Distinguishing dominant drivers on long-term vegetation dynamics across China considering time-lag and accumulation effects using machine learning techniques
文献类型：期刊论文
| 作者 | Wang, Qianxin1,2; Huang, Lin2; Yang, Meng2; Wang, Lan1,2; Cao, Wei2 |
| 刊名 | JOURNAL OF ENVIRONMENTAL MANAGEMENT
 |
| 出版日期 | 2025-09-01 |
| 卷号 | 391页码:126477 |
| 关键词 | Long-term vegetation dynamics Time-lag and accumulation effects Attribution analysis Machine learning China |
| ISSN号 | 0301-4797 |
| DOI | 10.1016/j.jenvman.2025.126477 |
| 产权排序 | 1 |
| 文献子类 | Article |
| 英文摘要 | Accurate attribution of vegetation dynamics is essential to ensure the conservation, restoration and sustainability of terrestrial ecosystems. However, due to the time-lag and accumulation effects of vegetation responding climate change and anthropogenic activities, traditional statistical methods often fail to capture the nonlinear impacts, and leading to ongoing debates about the relative contributions. In this paper, we explored the spatiotemporal dynamics of various vegetation types across China from 1982 to 2022 employing the Normalized Difference Vegetation Index (NDVI). Then we integrated machine learning methods with an improved residual trend approach to precisely quantify the contributions of anthropogenic activities and climate change on vegetation dynamics. Our results indicated that (i) the annual increase in NDVI has amounted to 0.012 per decade across all vegetated areas in China, and 57.0 % of the vegetated areas underwent notable greening trends in the past four decades. (ii) Over 92 % of vegetation areas exhibited climatic temporal effects in China, mainly with 1 to 2-month accumulation in response to temperature and precipitation, and 1 to 3-month lag in responses to sunshine duration. Furthermore, forest exhibited the longest time-lag (1.36 months) and accumulation months (0.93 months) to precipitation, whereas grassland responded the shortest time-lag (0.13 months) and accumulation months (0.59 months) to temperature. (iii) Anthropogenic activities predominantly influence 69.7 % of vegetation dynamics in China during 2000-2022. Our improved approaches can diminish quantization uncertainty and show higher accuracy (R2 = 0.97 and RMSE = 0.02). Our study provides valuable insights for understanding vegetation dynamics and informs ecological protection and restoration efforts in China. |
| URL标识 | 查看原文 |
| WOS关键词 | ECOLOGICAL RESTORATION PROJECTS ; PREDICTION ; GREENNESS ; PLATEAU ; TRENDS ; GROWTH ; FOREST ; KARST ; EARTH |
| WOS研究方向 | Environmental Sciences & Ecology |
| 语种 | 英语 |
| WOS记录号 | WOS:001548953600001 |
| 出版者 | ACADEMIC PRESS LTD- ELSEVIER SCIENCE LTD |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/215569]  |
| 专题 | 陆地表层格局与模拟院重点实验室_外文论文 |
| 通讯作者 | Cao, Wei |
| 作者单位 | 1.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Key Lab Land Surface Pattern & Simulat, Inst Geog Sci & Nat Resources Res, Beijing 100101, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Wang, Qianxin,Huang, Lin,Yang, Meng,et al. Distinguishing dominant drivers on long-term vegetation dynamics across China considering time-lag and accumulation effects using machine learning techniques[J]. JOURNAL OF ENVIRONMENTAL MANAGEMENT,2025,391:126477. |
| APA | Wang, Qianxin,Huang, Lin,Yang, Meng,Wang, Lan,&Cao, Wei.(2025).Distinguishing dominant drivers on long-term vegetation dynamics across China considering time-lag and accumulation effects using machine learning techniques.JOURNAL OF ENVIRONMENTAL MANAGEMENT,391,126477. |
| MLA | Wang, Qianxin,et al."Distinguishing dominant drivers on long-term vegetation dynamics across China considering time-lag and accumulation effects using machine learning techniques".JOURNAL OF ENVIRONMENTAL MANAGEMENT 391(2025):126477. |
入库方式： OAI收割
来源：地理科学与资源研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。