Integrated Orderly Classification And Remote Sensed Monitoring Of Natural Grassland On Tibetan Plateau

As the main body of the terrestrial ecosystem, the grassland ecosystem in the global energy flow, substance and information flow plays an important role in circulation. Comprehensive, timely and accurate understanding the changing of the regional grassland resource and ecological environment, is important to achieve sustainable development of grassland agriculture and ecological environment.As a sensitive and fragile area, it has significant value to study the climate change, environment protection and ecologic construction on Tibetan Plateau. Based on the integrated orderly classification, this thesis studied the natural grassland using remote sensing technique, set up an integrated orderly grassland class map through the selection of right interpolation model using GIS geostatistic approach, then analyzed the changing characteristics of EVIindex and the correlation between EVI and corresponding observed data, built different aboveground biomass (AGB) models, and discussed the effect and value of the integrated orderly classification and remotely sensed monitoring in the Tibetan Plateau grasslands. Among them, the interpolation model selection of climatic data and analysis of AGB models are the keys to this study.Integrated orderly classification approach has similar advantages with Linnaean binomial taxonomic system of law and the Mendeleev periodic table of chemical elements; both of them contain not only the known types, but the unknown type as well, and reveal the genesis relationship between the various types. The classification system quantizes the dominant factors-the values of water and calorie, reduces the influence of subjective factors in a large part, and substantially increases the accuracy and practicality in grassland classification. As an advanced technology, satellite remote sensing has some advantages in wide range, multi-period, low-cost, high efficiency, etc. Combining remotely sensed monitoring and integrated orderly classification could get hold of the processes of the grassland genesis and development fully and quickly, and make it possible to study on types and changes of the grassland on greater spatial scales.The main conclusions:(1) When conducting the spatial interpolation for annual>0℃accumulated temperature and annual precipitation in Tibetan Plateau, the overall performance of radial basis function is better than inverse distance weighting, ordinary kriging and co-kriging. (2) Natural grassland in Tibetan Plateau can be classified by Integrated Orderly Classification approach with certain degree of technical feasibility and a higher level of discrimination; the natural grassland in Tibetan Plateau can be divided into 35 categories and 9 category groups; the spatial distribution of these categories reflected the trend of zonality of the decreasing of hydro-thermal conditions from southeast to northwest area. (3)The condition of growth and changes of grassland are well presented by Enhanced Vegetation Index (EVI), which kept significant level of consistency with the growth and changes of grassland. (4) The exponential regression models is overall better than simple liner regressing models for the estimation of aboveground biomass of categories of natural grassland.