Water Management in Tarbela Dam by Bayesian Stochastic Dynamic Programming

Abstract

Pakistan stores less than 11% of its average annual river flows against 40% of world average. Low storage hurts economy in two ways: (1) It gives birth to more floods; and (2) keeps hydro power generation low. According to Water and Power Development Authority (WAPDA) construction of structures across rivers, especially dams, usefulness of dams and barrages cannot just be ignored in mitigating floods and their role in agriculture and cheap electricity generation. Construction of dams, barrages, dykes and embankments etc. needs huge funding, but cost is paid back over time. Although more dams are needed but we also need to operate our existing dams with best optimization policy. Dams have traditionally proven their usefulness with their multipurpose role in water storage so critical to economy, population and industry, inexpensive power generation and mitigating flood hazards. Tarbela dam was constructed in the 1970‟s on the river Indus in north central Pakistan. Location of Tarbela Reservoir is such that about 90% of annual river inflow is derived from snowmelt. The remainder of runoff is derived from rainfall in the Catchment area of 4,000 sq. miles lying immediately upstream of dam site