Sensitivity of outflow peaks and flood stages to the selection of dam breach parameters and simulation models

Abstract Important breach parameters were identified and their ranges were estimated from a detailed study of historical earthdam failures due to overtopping. The U.S. Army Corps of Engineers Hydrologic Engineering Center (HEC) and the National Weather Service (NWS) dam breach models were chosen for evaluation and simulation. Both models use similar input data and breach descriptions, but the HEC uses the hydrologic routing method (modified Puls method), whereas the NWS uses the St. Venant equations for routing. Information on eight dams in Illinois was taken from the Corps of Engineers inspection reports, and surveyed cross-sections of the downstream channels were supplied by the Division of Water Resources of the Illinois Department of Transportation. Various combinations of breach parameters (failure time, TF; depth of overtopping, hf; and breach size, B) were used for breach simulations by both methods with the 1.00PMF, 0.50PMF and 0.25PMF (probable maximum flood) inflow hydrographs. In general, the flood stage profiles predicted by the NWS were smoother and more reasonable than those predicted by the HEC. For channels with relatively steep slopes, the methods compared fairly well, whereas for the channels with mild slope, the HEC model often predicted oscillating, erratic flood stages, mainly due to its inability to route flood waves satisfactorily in non-prismatic channels. The breach outflow peaks are affected significantly by B but less so by hf. The ratio of outflow peak to inflow peak and the effect of TF on outflow decrease as the drainage area above the dam and impounded storage increase. Flood stage profiles predicted with cross-sections taken from 7.5′ maps compared favorably with those predicted using surveyed cross-sections. For the range of breach parameters studied, the range of outflow peaks and flood stages downstream from the dam can be determined for regulatory and disaster prevention measures.