基于大气化学模式中气溶胶变量优化组合的多源观测资料同化研究

项目名称： 基于大气化学模式中气溶胶变量优化组合的多源观测资料同化研究

项目编号： No.41275128

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 臧增亮

作者单位： 中国人民解放军理工大学

项目金额： 80万元

中文摘要： 大气气溶胶的精细化分析预报是环境污染和气候变化中的关键问题，大气化学模式及其同化系统是解决这一问题的重要工具。但气溶胶的资料同化面临很大困难，其一是模式中包含气溶胶的多种化学组分及粒径段变量，变量数目非常多；其二是观测资料的类型十分多样，如PM2.5、硫酸盐及卫星光学厚度资料等,且一般不与模式变量直接匹配。本项目在气溶胶多变量直接同化和单变量间接同化方法的基础上，提出一种新方法，其思想是基于观测变量对模式变量进行优化组合，然后对组合后的多变量进行同化，同化后再分配至模式变量。基于该方法并利用三维变分理论，针对WRF-Chem中MOSAIC方案的64个气溶胶模式变量，设计面向多源观测的同化系统，并重点解决误差协方差矩阵中交叉相关的计算及卫星光学厚度资料的同化问题。该方法可以针对不同的观测系统和模式系统进行变量的优化组合，实现对观测信息的最大化利用，为气溶胶的精细化分析提供新的技术和方法。

中文关键词： 资料同化；大气气溶胶；WRF-Chem 模式；背景误差协方差；气溶胶光学厚度

英文摘要： The analysis and forecast of atmoshheric aerosol are the key problems in the environmental pollution and climate change. Atmospheric chemistry model and its assimilation system are important tools to solve these problems. However, aerosol data assimilation faces great difficulties. Firstly, there are numbers of aerosol variables including different particle sizes and chemical components in the model. Secondly, the types of observation are very diverse, and most of them don't directly match with model variables, such as of PM2.5, sulfate and satellite optical thickness data. There are usually two methods of aerosol data assimilation. One is the indirect assimilation method for one variable, and the other is the direct assimilation method for all of the model variables. Based on the ideas of these two methods, we will develop a new intermediate method. At first, the model variables will be combinated optimally, then the combination of variables are assimilated as state variables. At last, the analyzed combination variables are distributed into the model variables. Based on this method and three-dimensional variational theory, the assimilation system for 64 model variables of WRF-Chem with MOSAIC scheme is designed.In this system, we focus on solving the multivariate cross-correlation in error covariance matrix and

英文关键词： data assimilation；atmospheric aerosols；WRF-Chem model；background error covariance；AOD