Effects of precipitation events on colloids in a karst aquifer

Abstract

The effects of precipitation events on colloid mobilization were evaluated during several storms from six wells in a karstic aquifer at the Oak Ridge Y-12 Plant in eastern Tennessee (USA). Turbidity increases and rapidly recedes following rain events. Although the magnitude of the turbidity increases are relatively small (≤4.78 NTU), the increased turbidity suggests transient increases in colloid abundance during storm versus non-storm periods. During the larger storms (>19 mm), the increased turbidity is associated with increases in pH, total organic carbon (TOC) and temperature, and with decreases in dissolved oxygen (DO). These larger storms result in flushing of a greater proportion of higher pH, TOC (and lower DO) soil or matrix waters into the fractures and conduits than occurs during smaller storms. Smaller storms also result in increases in turbidity, but show increases in DO and decreases in pH reflecting less influence on the water chemistry from the longer residence time epikarst or and matrix waters, and greater impact from the more dilute, newly recharged waters. Due to the complexity of karst flow and temporal variations in flow and chemistry, controls on turbidity are not consistent through time and space at the wells. During smaller storms, recharge by lower ionic strength waters may promote colloid release and thus contribute to observed increases in turbidity. During larger storms, elevated turbidity may be more related to pH increases resulting from greater influx of matrix and soil waters into fractures and conduits. Chemical factors alone cannot account for the changes in turbidity observed during the various storms. Because of the complicated nature of flow and particle transport in karst aquifers, the presence of colloids during precipitation events is dictated by a complex interplay of chemical reactions and the effects of physical perturbations due to increased flow through the conduits and fractures. Simple trends in water quality parameters could not be identified, and broad generalizations cannot easily be made in karst settings, and some of the expected correlations between chemical parameters during the storms were not observed in this work.