2000—2024年中国植被生态水分利用效率变化及归因分析

Spatiotemporal evolution and attribution analysis of vegetation water use efficiency in China from 2000 to 2024

关键词:  生态用水效率  Theil-Sen趋势分析  Mann-Kendall检验  XGBoost  SHAP

Key words:  WUE  Theil-Sen trend analysis  Mann-Kendall test  XGBoost  SHAP

摘要:

生态用水效率（WUE）是表征陆地生态系统碳-水耦合关系的重要指标。文章基于2000—2024年遥感总初级生产力（GPP）、蒸散发（ET）及多源环境因子数据，采用Theil-Sen趋势分析、Mann-Kendall显著性检验和XGBoost-SHAP模型，分析中国植被WUE时空变化及驱动机制。结果表明：中国WUE空间格局总体稳定，呈“东高西低”分异特征，东部湿润、半湿润区多为1.5—2.7 g/kg（以碳计，下同），局部超过3.0 g/kg；西北干旱区和青藏高原大部低于1.2 g/kg。2000—2024年全国年均WUE在1.28—1.41 g/kg波动，整体显著上升，变化斜率为0.0015 g/(kg·a)（p<0.05）。趋势检验显示，46.61%的区域WUE呈增长趋势，其中显著增长区占15.09%；27.04%的区域呈下降趋势，26.35%的区域基本稳定。不同植被类型中，栽培植被、阔叶林和针阔叶混交林显著增长比例分别为24.63%、22.77%和25.19%，高山植被则以基本不变为主，占45.24%。驱动因子贡献排序为核归一化植被指数（kNDVI）>气温（TMP）>水汽压亏缺（VPD）>细颗粒物（PM_2.5）>降水（PRE）>太阳辐射（SRAD）>相对湿度（RH）>人口密度（POP）>二氧化碳（CO₂）>土壤湿度（SM），其中kNDVI和TMP为主导因子。总体来看，中国植被WUE呈稳中有升态势，其变化受植被生长状况、热力条件和水分胁迫共同调控，且具有明显区域和植被类型差异。

Abstract:

Background, aim, and scope Water use efficiency (WUE) is a key indicator of carbon-water coupling in terrestrial ecosystems and reflects the capacity of vegetation to assimilate carbon per unit water consumption. Clarifying the long-term spatiotemporal dynamics and driving mechanisms of vegetation WUE is essential for understanding ecosystem responses to environmental change and supporting regional ecohydrological management. Accordingly, this study aims to characterize the spatiotemporal evolution of vegetation WUE across China from 2000 to 2024 and quantify the contributions of its major environmental drivers. Materials and methods Based on remotely sensed gross primary productivity (GPP), evapotranspiration (ET), and multi-source environmental datasets from 2000 to 2024, this study estimated vegetation WUE across China. The Theil-Sen trend analysis, Mann-Kendall significance test, and an interpretable eXtreme Gradient Boosting-Shapley Additive Explanations (XGBoost-SHAP) framework were used to identify WUE trends, vegetation-type differences, and dominant driving factors. Results The results showed that the spatial pattern of WUE in China remained generally stable from 2000 to 2024, characterized by an evident east-high and west-low gradient. WUE values in humid and semi-humid eastern China were mostly 1.5—2.7 g/kg (calculated as carbon, same below), with local values exceeding 3.0 g/kg, whereas most areas in northwestern arid regions and the Qinghai-Xizang Plateau were below 1.2 g/kg. National annual mean WUE fluctuated between 1.28 and 1.41 g/kg and showed a significant increasing trend, with a slope of 0.0015 g/(kg·a) (p<0.05). The trend analysis indicated that 46.61% of China experienced WUE increases, including 15.09% with significant increases, while 27.04% showed decreasing trends and 26.35% remained basically stable. Across vegetation types, cultivated vegetation, broadleaf forests, and mixed coniferous-broadleaf forests showed relatively high proportions of significant WUE increases, accounting for 24.63%, 22.77%, and 25.19%, respectively, whereas alpine vegetation was dominated by stable trends, with 45.24% showing no obvious change. The contribution of driving factors followed the order of kernel normalized difference vegetation index (kNDVI) > temperature (TMP) > vapor pressure deficit (VPD) > fine particulate matter (PM_2.5) > precipitation (PRE) > solar radiation (SRAD) > relative humidity (RH) > population density (POP) > carbon dioxide (CO₂) > soil moisture (SM), with kNDVI and TMP being the dominant factors. Discussion The east-west spatial contrast of WUE was closely associated with regional differences in hydrothermal conditions, vegetation productivity, and ecosystem constraints. Higher WUE in eastern humid and semi-humid regions was mainly related to favorable water availability and stronger vegetation growth, whereas lower WUE in arid and alpine regions was constrained by water scarcity, low vegetation coverage, and fragile ecological conditions. The XGBoost-SHAP results further revealed that WUE responses to environmental factors were nonlinear, indicating that vegetation growth status, thermal conditions, and atmospheric water stress jointly regulated WUE variation. Conclusions From 2000 to 2024, vegetation WUE in China generally showed a stable spatial pattern and a significant upward temporal trend, with marked regional heterogeneity and vegetation-type differences. The observed WUE dynamics were primarily controlled by vegetation greenness and temperature, while VPD, PM_2.5, precipitation, and solar radiation exerted secondary influences. Recommendations and perspectives Future ecosystem management should adopt region-specific strategies. In eastern and intensively managed vegetation regions, improving vegetation structure and management practices may further enhance WUE, whereas in western arid and alpine regions, maintaining ecosystem stability and preventing degradation should be prioritized. Further studies should focus on the mechanisms underlying WUE declines in local areas and strengthen differentiated analyses across climatic zones and vegetation types.