TEMPERATURE: A USEFUL TRACER FOR SEPARATING CONDUIT AND DIFFUSE FLOW IN THE KARSTIC MADISON AQUIFER

Continuous water temperature data can be recorded at high precision and low cost for wells, springs, and streams. Groundwater temperature is affected by heat exchange with aquifer material, which is affected by the geothermal gradient, and thus temperature is a tracer that may require numerical heat transport-modeling in groundwater. This type of model was applied to the carbonate karst Madison aquifer in the Black Hills of South Dakota, which contains pipe-like openings called conduits into which water from sinking streams enter to recharge the aquifer. The model accounted for several variables including the heat exchange between aquifer rocks and groundwater, variable groundwater velocity in the conduit, conduit roughness and turbulence, the geothermal gradient, and a proportionality constant weighted by the rate of sinking streamflow to quantify the relative fractions of conduit and diffuse flow. The model was applied to 7 months of continuous temperature measurements for the well, which withdraws fast-moving groundwater from a conduit as well as slow-moving diffuse flow from smaller pore spaces. Temperature fluctuations of the sinking stream water are heavily damped by contact with conduit walls at nearly constant temperature, which results in near equilibrium between rock and water temperatures inside the conduit. Similarly, the model was insensitive to temperature changes in the sinking stream. Results of the simulation indicated that within the well’s capture zone, groundwater in the conduit generally is about 0.5°C cooler than diffuse flow as a result of prior contact with conduit walls upgradient from the well and closer to the land surface than small pore spaces within the well’s capture zone. Thus it was possible to estimate from model calibration that the conduit-flow fraction to the well was in the range of 2-31% of total flow depending on the recharge rate, and that simulated conduit velocity was in the range of 44-353 m/d, which is consistent with velocities determined from previous dye tracing results. Temperature fluctuations for the well with periods ranging from 0.5 to 1.5 days could not be simulated but probably resulted from rapid changes in the mixture of conduit and diffuse flow, which may have resulted from factors such as pumping of proximal wells or earth tides.