The Study Of Simulation On Land Use Spatial Pattern Based On CLUE-S Model

Regional land use is a more complex process, establish an important means of landuse change models are in-depth understanding of the causes of land use change, processesand future trends. The hexi corridor region in Gansu province as the country’s major grainproducing areas, relatively fragile ecological environment, and unreasonable land use waysin recent years makes the ecological environment worsens, and constituted a certainthreat.in hexi corridor even the whole Gansu province ecological security and sustainabledevelopment. Based on this background, this paper selects the zhangye city is located in thehexi corridor as the study area,on the basis of2000、2010and2013third phase of the landuse data，Introduction of changes in land use cover widely used analog CLUE-S land usedynamic change model，land use dynamic evolution based on the analysis of the pastempirical, simulate the future of land use spatial pattern,the results prove that:(1) From the number of structural change, between2000-2010, area of constructionland and arable land has increased, the woodland、grassland、water and unused land isreducing，Land diversity index of each county at around0.6indicates that the counties inthe city land use type is relatively stable,from the intensity of land use dynamic degree anduse area, LinzeXian is higher than the other counties; From the point of space structurechanges, Zhangye city various types of land use patch shape index and aggregation indexbetween2000and2010year, reflects the various land use types in the shape of patch didn’thappen big changes，But due to the effect of human, compared to other land use types ofcultivated land and construction land concentration increased.(2) In the non-spatial CLUE-S model analysis module, build the GM gray predictivemodels and DGM discrete gray model, And build a combined model of gray entropy theoryfuture land requirements for validated measure, The average prediction error from the errorterm entropy model portfolio0.79%, lower than the error of gray prediction model DGM0.95%and the average error of1.43%discrete gray model; From the verification results,the difference between the posterior Entropy Portfolio Model C and the minimum errorfrequency ratio P respectively0.1161and0.998, the prediction accuracy is better. Entropycan be seen that the combination model can be well applied in the analysis of land use demand estimates.(3)Taking into account the distribution of land use are not the same at differentscales,This paper selects the100m,200m,300m,500m,700m five kinds of spatial scaleuse Logistic regression analysis each land use type and driving factors, Found show that thevarious types of landuse in Zhangye City of ROC is the largest when the time scale is500m,the research show that when the spatial scale is500m the Logistic regression can betterexplain Zhangye City land use spatial distribution(4)With the optimal scaling, through a combination of gray entropy model to calculatethe future land requirements in Zhangye City, and simulate2013and2020the land usespatial pattern with CLUE-S model, test and verify the2013land use data after thesimulation, the kappa index reached87.44%, ROC values for each type of land use aregreater than0.5, research show that the CLUE-S model can be better applied in the oasisregion of Hexi Corridor arid zone of the municipal-scale land use spatial simulation.