Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning
Under the combined influence of global change and human activity, wetland ecosystem health (WEH) faces considerable challenges. Understanding spatiotemporal evolutionary patterns and the key driving mechanisms is critical for promoting scientific wetland conservation and sustainable development. How...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Ecological Indicators |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1470160X25008088 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839630501961269248 |
|---|---|
| author | Walian Du Shouzheng Tong Mingye Zhang Yuan Xin Dongjie Zhang Xianglong Xing Yu An Geng Cui Guangtao Liu |
| author_facet | Walian Du Shouzheng Tong Mingye Zhang Yuan Xin Dongjie Zhang Xianglong Xing Yu An Geng Cui Guangtao Liu |
| author_sort | Walian Du |
| collection | DOAJ |
| description | Under the combined influence of global change and human activity, wetland ecosystem health (WEH) faces considerable challenges. Understanding spatiotemporal evolutionary patterns and the key driving mechanisms is critical for promoting scientific wetland conservation and sustainable development. However, prior studies have often overlooked the complex nonlinear relationships and interactions among multiple driving factors. In this study, we constructed a Wetland Ecosystem Health Composite Index (WEHI) by integrating the Driving Forces-Pressure-State-Impact-Response-Management (DPSIRM) and VORS models to assess the dynamic evolution of wetland health in Northeast China from 2000 to 2020. Extreme Gradient Boosting (XGBoost) combined with SHapley Additive exPlanations (SHAP) artificial intelligence techniques were used to analyze the nonlinear effects, interactions, and threshold effects of the driving factors. Partial least squares structural equation modeling (PLS-SEM) was applied to examine the direct and indirect pathways through which drivers influenced the WEHI. Swampland is the dominant wetland type in Northeast China, with frequent transitions among wetland types. The average WEHI is 0.4448, reflecting a moderate ecosystem health level, with poorer conditions in the south and better conditions in the northwest. Total water resources and cropland area are the primary drivers with significant nonlinear and threshold effects. The interaction between total water resources and natural population growth is most pronounced – when water resources fall below 15 billion m3, population growth significantly exacerbates wetland degradation. The management level acts as a mediating variable, indirectly weakening the positive effects of other drivers. Integrating XGBoost-SHAP and PLS-SEM provides an innovative and complementary analytical framework for exploring ecosystem health drivers, offering robust theoretical support for the scientific formulation of wetland protection and management policies and facilitating precise identification and application of multidimensional driver effects. |
| format | Article |
| id | doaj-art-736679df9f774493b2ee858dce1091d2 |
| institution | Matheson Library |
| issn | 1470-160X |
| language | English |
| publishDate | 2025-09-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Ecological Indicators |
| spelling | doaj-art-736679df9f774493b2ee858dce1091d22025-07-14T04:14:52ZengElsevierEcological Indicators1470-160X2025-09-01178113878Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learningWalian Du0Shouzheng Tong1Mingye Zhang2Yuan Xin3Dongjie Zhang4Xianglong Xing5Yu An6Geng Cui7Guangtao Liu8Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaKey Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, China; Corresponding author at: Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China.Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaKey Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaShandong Key Laboratory of Eco-Environmental Science for the Yellow River Delta, Shandong University of Aeronautics, Binzhou 256600 Shandong Province, ChinaKey Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaKey Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaKey Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China; University of Chinese Academy of Sciences, Beijing 100049, ChinaWuyuerhe-Shuangyanghe Provincial Nature Reserve Development Center of Heilongjiang, ChinaUnder the combined influence of global change and human activity, wetland ecosystem health (WEH) faces considerable challenges. Understanding spatiotemporal evolutionary patterns and the key driving mechanisms is critical for promoting scientific wetland conservation and sustainable development. However, prior studies have often overlooked the complex nonlinear relationships and interactions among multiple driving factors. In this study, we constructed a Wetland Ecosystem Health Composite Index (WEHI) by integrating the Driving Forces-Pressure-State-Impact-Response-Management (DPSIRM) and VORS models to assess the dynamic evolution of wetland health in Northeast China from 2000 to 2020. Extreme Gradient Boosting (XGBoost) combined with SHapley Additive exPlanations (SHAP) artificial intelligence techniques were used to analyze the nonlinear effects, interactions, and threshold effects of the driving factors. Partial least squares structural equation modeling (PLS-SEM) was applied to examine the direct and indirect pathways through which drivers influenced the WEHI. Swampland is the dominant wetland type in Northeast China, with frequent transitions among wetland types. The average WEHI is 0.4448, reflecting a moderate ecosystem health level, with poorer conditions in the south and better conditions in the northwest. Total water resources and cropland area are the primary drivers with significant nonlinear and threshold effects. The interaction between total water resources and natural population growth is most pronounced – when water resources fall below 15 billion m3, population growth significantly exacerbates wetland degradation. The management level acts as a mediating variable, indirectly weakening the positive effects of other drivers. Integrating XGBoost-SHAP and PLS-SEM provides an innovative and complementary analytical framework for exploring ecosystem health drivers, offering robust theoretical support for the scientific formulation of wetland protection and management policies and facilitating precise identification and application of multidimensional driver effects.http://www.sciencedirect.com/science/article/pii/S1470160X25008088DPSIRM-VORS modelEcosystem health assessmentPLS-SEMWetlandXGBoost-SHAP |
| spellingShingle | Walian Du Shouzheng Tong Mingye Zhang Yuan Xin Dongjie Zhang Xianglong Xing Yu An Geng Cui Guangtao Liu Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning Ecological Indicators DPSIRM-VORS model Ecosystem health assessment PLS-SEM Wetland XGBoost-SHAP |
| title | Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning |
| title_full | Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning |
| title_fullStr | Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning |
| title_full_unstemmed | Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning |
| title_short | Revealing the spatiotemporal dynamics and nonlinear interaction-driven mechanisms of wetland ecosystem health in Northeast China using interpretable machine learning |
| title_sort | revealing the spatiotemporal dynamics and nonlinear interaction driven mechanisms of wetland ecosystem health in northeast china using interpretable machine learning |
| topic | DPSIRM-VORS model Ecosystem health assessment PLS-SEM Wetland XGBoost-SHAP |
| url | http://www.sciencedirect.com/science/article/pii/S1470160X25008088 |
| work_keys_str_mv | AT waliandu revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT shouzhengtong revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT mingyezhang revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT yuanxin revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT dongjiezhang revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT xianglongxing revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT yuan revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT gengcui revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning AT guangtaoliu revealingthespatiotemporaldynamicsandnonlinearinteractiondrivenmechanismsofwetlandecosystemhealthinnortheastchinausinginterpretablemachinelearning |