Urban Land Use Change Simulation And Multi-Objective Spatial Optimization At The Land Parcel Scale

Urbanization is the most significant phenomenon that occurs in human society,which directly affects global land use and landscape patterns.In the process of urbanization,urban lands have been gradually expanding,the natural lands are continuously converted to built-up areas,and the land use structures have become complex.In the past 40 years,unprecedented advancements in the scale and rate of urbanization took place in China.The urban expansion raises a series of problems,arable land degradation,a sharp decrease in the forest,soil erosion,and the greenhouse effect.The New-type Urbanization is proposed in China as a development strategy that aims to improve the quality of urban development.Thus,exploring the dynamic mechanism and reasonable allocation of urban land use patterns are very important.In the process of urban development,land use types are changed with the land parcel as the basic unit,and then produce more complex changes in land use spatial structure changes,and urban functions.Also,the land parcel is the basic unit for urban functional planning and practice management.The implementation of land use regulations is based on the land parcel.Based on this,the government needs to make urban master plans and maintain sustainable development by exploring the mechanisms of land use changes,predicting future land use trends,and realizing multi-objective land use layout optimization at the land parcel scale.Cellular automata(CA)model is one of the most widely used urban development models.The vector-based cellular automata(VCA)model,which takes the land parcel as the basic unit,has been applied by many scholars to simulate urban land use changes.In CA models,the transition probability of each cell consists of four parts: the transition suitability,the neighborhood land-use condition,the constraint coefficient and the stochastic factor.In previous VCA models,the transition suitability of a cell is usually regarded as a local condition,meaning that only the attributes of the driving factors of the cell of interest are considered.However,the attributes of the driving factors within a cell’s neighborhood also have impacts on the suitability of a land use type of the cell,and such neighborhood effects are largely ignored.Ecological control line is one of the important policies for promoting ecological progress in China,which represent the boundaries of ecological control areas(ECAs).Generally,the lands within ECAs are prohibited from being converted to urban construction lands.Later,local governments have promulgated regulations for adjusting ECAs.The plan for adjusting ECAs consists of two parts: “transfer out” and “transfer in”."Transfer out" refers to transferring lands out of ECAs for national,provincial,and city construction projects;"Transfer in" refers to the transferring of lands into ECAs to strengthen ecological services functions.Some studies have identified suitable "transfer-out" lands by screening out crucial ecological areas.However,these studies have only considered the current land-use type of land parcels,while the future urbanization trends have largely been ignored.The purpose of transferring out is to provide more spaces for future urban development where available land resources are too scarce.the future urbanization possibilities of land parcels must be taken as an essential basis when formulating an identification scheme of "transfer-out" lands.Land use spatial optimization aims to allocate land resources and realize the specific planning objectives reasonablely,which is of great significance to urban master plans.Most research on land use spatial optimization lack the comprehensive consideration of land use intensity and urbanizaiton suitability.If only considering land use intensity would lead to the optimization result of a specific area is completely opposite to the actual urbanizaiton trends,which is not desirable in practice.And if only the urbanizaiton suitabilities of lands is considered,the optimization results will not satisfy the objective of intensive development in urban planning,especially for mega-cities with scarce land resources.To address the above issues,this study takes Shenzhen,a mega-city with the most constrained land resources,as the study area,and carries out the following three aspects of work:1)An urban land use change simulation model(CNN-VCA)is proposed that takes into accounts the driving forces driving factors within the land parcel’s neighborhood.By incorporating the CNN method in the VCA model when computing the transition suitability for each parcel,the model can effectively explore the relationships between multiple land use changes and the driving factors at the neighborhood level,thus improving the land use simulation accuracy.2)A framework is proposed for identifying "transfer-out" lands to adjust ECAs based on on the perspective of "high urban development benefit + low ecological value loss".This framework integrates the exploration of future urbanization probabilities and an evaluation of the ecosystem services values of land parcels,which not only provides effectively practical solutions for ECAs adjustments,also serves recommendations for mitigating urbanization on ecological efficiency.3)A land use spatial optimization framework that integrates land use intensity and urbanizaiton suitability at the land parcel scale is proposed.This framework can realize land use spatial optimization for multiple development objectives.Comparing the results of different land use optimization soultions,this study can provide sufficient supports for local urban planning departments.The main research contents and conclusions are as follows:(1)To estimate the effect of the driving factors within the land parcel’s neighborhood on the land use change of the parcel,this study incorporates the CNN method in the VCA model when computing the transition suitability for each parcel.The proposed CNN-VCA model is applied to urban land use changes simulation in Shenzhen from 2009 to 2012.Our results show that the proposed CNN-VCA model can achieve the best simulation results compared with those of other existing VCA models(LR-VCA,ANN-VCA,and RF-VCA)with a result that is 11.8% higher than that of the RF-VCA model.The results indicated that the CNN-VCA model can effectively uncover the neighborhood effects of multiple driving factors on the developmental potential of land parcels and obtain more details on the morphological characteristics of land parcels.(2)This study proposes a framework for identifying "transfer-out" lands that are suitable to be transferred out of ECAs for urban development.This framework integrates the exploration of future urbanization probability and the evaluation of ecosystem services value at the land parcel scale.The identified "transfer-out" land parcels have the characteristics of both high future urbanization probabilities and low ecosystem services values.Our results show that the lands currently available for the urban development of Shenzhen will be exhausted due to the ECL policy by 2041.The identified "transfer-out" lands based on the proposed framework are relatively concentrated in their spatial distribution.Several large areas of contiguous "transfer-out" land parcels are located in the Nanshan and Guangming districts.This study can provide local land resource-concerned governments with scientific support at the practical level for ECA adjustments.The proposed framework for identifying suitable "transfer-out" lands in our study is is flexible and general.(3)This study proposes a land use spatial optimization framework that that integrates land use intensity and urbanizaiton suitability as planning objetives.Firstly,a land use change simulation model CNN-VCA was used to calculate the future urbanizaiton suitability of land parcels.Then,an index of land use intensity considering the spatial aggregation degree of land parcels was constructed,which includes the population density,the aggregation index and the density of points of interests(POI).Finally,both the land use intensity and urbanizaiton suitability are taken as multi-objective spatial optimization.And the 2014 land use was input as the initial state to realize the spatial optimization of Shenzhen in 2030 using the genetic algorithm.Through the comparison of the results of different land use spatial optimization schemes,it is found that the optimization scheme integrating land use intensity and urbanizaiton suitability can improve the intensity of urban land use and functional efficiency on the basis of the current development trend of land parcels.