Paddy rice is one of the main grain crops in the world. Accurate estimations of the gross primary production (GPP) of paddy rice are essential for assessing rice grain production and monitoring the carbon cycle in paddy fields with the aim of providing ideal conditions for crops throughout the growing season. Several studies have demonstrated the advantages of combining the eddy covariance technique with remotely sensed data to model GPP at CO₂ eddy flux tower sites. As paddy rice continuously changes during its growth and development, and important growth events frequently occur, it is critical to observe the growing conditions at various stages of the process. To better understand the variations in GPP at different growth stages, two key parameters that drive the vegetation photosynthesis model (VPM) are analyzed and estimated at various phenological phases. Specifically, general piecewise logistic functions are used to extract phenological transitions from data at four paddy rice flux tower sites. The maximum light-use efficiency (LUE) and optimum temperature are estimated from these phenological transitions, and these indicators are used to drive the VPM to simulate GPP over multiple years at the four sites. The simulation results show that GPP based on our phenological transition-based VPM (GPP_PVPM) agrees reasonably well with the variations of GPP estimated from CO₂ flux data (GPP_EC) (R² > 0.9). In addition, a comparison indicates that GPP_PVPM tracks the seasonal dynamics of GPP_EC better than GPP estimated from the original VPM. Furthermore, GPP based on the improved maximum LUE is lower than GPP_EC at most flux sites and GPP based on the improved optimum temperature is higher than GPP_EC. These comparisons imply that the maximum LUE and optimum temperature estimated in the phenological transitions of paddy rice are beneficial to enhance the accuracy of GPP estimation. The improved estimation of GPP provides phenological insights into the temporal dynamics of vegetation photosynthesis in paddy fields.

水稻是世界上主要的粮食作物之一。准确估算水稻的初级总产值（GPP）对于评估稻米产量和监测稻田的碳循环至关重要，目的是为整个生长季节的农作物提供理想的条件。多项研究已经证明了将涡动协方差技术与遥感数据相结合以在CO _2下建模GPP的优势。涡流塔站点。由于水稻在其生长和发育过程中不断变化，并且频繁发生重要的生长事件，因此，观察该过程各个阶段的生长状况至关重要。为了更好地理解GPP在不同生长阶段的变化，在植物的各个物候期分析和估算了驱动植被光合作用模型（VPM）的两个关键参数。具体而言，一般的分段逻辑函数用于从四个水稻通量塔站点的数据中提取物候转换。从这些物候过渡过程中可以估算出最大的光利用效率（LUE）和最佳温度，并且这些指标用于驱动VPM在这四个站点上模拟GPP多年。_PVPM）与根据CO ₂通量数据（GPP _EC）估计的GPP的变化相当吻合（R ² > 0.9）。另外，比较表明，与从原始VPM估计的GPP相比，GPP _PVPM更好地跟踪了GPP _EC的季节动态。此外，基于改进的最大LUE的GPP在大多数通量位置均低于GPP _EC，基于改进的最佳温度的GPP高于GPP _EC。这些比较表明，在水稻物候变化中估算的最大LUE和最佳温度有利于提高GPP估算的准确性。GPP的改进估计为水稻田中植物光合作用的时空动态提供了物候方面的见识。