| 杜佩昱,唐秀美,淮贺举,郜允兵,葛荣凤,刘雯.机器学习支持下的京津冀地区农田生态系统服务交互作用及时空演变分析[J].农业环境科学学报,2025,44(11):2820-2834. |
| 机器学习支持下的京津冀地区农田生态系统服务交互作用及时空演变分析 |
| Analysis of interactions and spatiotemporal evolution of farmland ecosystem services in the Beijing-Tianjin-Hebei region supported by machine learning |
| 投稿时间：2025-05-18 |
| DOI：10.11654/jaes.2025-0459 |
| 中文关键词: 农田 生态系统服务 机器学习算法 权衡/协同 生态系统服务簇 |
| 英文关键词: farmland ecosystem services machine learning algorithms trade-offs/synergies ecosystem service bundles |
| 基金项目:国家重点研发计划项目（2023YFD1702400）；北京市农林科学院科技创新能力建设专项（KJCX20230309） |
| 作者 | 单位 | E-mail | | 杜佩昱 | 山东农业大学资源与环境学院, 山东 泰安 271018
北京市农林科学院信息技术研究中心, 北京 100097
国家农业信息化工程技术研究中心, 北京 100097 | | | 唐秀美 | 北京市农林科学院信息技术研究中心, 北京 100097
国家农业信息化工程技术研究中心, 北京 100097 | | | 淮贺举 | 北京市农林科学院信息技术研究中心, 北京 100097
国家农业信息化工程技术研究中心, 北京 100097 | huaihj@nercita.org.cn | | 郜允兵 | 北京市农林科学院信息技术研究中心, 北京 100097
国家农业信息化工程技术研究中心, 北京 100097 | | | 葛荣凤 | 中国国际工程咨询有限公司, 北京 100048 | | | 刘雯 | 山东农业大学资源与环境学院, 山东 泰安 271018
北京市农林科学院信息技术研究中心, 北京 100097
国家农业信息化工程技术研究中心, 北京 100097 | |
|
| 摘要点击次数: 1026 |
| 全文下载次数: 819 |
| 中文摘要: |
| 为揭示农田生态系统服务之间的交互关系及其时空演变特征，支持区域耕地资源管理与生态系统可持续性提升，本研究以京津冀地区2000、2015年和2023年农田生态系统为研究对象，基于InVEST模型和粮食产量估算模型评估了粮食供给、碳储存、土壤保持、产水服务4种农田生态系统服务功能，采用Spearman相关性和地理加权回归模型量化其权衡/协同关系，并在自组织映射算法的支持下识别了农田生态系统服务簇。结果显示：2000—2023年，京津冀地区农田生态系统服务发生了明显的时空变化，其中产水服务和粮食供给能力逐步增强，尤其在南部和沿海地区，土壤保持和碳储存呈波动变化。产水服务与土壤保持、粮食供给服务协同增强，碳储存与其他服务之间则在协同与权衡之间反复变化。农田生态系统服务簇表现出显著变化，粮食主导型和产水-粮食复合型簇扩大，土壤-粮食复合型和碳储主导型簇缩小。本研究构建了融合农田生态系统服务交互关系识别与空间聚类分析的评估框架，结果表明该框架能够有效识别服务之间的权衡/协同关系及其空间演变特征。 |
| 英文摘要: |
| The purpose of this study was to reveal the interaction between farmland ecosystem services and their spatiotemporal evolution characteristics, so as to support the improvement of regional cultivated land resource management and ecosystem sustainability. In this study, the farmland ecosystems in the Beijing-Tianjin-Hebei region were studied in 2000, 2015 and 2023, and four farmland ecosystem service functions, namely, food supply, carbon storage, soil conservation, and water production services, were assessed based on the InVEST model and the grain yield estimation model, and the trade-offs/synergies were quantified using Spearman's correlation and geographically-weighted regression models, and the bundles of farmland ecosystem services were identified with the support of a machine-learning algorithm support identified bundles of farmland ecosystem services. The results showed that the significant spatial and temporal variations in farmland ecosystem services occurred in the Beijing-Tianjin-Hebei region from 2000 to 2023, in which water-producing services and food production capacity were gradually enhanced, especially in the southern and coastal regions, and soil retention and carbon storage showed fluctuating changes. Water production services synergized with soil conservation and food supply services increased, while carbon storage and other services changed repeatedly between synergies and trade-offs. The farmland ecosystem service bundles showed significant evolution, with the expansion of food-dominant and water-producing-food-complex bundles, and the shrinkage of soilfood-complex and carbon storage-dominant bundles. This study developed an evaluation framework that integrates the identification of interactions among farmland ecosystem services with spatial clustering analysis. The results demonstrate that the framework can effectively capture the trade-offs and synergies among services as well as their spatiotemporal evolution patterns. |
| HTML
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
