Observations for Air Quality Modeling Application and Evaluation

Air quality models and their application require observational data from a variety of sources. The quality and availability of these data are important in both the decisions and confidence resulting from these applications. Models are not perfect, nor are the data that drive the models. This paper will focus on the past, present and future atmospheric data available for use in the application and evaluation of air quality models. It will also describe efforts in progress for improved cooperation between observational scientists and the scientists who model the physical and chemical processes of air quality. Quality and homogeneity of data in both time and space are important in applying models for decisions. Establishing the confidence (or conversely, the uncertainty) in observational data is important to the credibility of the models. Model reanalysis is one way to approach this issue, but is a very demanding computational task. New capabilities such as data grids and grid computing are being investigated as tools to be used. Enhanced use of existing and future data from remote sensing platforms needs to be supported by better access (both timely and in forms that can be readily used) for both model improvements and model evaluation. Cooperative working relationships are being established to promote these goals.