Figure 1. Domains on which ALADIN output surface wind fields are dynamically adapted.
Figure 2. ALADIN/HR: 12h forecast of 10 m agl wind field for 7th November 1999 00 UTC run
Stjepan Ivatek-Sahdan, Martina Tudor and Alica Bajic
Meteorological and Hydrological Service of Croatia
Planning, constructing and using roads and bridges assume insight into behaviour of those meteorological parameters that influence their safety and functionality. The most important meteorological parameter that affects traffic along the Adriatic coast is wind. Each winter several damaging Bura storms hit the coastal region and the islands of Croatia strongly affecting road transport and life in general. During the past years frequent severe winds produced serious difficulties (cars overturning) in traffic across the Dalmatian bridge Maslenica that lies on the main route between northern Croatia and its southern parts. So there is a strong necessity of having proper warning service that enable fast action (stop the traffic across the bridge in order to prevent the accidents and traffic jams). The basis for the warning system is adequate local weather forecast.
Croatian coast is lined with mountains that are relatively small in area, but rather steep with tops above 1500 m that are very close to the coastline. Orographically induced weather patterns in that part are poorly resolved or not resolved at all in large-scale models. Thus, forecast using high resolution models like the ALADIN become crucial.
In NWP practice in Croatia products from integration on the LACE domain with 12 km resolution are used for weather forecasting, but also as input for integration on HRv8 domain with 8 km resolution (SW corner 41.79° N, 8.93° E; NE corner 49.53° N, 21.98° E). Output surface wind fields are dynamically adapted to orography with 2 km resolution. Area around the Maslenica bridge is one of the domains on which dynamical adaptation of surface wind a field is done (Figure 1).
The possibility of the Aladin model to forecast the strength and onset time of Bura wind is analysed in the MAP IOP 15 situation 5-10 November 1999 characterized by the strong Bura wind along the northern and middle Adriatic.
Forecast of the 10 m wind field obtain using ALADIN/HRVATSKA (Figure 2) without and with dynamically adapted wind field show great influence of orography and consequently great variability in wind velocity. Vertical cross-section indicates the downstream low level jet characteristic for every Bura layer. The Bura layer (NE wind) extends to 1500-2000 m. The SE wind aloft decouples the upper and lower regions and prevents disturbances aloft. The obtained presence of stable layer descent from upstream to downstream region is consistent with upstream acceleration.
Spatial changes in wind speed observed in the dynamically adapted wind field confirm to give realistic pattern of wind speed in that region. This could be seen from the comparison of the ALADIN model output data with wind measurements on Maslenica location (indicated on Figure 2) that are given on Figure 3. It is obvious that in considered case the ALADIN/HR forecast (8 km resolution) produces too weak winds for that specific location. Dynamical adaptation gives wind speeds that significantly better correspond to the observations.
Although more Bura cases have to be analysed presented analysis and existing forecast practice show that the applicability of dynamically adapted wind fields is big and especially important in issuing the strong wind warnings for a number of special users.
Figure 1. Domains on which ALADIN output surface wind fields are dynamically adapted.
Figure 2. ALADIN/HR: 12h forecast of 10 m agl wind field for 7th November 1999 00 UTC run
Top left : vectors and speed in m/s as shaded areas, resol. 8 km
Top right : vertical cross section along A-B line, resol. 8 km
Bottom : ALADIN/HR + dynamical adaptation vectors and speed in m/s as shaded areas resol. 2 km
Figure 3.
Wind speed measured data and results of ALADIN simulation with 8 km and 2 km grid models are compared for Zadar and Maslenica locations : period from 7th November 1999 00 UTC to 9th November 1999 00 UTC. 10 min wind - wind speed averaged over a period of 10 minutes; gust - maximum value of the momentary turbulent wind (measuring period of 1 second); model output is wind speed in nearest time step of model
 |
Home |