Dynamic Monitoring Of Grassland Degradation Based On Landsat Image

As an important part of terrestrial ecosystem and basic data for the sustainable development of animal husbandry, grassland ecosystem has enormous ecological and economic value. In recent years, the situation of grassland degradation is getting worse, it has become an important factor for restricting our animal husbandry development and ecological balance. Therefore, it is especially important for us to timely and accurately grasp the degradation situation and diagnose the degradation degree. With the continuous development of 3S technology, remote sensing has become the main method for study on large area of grassland degradation.This paper took Bashang grassland as the study area. In this study, Landsat TM from 2009 to 2011 and Landsat OLITIRS from 2013 to 2014 were used to calculate a variety of vegetation index. This paper compared and analyzed the correlation between vegetation index of 2014 and the grassland degradation index which were based on the measured data. The study tried to find out the most suitable vegetation index and evaluation model. And then, the remote sensing monitoring grassland vegetation degradation model was established. The images from 2009 to 2014 were classified by the model in order to transform remote sensing image into the grassland vegetation degradation image. This article analyzed and clarified spatial distribution and developmental direction of various grades of degraded grassland from four aspects that were change range, trend, strength and direction. The results were as follows:(1) This paper used the method of factor analysis to determine the weight of the index (biomass, coverage, height) which reflected the actual changes of the grass. It was concluded that the coverage was the largest weight of indicators, and the weight was 47%, followed by biomass, was 37%, the smallest for height, accounted for 16%.(2) RDVI of grass in the study area were most closely correlated with grassland degradation index. Their regression models took the following form: RDVI=-0.356GD/2+0.016GD/+0.221 (R2= 0.734)After inspection of three kinds of correction model, what were relative error, the root mean square difference (RMSE) and the determination coefficient R2, the accuracy of the inversion model was obtained:the relative error is 1.53%; RMSE= 0.071; R2= 0.7146, the accuracy was high enough.(3) The degraded grassland in the study area was divided into four grades, namely inconspicuous degradation, mild degradation, moderate degradation and severe degradation. Bashang area of degraded grassland grading standards were obtained by cluster analysis method. When GDI>0.65 or RDVI>0.382, the grass was defined as inconspicuous degradation, when 0.40<GDI<0.65 or 0.284<RDVI<0.382, the grass was defined as mild degradation, when 0.20<GDI<0.40 or 0.238<RDVI<0.284, the grass was defined as moderate degradation, when GDI<0.20 or RDVI<0.238, the grass was defined as serious degradation.(4) This article analyzed the grassland vegetation change from four aspects that were change range, trend, strength and direction. The condition of grassland vegetation in different historical periods from 2009 to 2014 was analyzed in multi-angle.From the perspective of the space, regressive type and restoring type of grassland vegetation existed at the same time type in the study area. Grassland vegetation degeneration had taken place in some areas, but other areas of grassland was restored. The degraded condition of grassland presented inconsistency in the study area. The grass was given priority to medium and severe degradation in the southwest, and the northeast region was given priority to inconspicuous and mild degradation. Degenerated grassland was mainly distributed in central and southern areas, and restorative grassland was distributed dispersedly in each area.From the time point of view, the grassland vegetation was restored significantly in 2013, while there were not obvious change or degradation increased in the other years. Especially in 2014, the grassland vegetation presented the obvious degradation trend in the study area. The changes of grassland vegetation were slow in each year. Slowly change accounts for absolute advantage and the skip level of the grass was small, unchanged area was the biggest, which had reached more than half. It was a change process which was slow and degenerated partly.Generally speaking, the grassland degradation was serious in the west, and was less serious in the east. It presented a trend of deterioration with fluctuation.