An automated and rapid method for identifying dam wall locations and estimating reservoir yield over large areas

Global demand for water, food and energy has seen the construction and planning of large dams continue at a steady pace in many parts of the world. However, the process of 'exhaustively' examining all potential dams sites within an area as part of an initial scoping study is still largely undertaken using manual methods. This paper describes DamSite, a series of novel algorithms that construct 'virtual' dam walls at every pixel along every channel within a catchment, including saddle dams where required by the terrain. By repetitively calculating dam and reservoir dimensions, reservoir yield and dam costs along a river network and for incrementally higher dam walls at each location it is possible to identify both optimal dam wall locations and optimal dam wall height at a given location. The DamSite model was tested in two catchments in northern Australia and accurately pin-pointed previously identified potential dam locations.