Two classes of field evidence firmly establish that late Wisconsin glacial Lake Missoula drained periodically as scores of colossal jökulhlaups (glacier-outburst floods). (1) More than 40 successive, flood-laid, sand-to-silt graded rhythmites accumulated in back-flooded valleys in southern Washington. Hiatuses are indicated between flood-laid rhythmites by loess and volcanic ash beds. Disconformities and nonflood sediment between rhythmites are generally scant because precipitation was modest, slopes gentle, and time between floods short. (2) In several newly analyzed deposits of Pleistocene glacial lakes in northern Idaho and Washington, lake beds comprising 20 to 55 varves (average = 30–40) overlie each successive bed of Missoula-flood sediment. These and many other lines of evidence are hostile to the notion that any two successive major rhythmites were deposited by one flood; they dispel the notion that the prodigious floods numbered only a few.
The only outlet of the 2,500-km3 glacial Lake Missoula was through its great ice dam, and so the dam became incipiently buoyant before the lake could rise enough to spill over or around it. Like Grímsvötn, Iceland, Lake Missoula remained sealed as long as any segment of the glacial dam remained grounded; when the lake rose to a critical level ∼600 m in depth, the glacier bed at the seal became buoyant, initiating underflow from the lake. Subglacial tunnels then grew exponentially, leading to catastrophic discharge. Calculations of the water budget for the lake basin (including input from the Cordilleran ice sheet) suggest that the lakes filled every three to seven decades. The hydrostatic prerequisites for a jökulhlaup were thus re-established scores of times during the 2,000- to 2,500-yr episode of last-glacial damming.
J Harlen Bretz's “Spokane flood” outraged geologists six decades ago, partly because it seemed to flaunt catastrophism. The concept that Lake Missoula discharged regularly as jökulhlaups now accords Bretz's catastrophe with uniformitarian principles.