It is of importance when building infrastructure in relation to rivers to be able to predict the maximum discharge, and the maximum water level, of the river, as well as the duration and return period of these maxima. More often than not, the prediction should be made for a longer period than what data is available for. The statistical analysis of data may therefore include also a regional classification, to increase the accuracy in the predictions. Furthermore, changes in the drainage area have to be considered, since obviously the prediction should be made for the future, but the data are from the past.
Example: Analysis of Extreme Flows, Costa Rica
At present a project is under way in Costa Rica for developing better prediction methods for extreme flows for design purposes. The client is ICE, Instituto Costaricence de Electricidad, who will use the resulting prediction model for the safe design of future hydropower projects.