In calculations of building heat loss via the ground, the coupling with soil moisture transfer is generally ignored, an important hypothesis which will be falsified in this paper. Results from coupled simulations – coupled soil heat and moisture transfer equations and complete surface heat and moisture balances – are compared to results from linear simulations. It is shown that the coupled calculations give notably higher heat losses. Surface temperature, the driving force for heat loss via the ground, is identified as a first important source for these deviations: it is shown that while the averages of the surface temperature are almost equal in coupled and linear calculations, the amplitude in the coupled simulation is considerably higher. A further study reveals that variation of the thermal properties with moisture content also contributes partially to the observed differences. Transfer and storage of sensible heat linked to moisture in the liquid phase was shown to be the last important influencing factor.