Simulation Of Spatial Change Of Land Use Pattern In Yu-Shen-Fu Area Based On CLUE-S Model

Land use and land cover change (LUCC) is an important part of global change research, and analysis and simulation of regional land use change is becoming a hot spot in the study on LUCC. Regional land use change is a complex geographic process, through the analysis of driving force, land use change model is set up. It can illuminate the cause, process and development trend of regional land use change, which is also a important means to study the effects of land use and land cover change.Based on the current widely used CLUE-S model in the world, and taking Yu-Shen-Fu area of the resource-based city Yulin as the research area, the studies have been carried out and the main ideas are as follows:first of all, the trend prediction method, gray GM (1,1) prediction method and linear regression prediction method have been taken for the fitting on the demand of land use from 2000 to 2010 in research area, and the minimum error of the simulation demand of land use from 2010 to 2020 has been selected. Then the Logistic Bivariate Regression Model is used to calculate regression equation between the land type in research area and driving factors which include nature, economy, society and resources. By the use of ROC curve to test the accuracy of multi-scale regression, and it make out the optimum scale of simulation effect. Finally, according to the input rules of CLUE-S model, the spatial distribution of land use in study area in 2020 can be simulated. The results shows that model’s simulation effect of spatial distribution of land use in study area in 2020 is good, the results can also be a more scientific basis for the land use decision-making in study area. Main research work and conclusions are as follows:(1) For the land demand module of the model, using the trend prediction method, gray GM (1.1) prediction method and linear regression prediction method to simulate and predict land demand in the study area, the forecast amount of land demand from 2010 to 2020 can be determined. The results show that for the land demand in study area, the arable land, grassland, land for construction of urban and rural residents and unused land can obtain a better effect of forecast by linear regression model, and the forecast method of better effect for water is grey GM (1,1) prediction model, for forest land is quadratic polynomial prediction.(2) With the interpreted land use map, the driving factors of land use change have been extracted which conclude farmland, rivers, reservoirs, such as road, industrial and mining areas, vegetation, towns and villages. Then coupled with slope, slope direction of DEM and population, GDP and other factors of social economic data. Through the binary Logistic regression analysis, it is concluded that the ROC inspection result of land type is ideal, and the driving factors and the land type has a high correlation, so the fitting effect is good.(3) The multi-scale analysis method has been used to calculate the simulation precision of land use in study area, the results show that the ROC values are higher than 0.8 on each scale which is a relatively high accuracy, it illustrates that the selection of driving factors can effectively simulate the land use change, and for the study area, the high precision can be obtained on the scale of 500* 500.(4) Further comparing the simulation results and interpretation chart, a certain deviation of displacement will be find, which can explicate two concepts of the total amount of increase or decrease in quantity and total transfer changes in quantity. The simulation results show that the CLUE-S model can better remove the displacement deviation caused by error of interpretation which is helpful for achieve more accurate simulation results.(5)The simulation results show that the cultivated land, grassland, forest land, water, residents of urban and rural construction land and unused land class in the spatial layout on the simulation results conform to the development of resource-based cities characteristics. Simultaneously, there are also exist some problems to be solved, so the area of land use in the future should focus on the following three points:the first is to promote the land management work; the second is scientific research on land development; the last is to actively perfect ecological construction.