Data denial experiments for extratropical transition

Data denial experiments using the European Centre for Medium-Range Weather Forecasts (ECMWF) model are designed to investigate the value of targeted observations for historical extratropical transition (ET) cases over the Atlantic. The impact of removing data from specified locations linked to the ET development is therefore examined. It is shown that the impact of denying data in the near tropical cyclone (TC) environment is, on average, as important as denying data in mid-latitude sensitive regions determined using extratropical singular vectors (SV). The impact of data denial over TC regions propagates downstream from the Atlantic towards Europe, with a maximum degradation at day 4. This degradation is mainly attributed to the data denial at the TC stage, i.e. before ET is completed. When data are denied on mid-latitude sensitive regions, the largest degradation is found around day 2 and also after the day 4 forecast. In general, the loss of information content is larger when data are denied in mid-latitude sensitive areas because these identify dynamically active regions. In both denial experiments, aircraft and satellite radiance data are the most influential observations. For the selected case of Hurricane Irene, the largest degradations are found for forecasts initialised while Irene reached its peak intensity. If observations are denied in the near storm environment, the TC mostly disappears from the analysis and the subsequent forecast. This allows the impact of Irene on the formation of the downstream cut-off low to be investigated.