Data assimilation is used in atmospheric chemistry models to improve airquality forecasts, construct re-analyses of three-dimensional chemical(including aerosol) concentrations and perform inverse modeling of inputvariables or model parameters (e.g., emissions). Coupled chemistrymeteorology models (CCMM) are atmospheric chemistry models that simulatemeteorological processes and chemical transformations jointly. They offerthe possibility to assimilate both meteorological and chemical data;however, because CCMM are fairly recent, data assimilation in CCMM has beenlimited to date. We review here the current status of data assimilation inatmospheric chemistry models with a particular focus on future prospects fordata assimilation in CCMM. We first review the methods available for dataassimilation in atmospheric models, including variational methods, ensembleKalman filters, and hybrid methods. Next, we review past applications thathave included chemical data assimilation in chemical transport models (CTM)and in CCMM. Observational data sets available for chemical dataassimilation are described, including surface data, surface-based remotesensing, airborne data, and satellite data. Several case studies of chemicaldata assimilation in CCMM are presented to highlight the benefits obtainedby assimilating chemical data in CCMM. A case study of data assimilation toconstrain emissions is also presented. There are few examples to date ofjoint meteorological and chemical data assimilation in CCMM and potentialdifficulties associated with data assimilation in CCMM are discussed. As thenumber of variables being assimilated increases, it is essential tocharacterize correctly the errors; in particular, the specification of errorcross-correlations may be problematic. In some cases, offline diagnosticsare necessary to ensure that data assimilation can truly improve modelperformance. However, the main challenge is likely to be the paucity ofchemical data available for assimilation in CCMM.
展开▼