Quantifying uncertainty in projected temporal variations of reservoir releases for crop water requirement

Abstract

Climate change brings about significant uncertainties in reservoir operation and agricultural water management. This study aims to assess the temporal variability and associated uncertainties in reservoir release projections to meet the agricultural water demand for maize in the Beydag Reservoir, located in İzmir, Türkiye. Agricultural water demand was estimated based on the FAO Penman-Monteith potential evapotranspiration method, employing crop-specific coefficients for maize. Reservoir operation was carried out by means of the Standard Operating Policy (SOP). Future runoff projections were derived from a 140-member ensemble constructed by combining five general circulation models (GCMs), two emission scenarios (RCP4.5 and RCP8.5), two downscaling methods (statistical and dynamical), and seven calibrated hydrological models. Using these projections, annual reservoir releases under SOP were simulated and evaluated in terms of anomalies relative to a historical baseline. Uncertainty analysis based on variance decomposition revealed that GCMs exert a rather dominant influence on total projection uncertainty in reservoir releases. Moreover, the SOP approach demonstrated limited responsiveness to the increasing temporal variability imposed by climate change because it prioritizes meeting current demand without reserving water for future use. The results emphasize the critical need to consider uncertainty assessments for the planning of water-sensitive cropping strategies such as maize cultivation.