| Land, as the most important part of the land system, is the link and bridge in the study of relationship between human and environment. As many global problems are the intensifying such as the growing population, the decreasing biodiversity, ozone depletion and loss, the crisis of water resources and the destruction of the marine geology, the land has become the most active field in many international researches. And the study of land use and land cover change is the key target which has become one of research hot spots in global academy. Besides, the accelerating global urbanization cause the change of land use or land cover which leads to the change of the regional society, economy and environment and thus producing more complex evolution process of land use pattern and function that causes more change of the natural phenomena and ecological processes such as soil properties, surface runoff and erosion, etc. Therefore, on the basis of the comprehensive analysis of the change of regional land use, it is of realistic and theoretical significance to explore the driving force factors affecting the change of land use, construct the dynamic simulation model of land use pattern, reconstruct the past and predict the future land use pattern and explore the best model of regional land use and in order to grasp the future trend of the regional land use and develop ecological construction of regional land.This thesis takes Hunchun City as its research area and makes a multi-angle analysis of the land use changes of Hunchun City in recent years based on the model of local spatial autocorrelation, the transfer matrix and the space gravity migration by the data sources of the land use change, DEM and grid GDP and so on. On this basis, this thesis has made an overlay analysis of the land use change data and built the matrix of sample data and driving factors of the land use change by using the method of typical corresponding analysis to analyze the relationship between the land use change and driving factors, mainly to explore the extent of the effects of driving factors on the land use change. According to the selecting and setting of driving factors, the ANN- CA model has been constructed. On the basis of the model calibration and accuracy test, the thesis, taking the land use data in 2014 as the initial value, simulates the land use pattern of Hunchun City in 2020 in order to provide theoretical reference for the future land use manner of Hunchun City. The main research conclusions are as follows:(1) Through the spatial autocorrelation model, land use transfer matrix model and space center migration model, an analysis and study of the present situation of the land use in Hunchun City has been made and it draws a conclusion that all types of land of Hunchun City is positively correlated on the spacial location. The related degree is the forest land, the cultivated land, the construction land, the other land, water, grassland, the garden land. All types of land in space is mainly distributed as that the cultivated land is scattered, the forest land is concentrated in the northeast and north and the garden land is distributed randomly. Nearly five years, all types of land are converted to construction land actively. Especially the cultivated land and the forest land are larger, and the effective complement of the cultivated land is insufficient, causing the quantity of the cultivated land of Hunchun City presents the situation of “living beyond its means” in recent years. The gravity center of the construction land shifts from the southeast to the northwest gradually with the migration distance of 395 meters and a particularly evident direction.(2) The thesis researches correlation between the land use type and driving factors by the typical correspondence analysis and the analysis results show that the GDP, the slope, the policy and the DEM has become the main driving factors of the land use change of Hunchun City, followed by distance from the road, distance from the river and distance from the town. Besides, under the condition of the limited information and the shorter year interval, the CCA distribution can better reflect, under the restriction of driving factors, the distribution pattern of land use types, and directly display the influence of driving factors and land use change.(3) After the integration of the ability to deal with nonlinear relationship by the ANN model and CA model space expression ability, the thesis builds the ANN-CA model, sets up six kinds of different parameter combinations and simulates the land use of Hunchun City in 2014. The results show that under the parameter combination of a=1 and T=0.95, all types of land has a higher simulation precision. By the quantity precision testing, Kappa coefficient verification and Kappa index series of verification, the verification results show that the quantity precision were all higher than 75%, the comprehensive index of the Kappa is 0.9840 and the Kappa series of index is respectivelyKstandard=0.992,Kno=0.995,Klocation=0.998,KlocationStrata=0.998, which are higher than 75%, indicating that the simulation grid figure in the study area is in a high consistency with the actual grid figure which has higher credibility.(4) By using the trained ANN-CA model, this thesis simulates the 2020 land use situation in the studied area which shows that the forest land, the arable land, the grassland and the other land all present the decreasing trend, of which the forest land and the cultivated land decrease much, reducing 1087.92 hectares and 929.97 hectares respectively. The garden land and the construction land show the increasing trend, of which the construction land increase more by 3094.92 hectares. From the point of spatial pattern of the land use change, the newly-increased construction land is mainly distributed in the relatively flat terrain near the river in Hunchun City. The construction land is distributed largely in the new district, the border zone and the Jingxin town. |
PDF Full Text Request