Researches On Land-Use Change Of Hangzhou City Based On CA-Markov Model

Today, China is in the stage of rapid urbanization. The urban land expansion has brought a series of problems such as increasing arable land, reducing biodiversity and environmental pollution,etc. It is the premise and foundation of realizing sustainable land use to analyze and forecast regularity of urban land expansion and evolutionary by making research on urban land using changes. With Hangzhou urban area defined as the research object, cellular automata and Markov model related theory as the main method, this paper makes reasonable prediction of land for construction land by using ERDAS9.2 software to preprocess and classify the remote sensing image data of Hangzhou city in the year 1994,2004 and 2013,by using ArcGIS9.3 software to do area statistics and data transformation, by using CA-Markov model realized by IDRISI 17.0 to study development evolution law of Hangzhou urban construction land.This paper’s main aspects go as follows:（1） The data processing for land use type.By using ERDAS, the remote sensing images of Hangzhou in 1994,2004 and 2013 are unified coordinate system for WGS₁984_UTM_Zone₅0N, improved the geometric precision correction through quadratic polynomial correction method. By using ArcGIS software, Hangzhou urban area administrative boundaries are converted as vector quantization. With the administrative boundaries graph defined as mask, the images after correction are cut by using ArcToolbox toolkit. And then, six types of classification templates for forest land, cultivated land, garden land, other land, land for residential areas and mining and water are extracted by image interpretation through maximum likelihood method in ERDAS human-computer interaction supervised classification. Six types of classification templates for land use status quo are generated after precision evaluation. After classifying, the land utilization present situation chart of the year 1994,2004 and 2013 is converted into Coverage format firstly and the coverage format is then converted into shp format by ArcGIS software.（2）The analysis for land use changes.By using the ArcGIS software, six kinds of land use types for forest land, cultivated land, garden land, other land, land for residential areas and mining and water in Hangzhou city are analysed from the aspects of quantity, dynamic and spatial to study the trend of land use change.（3） The simulation study for land use.Various types of suitability figures are acquired by using ArcGIS to convert shp date into raster data which will be made RECLASS、DISTANCE and FUZZY processing in IDRISI. And then,the corresponding land use transfer matrix area is made by Markov modules in IDRISI.Finally, on the basis of the images in 2004,the land use status of Hangzhou in 2013 is predicted by using CA-Markov model through the transfer matrix of land use area from 1994 to 2004 and the land use change suitability of atlas in 2004.From two aspects of quantity and spatial accuracy, this paper verifies the accuracy between the forecast figure in 2013 and the actual area figure to check the CA-Markov model prediction accuracy.Among them, the key is the fabrication for land suitability atlas. With construction land, water, and the basic farmland chosen as the constraint factor, construction land, water, communication, towns and slope as a limiting factor, six types of land use suitability images for forest land, cultivated land, garden land, other land, land for residential areas and mining and water are made respectively, and then, these images will be converted into land use change image synthesis suitability atlas.（4） The prediction research for land use.On the basis of the images in 2013,the land use status of Hangzhou in 2020 which will provide scientific basis for rational planning of land use pattern in Hangzhou is predicted through the transfer matrix of land use area from 2004 to 2013 and the land use change suitability of atlas in 2013.