Using Streambed Temperature Time Series to Calculate Natural Changes in Streambed Elevation: the Influence of Rain Events

Abstract:

The phase shift and amplitude ratio of the diurnal temperature signal of paired streambed temperature sensors have already been used successfully to calculate vertical groundwater fluxes assuming a constant vertical distance between the sensors. Sedimentation and scouring however, can change the position of temperature sensors relative to the streambed influencing calculated groundwater fluxes. Here we report on a field investigation, where streambed sediment temperature time series were used to; (i) determine the range of natural sedimentation and scour over a two month period, (ii) study the influence of rain events on the calculated streambed elevation.

The study was carried out at a field site located along Holtum stream, in Western Denmark. The 3 m wide stream has a sandy streambed, a mean discharge of 200 l/s and a mean depth of 0.3 m. Streambed temperatures were measured at two locations in a straight stream section with a vertical high-resolution Distributed Temperature Sensing (DTS) system installed 0.7 m deep in the streambed. Temperature measurements were taken with 30 min intervals between 16 April and 25 June 2013 and streambed elevation was recorded on 24 May and 17 June 2013. The effective thermal diffusivity (κ_e ) of streambed sediments was measured on 17 June 2013 by a KD2 Pro SH-1 sensor (Decagon Devices).

The phase and amplitude of the diurnal temperature signal of paired sensors were extracted by VFLUX. The κ_e of streambed sediments was calibrated based on measured streambed elevations and validated by in-situ measurements. The range of potential streambed elevation over the two month period was calculated by assuming; (i) sedimentation, using the calibrated κ_e for saturated streambed sediments, (ii) scouring, using κ_e for stream water between the temperature sensors.

Under natural conditions, streambed elevation changes by 5-12 cm during the two month period and by 1-3 cm even during base flow conditions, emphasizing the importance of sensor location relative to the streambed. During rain events however, streambed elevation was overestimated by up to 8 cm due to the rain-induced disturbance in the diurnal streambed temperature signal which could potentially also lead to uncertainties in calculated groundwater fluxes.