利用模拟退火算法优化Biome-BGC模型参数  

Using simulated annealing algorithm to optimize the parameters of Biome-BGC model

作　　者：张廷龙[1,2,3] 孙睿[1,2,3] 胡波 冯丽超[1,2,3] 

机构地区：[1]遥感科学国家重点实验室(北京师范大学/中国科学院遥感应用研究所),北京100875 [2]北京师范大学地理学与遥感科学学院,北京100875 [3]北京市环境遥感与数字城市重点实验室,北京100875 [4]宁波市气象局,浙江宁波315012

出　　处：《生态学杂志》2011年第2期408-414,共7页Chinese Journal of Ecology

基　　金：国家自然科学基金项目(40971221)资助

摘　　要：生态过程模型建立在明确的机理之上,能够较好地模拟陆地生态系统的行为和特征,但模型众多的参数,成为模型具体应用的瓶颈。本文以Biome-BGC模型为例,采用模拟退火算法,对其生理、生态参数进行优化。在优化过程中,先对待优化参数进行了选择,然后采取逐步优化的方法进行优化。结果表明,使用优化后的参数,模型模拟结果与实际观测更为接近,参数优化能有效地降低模型模拟的不确定性。文中参数优化的过程和方法,可为生态模型的参数识别和优化提供一种实例和思路,有助于生态模型应用区域的扩展。Ecological process model based on defined mechanism can well simulate the dynamic behaviors and features of terrestrial ecosystem,but could become a bottleneck in application because of numerous parameters needed to be confirmed. In this paper,simulated annealing algorithm was used to optimize the physiological and ecological parameters of Biome-BGC model. The first step was to choose some of these parameters to optimize,and then,gradually optimized these parameters. By using the optimized parameters,the model simulation results were much more close to the observed data,and the parameter optimization could effectively reduce the uncertainty of model simulation. The parameter optimization method used in this paper could provide a case and an idea for the parameter identification and optimization of ecological process models,and also,help to expand the application area of the models.

关 键 词：碳通量 Biome-BGC模型 参数优化 

分 类 号：Q14-3[生物学—生态学]