Predicting and evaluating extreme weather events

Barbara Brown

Short Abstract
High impact weather events that have large societal impacts often are extreme or rare. Examples of such events include hurricane landfall events which can lead to evacuation of cities and extensive damage from rain, wind, and severe weather; rainfall events that lead to severe flooding; and wind events that lead to tree and roof damage. Because of these impacts, it is highly desirable to be able to forecast the occurrence of such weather events with a large degree of accuracy and reliability to allow decision makers to take appropriate actions in anticipation of the forecasted event. Unfortunately, the extreme nature of these weather events typically also makes their prediction difficult. High temporal and spatial variability as well as limited frequencies of occurrence and poor or limited observations contribute to these predictive limitations. Moreover, evaluation of forecasts of these events - which can help lead to improvements in the forecasts - is also difficult due to many of the same factors that make the forecasting process difficult. Atmospheric scientists have striven to overcome these limitations using several strategies. These strategies have included (i) the evolution of numerical weather prediction (NWP) models (which provide gridded predictions of the translation and development of weather systems) toward higher temporal and spatial resolution, (ii) the inclusion of more realistic parameterizations of subgrid scale processes such as cloud microphysics, (iii) the integration of observations with NWP forecasts to provide more realistic representations of weather in subsequent short time periods, (iv) the use of statistical post-processing methods to create more localized predictions, and (v) the use of techniques to assess the uncertainty associated with the predicted events. The evolution of these methods of forecasting, the methods used for their evaluation, and their implications for extreme weather events will be considered in this talk.