Simulation Of City Spatial Expansion Based On Geographically Weighted Regression And CLUE-S Model

City is a complex dynamic system,with open,dynamic,self-organization,dissipative structure characteristics of non-balance etc.City is the inexorable trend of social economic development,the rapid expansion of city has brought the city space,a large number of agricultural land occupied,also brought a series of problems such as the urban heat island effect,environmental pollution,environment worsening and so on.It's a hot question for researchers studying experts at home and abroad.The development of a city is influenced by many aspects,such as society,economy,culture,science and technology.City development process with high complexity.The traditional prediction method for urban land use scale is based on historical data.According to the data obtained to forecast urban spatial scale,but can't get the direction of city expansion.The traditional model visible already could not meet the needs of current research,and research on dynamic model of the city has become a current research trend.Hanchuan city is located in Wuhan City Economic Circle of the core layer.It's an important base to undertake the Wuhan Economic radiation,with the industrial base,better comparative advantage,external conditions and development space.In recent years,Hanchuan city is in an important period of resource transformation.With the rapid increase in population and the development of the city,urban area expanded rapidly.So the expansion of the city of the future to make planning decision reasonable,has the significant significance.This paper takes Hanchuan city as the research object,take the second national land survey data and change data as the basic data,according to the data extracted from the land and the non-construction land data construction.Using CLUE-S model to simulate the expansion of city space.At first,the simulation process is on the Logistic regression model and GWR model are compared with the calculation of probability distributions.Then use the GM(1,1)model to forecast the demand for land,and set the related parameters.Finally,the model operate successfully.Comparing the simulation results and the actual data,get the direction of city expansion,provide reference for the future direction of development of Hanchuan city.Draw the following conclusions:(1)For city spatial expansion model choice,most experts selected the CA(Cellular Automata)model and its extended model.The CA model can simulate the dynamic changes in the time space of the city,have better spatial expression.But the model attaches importance to biophysical factors influence on city expansion,played down the impact of human activities on the development of the city.CLUE-S model in the research area according to the size of the grid,through spatial analysis module and non-spatial analysis module matching to predict study area city spatial expansion.CLUE-S model has more advantages than the CA model.The model can integrate regional biophysical,population,technology,affluence,market,economic conditions and the attitude of value orientation factor of human activity as the driving factor.Policy factor can also be difficult to consider the general model into the model.This study attempts to predict attempts to extend the space has certain reliability and higher resolution capability using CLUE-S model.(2)This paper uses the Logistic model and GWR model to calculate the land distribution and driving factor analysis(the distance to the town,away from the highway,the distance to the railway station,away from the main road distance,distance to the waters,population distribution,GDP index,height,slope,aspect),and calculate the land use and non-construction land probability distribution map construction.The results will be imported into the SPSS software and tested for consistency,the former ROC value is 0.681 and the latter ROC is 0.896.The curve of ROC value is 0.5-1,its value is close to 1 indicates the consistency is better,if the value of ROC equal to 0.5 is meaningless.The ROC values calculated by GWR model are larger and closer to the 1.The GWR model can relationship between distribution and driving factors for better analysis.Logistic regression analysis model based on global regression,but the effect of driving factors on land use pattern with spatial instability.GWR model is different.The regression coefficients of independent variables is along with the spatial position change and the local analysis ability for the spatial data is very.strong.In conclusion we can see that the GWR model is better than Logistic regression model.Probability distribution calculated by the GWR model,will be taken as input.(3)The paper constructs the CLUE-S model and realized the simulation of Hanchuan city spatial expansion.The CLUE-S model was adopted to the raster data to calculate.Taking the area of land use change in the future as the goal,through the iterative implementation of future land use spatial distribution area,implement the land use spatial visualization expression predicts.City spatial expansion analysis chart obtained by overlay analysis on the map and current situation of simulation graph with raster calculator tool of ArcGIS.Thus can be seen that the spatial expansion direction of Hanchuan city in the future:the main groups in the old urban fairy Street office,New River Development Zone,Ma town and along the 106 provincial way Chenhu town and Fenshui town.