On the use of surface and ground temperature data to recover soil water content information

Ground temperature measurements are largely used in agronomy, hydrology, or climatology. As temperature variations in the subsurface depends on its thermal diffusivity which itself is a function of water content, temperature measurements are also relevant in hydrogeophysics. In this paper, we model surface and underground (60 cm depth) temperature data to analyze the time variations of soil water content over one year. We have considered two modelling approaches. We first invert the data using a homogeneous half-space model to recover the apparent effective soil water saturation. For that, we examine several thermal diffusivity-water content relations and conclude that among them only Johansen's relation is physically acceptable. As expected, the apparent effective saturation is maximal at the beginning of the spring, and then it decreases until the end of the summer. In the second approach, we use a two-layer model with fixed thermal parameters in both layers (top: unsaturated zone; bottom: saturated zone), and invert the data to monitor the time variations of the interface depth. The obtained depths are coherent with the results of the first approach and show that the soil is unsaturated over more than ten centimetres by the end of the summer. Even though more complex multi-layer models including capillary effects would probably be constrained by using temperature data measured at more than two depths, our results on this data set show the usefulness of introducing apparent parameters that provide different simple means to characterize the complex heterogeneous subsurface.