Coupled thermo-geophysical inversion for permafrost monitoring

This study explores an alternative way of deriving soil thermal properties from surface geophysical measurements. We combined ground surface temperature time series with time lapse geoelectrical acquisitions measured from the ground surface in a fully coupled inversion scheme to calibrate a heat conduction model. The quantitative link between the thermal and geoelectrical parts of the modeling framework is the temperature-dependent unfrozen water content, which is also the main factor influencing electrical response of the ground. The apparent resistivity data were incorporated into the coupled framework without being inverted separately, thus reducing the uncertainty inevitably associated with inverted resistivity models. We show that geoelectrical time lapse data are useful as alternative calibration data and can provide as good results as borehole temperature measurements. The fully coupled modeling framework using field data achieved performance comparable to calibration on borehole temperature records in terms of model fit within 0.6 °C, inversion convergence metrics, as well as the predictive performance of the calibrated model.