Modeling urban encroachment on ecological land using cellular automata and cross-entropy optimization rules

Chen Gao; Yongjiu Feng; Xiaohua Tong; Yanmin Jin; Song Liu; Peiqi Wu; Zhen Ye; Cairong Gu

doi:10.1016/j.scitotenv.2020.140996

Modeling urban encroachment on ecological land using cellular automata and cross-entropy optimization rules

Sci Total Environ. 2020 Nov 20:744:140996. doi: 10.1016/j.scitotenv.2020.140996. Epub 2020 Jul 18.

Authors

Chen Gao¹, Yongjiu Feng², Xiaohua Tong¹, Yanmin Jin³, Song Liu⁴, Peiqi Wu¹, Zhen Ye⁵, Cairong Gu¹

Affiliations

¹ College of Surveying & Geo-Informatics, Tongji University, Shanghai 200092, China; The Shanghai Key Laboratory of Space Mapping and Remote Sensing for Planetary Exploration, Tongji University, Shanghai 200092, China.
² College of Surveying & Geo-Informatics, Tongji University, Shanghai 200092, China; College of Architecture & Urban Planning, Tongji University, Shanghai 200092, China. Electronic address: yjfeng@tongji.edu.cn.
³ College of Surveying & Geo-Informatics, Tongji University, Shanghai 200092, China.
⁴ College of Architecture & Urban Planning, Tongji University, Shanghai 200092, China.
⁵ College of Surveying & Geo-Informatics, Tongji University, Shanghai 200092, China; The Shanghai Key Laboratory of Space Mapping and Remote Sensing for Planetary Exploration, Tongji University, Shanghai 200092, China. Electronic address: 89_yezhen@tongji.edu.cn.

PMID: 32947762
DOI: 10.1016/j.scitotenv.2020.140996

Abstract

Rapid urban expansion often leads to substantial encroachment on ecological lands and destruction of natural environments. We developed a new cellular automata model (named CA_CEO) that uses cross-entropy optimization (CEO) to reproduce and project urban expansion into coastal areas and to assess urban encroachment on ecological lands. The CEO algorithm automatically searches for the near-optimal CA parameters and is capable of objectively parameterizing CA models to predict multi-objective scenarios. We calibrated CA_CEO by simulating urban expansion at Wenzhou from 1995 to 2005, validated the model from 2005 to 2015 using real data, and then predicted urban expansion for 2025 and 2035. End-state overall accuracies were 93.8% for 2005 and 94.4% for 2015, while figure-of-merit metrics were 27.9% for 2005 and 19.1% for 2015. We predicted four different scenarios to year 2025 and 2035: (1) a business-as-usual (BAU)-scenario using benchmark settings; (2) a District-scenario based on a district-oriented urban development strategy; (3) a Road-scenario based on a road network-oriented urban development strategy; and (4) a Coast-scenario based on a coast-oriented urban development strategy. Each scenario predicts a substantially different pattern of urban encroachment on ecological land and significant loss of farmland, forest, wetland and grassland. These scenarios should be useful in adjusting urban development strategies at Wenzhou and elsewhere.

Keywords: Cellular automata (CA); Coastal areas; Cross-entropy optimization (CEO); Scenario prediction; Urban encroachment.

MeSH terms

Algorithms
China
Conservation of Natural Resources*
Entropy
Forests
Urban Renewal*