Articles | Volume 24, issue 4
https://doi.org/10.5194/nhess-24-1539-2024
https://doi.org/10.5194/nhess-24-1539-2024
Research article
 | Highlight paper
 | 
02 May 2024
Research article | Highlight paper |  | 02 May 2024

Improving seasonal predictions of German Bight storm activity

Daniel Krieger, Sebastian Brune, Johanna Baehr, and Ralf Weisse

Data sets

MPI-ESM-LR1.2 decadal hindcast ensemble 3-hourly German Bight MSLP Daniel Krieger and Sebastian Brune http://hdl.handle.net/21.14106/04bc4cb2c0871f37433a73ee38189690955e1f90

MPI-ESM-LR1.2 decadal hindcast ensemble yearly German Bight storm activity Daniel Krieger and Sebastian Brune http://hdl.handle.net/21.14106/e14ca8b63ccb46f2b6c9ed56227a0ac097392d0d

ERA5 monthly averaged data on pressure levels from 1940 to present H. Hersbach et al. https://doi.org/10.24381/cds.6860a573

MPI-ESM-LR_1.2.01p5 decadal predictions localEnKF large ensemble: monthly mean values members 17 to 80 Sebastian Brune et al. https://hdl.handle.net/21.14106/c69ceecb1584cc50247ae6e492fb1ef33e65ac37

Download
Executive editor
Winter storms in the German Bight are a significant coastal hazard in the southeastern North Sea. The authors use a model for enhanced seasonal forecasting accuracy. This enhancement focuses on storms identified by the highest wind speeds, determined using sea-level pressure data, during winter. The forecasting system, comprising 64 simulations initiated each November, aims to forecast these storms for winters spanning from 1960 to 2018. Initial forecasts for the first winter proved inaccurate. However, by concentrating on specific weather patterns in September and November linked to these storms, the authors refined their forecasting method. Selecting the most reliable simulations based on these patterns significantly improved the forecasting accuracy for winter storms, indicating enhanced predictability of major atmospheric changes. Their simulations might be applied to other similar applications.
Short summary
Previous studies found that climate models can predict storm activity in the German Bight well for averages of 5–10 years but struggle in predicting the next winter season. Here, we improve winter storm activity predictions by linking them to physical phenomena that occur before the winter. We guess the winter storm activity from these phenomena and discard model solutions that stray too far from the guess. The remaining solutions then show much higher prediction skill for storm activity.
Altmetrics
Final-revised paper
Preprint