Dynamic Monitoring And Simulate Prediction On LUCC Based On None-Liner Least-Sequare Method

Oasis,being a kind of specially ecological item in arid and semi-arid area,is a basic area which sustains the lives and developments of the local people. Land use and cover change (LUCC) forms major aspect of global ecosystem on both environmental change and one of the hot point of research to current world's change , and monitoring and finding its change regulation has become the premise that to predict LUCC change trend in the future. This paper discussed the pre-processing of the image's object extraction , geometric correction and transform projection base on the characters of TM and MODIS images.The data processing has used the MODIS processing software of national satellite weather center and PCI, ENVI software, and had acquired good results.and use the maximum likely hood classification method to more process classify the remote sensing images of study erea ,the result shows that the class precision is already enough to analysis(98%).The paper as the hotan oasis (three county and one city), with the weak ecosystem environment, for the study area, and carryied on monitor to the Land Use Land Cover of Hetian oasis. According to each pixel's variety trend and its variety regulation,it is carried on the first step estimate to the future changing trend of the whole study area. The paper use none-linear Least-square method and predicted the LUCC of study area,and both the result data and the change of study erea is more inosculate from 2002 to 2006 . It is prove that if the overall drive force is fix under present condition ,the method used for LUCC is viable. This paper is mainly consisted of three parts, First part is mostly narrated the basic theoretics of LUCC and its domestic and overseas research progress, discussed the paper's research purpose and meaning, research methodology and its theory, the innovation of research. Second part introduced the preprocessing of remote sensing image in detailed, attained the value of after-classification image colour (R, G, B) via emluator, and the Land Use Land Cover changing process. Third part is the application of experiment validation algorithm model and the conclusion, mainly introduced the dynamic monitoring of LUCC by using none linear Least-square method, then according to its numerical value of analyzed result, carried out dynamic simulation predicting of study area and obtained the predicting result picture and curve, finally put forward some existing problems of this paper and some relevant possible solving methods .Take this none linear Least-square method as a powerful regression mathematic model and use to environment field is viable. The study result shows that, while adopting this model to predict the future 10 years variety trend of the study area, and colligation analysis both of the background and the trend of each class of the study erea ,believe that the ecotone in the west part of karkax oasis ,and the bank erea of the two river and east part of the lop oasis is a segment part in very easy change . It is find that oasis, desert area and water body are on the increasing trend, and the ecotone is on the decreasing trend. Acording to the result of that study, the land which important variational would be happens in the future can be find. So the study results have very good using values and meanings to finding the variance is most sensitivity area which in the climate is extreme and the vegetation covering is lower lands in dry lands, analysis the positive and negative influence of the variation, take various protection measures, thus, reasonable exploitation and develop the land and the water resources, keep and optimize the ecological environment, attaining the sustainable development of agriculture and economic.