DELINEATING KARST CONDUIT LOCATION USING GEOPHYSICAL SIGNALS GENERATED DURING RECHARGE EVENTS

Flow and transport in karst aquifers primarily occur through conduits, but locations of preferential flow paths are largely unknown. To identify if geophysical signatures generated from the conduits themselves during recharge events can be detected at the surface to facilitate mapping of subsurface flow paths, we monitored surface deformation during two controlled recharge experiments and a natural recharge event near Bear Spring, west of Eyota, MN. In water-filled conduits, recharge pressure waves travel on the order of the speed of sound in open water (~1500 m/s). This phenomenon is often observed with recharge events in karst aquifers, where spring discharge increases shortly after a recharge event commences and long before changes in chemical or thermal parameters that generally indicate event water. Poroelasticity theory suggests that pressure waves in flow paths with full pipe flow may cause deformation at the surface. The controlled recharge experiments involved injecting a pool full of water (~13,000 L) into a dry overflow spring, which then flowed underground until it was discharged at Bear Spring. To monitor surface signals, we deployed twelve seismometers (11 short period and 1 broadband) and 1 GPS instrument between the dry overflow spring and Bear Spring. In addition, we added 68.82 kg of salt to the first pool to track the flow of the pool water in the subsurface with a resistivity survey, where electrodes were installed to facilitate monitoring along three transects perpendicular to flow. After the recharge experiments, rain fell the next morning (2.1 inches at the nearby Chester Woods Park), which caused the overflow spring to start flowing and total discharge (overflow spring and Bear Spring) to increase from a background of ~100 L/s to ~300 L/s. Preliminary results suggest all three recharge events produced large amplitude seismic signals on all of the seismic stations, with the rain-induced flow producing longer duration signals. Initial analysis of the 20Hz GPS data shows no signal due to the recharge experiments, but some subsidence in response to the rain event may have been recorded. No signal was observed with the resistivity survey, likely due to fences and buildings that limited electrode spacing and location. Ongoing research is determining information that may be derived from these signals.