Application of geographical information systems and stochastic hydrology to the siting and design of water reservoirs

The design of water reservoirs requires the application of techniques related to probabilistic analysis of hydrologic variables. Siting and dimensioning of water reservoirs require the forecasting of future river water discharges based on registered past streamflows. Stochastic hydrologic time series modeling allows the generation of synthetic series (Monte Carlo simulation) that preserve the main statistical characteristics from the historic series. These generated series and the previewed water demands may be used for defining the necessary reservoir storage capacity for a fixed risk of the available water not fulfilling the demands. Geographical information systems (GIS) techniques are very useful for the selection of the best river transversal sections for building dams to create reservoirs. GIS allow a fast procedure for selecting the best sites taking into account the environmental impacts from inundation. This paper describes the joint application of GIS and stochastic hydrology to the choice of location and sizing of water reservoirs in the southern branch of Jucu watershed, located in the Espirito Santo State, southeast of Brazil.