| Regional environmental change research is now one of the hot spots of the ecological environment change, and regional scale of land use/cover change is an important part of global land use/cover change. With the development of economy, increase of population, as well as the advancement of urbanization, the unreasonable land use caused a lot of land waste, vegetation reduction, soil erosion, and land desertification, which makes the ecological environment increasingly hostile and threatens human being's living environment. Therefore, it is significant to study and analyze the influence of land use/cover change on the local ecological environment from the perspective of regional scale.In this paper, Huidong County as the study area, we used satellite remote sensing images which got on the USGS include landsat5 images in 1990?landsat7 images in 2002?landsat8 images in 2015 as data source and obtained three phases land use data after processed.Based on transfer matrix, annual change rate of single-type land use, annual change rate of multi-type land use, and comprehensive index of land use,we analyzed the effects and driving force of the change of land use. We also analyzed the change of land use change on the ecological environment through using regional ecological environment quality index model of land use. We uesd logistic regression analysis method to analysis the probability distribution of land use change throug the driving factors that we chose. Based on the land use data from 2002 to 2015,the response functions of land use type area was obtained by using grey system prediction model GM(1,1) in MATLAB and predicted the area of land use types from 2016 to 2020.We simulated the spatial distribution of land use type of 2015 by using CLUE-S model and predicted the spatial and temporal patterns of land use type of 2020 in the study area.The results are as following:(1) By using the landsat5, landsat7, and landsat8 remote sensing data obtained on USGS, we obtained data of land use during the three periods through supervised classification and visual interpretation and the accuracy of the interpretation results was verified.(2) In terms of single-type land use, from 1990 to 2002, annual change rate of building land for urban and rural residents, which is 17.55%, is the biggest. Change rates of cultivated land and water area, which are 8.53% and 6.03% respectively, rank the second. From 2002 to 2015, with the rate of 20.02%, annual change rate of building land for urban and rural residents also witnessed the biggest change, but the change rate of water area increased slowly, with a change rate of only 1.57%. In terms of multi-type land use, the annual change rate is 1.57% from 1990 to 2015. In terms of the degree of land use, comprehensive indexes for land use are 225.56, 252.55, and 260.79 respectively.(3) Ecological environment quality was analyzed through using ecological environment quality index model and ecological contribution rate for change rate of land use. The results show that the ecological environment of Huidong County was serious since 1990,12 years later. However, it was improved since 2002,13 years later due to the “returning farmland to forests” policy in Huidong County.(4) In terms of Logistic regression analysis of the driving factor for land use, on the basis of an analog scale of 100×100m, we selected 7 driving forces, which are elevation, gradient, distance from the cities, distance from the main roads, distance from the rivers, distance from ponds, and population density, and got the different degrees of influence of these factors on land use. It is suggested that gradient and population density have the greatest influence on land use. Using ROC Curve and Logistic regression analysis in SPSS, we got the ROC values for cultivated land, grass bush, forest, water area, building land for urban and rural residents, and unusable land, they are: 0.702, 0.723, 0.749, 0.712, 0.864, and 0.789.(5) In terms of the need for land use, we got the areas for different types of land use from 2002 to 2015 through using liner interpolation. And on this basis, we also got response function of different types of land use for 2016 to 2020 through using gray system prediction model GM(1, 1) in MATLAB.(6) We stimulated the space distribution of the types of land use through using CLUE-S model based on the land use of 2002. The result was checked by Kappa index, which is 0.951, indicating that it conforms to the stimulation result. As a result, we can conclude that CLUE-S model can stimulate the space distribution of land use in the studied area. A predication of the land use in 2020 was also made. |
PDF Full Text Request