Ninth International Conference on Permafrost ... - IARC Research

Modeling Potential Climatic Change Impacts on Mountain Permafrost Distribution,Wolf Creek, Yukon, CanadaPhilip P. Bonnaventure, Antoni G. LewkowiczDepartment of Geography, University of Ottawa, Ottawa, CanadaIntroductionDifferences in air temperature, precipitation, andvegetation cover that develop across hundreds of kilometresin lowland areas can be generated by a few hundred metresof elevation change in mountainous regions. Consequently, amountain basin located in the discontinuous permafrost zonemay span the entire range of permafrost conditions, fromisolated patches at low elevations on north-facing slopes tocontinuous permafrost on summits (Lewkowicz & Ednie2004). It is reasonable to infer that within such a basin therewill be terrain present at temperatures close to 0°C that canbe affected by changes in climate. If unfrozen, this terrainmay become permafrost during sustained climate cooling,or if permafrost, it may thaw during sustained warming orlong-term increases in snow depths. Consequently, mountainbasins with permafrost may be particularly sensitive toclimate change. However, the complexity of permafrostdistribution within them means that there have been relativelyfew attempts to model climate change impacts (e.g., Janke2005). Our goal here is to present a method that can be usedto explore the potential effects of past and future climatechange on mountain basins with permafrost.Study AreaWolf Creek Basin (60°30′N, 135°10′W) is a mountainouswatershed of approximately 190 km 2 , with elevations rangingfrom 700–2080 m a.s.l. and located 20–30 km south ofWhitehorse in the Yukon Territory. The climate is continentalwith dry, cold conditions (Wahl et al. 1987), and the basin fallswithin the zone of sporadic, discontinuous permafrost zoneaccording to the Permafrost Map of Canada (Heginbottomet al. 1995). Basin vegetation comprises boreal forest, withsub-alpine forest, a shrub zone, and an alpine tundra zoneat progressively higher elevations. Under current conditions,permafrost probability models in Wolf Creek indicate that 38to 43% of the area is underlain by permafrost (Lewkowicz &Ednie 2004, Lewkowicz & Bonnaventure 2008).MethodologyBTS measurements were collected in Wolf Creek duringthe winters of 2001 and 2002 (Lewkowicz & Ednie 2004).The spatial field of BTS was modeled in a GIS using elevationand Potential Incoming Solar Radiation (PISR) as independentvariables. Logistic regression was used to relatethe modeled BTS temperatures to the presence or absence ofpermafrost at numerous sites within the basin in late-summer.The end result of this procedure is a map of permafrostprobability at a grid cell resolution of 30 x 30 m.The effects on permafrost distribution of climate coolingor warming scenarios can be simulated by respectivelyincreasing or decreasing the values of the elevationthroughout the study area (Janke 2005). This alters themodeled BTS field which in turn affects the predictedpermafrost probabilities. We used a standard environmentallapse rate of 6.5°C/1000 m to calculate the necessary changeof elevations, a value which is less than the BTS lapserate (8.2°C/1000 m) (Lewkowicz & Bonnaventure 2008).Temperature changes of -2 to +5°C were used in order toexamine how permafrost distributions might have appearedunder equilibrium conditions similar to those of the LittleIce Age, when temperatures in the basin were lower (e.g.,Farnell et al. 2004), and for future changes through to themost aggressive temperature warming scenarios proposedby the Intergovernmental Panel on Climate Change (IPCC2007). As in previous work, we assume that permafrostprobability can be equated over many grid cells to permafrostextent.Results and DiscussionIt should be emphasized that model predictions are forequilibrium states; the model does not account for lag timesassociated with permafrost formation and degradation.Given these lag effects, the model outputs are best thoughtof as referring to the upper few metres of permafrost only.Under cooler-than-present conditions, permafrost areawithin the basin expands, doubling to about 75% for atemperature reduction of 2°C. The form of the change isapproximately linear within this range (Fig. 1). Spatially,permafrost zones become more extensive at intermediateelevations and the boundary between continuous andextensive discontinuous permafrost (90% probability)descends from about 1600 m to 1250 m, while that betweenextensive and sporadic permafrost (50% probability) changesfrom 1400 to 1100 m.Under warming conditions, such as those expectedunder IPCC projections, permafrost extent is substantiallyreduced in a nonlinear fashion. An increase of only 1°Chalves the permafrost extent in the basin, and a further1°C change reduces it to less than 10% of the basin area(Fig. 1). Boundaries of permafrost zones within the basinmove upwards so that a 1°C change causes the continuouspermafrost boundary to rise to 1700 m. At an increase of2°C, permafrost is present only on high elevation mountaintops and upper elevation north-facing slopes. A 4°C increasereduces permafrost extent to less than 1% of the basin area,and probabilities exceed 10% only on the highest mountainpeaks above 1850 m.31