A statistical model for presence of late-season frozen ground in discontinuous permafrost at Dublin Gulch, Yukon

作者：Quinn P E^*

来源：Canadian Geotechnical Journal, 2013, 50(8): 889-898.

DOI：10.1139/cgj-2012-0318

• permafrost
• weights of evidence
• susceptibility modelling
• geographical information system (GIS)
• Yukon

This paper describes geostatistical analyses completed at a discontinuous permafrost site in central Yukon to develop a predictive model for the presence of late-season frozen ground in support of planning and design for potential site development. The most important factors in the bivariate statistical model were soil type, as determined through terrain analysis, and slope aspect, as inferred from available topographic data. The other three factors included in the final model were profile curvature, slope angle, and ground elevation, each interpreted from available topographic data. The resulting model subdivides the site into three broad classes of frozen ground likelihood: low, where frozen ground can be expected to be encountered in late summer at 15% of observation locations; medium, where 50% of the ground is expected to remain frozen; and high, where 85% of the ground is expected to remain frozen. New test pit and borehole data from the summer of 2012 were used to verify model performance. The inferred correlations between frozen ground and soil type, aspect, curvature, slope, and elevation obtained in this case study may provide useful information relative to expected permafrost occurrence at sites in central Yukon with similar geology and physiography.

• 出版日期2013-8