GSA Annual Meeting in Seattle, Washington, USA - 2017

Paper No. 300-3
Presentation Time: 8:40 AM

FROM SINK TO SPRING; INTERPRETING LOCAL RECHARGE AND CONDUIT FLOW USING HIGH-RESOLUTION DATA LOGGERS AND A FINITE ELEMENT HEAT-AND-MASS TRANSPORT FRACTURE FLOW MODEL


BERGLUND, James L., Earth and Environmental Sciences, Temple University, 326 Beury Hall, 1901 N. 13th St., Philadelphia, PA 19122, TORAN, Laura, Earth and Environmental Science, Temple University, Philadelphia, PA 19122 and HERMAN, Ellen K., Department of Geology and Environmental Geosciences, Bucknell University, 1 Dent Drive, Lewisburg, PA 17837, tug13021@temple.edu

The local recharge and flow to a Central Pennsylvania karst spring are described using data loggers in conjunction with the Finite Element subsurface FLOW system (FEFLOWTM). Tippery Spring discharges 1-10 cubic feet per second (cfs) from a carbonate bluff. It has been traced to a perennial stream feeding a sinkhole 1500 meters up-gradient, called Tippery Sink. High-resolution (15-minute) data loggers provide information on sink and spring water temperature, spring pool water level, and precipitation. Multiple regression analysis shows the best single predictor for temperature response at Tippery Spring is the pre-storm temperature contrast between sink and spring, believed to be the result of piston flow pushing out older warm water. Despite this relationship, the initial temperature difference alone is not sufficient to completely predict the temperature response, which depends on other factors such as precipitation duration, timing, and time since last rainfall. FEFLOW models were produced to better understand the interplay of these factors using data from multiple storm events during 2016-2017. Conduit flow was modeled between the sinking stream and spring in FEFLOW as a discrete feature defined to obey one of several flow laws with unique conduit parameters (e.g. cross-sectional area, thermal conductivity of conduit wall). Precipitation and sinking stream temperature data were used as model input parameters. Temperature and water level values from the spring loggers provided simulation targets for the flow model. As the simulations are highly sensitive to particular conduit geometries, a model which adequately simulates the discharge and thermal responses at the spring under varying input conditions is expected to have accurately represented the nature of recharge from the sinking stream and conduit geometry feeding the spring. Models produced and calibrated this way may be useful for further describing and verifying other transient spring behaviors such as storm hysteresis.