Glacier Outburst Floods From “Hazard Lake”, Yukon Territory, and the Problem of Flood Magnitude Prediction

• Published in: Journal of Glaciology
• Main Author: Clarke, Garry K. C.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Cambridge University Press (CUP) 1982
• Subjects: glacier*; Journal of Glaciology; Yukon
• Online Access: http://dx.doi.org/10.1017/s0022143000011746; https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S0022143000011746

Abstract In August 1978 “Hazard Lake” released 19.62 × 10 6 m 3 of water through a subglacial tunnel beneath Steele Glacier, Yukon Territory, Canada. The discharge during the outburst flood was measured by recording lake level changes with time, and a peak discharge of approximately 640 m 3 s –1 was estimated from the data. We have attempted to model the 1978 flood from “Hazard Lake” using an adaptation of Nye’s (1976) theoretical model for jökulhlaups from Grimsvötn. Our aim has been to calibrate the Nye model as a first step toward using it as a peak discharge estimator for other glacier–dammed basins. The agreement between our measured and simulated hydrographs is good, and we find that creep closure, though included in our analysis, appears to play an insignificant role in limiting the discharge of “Hazard Lake”. Release of thermal energy from the relatively warm lake water is the dominant factor contributing to tunnel enlargement. The Manning roughness of outlet channels from glacier–dammed lakes is not known a priori and must either be assumed or estimated after the fact from the flood hydrograph. For “Hazard Lake” our fit implies Manning roughness in the range n ′ = 0.105 m –1⁄3 s, consistent with Nye’s estimate of n ′ = 0.12 m l⁄3 s for the 1972 Grimsvötn flood and our estimate of n ′ = 0.12 m 1/3 s for the 1967 Summit Lake flood. If the Manning roughness for flood conduits can be shown to lie within a narrow range, this would constrain one of the least certain variables of the Nye model. By making several simplifying assumptions, we have succeeded in reducing our adapted version of Nye’s model to a simple mathematical description involving dimensionless numbers characterizing reservoir geometry and the relative magnitudes of creep closure and tunnel enlargement by melting. In this simplified form, the influence of lake temperature, reservoir geometry, and creep closure on the character of flood hydrographs can be conveniently studied.