Alan E. Kehew

Origin and large-scale erosional features of glacial-lake spillways in the northern Great Plains

Deglaciation of the northern Great Plains produced two basic types of fluvial systems: (1) depositional regimes produced by the action of glacial melt water that formed outwash plains and (2) highly erosive systems produced by the sudden, rapid drainage of glacial lakes that formed spillways. Spillways are huge, deeply incised trenches that display all of the characteristics of fluvial channels. Glacial lakes drained in weeks or months, and discharges of ∼10 5 m 3 /s were attained at bankfull stage. Seven spillway systems studied were formed as a consequence of the sudden, rapid draining of glacial lakes; many other large valleys in the mid-continent area and other glaciated regions were probably formed by the same process. Spillways consist of deeply incised inner channels commonly flanked by broad, scoured outer zones. Initial stages of flow produced shallow, anastomosing channels; continued erosion produced longitudinal grooves, streamlined erosional residuals, and surficial boulder-lag deposits. Inner channels developed by the enlargement of one or more centrally located longitudinal grooves. The development of erosional residuals was initiated by the incision of shallow anastomosing channels. The interchannel areas were progressively modified from irregular, quadrilateral, or elliptical shapes to an equilibrium, minimum-drag lemniscate shape as glacial-lake drainage proceeded. Correlations between length and width, length and area, width and area, and length/width ratio and k (a shape parameter of the lemniscate loop) are very high for all erosional residuals studied, although only ∼20% of the residuals display an equilibrium lemniscate shape. Stable relationships between length, width, and area of erosional residuals are therefore established very early and are maintained throughout the erosional modification to the lemniscate shape. The only shape factor tested which can be used to quantitatively differentiate lemniscate from nonlemniscate erosional residuals is the position of maximum width, as measured by the ratio of the length from the lee to the point of maximum width to the total length, Xm/L . The mean Xm/L ratio for erosional residuals in a spillway segment is an indication of the relative duration of flood flows that produced the spillway.

Catastrophic flood hypothesis for the origin of the Souris spillway, Saskatchewan and North Dakota.

Glacial-lake spillways in the midcontinent region consist of narrow, deep trenches connecting proglacial lake basins. Geomorphic and sedimentologic evidence from the Souris and Des Lacs spillways in Saskatchewan and North Dakota indicates a catastrophic drainage of Glacial Lake Regina. The Souris spillway in Saskatchewan is composed of a broad, upper scoured surface 5 to 10 km wide, which merges with the floor of Glacial Lake Regina. The upper surface is longitudinally grooved and mantled with coarse lag deposits. An inner trench, 1 km wide and 20 to 45 m deep, occupies the center of the spillway. The two-level morphology of the spillway probably developed by initial erosion of the upper level, followed by incision of the inner trench, a sequence of events shown by flume experiments performed by Shepherd and Schumm. Near the Saskatchewan and North Dakota border, the flood water bifurcated into two branches as part of the water spilled over a divide into another spillway channel. The two spillways in North Dakota, pre-existing meltwater channels, were occupied and enlarged by the discharge from Glacial Lake Regina. Along the path of the spillways leading to the Glacial Lake Souris basin, water repeatedly breached the sides of the channels, eroding shallow channels leading away from the spillway. Near Minot, North Dakota, a significant amount of the water spilled out toward the east and eroded a plexus of anastomosing channels with streamlined erosional remnants between the channels. Deposits of the flood include huge point bars located at the inside of each channel meander. Internally, the bars generally consist of poorly sorted, indistinctly bedded, coarse gravel containing boulders of resistant and nonresistant lithologies as much as 3 m in diameter. Peak discharge of 1 × 10 5 m 3 /s or more may have been achieved at the outlet of the lake. At this rate, Glacial Lake Regina would have drained rapidly, probably in less than a month. The most likely cause for the rapid drainage of such proglacial lakes as Glacial Lake Regina was high discharge inflow of water from an upstream source along the ice margin.