STREAMBED TEMPERATURES ARE TRANSIENT. HOW RELIABLE ARE VERTICAL FLUXES ESTIMATED FROM METHODS THAT ASSUME THAT THEY ARE CONSTANT?

Estimating groundwater discharge to surface water bodies is critical for ecological and hydrogeological characterizations, as discharge impacts surface water quality and temperature. The use of ambient heat as a tracer to quantify these fluxes is attractive because temperature data are relatively simple and cost efficient to collect. Additionally, there are several analytical solutions to quantify vertical fluid fluxes from subsurface temperatures. In particular, steady state analytical solutions based on temperature-depth (T-z) profiles are appealing, as they can be applied using point-in-time data, allowing the quantification of groundwater discharge at many points within a short period of time. While the requirement of a temperature gradient to apply T-z profile methods is well understood, the influence of diurnal temperature variations on gradient-based models has not been rigorously assessed.

For this study, numerically generated synthetic datasets were used to investigate the influence of transient stream temperatures on the use of T-z methods. The use of synthetic data allows the influence of various model assumptions (e.g. transient stream temperatures) to be tested where the true vertical fluxes are known. Fluxes were quantified using point-in-time and daily averaged T-z profiles, as well as with diurnal signal methods using VFLUX 2.

Results highlight the importance of the upper boundary condition for T-z methods, with estimated fluxes generally oscillating around the known flux when T-z profiles were taken at various times throughout the day. However, when time series data are available, the use of daily averaged T-z profiles provided more accurate flux estimates than point-in-time data. However, point-in-time data can be used to produce reasonable flux estimates when the upper boundary condition at the water-bed interface is close to its daily mean temperature. At stronger discharge rates (upwards flows ≥1 m d^-1), the diurnal signal methods tended to be more accurate than T-z approaches as the temperature gradients become strongly concave in shape.

These findings highlight the approaches to take to obtain the most reliable estimates of groundwater discharge to streams using temperature data, highlighting where point-in-time T-z can best be used to produce reliable vertical flux estimates.