A rainfall loading response recorded at 300 meters depth: Implications for geological weighing lysimeters

Abstract

Static pore water pressures in confined aquifers vary in response to ground surface loading changes, including precipitation and evaporation. Under certain hydrogeological conditions such aquifers can function as giant natural weighing lysimeters, referenced here as ‘geological weighing lysimeters’. The extent of the land area ‘weighed’ increases with aquifer depth and it is of interest to establish at what depth it is still possible to monitor surface water budgets. An 86 mm rainfall event produced a clear loading signal in a well in western Kansas at 300 m depth. The loading effect is quantitatively consistent with elastic deformation induced by the rainfall mass and suggests that geological weighing lysimeters could operate at considerably greater depths, thereby monitoring water budgets over a significant land area.

Introduction

Static pore water pressures in elastic confined aquifers may change measurably in response to ground surface loading. The change in pore water pressure is determined by the loading coefficient c=1−b, where b is the static confined barometric efficiency (Jacob, 1940). The loading efficiency can be regarded as a time-invariant constant characterizing aquifer response to ground surface loading, provided the confining aquitard is thick and has low diffusivity so that there is minimal leakage of water pressure changes (Rojstaczer, 1988, Roeloffs, 1996). Confined aquifers in this situation can serve as giant natural weighing lysimeters for in situ monitoring of surface water budgets, subject to the additional condition that there are negligible dynamic effects influencing pore water pressures (van der Kamp and Maathuis, 1991, Bardsley and Campbell, 1994, Bardsley and Campbell, 2000). For convenience, such systems are referred to here as ‘geological weighing lysimeters’. The term is used in this paper in the context of confined aquifer measurements only. However, the geological weighing lysimeter concept has also been applied to aquitard measurements as well (van der Kamp and Schmidt, 1997).

The focus of geological weighing lysimeters is toward the surface water budget rather than as a means of deducing geological formation elastic properties and deformation modes. In essence, the approach is pragmatic in that pore water pressure response is simply assumed consistent for all surface loading—whether due to atmospheric pressure changes or surface water storage changes. Once calibrated by atmospheric pressure data, the water mass loading changes appear after atmospheric pressure correction of recorded pore water pressures. Surface loading effects on confined units may include horizontal strain effects (Rojstaczer and Agnew, 1989) but this is of no particular consequence to the operation of a geological weighing lysimeter.

The unique feature of geological weighing lysimeters is their ability to give a spatial average of surface water budget changes over an extended land area. The size of the averaged area is difficult to quantify exactly but must increase with both aquifer transmissivity and depth below ground surface. Even aquitard-based measurements indicate that a 30-m piezometer depth will result in an averaged ground surface monitored area in the order of hectares (van der Kamp and Schmidt, 1997).

It is of interest to establish recorded observations of surface loading effects at depth because this extends the potential operational range of geological weighing lysimeters and, by implication, the area of land surface monitored. Presently, the greatest depth of an operational geological weighing lysimeter is 160 m at a site in New Zealand (Bardsley and Campbell, 2000). The maximum functioning depth may in fact be considerably greater because a general model suggests that surface load deformation penetrates to a depth approximately twice the length scale of the loaded area (Pagiatakis, 1990). A significant rainfall event, for example, might extend a surface loading effect at least over several kilometers.

One approach to detecting deep loading would be to simply check for barometric effects in deep open wells, because the inverse barometric effect represents in part the effect of atmospheric ground surface loading. This is not entirely unambiguous, however, because some aquifer waters may accumulate gas bubbles which cause significant increases in pore water compressibility (Matsumoto and Roeloffs, 2003). Open wells in such situations would display an apparent inverse barometric effect even when there was no transmission of surface loading through the confining layers. This potential problem can be avoided in open wells by checking for the effect of loading by ocean tides or rainfall events, which do not impact directly on the well water surface. Alternatively, deep wells could be sealed from the atmosphere and a check made for a positive water level response to ground surface atmospheric loading variations.

The loading signal of individual precipitation events is of particular interest from the geological weighing lysimeter viewpoint because this represents inputs to the surface water budget. Rainfall events appear to have been previously detected at 500 m depth (Igarashi and Wakita, 1991) but the signal was poorly defined. Bardsley and Campbell (2000) reference the loading effect of a 40-mm rainfall event, detected at 240 m below ground surface in a seismic monitoring network near Parkfield, California. We report in this short communication a rainfall loading response at almost 300 m depth from a sealed well near WaKeeney, west-central Kansas. We believe this to be one of the clearest signals to date of rainfall loading at depth, suggesting that there is indeed scope for extension of geological weighing lysimeters to at least this depth and possibly considerably deeper.