DLM-FvCB模型光合能力关键参数季节性参数化方案及模拟效果评估

选取包含9种植被功能型的15个通量塔观测数据（60站点年），采用控制变量法，分别在月尺度和年尺度上对25℃时最大羧化速率（V_cmax25）进行优化，并利用未参与参数反演的站点年数据集定量评估这2种参数化方案的模拟效果。研究发现：① 在参数的季节变异方面：9种植被功能型的V_cmax25均呈现明显的季节性波动。V_cmax25的波动幅度为：冬季>秋季>春季>夏季，不同植被类型季节变化幅度相差不大，而寒带植被V_cmax25季节变化幅度接近温带植被的2倍；② 对生态系统初级生产力（GPP）估算的影响方面：V_cmax25季节性参数化方案显著提高了GPP的模拟能力和模拟精度，其中森林和灌木冬季提升最显著（R²提高了35.7%，RMSE降低了23.24%），春秋季次之，夏季最小。然而，对于C₃草地，DLM无论采取V_cmax25季节性参数化方案还是年尺度参数化方案，均为系统低估GPP。

Seasonal Parameterization Scheme of Key Photosynthetic Parameter and Simulation Effect Evaluation of DLM-FvCB Model

The key photosynthetic parameter of maximum carboxylation rate (V_cmax25) in Dynamic Land Surface Model (DLM) was parameterized at monthly and yearly time scales, respectively, for 15 flux towers (60 site-years data) including nine vegetation functional types (PFTs) by using variable-controlling iteration method. The GPP simulation results of the two parameterization schemes are compared and analyzed by using the observation data of the remaining station years not involved in parameterization. The results indicate that: 1) the V_cmax25 varied seasonally obviously for all PFTs. Specifically, the variation of V_cmax25 was largest in winter and least in summer. The fluctuation of V_cmax25 was similar among different biomes and was nearly twice in boreal climate region as much as that in temperate climate region. 2) The accuracy of GPP estimation using the seasonal parameterization scheme of V_cmax25 was obviously improved, among which the forest and shrub increased the most significantly in winter (R² increased by 35.7%, RMSE decreased by 23.24%), followed by spring and autumn, and the least in summer. Even with seasonal fluctuation of parameter considered, DLM-FvCB still difficult well capture the GPP variations for C₃ grassland. This study suggests an important consideration of temporal parameterization in further improvement of modelling GPP in land-surface model.