EFFECTS OF VERTICAL HEAD GRADIENTS ON THE INTERPRETATION OF CHEMICAL AND WATER LEVEL MEASUREMENTS IN OBSERVATION WELLS: IGNORED PHENOMENA OF CONSIDERABLE IMPORTANCE

Much of traditional ground water hydrology is based on horizontal flow concepts. However, it is well known that vertical gradients exist in natural systems, and that such gradients set up spontaneous vertical flows in observation wells, often called ambient flows. [Elci et al., Ground Water, 39, in press] Such gradients may lead also to a water level within a well that does not correspond to local hydraulic head averaged over the screen. The objective of this communication is to present information showing that vertical hydraulic gradients and the resulting ambient flows may make chemical data from long-screened observation wells highly misleading, and may also provide misleading hydraulic head data. Calculations are based on head, flow-meter, and hydraulic conductivity data obtained at the Savannah River Site (SRS). Ambient flow and transport simulations for well P26-M1 in the confined Gordon aquifer were performed using the GMS package. Simulated flows agreed well with measurements. Natural flow was upward, so water entered the well through high K layers in the lower portion of the aquifer and exited through similar layers in the upper portion. The maximum upward discharge was about 0.4 m³/day, implying an induced exchange of 12m³/month from the bottom half of the aquifer to the upper half. Transport simulations illustrate how a contaminant located initially in a lower portion of the aquifer moves continuously into the upper portion and is diluted in the well by in-flowing water. Tracers released in the upper portions of the aquifer may by-pass the observation well entirely, and ambient flow magnitude is sensitive to screen length. Additional studies at the SRS R-Reactor Seepage Basins indicate that in the presence of the steep vertical hydraulic gradients that exist there, conventional interpretation of well water levels may be grossly inappropriate. More specifically, a well may be screened below the formation water table, and yet exhibit a water level that is below the top of the screen. This phenomenon was inferred at the seepage basins using cone penetration testing with soil moisture and resistivity sensors, and concurrent water level data from wells. Subsequent analytical and numerical analyses demonstrate that the phenomenon is physically possible, and define the necessary hydrogeologic conditions.