MODELING CATACLYSMIC OUTBURST FLOODS FROM PLEISTOCENE GLACIAL LAKE MISSOULA

We have modeled with the shallow water code of GeoClaw (www.geoclaw.org) six different scenarios of Missoula flooding reflecting different plausible boundary conditions. These scenarios test combinations of different positions of the Okanogan ice lobe (and thus different degrees to which Columbia River is blocked by ice or upper Grand Coulee is blocked by rock or ice). All scenarios assume that the maximum stage for glacial Lake Missoula is 1265 m and that the ice dam failed instantaneously, and results in rapid flooding of vast areas of eastern Washington. Results show flood discharge restricted by the narrow Columbia River gorge downstream of the Channeled Scabland within 38 to 55 hours. This channel constriction backfloods Umatilla and Pasco Basins, and the Yakima and Walla Walla valleys connected to Pasco basin. Flood timing and maximum flood stages depend on flood pathways into these basins. And these pathways, in turn, depend on the extent to which the channel of the Columbia River is blocked by the Okanogan lobe, and whether or not water can flow into channels feeding either Grand Coulee or Moses Coulee. Simulations to date show that maximum stages in Pasco Basin occur when channels into upper Grand Coulee, Moses Coulee, and the Columbia River valley are all blocked by ice, and most Missoula floodwater drains south through the Cheney-Palouse scabland tract. If Moses Coulee is open and the entrance to Grand Coulee is blocked by rock, then Moses Coulee becomes a major flood route that feeds water back into Grand Coulee south of the blockage. The lowest stages in Pasco Basin occur when advancing Okanogan ice does not block the Columbia River valley, allowing most flow to follow the river channel through north-central Washington. These varied model scenarios illustrate that flooding from the gigantic Missoula floods does not reach maximum stage everywhere for any one scenario. Rather, different scenarios each produce different maximum stages in different places, and provide a focus for future field studies.

Denlinger, R.P. and D.R.H. O’Connell, doi: 10.1130/B26454.1