A radiative-conductive-convective approach to calculate thaw season ground surface temperatures for modelling frost table dynamics

The frost table depth is a critical state variable for hydrological modelling in cold regions as frozen ground controls runoff generation, subsurface water storage and the permafrost regime. Calculation of the frost table depth is typically performed using a modified version of the Stefan equation, which is driven with the ground surface temperature. Ground surface temperatures have usually been estimated as linear functions of air temperature, referred to as ‘n-factors’ in permafrost studies. However, these linear functions perform poorly early in the thaw season and vary widely with slope, aspect and vegetation cover, requiring site-specific calibration. In order to improve estimation of the ground surface temperature and avoid site-specific calibration, an empirical radiative–conductive–convective (RCC) approach is proposed that uses air temperature, net radiation and antecedent frost table position as driving variables. The RCC algorithm was developed from forested and open sites on the eastern slope of the Coastal Mountains in southern Yukon, Canada, and tested at a high-altitude site in the Canadian Rockies, and a peatland in the southern Northwest Territories. The RCC approach performed well in a variety of land types without any local calibration and particularly improved estimation of ground temperature compared with linear functions during the first month of the thaw season, with mean absolute errors