Rainfall forecasts using weather radar
With actual radar data and rain gauge measurements, the SCOUT software is able to create Short-term precipitation forecast over a forecast period of 1 to 3 hours. Superior to numeral weather forecasts (NWP), the nowcasts are based on image recognition and extrapolation from radar measurements. With this, the nowcasts fulfill the function similar to a human observer, who would create the movements of rainfall cells like a motion picture made of radar pictures within his mind`s eye. The time necessary for computing is substantially less than the time that is needed for numeral weather forecasts. Forecasts are available in near real time.
The development of rainfall in the atmosphere is the result of a complex interaction of humidity, temperature, wind, the availability of condensation nuclei and other atmospheric parameters. These data cannot be measured in the whole atmosphere. Therefore, simulation and forecast of rainfall are, as a matter of principle, prone to uncertainty. Especially convective rainfall cells only have a dimension of a few kilometers. Their life time may be less than half an hour and they change their shape and size very fast, which makes the prediction of their development quite difficult. An ensemble forecast contributes to describing forecast uncertainty for users in a more visible and quantifiable way.
An ensemble of forecasts is made of several forecast trials, based on different kinds of development possibilities. Input data with their specific uncertainty bandwidth will be used for the calculation and lead to a variety of forecasts. With the SCOUT software ensemble nowcasts are calculated, where the uncertainty for the identification of the cell and the development of rainfall cells are considered.
Due to weather situations the results from ensembles can be very similar to each other or clearly different. This can be used, for example, to derive the probability of exceeding a threshold value.