A NEW LOOK AT DISCHARGE ESTIMATES FOR GLACIAL RIVER WARREN, MINNESOTA, USA (Invited Presentation)

The prevailing hypothesis for the formation of glacial River Warren is that it was the spillway incised during initial catastrophic drawdown of Lake Agassiz, 13,000 ybp. Although there is debate about the number and size of subsequent drainage events, it appears that comparably large later discharges may not have occurred. This has significant archaeological implications because a large later discharge would have destroyed evidence for early Paleoindian occupation of the trench. If there were no subsequent large occupation, intact Folsom or Clovis-aged sites may exist. Other researchers are concurrently surveying new archeological sites in the upper valley, one of which may be Clovis-aged. We focus here on new discharge estimates for the channel-forming flow.

Former estimates, one by C. Matsch, a former student of H.E. Wright, were based on valley morphology and boulders that were presumed to have moved. We were fortunate to discover a new exposure with preservation of the bedload of glacial River Warren deposited among boulders interpreted as corestones. This allows us to review and constrain previous discharge estimates.

The exposure lies in a straight reach downstream of the spillway transition from a multi-thread to single-thread channel. The stream cut through layers of dense, clayey glacial sediment and typically did not leave recognizable bedload. At this location, however, the glacial river broadened because downcutting was impeded by a high in the bedrock surface. The exposure of imbricated cobble-to-boulder gravel lies in the lee of a bedrock knob. The stream removed the glacial sediment and a thick saprolith and exposed the underlying crystalline rock. The gravel lies directly on the undulating, jointed bedrock surface. Oversized corestones of the local bedrock remain among the gravel deposits.

We estimate paleoflow based on the median grain size of the gravel, channel morphology extracted from a LiDAR DEM, and depth constraints from a nearby suspended-load exposure and maximum depth of the valley. We compute paleodischarge based on simple open-channel flow and initiation of sediment motion estimates and with hydrodynamic modeling. We conclude that, even with the valley completely filled with water, it is unlikely that the corestones were mobilized as part of the bedload.