<p>A novel two-stage adaptive multi-objective (M-O) parameterization-simulation-optimization (P-S-O) framework for long-term reservoir operation is proposed to obtain the optimal trade-off between the deficit performances of Irrigation (human needs) and E-flows, while ensuring the health of the river ecosystem. The first stage involves deriving the optimal monthly E-flow targets, by trading off between the modified shortage index (MSI) of Irrigation and HA. The novelty concerns the stage-2 formulation that aims to achieve an acceptable trade-off between MSIs of Irrigation and E-flows at the reservoir level, by transferring the E-flow targets derived from stage-1 and the pre-defined Irrigation demands and including a constraint on HA. The reservoir operation employs zone-based and transition rule curves besides fuzzy hedging factors with the hedging being triggered by monthly reservoir water availability. The Histogram Comparison Approach (HCA) is employed for the estimation of HA and the robust multi-objective evolutionary algorithm BORG is adopted as the search engine. The application of the proposed framework to the Bhadra river-reservoir system in Southern India illustrates its superiority over the conventional single-stage model in yielding a better trade-off between Irrigation and E-flow releases at the reservoir level for long-term as well as critical drought period operations. The robustness of the operating policy obtained from the proposed two-stage model over that from the single-stage model is brought out through uncertainty analysis as well. The applicability of the proposed model to varied hydrologic setting is demonstrated through an additional case example presented on the Hemavathy Reservoir system.</p>