Hostname: page-component-77c89778f8-vsgnj Total loading time: 0 Render date: 2024-07-19T06:12:47.279Z Has data issue: false hasContentIssue false

Paleohydraulic Modelling of Outburst Floods From Neoglacial Lake Alsek, Yukon Territory, Canada

Published online by Cambridge University Press:  20 January 2017

Garry K.C. Clarke*
Affiliation:
Department of Geophysics and Astronomy, University of British Columbia, Vancouver, British Columbia V6T 1W5, Canada
Rights & Permissions [Opens in a new window]

Abstract

Type
Abstracts of Papers Presented at the Symposium but not Published in this Volume
Copyright
Copyright © International Glaciological Society 1989

Neoglacial Lake Alsek is a glacier-dammed lake that forms when surge advances of Lowell Glacier block the normal flow of Alsek River. When filled to the highest level, the reservoir held approximately 30 km3 of water and is believed to have been the largest ice-dammed lake to form since the end of the Wisconsin glaciation. One or more fillings in the twentieth century have been comparatively minor, impounding less than 2 km3 of water. Future fillings of Neoglacial Lake Alsek could have two main human impacts: (1) the level of impounded water could rise to inundate the town of Haines Junction and sections of the Haines and Alaska Highways; (2) catastrophic outburst floods could endanger down-stream development.

Over the past decade we have collected data on the reservoir geometry, the elevation of staircase beaches, the geometry of giant current ripple-marks, and the sedimentology of ponds that lie within the margins of Lake Alsek. These data cast light on various aspects of the history of cyclic filling and outburst flooding from the reservoir and can be synthesized to construct a paleohydraulic simulation model. Two past filling events (see Reference Clague and RamptonClague and Rampton, 1982) are singled out for analysis: a filling to 595 m a.s.l. occurred c. 1850 and impounded 4.7 km3 of water. A filling to 678 m a.s.l. occurred at some earlier time and impounded 2.84 km3 of water; this filling, though high, is not perhaps the maximum filling. Calculations using a previously described computer model (Reference ClarkeClarke, 1982; Reference Clarke, Mathews and PackClarke and others, 1984) place the peak discharge of these two outbursts at roughly 3.0 × 104 and 4.7 × 105 m3 s−1, respectively. If accurate, the latter estimate would establish the maximum floods from Neoglacial Lake Alsek as the greatest floods that have occurred since the end of the Wisconsin glaciation. For comparison, the mean discharge of all the world's rivers to oceans is 1.2 × 106m3s−1.

Dunes constructed from cobble-sized material are found in the present (subaerial) lake bed. These bed forms yield constraints on the hydraulic conditions that must have existed during past floods: (1) Water level must have exceeded the height of the dune features at the time they were formed. (2) Bed shear stress must have been sufficiently high to mobilize the coarse-grained material from which the dunes are constructed. (3) The Froude number must have been near 1.0 (actually somewhat less than 1.0 but this nicety adds nothing to a rough model) when the dunes were formed. These constraints on dune development can be used to test the model predictions against field observations that dunes can be observed as far up-stream as 47.2 km from the former ice dam. The approach followed in the paleohydraulic modelling is to generate a computer-predicted discharge hydrograph Q(0,t) for each flood and from this to calculate Q(l,t), the discharge Q as a function of time t, and up-stream distance l from the ice dam located at l = 0. let A(l,zw) be the up-stream reservoir area (for a water level zw at a distance i up-stream from the ice dam) and S(l,zw) be the cross-sectional area of the channel. From the model-generated hydrograph Q(0,t) and water level zw(t), the following can be calculated:

(discharge at up-stream distance l)

(mean velocity)

(Froude number)

(bed shear)

(grain diameter trom Shield's

where zfloor is elevation of valley floor; cf is frictional drag coefficient; ρw is density of water; D is grain diameter (mm); ρs is grain density; g is gravitational acceleration. The paleohydraulic calculations of h(l,t) = zw(t) - zfloor(l), Fr(l,t) and D(l,t) reveal that the predicted flood for the 678 m a.s.l. filling could have generated the existing bed forms but the 595 m a.s.l. flood could not. This test indicates that only a very great flood could have created the bed forms now found in the Lake Alsek bed.

References

Clague, J.J. Rampton, V.N. 1982 Neoglacial Lake Alsek Can. J. Earth Sci. 19(1), 94117.CrossRefGoogle Scholar
Clarke, G.K.C. 1982 Glacier outburst floods from “Hazard Lake” Yukon Territory, and the problem of flood magnitude prediction. J. Glaciol. 28(98), 321.CrossRefGoogle Scholar
Clarke, G.K.C. Mathews, W.H. Pack, R.T. 1984 Outburst floods from glacial Lake Missoula Quat. Res. 22(3), 289299.CrossRefGoogle Scholar