Reservoir performance and dynamic management under plausible assumptions of future climate over seasons to decades

An analysis procedure is developed to explore the robustness and overall productivity of reservoir management under plausible assumptions about climate fluctuation and change. Results are presented based on a stylized version of a multi-use reservoir management model adapted from Angat Dam, Philippines. Analysis focuses on October-March, during which climatological inflow declines as the dry season arrives, and reservoir management becomes critical and challenging. Inflow is assumed to be impacted by climate fluctuations representing interannual variation (white noise), decadal to multidecadal variability (MDV, here represented by a stochastic autoregressive process) and global change (GC), here represented by a systematic linear trend in seasonal inflow total over the simulation period of 2008-2047. Stochastic (Monte Carlo) simulations are undertaken to explore reservoir performance. In this way, reservoir reliability and risk of extreme persistent water deficit are assessed in the presence of different combinations and magnitudes of GC and MDV. The effectiveness of dynamic management is then explored as a possible climate change adaptation practice, focusing on reservoir performance in the presence of a 20 % downward inflow trend. In these dynamic management experiments, the October-March water allocation each year is adjusted based on seasonal forecasts and updated climate normals. The results illustrate how, in the near-term, MDV can be as significant as GC in impact for this kind of climate-related problem. The results also illustrate how dynamic management can mitigate the impacts. Overall, this type of analysis can deliver guidance on the expected benefits and risks of different management strategies and climate scenarios.