PEAK DISCHARGE ESTIMATES OF A PLEISTOCENE LAVA-DAM OUTBURST FLOOD, WESTERN GRAND CANYON, ARIZONA, USA

An unsteady flow, dam-break model is used to reconstruct the flood wave that emplaced the a lava-dam outburst-flood deposit (Qfd4) along the Colorado River in western Grand Canyon, Arizona. Sedimentological evidence supports the hypothesis that the deposit was emplaced by a large-scale flood that resulted from the catastrophic failure of an upstream lava dam. The water-surface profile of the flood is reconstructed by means of cosmogenic 3He ages and chemical compositions of basalt clasts in the Qfd4 unit. The Qfd4 unit was deposited by a flood that occurred 165 ± 18 ka upon the catastrophic failure of the Hyaloclastite Dam at river mile (RM) 188.5. The dam was approximately 280 m high and impounded a maximum of 9×109 m3 in its reservoir. The exact geometry of the dam is uncertain, as only remnants remain. Preliminary results do not exactly match the water-surface profile preserved in flood deposits. By using a range of dam heights constrained by field evidence that represent various reservoir volumes, a range of discharges for the Qfd4 flood were determined that encompass all high-water marks. Peak discharges (Qp) of 2.8-4.8×105 m3/s for the Qfd4 flood were produced by model with dam heights of 210 and 300 m. Qp dissipated to a maximum of 1.6×105 m3/s at a distance of 59 km downstream from the dam. Qp estimates are an order of magnitude larger than both the largest Holocene runoff flood (1.4×104 m3/s) in the geologic and historic records and the predicted probable maximum flood (2.0×104 m3/s) on the Colorado River near Lee’s Ferry. Present-day geometry was used to route the flood wave and to calculate reservoir volumes. Manning’s n values ranged from 0.03 to 0.1, where the highest values occurred nearest the dam breach. Increased sediment and debris as well as energy losses associated with free-surface deformation and waves, like that which would accompany a dam failure, require higher n values in the model in order to match preserved water-surface profiles. In addition, there is no certain way to account for aggradation of the river channel nearest the dam breach due to the rapid injection of rubble and debris into the river; Qp may be overestimated at the breach. The greatest uncertainty, however, is in the geometries of the dams and their breaches.