engleski

Heavy rainfall and flash flood in Dubrovnik on 22nd November 2010

A case with the extreme precipitation in the morning on 23rd November 2010 in Dubrovnik, Croatia, is used as a testbed for the numerical weather prediction model ALADIN (Aire Limiteé Adaptation Dynamique dévelopement InterNational). In addition, the study includes sensitivity simulations with the WRF (Weather Research and Forecasting) model. The synoptic situation was characterized by a huge low moving over the Alps with the frontal system covering most of Central Europe and Mid-Mediterranean. During the passage of the cold front over the southern Italy and Adriatic Sea, a secondary cyclone developed and a meso-scale convective system grew within. The rain gauge measurements exceeded 100 mm/24hr in the area and the one in Dubrovnik measured 161.4 mm/24hr, with a peak intensity of 71.5 mm/h. This was the measured maximum for the event. The traffic and electricity supply were interrupted, water flow invaded and damaged homes and roads and the famous Stradun street in the old part of Dubrovnik was flooded by 40 centimetres of water. The disaster was reported in most of the Croatian media that day. The operational ALADIN forecast in the Croatian Meteorological and Hydrological Service (DHMZ) is run with 8 km horizontal resolution 72 hours in advance, twice per day, starting from 00 and 12 UTC analyses. It uses ARPEGE initial and boundary conditions and digital filter initialization (DFI). The parallel suite uses data assimilation (DA). The operational and parallel suite model results forecast large 24 hour accumulated rainfall amounts in the area around Dubrovnik, but the maxima over 100 mm are not situated above Dubrovnik, but above the surrounding areas of Montenegro and Bosnia and Hercegovina. The operational and data assimilation forecast runs that cover the 24 hour period form 06 UTC on 22nd to 06 UTC on 23rd November 2010 have forecasted the 24 hour accumulated precipitation ranging from 20 to 100 mm for Dubrovnik with the amount exceeding 100 mm/24hr in the neighbouring Montenegro. Several experiments using high-resolution (2 km) non-hydrostatic ALADIN and WRF model runs have been performed. The results of these runs show strong dependency on the model, input data from initial and lateral boundaries. The position of the precipitation maxima in the high resolution ALADIN output fields is very similar to the lower resolution (8 km) run used for initial and lateral boundary conditions. Therefore, different options for initial and boundary conditions will be tested. In DHMZ, two options are operationally available. The first set is from ARPEGE (Action de Recherche Petite Echelle Grande Echelle) and the second set is obtained from the IFS (Integrated Forecast System) model run operationally in ECMWF (European Center for Medium-Range Weather Forecast). A possibility of improving the initial and lateral boundary conditions via the data assimilation is investigated. In the high-resolution experiments of the ALADIN model, significant part of the precipitation was given by the convection scheme, when it was switched on. This result suggests the importance of using the convective parameterization in the ALADIN model even for resolutions in which it is assumed that the convection is resolved. Further experiments with a diagnostic convection scheme and without any deep convection scheme will be performed in order to investigate this problem.

flash flood; heavy rain

nije evidentirano