Risks and uncertainties in reservoir yield in highly variable intermittent rivers: case of the Castanhão reservoir in semi-arid Brazil

Abstract

We analyse the risks and uncertainties in the assessment of reservoir yields in Northeast Brazil using Monte Carlo simulation. The case study considers the Castanhão Reservoir, which is located in the Jaguaribe River in the State of Ceará. The hydrologic regime of the reservoir inflow was obtained from a 37-year historical series. From the data, 600 series of annual inflows for 30 years were generated. The operation of the reservoir was simulated to estimate the uncertainties and risks regarding the water supply by the reservoir. The variability of regulated streamflows with 30-year horizons, measured by their coefficients of variation, is 27%. It was also found that the mean probability of failure for the 600 traces, with a mean release equal to the mean yield and an assumed 10% risk of failure, was 12.5%, which is 25% greater than the value of 10% used to estimate the yield.