The volume of aquifer system that can be characterized with aquifer tests is controlled by the distance at which drawdown can be detected. Drawdown detection typically is limited to distances of less than 1 km, because environmental water-level fluctuations frequently exceed the maximum displacement from pumping (Risser and Bird 2003).
Environmental fluctuations in measured water levels are described here as nonpumping (natural or anthropogenic) stresses on the aquifer system. These fluctuations include short-term, seasonal, and long-term stresses such as barometric pressure, tidal signals, natural and artificial recharge, and surface-water stage changes and diversions. Barometric pressure and tidal signals acting on the aquifer system can induce water-level changes of more than 0.3 m during periods of less than a few days (Fenelon 2000). Individual recharge events also can cause episodic water-level rises that exceed 0.5 m over a few days (O'Reilly 1998). Recharge likewise can induce long-term rising trends of more than 1 m/year that affect detection of small pumping signals (Fenelon 2000; Elliott and Fenelon 2010). Stage changes of a fully penetrating river can cause daily and event-based fluctuations in local groundwater levels (Criss and Criss 2011).
Environmental fluctuations from recharge responses, surface water stage changes, or any other external stress can be modeled explicitly by using water levels from background wells that are affected by these environmental stresses (Halford 2006; Criss and Criss 2011). A useful background well is one in which water levels are affected by tidal potential, imperfect barometric coupling between the atmosphere and water table, and all other stresses that affect water levels in observation wells excluding pumping. Although the need for background water levels to characterize environmental fluctuations has long been recognized (Stallman 1971), trends and corrections characterizing environmental fluctuations typically have been estimated qualitatively.
One method to reduce drawdown obscurity by environmental fluctuations is to interpret water-level changes from long-term (i.e., years) pumping of water-supply wells (Gonthier 2011; Harp and Vesselinov 2011). Long-term pumping can generate substantial drawdown over time, which will exceed environmental fluctuations during the period of analysis. Despite the utility of this method, using years of water-level changes from pumping of supply wells is an opportunistic approach that generally cannot be applied to aquifer tests.
Alternatively, environmental fluctuations have been distinguished from drawdown by modeling and directly removing barometric and tidal effects from measured water levels (Erskine 1991; Rasmussen and Crawford 1997; Toll and Rasmussen 2007). However, this approach does not remove environmental fluctuations caused by regional trends such as long-term recharge and is difficult to automate because all significant stresses that affect water levels other than pumping are not simulated explicitly.
The above approach was expanded upon by modeling and removing all nonpumping stresses from the water-level record (Halford 2006). Environmental fluctuations caused by barometric pressure and tidal signals, and regional trends captured in background water levels are modeled prior to pumping, projected forward during pumping and recovery periods, and removed from measured water levels. This approach requires antecedent monitoring periods that are more than three times longer than the pumping and recovery period in order to capture short and long-term environmental trends (Halford 2006). Considering recovery periods are often six times longer than pumping periods (Neville and van der Kamp 2012), this approach rapidly becomes unreliable where pumping periods exceed a week.
This paper describes and tests a method presented by Halford et al. (2012) for estimating observation-well drawdown response to aquifer-test stress where (1) environmental fluctuations mask the pumping signal, and (2) the period of record is limited. Drawdown is distinguished from environmental fluctuations by analytically simulating all water-level stresses (nonpumping and pumping) simultaneously during the period of record. Simultaneous modeling of environmental fluctuations and pumping signals overcomes the limitations of long-term interpretation and antecedent water-level monitoring. This approach draws on previous approaches where all environmental fluctuations are simulated and the pumping signal is modeled analytically using Theis (1935). Changes in pumping rates, multiple pumping wells, and lithologic variability are simulated by transforming multiple pumping schedules into water-level changes with Theis (1935). Transmissivity and storage coefficient are curve fitting parameters and are not interpreted as aquifer properties because the underlying assumptions of the Theis solution frequently are not met.
The described method was tested using “known” drawdowns generated from a hypothetical aquifer test that was simulated with a three-dimensional MODFLOW model. The utility of the approach also was demonstrated using a real aquifer test. The hypothetical model was conceptually based on the complex hydrogeology at the Nevada National Security Site (NNSS) at Pahute Mesa whereas the real aquifer test was conducted in a deep fractured volcanic rock aquifer at Pahute Mesa. The hypothetical model demonstrated that the drawdown estimation approach described here can closely match known drawdowns.
Pahute Mesa Study Area
Pahute Mesa is located in southern Nevada, within the NNSS (Figure 1). The aquifer system beneath Pahute Mesa comprises layered sequences of volcanic rocks that have been faulted into distinct structural blocks (Warren et al. 2000). Rhyolitic lavas or welded ash-flow tuffs such as in the Benham and Topopah Springs Aquifers, respectively, comprise aquifers. Bedded and nonwelded, zeolitized tuffs typically comprise confining units (Blankennagel and Weir 1973; Prothro and Drellack 1997; Bechtel Nevada 2002). More than a half dozen faults with offsets in excess of 200 m have been mapped previously in Pahute Mesa (McKee et al. 2001) and additional faults are mapped as well drilling continues (e.g., National Security Technologies, LLC 2010).
Distant drawdown detection from multiwell aquifer tests is necessary to quantify the hydraulic properties of the stratigraphic and structural features on Pahute Mesa so that radionuclide migration can be evaluated (Laczniak et al. 1996). The depth to water exceeds 600 m and sparsely distributed wells (generally more than 1 km apart) penetrate more than 1 km of a complex, volcanic rock-dominated, hydrogeologic system (Fenelon et al. 2010). Environmental water-level fluctuations are substantial beneath Pahute Mesa because of the thick unsaturated zone and high pneumatic and hydraulic diffusivity of the volcanic rocks.