CONTROLS OF SULFATE VARIABILITY IN GROUNDWATER AT A LIGNITE MINE, NORTHEAST TEXAS

Chemical concentrations in the groundwater at the TXU Monticello lignite mine in northeastern Texas vary spatially and temporally. In particular, sulfate (SO42-) is monitored closely because it is a product of chemical reactions that can lead to acid mine drainage. Although this has not been a problem at Monticello, SO42-concentrations in some areas are high and correspond to high total dissolved solids (TDS) and low pH. Chloride, total dissolved iron, and total dissolved manganese concentrations are also variable. To examine this variability, 46 long-term monitoring wells are divided into those screened in: 1) the underburden (below the unmined lignite seams); 2) the unmined overburden; and 3) the reclaimed sections. There is a wide range of SO42-concentrations across the mine with statistically significant differences between the SO42-distributions for each well category. One hypothesis explaining spatial SO42- variability is the heterogeneity of the mined material. Two other hypotheses may explain the high SO42- in the unmined overburden: 1) flushing of SO42- from the reclaimed section into the overburden and 2) exposing iron sulfides to oxidizing conditions when the water table is dropped by mine dewatering operations. Temporal SO42- variability is evident when examining time series for each well. In the underburden, SO42- seems to be approaching a steady state; however, some wells in the unmined overburden and reclaimed area have SO42- trends that are increasing or are highly variable with time. Water table fluctuations may control these trends. Correlation analyses show that for many wells, SO42- does not correlate with pH, which indicates either that the source of SO42- may be gypsum dissolution, rather than pyrite oxidation, or that the SO42- may be advected from nearby reclaimed areas. Mineralogical XRD analyses of surface soil samples and filtered suspended sediment from surface water and groundwater samples also tested these hypotheses. Time series analyses show that most SO42- trends lack a seasonal component; comparing water level and SO42- time series show that advection from reclaimed areas cannot account for most high or increasing SO42- concentration in the overburden. The second hypothesis, exposure to oxidizing conditions, is generally suspected as the reason for most SO42- variability.