In the last few years catastrophic rainfall events have occurred in the Mediterranean area, leading to floods, flash floods and shallow landsliding.
These recent events have outlined the urgent need for:
· the implementation of forecasting systems able to predict meteorological conditions leading to disastrous runoff occurrences;
· some policies for issuing warnings, or alarms, to local authorities and the population.
Indeed, early warning systems in urban areas appear to be the only non structural measure suitable for reducing flood risk, if diffused with enough lead time and adequate reliability.
With the arrival of new measuring systems (Meteorological Radar and Meteosat Satellites), and, especially, with the new possibilities offered by Limited Area Models (LAM's), it has finally become apparent that it is possible to create complex forecasting systems on a series of different time scales.
Hydrological basins can be classified in three classes, regarding the flood forecasting problem:
1. large basins (>10,000 km2), for which the flood forecast with up to 12 hours advance warning in one section can basically be made on the basis of the water levels;
2. intermediate basins (1,000-10,000 km2), for which the flood forecast with up to 12 hours advance warning can be made on the basis of the precipitation measurements alone both from land networks and radar maps;
3. small basins (100 - 1,000 km2), for which the flood forecast with up to 12 hours advance warning can be made only when a precipitation forecast is available, such as one that might for example be generated by a LAM;
Downstream of the meteorological forecasting, the various levels at which it is necessary to organise the system comprising the monitoring, early warning, forecasting and decision-making support to the management of flood event emergencies, must adhere to the following scheme:
1. real-time data acquisition;
2. real-time analysis, verification and, if necessary, reconstruction of the missing data;
3. automatic rainfall forecasts based on the latest measurements and on any forecasts that may have been generated by the Limited Area Models;
4. automatic flood forecasts, based on the observed measurements and the rainfall forecasts;
5. automatic flood forecasts and routing along the river reaches using hydraulic models which provide for the identification of the key critical areas.
6. implementation of possible intervention simulations and verification of their effects; in this last phase of analysis which corresponds to a declared state of emergency, all of the activities included in points 2,3,4 and 5 above, which represented the early warning phase and which necessarily had to be implemented in automatic form, can be reviewed and modified in assisted semi-automatic form, in order to verify the effects of possible alternative courses of action.