关键词: 土地生态安全, 机器学习, 优化策略, 武汉城市圈

Abstract: The purpose of this research is to analyze the land ecological security pattern and the influencing factors in Wuhan Metropolitan Area, and to provide comments on the regional land ecological construction and sustainable development. Based on the evaluation system of land ecological security, BP neural network in the machine learning and GIS spatial analysis method, this paper explored and identified the status of land ecological security from 2010 to 2017 in Wuhan Metropolitan Area, and to propose the optimization of land ecological security pattern. Results show that:(1) The overall level of land ecological security of Wuhan Metropolitan Area declined slightly from 2010 to 2017, and had been in a critical safe state for a long time. Among them, the status of land natural ecological security was deteriorating, the status of land economic ecological security continued to improve, and the level of land social ecological security changed differently. (2) There were obvious spatial differences in the changes of land ecological security in Wuhan Metropolitan Area:the areas where the level of land natural ecological security decreased greatly were mainly concentrated in the west and south; the level of land economic ecological security of Huanggang and Huangshi decreased slightly, and the security level of other cities increased at different degrees; the level of land social ecological security changed greatly:Wuhan and Qianjiang increased by one security level, Xiantao and Xiaogan decreased by one security level, and Xianning decreased by two security levels. (3) The main reasons for the deterioration of the land natural ecological security are that the per capita cultivated land, the proportion of cultivated land, and the forest coverage rate continue to decline, while the total discharge of industrial and agricultural wastewater and wastes continue to increase; the main reasons for the continuous improvement of land economic ecological security are that industrial and agricultural production efficiency, residents' income, and the proportion of tertiary industry continued to rise; there are many factors affecting the land social ecological security system.

Key words: land ecological security, machine learning, optimization, Wuhan Metropolitan Area