Groundwater flux characterization using distributed temperature sensing: Separating advection from thermal conduction

Abstract:

Direct measurement of groundwater flux is difficult to obtain in the field so hydrogeologists often use easily-detectable environmental tracers, such as heat or chemicals, as an indirect way to characterize flux. Previously, we developed a groundwater flux characterization (GFC) probe by using distributed temperature sensing (DTS) to monitor the temperature responses to active heating in a well. The temperature responses were consistent with the hydraulic conductivity profiles determined at the same location, and provided high-resolution information (approx. 1.5 cm) about vertical variations in horizontal flux through the screen. One of the key assumptions in the previous GFC approach was that the vertical variations in the thermal conductivity of the aquifer materials near the well are negligible, so that the temperature differences with depth are primarily a result of groundwater flux instead of thermal conduction. Although this assumption is likely valid for wells constructed with an artificial filter pack, it might become questionable for wells with natural filter packs (such as the wells constructed by direct push where the sediments are allowed to directly collapse onto the well screen). In this work, we develop a new procedure for separating advection from thermal conduction during GFC measurement. In addition to the normal open-screen GFC profiling, an impermeable sleeve was used so that heating tests could be performed without advective flow entering the well. The heating tests under sleeved conditions were primarily controlled by the thermal conduction around the well, and therefore could be used to remove the impact of thermal conduction from the normal GFC results obtained under open-screen conditions. This new procedure was tested in a laboratory sandbox, where a series of open-screen and sleeved GFC tests were performed under different flow rates. Results indicated that for the tested range of rates (Darcy velocity 0 – 0.78 m/d), the relation between the temperature increase in open-screen GFC tests and flow rate appeared to be largely linear, while temperature responses in the sleeved tests did not change significantly with flow. Numerical simulations of these experiments have further confirmed the validity of the new procedure for removing the conduction impact from flux measurement.