Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China
Rapid urbanization has intensified pressure on regional ecosystems, constraining sustainable development. Constructing a scientific ecological zoning framework is essential for environmental protection and refined territorial spatial management. Taking Chengdu, China, as a case study, this study dev...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | Ecological Indicators |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1470160X25007174 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849427088233201664 |
|---|---|
| author | Xiaobin Huang Xiaosheng Liu Yuanhang Jin Xue Gao Youliang Chen |
| author_facet | Xiaobin Huang Xiaosheng Liu Yuanhang Jin Xue Gao Youliang Chen |
| author_sort | Xiaobin Huang |
| collection | DOAJ |
| description | Rapid urbanization has intensified pressure on regional ecosystems, constraining sustainable development. Constructing a scientific ecological zoning framework is essential for environmental protection and refined territorial spatial management. Taking Chengdu, China, as a case study, this study develops an ecological zoning framework based on the eXtreme Gradient Boosting-SHapley Additive exPlanations (XGBoost-SHAP) model. The framework integrates Ecosystem Service Value (ESV) and Landscape Ecological Risk (LER) as core indicators, applies Z-score standardization and quadrant classification to delineate four ecological zone types with distinct ecological functions, and further employs the XGBoost-SHAP model to identify key natural and anthropogenic drivers and explain their roles in spatial environmental evolution. The results show that: (1) Farmland and forest were the dominant land use types, accounting for over 87 % of the total area. From 2000 to 2020, farmland decreased by 11.64 %, while ecological land increased by 219.68 km2. (2) ESV followed a “rise–decline–rise” trend with a net decrease of 0.752 billion CNY and a spatial pattern of “high in the northwest, low in the center.” LER exhibited a “low in the northwest-central, high in the southeast” pattern, with low-risk areas gradually expanding. (3) The Ecological Control Zone (ECZ) and Ecological Improvement Zone (EIZ) dominated the study area, covering over 9,591 km2. At the same time, the Ecological Rehabilitation Zone (ERZ) and Ecological Conservation Zone (ECOZ) showed stable growth driven primarily by natural factors. (4) The XGBoost-SHAP model demonstrated high interpretability and attribution accuracy, effectively revealing the driving mechanisms behind zoning evolution. This ecological zoning framework is refined, interpretable and data-driven, providing scientific support for spatial planning and sustainable ecosystem management in rapidly urbanizing regions. |
| format | Article |
| id | doaj-art-144d91babe5a4ca8b465307af4e99465 |
| institution | Kabale University |
| issn | 1470-160X |
| language | English |
| publishDate | 2025-08-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Ecological Indicators |
| spelling | doaj-art-144d91babe5a4ca8b465307af4e994652025-08-20T03:29:09ZengElsevierEcological Indicators1470-160X2025-08-0117711378710.1016/j.ecolind.2025.113787Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, ChinaXiaobin Huang0Xiaosheng Liu1Yuanhang Jin2Xue Gao3Youliang Chen4School of Civil and Surveying & Mapping Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, China; The Engineering & Technical College, Chengdu University of Technology, Leshan 614000, China; Southwestern Institute of Physics, Chengdu 610041, ChinaSchool of Civil and Surveying & Mapping Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, China; Jiangxi Province Key Laboratory of Water Ecological Conservation in Headwater Regions, Ganzhou 341000, China; Corresponding authors.School of Civil and Surveying & Mapping Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, ChinaThe Engineering & Technical College, Chengdu University of Technology, Leshan 614000, China; Southwestern Institute of Physics, Chengdu 610041, ChinaSchool of Civil and Surveying & Mapping Engineering, Jiangxi University of Science and Technology, Ganzhou 341000, China; Jiangxi Province Key Laboratory of Water Ecological Conservation in Headwater Regions, Ganzhou 341000, China; Corresponding authors.Rapid urbanization has intensified pressure on regional ecosystems, constraining sustainable development. Constructing a scientific ecological zoning framework is essential for environmental protection and refined territorial spatial management. Taking Chengdu, China, as a case study, this study develops an ecological zoning framework based on the eXtreme Gradient Boosting-SHapley Additive exPlanations (XGBoost-SHAP) model. The framework integrates Ecosystem Service Value (ESV) and Landscape Ecological Risk (LER) as core indicators, applies Z-score standardization and quadrant classification to delineate four ecological zone types with distinct ecological functions, and further employs the XGBoost-SHAP model to identify key natural and anthropogenic drivers and explain their roles in spatial environmental evolution. The results show that: (1) Farmland and forest were the dominant land use types, accounting for over 87 % of the total area. From 2000 to 2020, farmland decreased by 11.64 %, while ecological land increased by 219.68 km2. (2) ESV followed a “rise–decline–rise” trend with a net decrease of 0.752 billion CNY and a spatial pattern of “high in the northwest, low in the center.” LER exhibited a “low in the northwest-central, high in the southeast” pattern, with low-risk areas gradually expanding. (3) The Ecological Control Zone (ECZ) and Ecological Improvement Zone (EIZ) dominated the study area, covering over 9,591 km2. At the same time, the Ecological Rehabilitation Zone (ERZ) and Ecological Conservation Zone (ECOZ) showed stable growth driven primarily by natural factors. (4) The XGBoost-SHAP model demonstrated high interpretability and attribution accuracy, effectively revealing the driving mechanisms behind zoning evolution. This ecological zoning framework is refined, interpretable and data-driven, providing scientific support for spatial planning and sustainable ecosystem management in rapidly urbanizing regions.http://www.sciencedirect.com/science/article/pii/S1470160X25007174Characterization factorsEcosystem service valueEcological zoningLandscape ecological riskMachine learning |
| spellingShingle | Xiaobin Huang Xiaosheng Liu Yuanhang Jin Xue Gao Youliang Chen Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China Ecological Indicators Characterization factors Ecosystem service value Ecological zoning Landscape ecological risk Machine learning |
| title | Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China |
| title_full | Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China |
| title_fullStr | Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China |
| title_full_unstemmed | Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China |
| title_short | Identification and attribution analysis of integrated ecological zones based on the XGBoost-SHAP model: A case study of Chengdu, China |
| title_sort | identification and attribution analysis of integrated ecological zones based on the xgboost shap model a case study of chengdu china |
| topic | Characterization factors Ecosystem service value Ecological zoning Landscape ecological risk Machine learning |
| url | http://www.sciencedirect.com/science/article/pii/S1470160X25007174 |
| work_keys_str_mv | AT xiaobinhuang identificationandattributionanalysisofintegratedecologicalzonesbasedonthexgboostshapmodelacasestudyofchengduchina AT xiaoshengliu identificationandattributionanalysisofintegratedecologicalzonesbasedonthexgboostshapmodelacasestudyofchengduchina AT yuanhangjin identificationandattributionanalysisofintegratedecologicalzonesbasedonthexgboostshapmodelacasestudyofchengduchina AT xuegao identificationandattributionanalysisofintegratedecologicalzonesbasedonthexgboostshapmodelacasestudyofchengduchina AT youliangchen identificationandattributionanalysisofintegratedecologicalzonesbasedonthexgboostshapmodelacasestudyofchengduchina |