Arid Zone Research ›› 2025, Vol. 42 ›› Issue (2): 321-332.doi: 10.13866/j.azr.2025.02.12

• Ecology and Environment • Previous Articles     Next Articles

Ecological restoration zoning based on the GeoSOM network: A case study of the Shanxi section of the Yellow River Basin

LI Hao1,2(), ZHANG Lei1,2, LIANG Xiaolei1,2, LIU Geng2,3()   

  1. 1. School of Economics and Management,Taiyuan Normal University, Taiyuan 030619, Shanxi, China
    2. Shanxi Key Laboratory of Earth Surface Processes and Resource Ecology Security in Fenhe River Basin, Taiyuan Normal University, Taiyuan 030619, Shanxi, China
    3. School of Geography Science, Taiyuan Normal University, Taiyuan 030619, Shanxi, China
  • Received:2024-08-29 Revised:2024-10-17 Online:2025-02-15 Published:2025-02-21
  • Contact: LIU Geng E-mail:lihao666@tynu.edu.cn;liugeng9696@126.com

Abstract:

Ecological restoration is a major project to implement China’s ecological civilization construction. Delineating zoning units is a prerequisite and an essential foundation for the differentiated implementation of land remediation and ecological restoration; it holds significant theoretical and guiding value for formulating differentiated restoration measures. Taking the Shanxi section of the Yellow River Basin as an example, this study introduces the GeoSOM (Geographic Self-Organizing Map) algorithm to perform spatial clustering for ecological restoration in the study area based on grid units. The Dunn index evaluated the effectiveness of the clustering and selected the optimal scheme. Finally, the Support Vector Machine (SVM) identified the boundaries of the ecological restoration zones, resulting in the delineation of the ecological restoration areas. The results indicate that, after the GeoSOM network spatial clustering, the study area was divided into four major categories according to the Dunn index evaluation, with each category exhibiting significant spatial differentiation characteristics. Based on the clustering results, the SVM identified ten ecological restoration zones, and ecological restoration strategies were proposed for each zone. This study improves the traditional SOM network, which emphasizes thematic attributes, by using the GeoSOM algorithm that measures the similarity of thematic and spatial attributes, making it more suitable for spatial clustering. The findings provide a new reference for methods of ecological restoration zoning.

Key words: GeoSOM network, ecological restoration zoning, boundary identification, Shanxi section of the Yellow River Basin