Reducing the uncertainty of parameters controlling seasonal carbon and water fluxes in Chinese forests and its implication for simulated climate sensitivities
Reducing the uncertainty of parameters controlling seasonal carbon and water fluxes in Chinese forests and its implication for simulated climate sensitivities
复制标题
DOI：
10.1002/2017gb005714
复制
发表时间：
2017-08-01
影响因子：
5.2
通讯作者：
Piao, Shilong
中科院分区：
文献类型：
--
作者：
Li, Yue;Yang, Hui;Piao, Shilong
减少基于过程的陆地生态系统模型的参数不确定性是准确估计碳收支和预测生态系统对气候变化响应的主要目标之一。然而，中国森林生态系统是北半球陆地上重要的碳汇，其参数很少受到中国森林生态系统观测的限制。本文利用中国6个林地的涡动协方差资料，对动力学生态系统中的有机碳和水文学参数进行了优化。通过参数优化的模式资料同化大大减少了先前模式的误差，并改善了模拟的净生态系统交换、潜热通量、总初级生产力和生态系统呼吸的季节周期和夏季日周期。基于优化模型的气候变化试验表明，森林净第一性生产量在中国南部受到抑制，而在中国东北部受到刺激。在缺水地区，降水变化对森林净初级生产力的影响是不对称的，在落叶阔叶林立地，降水量变化导致净初级生产力对降水量下降的响应减少高达61%。我们发现，季节优化改变了森林碳循环对环境变化的响应，参数优化一致地降低了模拟的异养呼吸对气候变暖的正响应。来自独立观测的评估表明，改善模型结构仍然对长期碳储量及其变化最重要，特别是与营养和年龄相关的光合作用速率、碳分配和树木死亡的变化。
Reducing parameter uncertainty of process-based terrestrial ecosystem models (TEMs) is one of the primary targets for accurately estimating carbon budgets and predicting ecosystem responses to climate change. However, parameters in TEMs are rarely constrained by observations from Chinese forest ecosystems, which are important carbon sink over the northern hemispheric land. In this study, eddy covariance data from six forest sites in China are used to optimize parameters of the ORganizing Carbon and Hydrology In Dynamics EcosystEms TEM. The model-data assimilation through parameter optimization largely reduces the prior model errors and improves the simulated seasonal cycle and summer diurnal cycle of net ecosystem exchange, latent heat fluxes, and gross primary production and ecosystem respiration. Climate change experiments based on the optimized model are deployed to indicate that forest net primary production (NPP) is suppressed in response to warming in the southern China but stimulated in the northeastern China. Altered precipitation has an asymmetric impact on forest NPP at sites in water-limited regions, with the optimization-induced reduction in response of NPP to precipitation decline being as large as 61% at a deciduous broadleaf forest site. We find that seasonal optimization alters forest carbon cycle responses to environmental change, with the parameter optimization consistently reducing the simulated positive response of heterotrophic respiration to warming. Evaluations from independent observations suggest that improving model structure still matters most for long-term carbon stock and its changes, in particular, nutrient-and age-related changes of photosynthetic rates, carbon allocation, and tree mortality.