Evaluation of TDR for Quantifying Heat-Pulse-Method-Induced Ice Melting in Frozen Soils

Abstract

The heat pulse (HP) method is the standard for measuring thermal properties of unfrozen soils, but its application in partially frozen soils is confounded by melting and refreezing of ice that changes the properties being measured, and the most widely used data analysis models do not account for these latent heat fluxes. This study combined time domain reflectometry (TDR) and HP methods to quantify HP-induced ice melting through: (i) TDR-estimated liquid water content (θl) before and after the HP; and (ii) heat capacity (Cv) calculated from TDR-estimated θl and ice content (θi) and HP-estimated apparent heat capacity (Ca). Different heat pulse durations (8–60 s) and strengths (280–2100 J m−1) during soil thawing and freezing (−35 to 25 to −35°C) were investigated on loamy sand soil samples with a water content of 0.25 m3 m−3. The results showed that the first approach was restricted by the TDR measurement uncertainty and the difficulty in synchronizing TDR measurements with the peak ice melting. Comparison of Cv and Ca showed: (i) ice melting is limited below −5°C and the HP-estimated Ca may approach the real values; (ii) application of the HP method between −5 and 0°C is largely affected by ice melting, the amount of melting ice is largest between −1.5 and −0.5°C during soil thawing but small during freezing; (iii) the maximum estimated ice melting was 0.005 to 0.02 m3 m−3 but caused significant problems in estimating heat capacity; (iv) Cv and Ca are hysteretic because of different θi at the same temperature during freezing and thawing; and (v) a good relationship between Cv and Ca and energy input was found, which may guide the optimized heat application for frozen soil and extend the possibilities for estimating frozen soil heat capacity.