Journal of Cold Regions Engineering

The warming weakening of permafrost strength as a result of different scenarios of climatic warming ranging from 0 to $5°\text{C}$ over the next century has been predicted using a one-dimensional geothermal model. These predictions are based on the results of an intensive geotechnical program carried out in a warm ice-rich silty permafrost in Northern Quebec. The dependency of permafrost strength on temperature was assessed from cone penetration tests performed at regular intervals over a 2-month period during the seasonal warming of the upper permafrost layer in spring 2000. A thermomechanical subroutine taking into account this dependency has been then added to the one-dimensional geothermal model for the simulation of the weakening of permafrost strength as it warms. A warming rate of $0.02°\text{C}/\text{year}$ over the next century leads on a slow decrease in permafrost strength corresponding to a relative loss of strength of about 50%. For a warming rate of $0.05°\text{C}/\text{year}$ , the strength weakening is much more pronounced and almost reaches the unfrozen state at the end of the warming period corresponding to a relative loss of 98%.

Air and surface freezing and thawing indexes and prediction of ground temperatures from air temperatures were investigated. A new method for applying $n$ factors on a daily basis to capture localized temperature extremes is presented and compared to conventional seasonal $n$ -factor analysis. Measured air temperatures from five locations and measured air and surface temperatures from one location were used. Freezing and thawing indexes were determined using daily and monthly average temperatures, different time frames, and seasonal and daily applications of $n$ factors. Air and surface freezing indexes ( $I_{\text{af}}$ and $I_{\text{sf}}$ ) varied more than air and surface thawing indexes ( $I_{\text{at}}$ and $I_{\text{st}}$ ). Significant variations were observed in air and surface indexes due to the length of the time period used (1, 10, and 30 years) and frequency of temperature data used (daily and monthly). The surface indexes from seasonal $n$ factors (using daily average temperatures) and daily $n$ factors were similar (within 4%) and higher than the indexes from seasonal $n$ factors (using monthly average temperatures). The average surface temperatures were within $2°\text{C}$ , whereas the maximum and minimum surface temperatures were significantly different (up to $26°\text{C}$ ) between the seasonal and daily $n$ factors. Maximum variations between consecutive daily maximum and minimum temperatures were significantly higher using daily $n$ factors (up to $52°\text{C}$ ) than seasonal $n$ factors (less than $1°\text{C}$ ). Surface indexes from seasonal (using daily average temperatures) and daily $n$ factors can be used interchangeably. Daily $n$ factors are recommended to obtain representative surface and near-surface temperature variations, diurnal extremes, representative timing for temperature change events, and localized freezing or thawing indexes during change-over months.