Articles | Volume 9, issue 2
https://doi.org/10.5194/nhess-9-623-2009
https://doi.org/10.5194/nhess-9-623-2009
27 Apr 2009
 | 27 Apr 2009

A severe blizzard event in Romania – a case study

F. Georgescu, S. Tascu, M. Caian, and D. Banciu

Abstract. During winter cold strong winds associated with snowfalls are not unusual for South and Southeastern Romania. The episode of 2–4 January 2008 was less usual due to its intensity and persistence. It happened after a long period (autumn 2006–autumn 2007) of mainly southerly circulations inducing warm weather, when the absolute record of the maximum temperature was registered. The important snowfalls and snowdrifts, leading to a consistent snow layer (up to 100 cm), produced serious transport and electricity supply perturbations.

Since this atypical local weather event was not correctly represented by the operational numerical forecasts, several cross-comparison numerical simulations were performed to analyze the relative role of the coupler/coupling models and to compare two ways of process-scale uncertainties mitigation: optimizing the forecast range and performing ensemble forecast through the perturbation of the lateral boundary conditions. The results underline, for this case, the importance of physical parametrization package on the first place and secondary, the importance of the model horizontal resolution. The resolution increase is beneficial only in the local process representation; on larger scale it may either improve or decrease the accuracy effect, depending on the specified nudging between large-scale and small-scale information. The event capture is likely to be favored by two elements: a more appropriate time-scale of the event's physics and the quality of the transmitted large-scale information. Concerning the time scale, the statistics on skill as a function of forecast range are shown to be a useful tool in order to increase the accuracy of the numerical simulations. Ensembles forecasting versus resolution increase experiments indicate, for such atypical events, an interesting supply in the forecast accuracy through the ensemble method when applied to correct the minimum skill of the deterministic forecast.

Download
Altmetrics