Abstract

Air quality models can be used for a variety of applications including the prediction of high pollution episodes, to determine the suitability of new pollution source sites and to inform decisions on air pollution strategy and regulation. Thus, it is important that they are evaluated against observations regularly to determine their predictive capability over the range of spatial and temporal scales for which they are applied.

A new technique for the evaluation of air quality forecasts, SAL , will be demonstrated in this talk. This novel evaluation method provides more informative feedback on some of the physical aspects of forecast/observation differences, by taking into account the spatial correlation existing within pollution fields. The results of the evaluation provide an unbiased quantitative evaluation of air quality models that is motivated by a subjective visual assessment of the forecasts. This cannot currently be achieved with traditional air quality evaluation metrics.