Research On Remote Sensing Monitoring Method Of Grassland Degradation Using MODIS Data

As the largest terrestrial ecosystem in our country,grassland is an important material basis for the development of animal husbandry and the most important ecological barrier in the north.It is of great significance to discuss how to scientifically utilize grassland resources.At present,90%of the usable natural grassland in our country have been degraded,and the situation of grassland resource protection and grassland ecological security is severe.Timely and accurate acquisition of grassland degradation information is essential for scientifically making grassland resource utilization decisions.Remote sensing technology provides technical support for large-scale,rapid and accurate monitoring of grassland degradation.However,there are many types of remote sensing monitoring methods for grassland degradation,and the selected degradation evaluation indicators are quite different and lack a unified grassland degradation evaluation standard system.Aiming at the problems of low inversion accuracy of key parameters in remote sensing monitoring of grassland degradation,lack of comprehensive evaluation indicators,and inconsistent degraded climatic conditions,this paper proposes a remote sensing monitoring method for grassland degradation based on the comprehensive evaluation index of grassland vegetation.On the basis of summarizing related research,this paper takes the grassland degradation in the Ili area of Xinjiang in the past 20 years as the research object.Under the premise of full field investigation,combined with the terrain and grassland vegetation characteristics of the study area,the grassland range and grassland information are accurately extracted.Select the key indicators of grassland vegetation status:vegetation coverage,biomass and bare sand area ratio,etc.,to optimize the remote sensing inversion method based on MODIS data.On this basis,the above key indicators are used to construct a grassland vegetation comprehensive index model.Under the premise of chronological classification of precipitation conditions,the model was applied to grassland degradation evaluation,and relatively good accuracy was achieved.The main research contents and conclusions of this paper are as follows:1.Extraction of grassland range and classification of grassland types in the study area:Based on the mask processing of the ground cover of water,snow,forest,arable land and artificial surfaces in the study area,the object-oriented classification method based on Landsat8 30m multispectral data is adopted The grassland range information in the study area is extracted,and the total grassland area in the study area is 3.4628 million hectares(51.942 million mu);the grassland in the study area is classified by the grassland type based on the decision tree classification method based on the terrain and vegetation index of the study area,and the main research area Grassland type:the distribution range of mountain meadows,alpine meadows,temperate meadow grasslands,temperate grasslands,alpine grasslands,temperate desert grasslands,temperate deserts,and lowland meadows.2.Remote sensing inversion of key indicators of grassland vegetation status:using the pixel binary model based on grassland type division,the linear mixed pixel decomposition model and the three-band maximum gradient difference model to estimate the vegetation coverage of the grassland in the study area,and the estimated accuracy The accuracy of the three methods is 85.66%,81.81%,and 84.45%in order.Therefore,this paper chooses the pixel binary model based on the division of grassland type as the method of obtaining vegetation coverage;adopts the grassland biomass regression based on many years of ground measurement data.The statistical model estimates the grassland biomass in the study area and compares it with the MOD17A3H NPP product.The MOD 17 product with higher estimation accuracy is selected as the inversion result of the grassland biomass in the study area;the linear spectral decomposition model is used to compare the bare sand area in the study area The remote sensing inversion was carried out,and a high-precision estimation result of the bare sand area ratio was obtained on the basis of the optimized end member selection,and the equivalent accuracy reached 81.81%.3.Classification of precipitation types in the study area:The MK test method and the Standardized Precipitation Index(Standardized Precipitation Index)were used to analyze the climate change trends in the study area,and the precipitation types in the grassland area were explored in time and space.The classification results of the precipitation type in 2017:average year,drought-prone year and flood-prone year.4.Construction of the Grassland Vegetation Comprehensive Index(GVSI)model:Based on the three single factors of vegetation coverage,biomass and bare sand area ratio,the factor weight coefficients:0.775,0.603,and 0.226 were determined using the principal component analysis method,on this basis A grassland vegetation comprehensive index model was constructed,and the spatial distribution and temporal change characteristics of GVSI in the study area were analyzed and evaluated using the model.5.Grassland degradation evaluation method based on grassland vegetation comprehensive index:According to the classification results of precipitation types in the study area,the average GVSI value of the same precipitation type years is used as the corresponding grassland degradation background value,and the classification threshold of the national standards for grassland degradation is adopted.The coefficient weighted average method determines the GVSI grassland degradation evaluation threshold(GVSI reduction rate):no degradation(0-10%),light degradation(10%-20%),moderate degradation(20%-31.25%),Severe degradation(above 31.25%),using ground verification point grassland degradation degree information to verify the accuracy of the evaluation,the accuracy rate of degradation evaluation reached 70.09%.6.Grassland degradation in the study area:The GVSI-based grassland degradation assessment method established in this paper was used to evaluate the grassland degradation in the study area in 2014 and 2017.In 2014,due to the impact of drought,the grassland degradation was relatively serious,and the total degraded area reached 208.875 million mu,accounting for 40.21%of the total grassland area;in 2017,there was no large-scale degradation of grassland in the Yili area.The total degraded area was 3.564 million mu,accounting for only 6.86%of the total grassland area.The main grassland type that experienced severe degradation was alpine meadows,the slight degradation is mainly the degradation of mountain meadows.The innovations of the thesis are concentrated in the following two aspects:one is the systematic and in-depth analysis and study of the temporal and spatial changes of grassland degradation in the Ili area;the other is the construction of the grassland vegetation comprehensive index GVSI evaluation model,and the establishment of the Ili area based on MODIS data.The remote sensing monitoring technology and methods of grassland vegetation degradation of different grassland types provide the basis for remote sensing monitoring of grassland degradation in Ili area.