Climate change projections, essential for water resource management, are dominated by multiple sources of uncertainty, including scenario uncertainty and Internal Climate Variability (ICV). The latter is often classified as irreducible and thereby overlooked in traditional decision-making processes on regional and local scales. Our study translates these traditionally irreducible uncertainties into actionable insights by leveraging multiple initial condition ensemble (MICE) outputs, a physically based hydrological model and a multiobjective stochastic optimisation approach. We focus on the Sardar Sarovar Dam in Gujarat, India, a multipurpose reservoir integral to flood control, hydropower generation and meeting diverse water demands under two future climate scenarios (SSP245 and SSP585). We find that for the multipurpose reservoir under consideration, despite considering a broad spectrum of outputs, which are equally plausible due to the inherent variability of the climate, the system is highly reliable for meeting the drinking water supply for the region for the next
century, although at the cost of agricultural and industrial water supply. Incorporating ICV provides a robust assessment of system attributes, including reliability, vulnerability, and resilience, particularly when these objectives are in trade-off with each other. Our study emphasizes the critical role of accounting for ICV in water resource planning. This analytical approach allows stakeholders to pinpoint specific vulnerabilities, allowing targeted planning for more adaptable and resilient water resource management strategies, including sustainable water supply and flood control.