in an overall trend of gradual grassland degradation.(4) Grassland vegetation change in the study area has responded to the global climate change to certain degree. Nevertheless, climate variables are not the only ones that have negatively affected the vegetation. The influence exerted by many other natural and human factors should not be discounted in the degradation process.