Simulated flow of groundwater and brine from a flooded salt mine in Livingston County, New York, and effects of remedial pumping on an overlying aquifer

Abstract

Two ceiling collapses in the Retsof salt mine near Geneseo in upstate New York in spring 1994 resulted in the upward propagation of two columns of rubble through 600 feet of overlying shale and carbonate bedrock. This upward propagation formed a hydraulic connection between the lower confined aquifer (LCA) and the mine and allowed water from the aquifer and bedrock fracture zones that intersected the rubble columns to flow into the mine at a rate of 18,000 gallons per minute (gal/min) . All salt mining ceased in September 1995, and the mine was completely flooded by January 1996. The flow of water from the lower confined aquifer into the mine caused widespread drawdowns, and water levels in the aquifer declined by as much as 400 feet near the collapse area and by more than 50 feet at wells 7 miles to the north and south. Within 3 to 4 weeks of the collapses, water levels in about a dozen domestic and industrial wells had declined severely, and some wells went dry. Water levels in at least 58 wells in the lower and middle confined aquifers were affected by mine flooding. Groundwater in the upper unconfined aquifer and surface water in streams were unaffected by water-level drawdown, but channels of the Genesee River and Beards Creek were altered by land subsidence related to the mine collapse. Water levels recovered from 1996 through 2006, but the mine is now filled with about 15 billion gallons of saturated halite brine. The weight of the overlying rock and sediment is expected to cause the salt beds to deform and fill the mine cavity during the next several hundred years; this in turn could displace as much as 80 percent of the brine and cause it to move upward through the rubble chimneys, rendering the LCA unusable as a source of water supply. Saline water was detected in the LCA in 2002 but was found to be derived primarily from fractures in the limestone and shale units between the mine and the LCA, rather than from the mine. In September 2006, the mine company began a brine-mitigation project that entailed pumping five wells finished in limestone and shale units within the collapse areas to alter the flow gradient and thereby prevent further movement of brine and saline water into the LCA. The pumped brine was routed to an onsite desalination plant. At the same time, the U.S. Geological Survey (USGS) began a study in cooperation with the New York State Office of the Attorney General to construct numerical models to analyze the groundwater chemistry and delineate the directions of flow. Specific objectives of the study were to: * Assess the sources of salinity within the collapse area and identify the factors that control the movement and mixing of freshwater, saline waters from fracture zones, and brine; * Evaluate the likelihood that the pumping will induce anhydrite dissolution and lead to continued land subsidence; * Construct variable-density groundwater flow models to predict the effect of remedial pumping on salinity within the LCA; * Evaluate the effectiveness of remedial pumping in preventing the movement of saline water into the LCA; and * Predict the extent of brine migration 8 years after a hypothetical shutdown of all pumping in 2008. This report (1) summarizes the hydrogeologic setting and effects of mine flooding, (2) describes the geochemical and variable-density model simulations and their principal results, (3) discusses the implications of (a) continued pumping and desalination to protect the LCA and (b) a full shutdown of pumping after 2008, and (4) suggests further research that could lead to refinement of model predictions. Additional information may be found in Yager and others (2001 and 2009). These reports can be accessed at http://pubs.usgs.gov/pp/pp1611/ and http://pubs.usgs.gov/pp/pp1767/, respectively. A summary of simulation results can be accessed at http://ny.water.usgs.gov/projects/Coram/seawat/seawat.html.