Steven Dutch, Professor Emeritus, Natural and Applied Sciences, Universityof Wisconsin - Green Bay
Above: map of the Lake Pend Oreille area in northern Idaho. The area is northeast of Spokane and southwest of Glacier National Park. The large lake is Lake Pend Oreille. The west-flowing river joins the Spokane River and then the Columbia. The town of Sandpoint is on the northern arm of the lake just above the long causeway. The Navy, interestingly, has a research facility in Sandpoint because the lake is 700 feet deep and provides a good laboratory for equipment testing. Also, no enemy ship has ever made it beyond Spokane.
The ice dam at its maximum extent, with Lake Missoula impounded by it.
This map and those that follow are exercises in visualization. No attempt has been made to account for every detail of geomorphology. The small impounded lakes are plausible from the topography but no attempt has been made to reconstruct their drainages. There are some minor misregistrations between the hydrography and the digital elevation model used for elevation. It is possible that the ice had begun to retreat from its maximum extent when the Missoula Floods began, but the scenario presented here shows the initial flood breaching the intact dam. The partial moraine preserved at the southern end of the lake is shown hypothetically extending all the way across.
It is impossible to reconstruct the exact events of the dam failure. These maps are an attempt to show how it could have happened, but don't necessarily guarantee it did happen that way.
Ice floats, and as the impounded lake approached the top of the ice dam, hydraulic lifting would likely have begun to lift the ice off its bed, allowing high-pressure water to get under the ice. It seems plausible to guess that catastrophic calving of icebergs into the lake would have followed.
Having nibbled away at the ice sufficiently, it is possible that water could have found a drainage along the ice margin or over the ice.
Water beneath and within the ice could have enlarged channels by frictional heating, and erosion by surface drainage could also have eroded the ice front. It's possible that breakup of the ice front might have occurred and jokulhlaups began to pluck away at the terminus.
Eventually complete rupture of the ice dam would have occurred. In this scenario I picture all the likely failure modes operating in tandem and finally joining to create a complete breach, but one mechanism or another may have been dominant.
Peak discharge for the earliest and largest floods is estimated at 20 times the flow of the Amazon River, or more than the total discharge of all the rivers in the world. In places the floods were 300 meters deep and moving at 100 km/hr. That's water a thousand feet deep with the force of a fire hose.
Some geologists have claimed that there wasn't enough water in Lake Missoula to inundate the entire Scablands area, and have argued that subglacial outbursts from the north contributed, or even dominated. It's also a mystery how high velocities, clearly indicated by the field evidence, could have developed in deep parts of Lake Missoula.
After roughly a week, the lake has drained and the floods ebb.
When I first animated this process, I showed far more extensive breakup of the ice. But there's evidence for numerous floods spaced a few decades apart during the early retreat of the ice. That's a lot of ice to replace in only a few decades. At narrow spots in the Lake Missoula basin (like just below and left of the elevation scale), and along the lower Columbia River gorge, we know the water passed through gaps only a couple of kilometers wide, and historical failures of ice dams typically cut narrow gaps between the ice and the valley wall. So this scenario depicts only a narrow gap a couple of kilometers wide after the drainage of the lake. Something like that can be plugged in just a few decades. If the ice had already retreated close to the position shown above before the first flood, then failure of the dam would have been easier. In any case, the ice wouldn't have returned to the maximum position, but would merely have advanced enough to plug the gap.
Not only was Lake Missoula receiving precipitation and stream runoff, it was also getting meltwater from numerous retreating glaciers outside the map area, so refilling could have been pretty rapid.
Created 15 January 2005, Last Update 06 June 2020