多源观测信息与陆表水碳通量过程模型的数据同化研究

项目名称： 多源观测信息与陆表水碳通量过程模型的数据同化研究

项目编号： No.41301451

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 张廷龙

作者单位： 西北农林科技大学

项目金额： 25万元

中文摘要： 精确估算水碳通量对陆地水碳循环研究意义重大,但同时极具挑战性。目前的估算精度有待进一步提高。传统的模型模拟和站点观测两种估算方法各有优势和不足,二者需结合进行研究。数据同化将观测融合到基于物理规律的模型中，尽可能得到模型状态变量和参数的最优估计，为模型和观测的结合提供了一条有效的途径。本研究综合多源观测信息，通过数据同化的方法，将其与生态过程模型结合，对陆表水碳通量进行模拟。研究首先对模型参数进行敏感性和时空变异特性分析，然后针对不同特征的参数，采取分阶段多途径参数优化和状态变量更新的同化策略，最终实现多数时空变异性小的敏感参数静态优化直接同化，少数关键时空变异性大的敏感参数与状态变量更新同步动态优化同化。同时通过转化和建立规则，将先验知识引入到生态模型的数据同化过程当中。利用多源观测信息的多重约束，来提高生态过程模型对陆表水碳通量模拟的精度，加深对模型不确定性来源的认识。

中文关键词： 生态过程模型；水碳通量；数据同化；时空异质性；先验知识

英文摘要： Estimating water and carbon fluxes accurately, it is very important to terrestrial water and carbon cycle research, but is still challenging.The current estimating accuracy need to be further improved. Two traditional estimating methods model simulation and in situ observation have their own advantages and shortcomings,need to combine each other for better studying. Data assimilation fuses measurement into model based on physical laws, as much as possible to get the optimal estimation of the model state variables and parameters, provides an effective way for the combination of models and observations. In this study, through data assimilation methods, multi-source observations combined ecological process model to simulate land surface water and carbon fluxes. Firstly, sensitivity and spatial and temporal variability of the model parameters were studied; Then according to the different characteristics of the parameters, a multi phased-channel parameter optimization and state variable update assimilation strategy was adopted; Ultimately, most of little spatial and temporal variability of sensitive parameters were static optimized and assimilated directly, a few key large spatial and temporal variability of sensitive parameters were dynamic synchronous optimized with state variables update. Meanwhile, through tra

英文关键词： ecological process model；water and carbon fluxes；data assimilation；spatial-temporal heterogeneity；priori knowledge