Tens of successive, colossal Missoula floods at north and east margins of channeled scabland

Abstract

In deposits of Pleistocene glacial lakes in northern Idaho and Washington, beds comprising 20 to 55 varves (average = 35-40) separate each successive graded gravel or sand bed that was swiftly emplaced by a catastrophic flood from glacial Lake Missoula. The floodlaid beds are similar to rhythmic successions of 40 or more graded beds in backflooded tributaries of the lower Columbia River. This new field evidence corroborates a controversial hypothesis that the great Pleistocene floods from glacial Lake Missoula were 40 or more colossal, separate joekulhlaups, and refutes the conventional notion that any two successive graded beds were deposited by one flood. The only outlet of the 2000-km3 glacial Lake Missoula was through its great ice dam. Calculations show that each time the lake rose to about 600 m deep, it made the glacier buoyant and engendered a catastrophic discharge along the glacier bed (a joekulhlaup). A reconstructed water budget suggests that after a complete draining, the lake refilled in 3 to 6 decades; thus the hydrostatic prerequisites for a joekulhlaup were reestablished dozens of times during the late-Wisconsin episode of lake damming. Various intercalated tephra layers, radiocarbon dates, varve successions, and the Bonneville flood deposits in the region suggest that late-Wisconsin glacial Lake Missoula existed for about 2 millennia within the period 15,000 to 12,700(?) yr ago. Varve beds indicate that the mean period between Missoula floods was about 4 decades, but became shorter during the last several floods. Between 20 and 30 of the Missoula joekulhlaups occurred after the single great flood from Lake Bonneville, which according to 14C dating in the Bonneville basin by W. E. Scott and associates and by D. R. Currey occurred some time between 15,000 and 14,000 yr ago.