Scholarlink

<p>The eastern margin of the Qinghai-Xizang Plateau, as a critical transition zone between the plateau and the Sichuan Basin, poses substantial challenges for geographic regionalization, primarily due to its intricate terrain and climatic heterogeneity. Traditional spatial clustering methods often struggle to balance spatial continuity and attribute similarity, suffering from subjectivity and inadequate representation of topographic complexity. This study proposes a novel mountainous geographic regionalization framework that integrates topographic and climatic characteristics, using Kangding county as a typical case. Principal Component Analysis (PCA) was employed to perform dimensionality reduction on multiple environmental variables and assign relative weights. A Gaussian-weighted function was further applied to adjust attribute distances to capture spatial non-stationarity, while the geographic distance weight was systematically optimized. The partitioning outcomes were evaluated using clustering quality indicators (Davies-Bouldin index, Silhouette index, Calinski-Harabasz index) and spatial autocorrelation indicators (Moran’s I index, Moran’s Z-score). Results indicated that when the number of clusters was set to five and the geographic distance weight was 0.5, the clustering algorithm optimized the trade-off between spatial continuity and attribute similarity (Davies-Bouldin index = 1.14, Silhouette index = 0.30, Calinski-Harabasz index = 25150.91, Moran’s I = 0.97, Moran’s Z-score = 292.28). Compared to the traditional K-means clustering, this approach enhanced intra-cluster similarity (Sil) by 259% and improved spatial continuity (Moran’s I, Moran’s Z-score) by approximately 44%. This method effectively addresses the challenge of coordinating spatial constraints with attribute heterogeneity in mountainous environmental zoning, in a county scale, providing an automated, data-driven approach for geographic partitioning in complex terrains. The findings offer valuable insights for mountain ecosystem management and regional geographic studies. Our study provides a set of effective methods of demarcation of regional boundaries based on raster data, offering important insights for ecological zoning management and regional studies in mountainous environments at a small scale.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing demarcation in regionalization in the eastern Qinghai-Xizang Plateau through geographically weighted

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月19日

查看全文
  • Xianglong Liu,
  • Desheng Hong,
  • Hongyang Dong,
  • Hangtao Du,
  • Xiaoguo Wang,
  • Bo Zhu

摘要

The eastern margin of the Qinghai-Xizang Plateau, as a critical transition zone between the plateau and the Sichuan Basin, poses substantial challenges for geographic regionalization, primarily due to its intricate terrain and climatic heterogeneity. Traditional spatial clustering methods often struggle to balance spatial continuity and attribute similarity, suffering from subjectivity and inadequate representation of topographic complexity. This study proposes a novel mountainous geographic regionalization framework that integrates topographic and climatic characteristics, using Kangding county as a typical case. Principal Component Analysis (PCA) was employed to perform dimensionality reduction on multiple environmental variables and assign relative weights. A Gaussian-weighted function was further applied to adjust attribute distances to capture spatial non-stationarity, while the geographic distance weight was systematically optimized. The partitioning outcomes were evaluated using clustering quality indicators (Davies-Bouldin index, Silhouette index, Calinski-Harabasz index) and spatial autocorrelation indicators (Moran’s I index, Moran’s Z-score). Results indicated that when the number of clusters was set to five and the geographic distance weight was 0.5, the clustering algorithm optimized the trade-off between spatial continuity and attribute similarity (Davies-Bouldin index = 1.14, Silhouette index = 0.30, Calinski-Harabasz index = 25150.91, Moran’s I = 0.97, Moran’s Z-score = 292.28). Compared to the traditional K-means clustering, this approach enhanced intra-cluster similarity (Sil) by 259% and improved spatial continuity (Moran’s I, Moran’s Z-score) by approximately 44%. This method effectively addresses the challenge of coordinating spatial constraints with attribute heterogeneity in mountainous environmental zoning, in a county scale, providing an automated, data-driven approach for geographic partitioning in complex terrains. The findings offer valuable insights for mountain ecosystem management and regional geographic studies. Our study provides a set of effective methods of demarcation of regional boundaries based on raster data, offering important insights for ecological zoning management and regional studies in mountainous environments at a small scale.