Within the frame of the project “ATMO Forensic Prediction and Analysis“, methods and routines will be developed that allow to estimate total losses caused by severe winter storms over Europe, based on weather forecasts or measurements. This project is part of the CEDIM focus on near-real time Forensic Disaster Analysis (FDA).
From selected representative station data (wind gusts), different weighted or unweighted storm indices will be calculated for past storm events, which considers both the intensity and the spatial extent of the storm field. These analyses are complemented by examinations of GFS and CFS model data, which will be adjusted by application of Model-output statistics (MOS). Wind gust parametrizations that also considers the stratification of the atmosphere, should further enhance the relation between modelled and observed gusts.
Past storm events will be classified according to a new damage index, which considers, for example, the number of damaged buildings, the number of fatalities and total amount of loss. All past and future events including name, date, regions affected, model data, measurement data and loss data will be fed into a data base (rationale and case-based reasoning, developed in the frame of CEDIM). Based on forecasted storm fields, a storm index may be estimated before the arrival of the storm. By using ensembles model, also information about the probability, most likely intensity, affected area, and the expected damage shall be provided.
The calculations will lead into a kind of “storm-MOS” that describes a storm event with reasonable probability, uncertainty and precision and that allows to evolve a storm damage function. Once the storm occurs, the wind measurements at representative locations on the one hand and the model predicted wind gusts on the other hand provide the storm indices and the appropriate damage index.
Figure: Winter storm "Joachim"; 850 hPa wind forecast (GFS Model), 13 December 2011