Climate change significantly impacts vegetation gross primary productivity (GPP), yet uncertainties persist in the carbon cycle of tropical terrestrial ecosystems due to incomplete consideration of productivity drivers and lag effects. To address this, we developed a remote sensing-based process model by integrating high-resolution vegetation indices and multi-layer soil hydrological module, to simulate monthly GPP at a 30 m resolution across Hainan Island from 2000 to 2020. The finer GPP can capture more spatial details and show higher accuracy at site scales (R = 0.79 and NRMSE = 14.79 %). Trend analysis and Hurst exponent were used to reveal spatiotemporal dynamics and sustainability of GPP. Meanwhile, nonlinear Granger causality tests quantified both concurrent and lagged correlations between various environmental factors and GPP. The results indicated significant GPP increases across 98.5 % of vegetated areas, with an annual rise of 437.02 g C/m2, and a marked improvement in trends around 2011. Future projections suggest sustained high GPP sustainability (Hurst = 0.53), and reducing “positive-inconsistent” areas in the northeast and southwest is crucial for enhancing local carbon sinks. Furthermore, water availability, temperature, and radiation were primary drivers of GPP changes, affecting 53.55 %, 27.77 %, and 14.43 % of vegetated areas, respectively, with their compounded effects enhancing explanatory power by 35.84 %. Relative humidity dominated water availability impacts on GPP (10.02 % to 79.98 % variation), surpassing precipitation and soil moisture impacts. Lag effects were observed in 68.83 % of vegetated areas, with 1 to 4-month delays in responses to net solar radiation and surface temperature, especially in forest and shrubland ecosystems. This study provides deeper insights into fine-scale GPP simulations and analysis of climate interactions, which are crucial for effective carbon cycle management in tropical ecosystems.

中文翻译：

---

以更高分辨率定量评估热带生态系统总初级生产力的时空变化和驱动因素


气候变化对植被总初级生产力 （GPP） 影响显著，但由于对生产力驱动因素和滞后效应的考虑不完整，热带陆地生态系统的碳循环仍然存在不确定性。为了解决这个问题，我们通过集成高分辨率植被指数和多层土壤水文模块，开发了一个基于遥感的过程模型，以 30 m 分辨率模拟 2000 年至 2020 年海南岛的每月 GPP。更精细的 GPP 可以捕获更多的空间细节，并在站点尺度上显示更高的精度（R = 0.79 和 NRMSE = 14.79 %）。趋势分析和 Hurst 指数用于揭示 GPP 的时空动态和可持续性。同时，非线性 Granger 因果关系检验量化了各种环境因素与 GPP 之间的并发和滞后相关性。结果表明，98.5% 的植被区域的 GPP 显著增加，年增长率为 437.02 g C/m2,2011 年前后的趋势显着改善。未来的预测表明，GPP 持续保持高可持续性 （Hurst = 0.53），减少东北部和西南部的“正-不一致”区域对于提高当地碳汇至关重要。此外，水的可用性、温度和辐射是 GPP 变化的主要驱动因素，分别影响了 53.55 % 、 27.77 % 和 14.43 % 的植被面积，它们的复合效应将解释力提高了 35.84 %。相对湿度对 GPP 的可用水影响占主导地位（变化 10.02 % 至 79.98 %），超过了降水和土壤水分影响。在 68.83% 的植被区域观察到滞后效应，对净太阳辐射和表面温度的响应延迟 1 到 4 个月，尤其是在森林和灌木丛生态系统中。 本研究为精细的 GPP 模拟和气候相互作用分析提供了更深入的见解，这对于热带生态系统中有效的碳循环管理至关重要。