Most existing cellular automata(CA)models impose strict requirements on the number and spatial distribution of samples.This makes it a challenge to capture spatial heterogeneity in urban dynamics and meet the modeling...Most existing cellular automata(CA)models impose strict requirements on the number and spatial distribution of samples.This makes it a challenge to capture spatial heterogeneity in urban dynamics and meet the modeling needs of large and complex geographic areas.This paper presents a CA model based on geographically optimal similarity(GOS)transition rules and similarly sized neighborhoods(SSN).By comparing the similarity in geographical configuration between samples and predicted points,the model enables a comprehensive characterization of the driving mechanism behind urban expansion and its self-organizing scope.This helps to mitigate the impact of sample selection and assumptions about spatial stationarity on simulation results.The performance of GOS-SSN-CA simulation was tested by taking the urban expansion in the Changsha-Zhuzhou-Xiangtan urban agglomeration in China as an example.The results show that GOS can derive more accurate and reliable urban transition rules with fewer samples,thereby significantly reducing spatial prediction errors compared with logistic regression.Moreover,SSN selects different neighborhood sizes to represent the difference between the local self-organizing range and surrounding cells,thus further improving the simulation accuracy and restricting urban expansion morphology.Overall,GOS-SSN-CA effectively characterizes the geographical similarity of urban expansion,improves simulation accuracy while constraining the urban expansion form,and enhances the practical application value of CA.展开更多
基金National Natural Science Foundation of China,No.41971219,No.41571168Natural Science Foundation of Hunan Province,No.2020JJ4372+1 种基金Key Project of Philosophy and Social Science Foundation of Hunan Province,No.18ZDB015The Graduate Science and Innovation Project of Hunan Province,No.CX20230719。
文摘Most existing cellular automata(CA)models impose strict requirements on the number and spatial distribution of samples.This makes it a challenge to capture spatial heterogeneity in urban dynamics and meet the modeling needs of large and complex geographic areas.This paper presents a CA model based on geographically optimal similarity(GOS)transition rules and similarly sized neighborhoods(SSN).By comparing the similarity in geographical configuration between samples and predicted points,the model enables a comprehensive characterization of the driving mechanism behind urban expansion and its self-organizing scope.This helps to mitigate the impact of sample selection and assumptions about spatial stationarity on simulation results.The performance of GOS-SSN-CA simulation was tested by taking the urban expansion in the Changsha-Zhuzhou-Xiangtan urban agglomeration in China as an example.The results show that GOS can derive more accurate and reliable urban transition rules with fewer samples,thereby significantly reducing spatial prediction errors compared with logistic regression.Moreover,SSN selects different neighborhood sizes to represent the difference between the local self-organizing range and surrounding cells,thus further improving the simulation accuracy and restricting urban expansion morphology.Overall,GOS-SSN-CA effectively characterizes the geographical similarity of urban expansion,improves simulation accuracy while constraining the urban expansion form,and enhances the practical application value of CA.
