A Performance-Based Modified RRV Framework for Evaluating the Dynamic Operation of Urban Flood Control Storage Systems

Urban flood control basins without forced drainage operate as one-shot systems: they must perform within a single event and cannot empty themselves until it ends. The timing of gate closure is therefore decisive—closing early wastes storage, whereas closing late lets stored water drain back to the falling channel—yet the recovery-based Reliability–Resiliency–Vulnerability (RRV) framework does not capture this trade-off. We reformulated the RRV indices for non-drainable basins and applied them to a full-scale experimental channel reproduced with high-resolution Structure-from-Motion terrain and simulated with the two-dimensional Nays2D Flood model. The three indices provide complementary performance dimensions: Reliability describes temporal threshold attainment, Resiliency describes event-integrated retention, and Vulnerability describes the maximum instantaneous deficit. Five gate-closing scenarios were compared with the no-gate reference. For this single-site, single-hydrograph event on the calibrated 211 × 111 grid, the net inflow at the inlet reversed at about t = 2611 s, and closures made before this instant avoided backflow. Closing just before reversal (t = 2574 s) gave the highest Resiliency (0.995), the lowest Vulnerability (0.074), and the largest net inflow (1487 m3) among the tested scenarios; delaying closure reduced Resiliency to 0.809 and raised Vulnerability to 0.255. The exact values are specific to the adopted grid and event, whereas the principal result is the reversal-based closure window. The method identifies this window using observed or forecast water level or flow direction information, without requiring new forced-drainage or pumping infrastructure.