Ninth International Conference on Permafrost ... - IARC Research

Ni n t h In t e r n at i o n a l Co n f e r e n c e o n Pe r m a f r o s tbeneath thick organic mats associated with boreal forestcover.Surface probing behind the landslide headwall in latesummer 2004 indicated an active layer thickness of 42–86cm where the organic mat was at least 40 cm thick. Wherethe organic mat was thinner, the permafrost table was notencountered within 1.05 m of the ground surface, howeverit was assumed to be present at greater depth. Thermistormeasurements and lateral probing into the lower headwalldid not detect permafrost in the glaciofluvial unit at this time(ground temperatures 1 m into the face were ~1°C).DC resistivity surveys conducted in 2006 suggest thatpermafrost is up to 5 m thick on the gentle north-facing slopebehind the landslide headwall, and up to 15 m thick on thesteeper north-facing slope adjacent to the landslide sourcezone (Fig. 1). The resistivity surveys also confirmed theabsence of permafrost in the landslide headwall, implyingthat rapid lateral thaw has occurred since the failure occurredin 2002.Landslide Failure MechanismUnfrozen glaciofluvial sand and gravel exposed in thelower unit of the landslide headwall are highly permeable andporous, with a capacity to store and transmit a large amountof groundwater. This is confirmed by the presence of springsflowing from the floor of the landslide source zone scar.Based on aerial photograph interpretation, these sedimentsare part of a larger geological unit consisting of beddedglaciofluvial deltaic materials that extend laterally furtherupslope and capture groundwater from the upper drainagebasin. As a result of this geomorphic configuration, the siteof the landslide source area represents the location of greatestgroundwater convergence within that basin. We hypothesize,therefore, that the base of the impermeable permafrostlayer within these sediments confined groundwater flowand allowed high pore pressures to accumulate below thepermafrost during the spring snowmelt period in June 2002.During the same time period, no abnormal precipitationwas recorded and no earthquakes occurred in the vicinity.We suggest that pore pressures increased until a thresholdwas reached that allowed rupture or “blow-off” (Cavers2003) of the unfrozen materials and detachment of the entirepermafrost layer above.We infer that subsequent rapid drainage of storedgroundwater supplied a large volume of water whichfacilitated the catastrophic debris torrent. Similar failuremechanisms are common in non-permafrost areas throughoutwestern Canada where groundwater pressures are insteadconfined by surficial material stratigraphy rather than byfrozen ground (Cavers 2003). Alternatively, the initialmovement could have partially blocked a small streamchannel to the northeast allowing water to accumulate in apond. However, we were unable to locate any evidence ofsuch ponding in the field.Local and regional climatic data show that at least fourdecades of climate warming have occurred in central Yukonsince 1930. Any corresponding permafrost warming and/or thinning may have weakened the permafrost layer andlowered the pore pressure threshold required to initiate afailure and cause detachment. Dataloggers were installedin 2005 to monitor the long-term air and ground surfacetemperatures above the landslide source zone.ImplicationsThe results of this study have important implicationsfor future development and land use planning in the area,as the geomorphic setting of the landslide source zone iswidespread throughout much of central Yukon. This casestudy illustrates that significant hazards are associated withthis type of landslide. It also highlights the need to performdetailed evaluations of basin characteristics, permafrostconditions, and surficial material stratigraphy severalkilometers upslope of any area targeted for human land use.In particular, stream crossing designs and development onfans in permafrost regions should carefully consider the riskof debris torrents triggered by groundwater blow-off failuresbeneath frozen ground.AcknowledgmentsFunding was provided by NSERC and the Universityof Ottawa to A. Lewkowicz; by the University of Oslo toB. Etzelmüller; and by the Yukon Geological Survey to P.Lipovsky and C. Huscroft.ReferencesCavers, D.S. 2003. Groundwater blow-off and piping debrisflow failures. Proceedings, 3 rd Canadian <strong>Conference</strong>on Geotechnique and Natural Hazards, Edmonton,Alberta, June 9–10, 2003: 151-158.Lipovsky, P. & Huscroft, C. 2007. A reconnaissanceinventory of permafrost-related landslides in the PellyRiver watershed, central Yukon. In: D.S. Emond, L.L.Lewis & L.H. Weston (eds.), Yukon Exploration andGeology 2006. Yukon Geological Survey, 181-195.Lipovsky, P.S., Coates, J., Lewkowicz, A.G. & Trochim,E. 2006. Active-layer detachments following thesummer 2004 forest fires near Dawson City, Yukon.In: D.S. Emond, G.D. Bradshaw, L.L. Lewis & L.H.Weston (eds.), Yukon Exploration and Geology 2005.Yukon Geological Survey, 175-194.Lyle, R.R. 2006. Landslide susceptibility mapping indiscontinuous permafrost, Little Salmon Lake, CentralYukon. M.Sc.E. Thesis. Queen’s University, 351 pp.190