Nonlinear response of vegetation productivity to climate change and human activities in a water-source basin: implications for environmental health and water security.

INTRODUCTION: Gross primary productivity (GPP) is a key component of the global carbon cycle, and its dynamics are influenced by climate conditions and human activities. However, the nonlinear response of GPP to climate change and human activities remains unclear. The Hanjiang River Basin (HJRB), a crucial water-source region of the South-to-North Water Diversion Project in China, is vital to China's ecological security. Changes in vegetation productivity in this water-source basin may affect water conservation, soil erosion control, drought and flood regulation, and the environmental conditions to which downstream populations are exposed. In recent decades, the region has experienced significant climate change, intensified human activities, and notable increases in vegetation greenness. However, previous literature relying on the assumption of stationarity and traditional linear methods has limited capability in detecting nonlinear and gradual vegetation changes. METHODS: The long-term nonlinear trend in GPP in the HJRB was analyzed using ensemble empirical mode decomposition (EEMD), and the nonlinear response relationships among climatic factors, human activities, and GPP were further quantified using the LightGBM machine-learning method. RESULTS: From 2000 to 2020, the annual growth rate of vegetation productivity in the HJRB was 12.27 g C m-2 yr-1 (P < 0.01). Most vegetated areas exhibited an upward trend in GPP. However, 28.25% and 4.41% of the region experienced transitions from growth to decline and from decline to growth, respectively, primarily in the southern upstream and eastern downstream areas. Drought (25.28%), precipitation (10.75%), and relative humidity (10.69%) were identified as significant factors influencing vegetation productivity trends. Annual precipitation above approximately 1,000 mm was associated with lower SHAP values for the continuous-increase trend, which may reflect combined hydrothermal, radiation, and topographic constraints rather than a direct inhibitory effect of precipitation. Additionally, an increase in SPEI12 was associated with higher SHAP values for the continuous-increase trend. DISCUSSION: This study elucidates trends and underlying response patterns in vegetation productivity in the HJRB, reveals the driving factors behind regional vegetation productivity changes, and provides environmental evidence for vegetation monitoring, water-source ecosystem management, and climate-related health-risk prevention in the HJRB and similar water-source regions.