Validation of the Basal Temperature of Snow (BTS) method to map permafrost in complex mountainous terrain, Ruby Range, Yukon Territory and Haines Summit, British Columbia

Abstract

This study is the second attempt to use the Basal Temperature of Snow (BTS) method to map permafrost in mountainous regions of northwestern Canada. It differs from the first study which took place in Wolf Creek in terms of (1) the methodology used to evaluate BTS, (2) the strategy used to avoid spatial autocorrelation in residuals, and (3) the climatic regions investigated. Two study areas, part of the Ruby Range (61° 12' N, 138° 19' W) and Haines Summit (59° 37' N, 136° 27' W) were selected for BTS sampling based on differing climatic conditions and previous knowledge of permafrost elevations from active rock glaciers. A total of 30 BTS measurements were made in the Ruby Range in the winter of 2006 and a total of 77 BTS values were obtained in the Haines Summit area during 2005 and 2006. From these results, modeled BTS surfaces were created using elevation and potential incoming solar radiation as independent variables in a multiple linear regression. At Haines Summit, potential incoming solar radiation was not significant in the model and thus was dropped. The surface of modeled BTS was then combined with a physical validation of permafrost presence completed during the late-summer of 2005 in a logistic regression. The modeled results produced permafrost probability maps for both study areas. Based on modeled results, permafrost underlies an estimated 282 km2 or 66% of the Ruby Range study area and 23--236 km 2 or 43--44% of the Haines Summit study area. An attempt was made to use the linear model derived in the Ruby Range at Haines Summit in order to examine the possibility of expanding predictions into new areas. Although the results produced similar total amounts of permafrost, the spatial distribution differed: permafrost probabilities were reduced at high elevations while lower elevation sites exhibited increased probabilities. The results of the model transfer illustrate the importance of the pit data in determining the total amount of permafrost, while knowledge of BTS ranges contributes to the spatial distribution of permafrost. With further study it is likely that generic models can be derived for areas of similar climate.