Spatial-temporal Variation Simulation For Urban Land Growth At The Later Stage Of Urbanization

Urbanization is inevitable when a particular stage of social and economic development has reached.Since the reform and opening up,China's urbanization experiences rapid development and has achieved great progress,becoming the driving power to promote economic growth.Urbanization brings the industry innovation and the transformation of economic development mode,helps to promote balanced development among regions,is a major strategic choice of accelerating socialist modernization in our country.The increase of urban population and economy development during urbanization create a strong force to the expansion of urban land.However,early extensive growth mode has raised a series of problems about resources,environment and ecology,restraining the sustainable development of city.At present,some big cities have gradually entered the late stage of urbanization,the contradiction between urban development demand and resource supply is becoming more and more prominent.The urban development mode needs to pay attention from the scale expansion to the emphasis on the intensive utilization of resources.In the later period of urbanization,urban extroverted expansion lack of motivation,land urbanization process slow down,infrastructure such as transportation highlight the role of urban space.In this context,it is of great significance to master the basic characteristics and objective rules of urban expansion and carry out multi development scenarios,to effectively guide the future layout of urban land and guide the coordinated development of cities.This study takes Shenzhen as a example,which is the late urbanization are,integrates land use data,planning data and related economic and social data,bases on the principles and methods of GIS,spatial syntax,spatial statistical analysis and neural network cellular automata model,carry out urban growth characteristics analysis and simulation research.The main contents and conclusions are as follows:(1)Spatial and temporal change analysis of urban land and traffic accessibilityThis paper analyzes the spatial growth characteristics of urban land in Shenzhenfrom three perspectives:scale,structure and form.From 2000 to 2014,Shenzhen urban land transited from rapid expansion to low growth mode.The growth space mainly distributed outside the original special zone,the degree of heart reduced and plaque shape became more complex.During this period,the urban traffic network structure is more perfect,and the road network accessibility index based on spatial syntax is improved,and the city reachability is promoted.Along with the layout of road network,regional accessibility also shows the characteristic that the western higher than east.The spatial and temporal variability of urban land use intensity and accessibility are,explored by bivariate spatial correlation analysis.The results show that there is a certain spatial positive correlation between them,and the relevant characteristics are significant.The data envelopment analysis of land use and traffic accessibility shows that outside the original special zone,especially the urban new area,land and traffic coordination is better.(2)ANN-CA urban land growth model constructionCombining artificial neural network algorithm and cellular automata model,the simulation model of urban land use growth is constructed.Considering traffic accessibility,natural conditions,location conditions,planning factors,social activitiesand neighborhood effects,automatic acquisition of land use conversion rules through neural network training is beneficial to improve model efficiency and simulation accuracy.Based on the land use data from 2009 to 2014,the layout of urban land use in Shenzhen City in 2015 was simulated and compared with the land use data of real year.The combination of the random interference factor and the transition probability threshold is used to evaluate the model and the accuracy is evaluated.The simulation results are in good agreement with the real layout,verifying the validity of the ANN-CA model.The best combination of parameters(? ?1,T=0.8)can be applied to future urban land prediction research.(3)Scenario simulation for urban land growthBased on the ANN-CA urban land growth model,considering land use change characteristics and urban development planning,the paper designed three urban development scenarios as trend continuation(scenario 1),traffic guidance(scenario 2)and spatial optimization mode(scenario 3),respectively through the urban self-organization law,road accessibility and planned rail transit network to achieve different future space growth.The simulation results of urban land use growth in 2020 and 2030 are compared and analyzed from the scale,layout and landscape pattern.In the recent simulation stage(2020),Scenario 1 and Scenario 2 urban land use spatial growth characteristics are similar,land grow preferentially in the midland,while the scenario 3 focuses on the eastern part of the city.In the long-term simulation(2030),scenario 1 is consistent with scenario 3,with the largest growth in the eastern region and the scenario 2 still in the midland.It is concluded that the comprehensive traffic system has an important supporting role in urban spatial structure.Urbanization in the late,urban growth on the one hand by the city planning guidance,on the other hand by the spatial resources constraints,the two combined role formulate a certain spatial development characteristics.