IODP Proceedings    Volume contents     Search


Overview of hydrogeologic tracer experiments and concepts

This paper describes the preparation and operations associated with tracer injection and sampling operations during Integrated Ocean Drilling Program (IODP) Expedition 327. Results from this experiment will not be known for several years, when data and samples will be collected and analyzed from several subseafloor borehole observatories (“CORKs”), but this paper provides background information on the motivation for this experiment and will be useful for researchers considering similar experiments on future drilling expeditions. Discussion of the configuration and installation of CORKs during Expedition 327 (and related IODP Expedition 301) is presented elsewhere (Fisher et al., 2005; Fisher, Wheat, et al.), as is a comprehensive summary of Expedition 327 shipboard operations (see “Operations” in each site chapter of this volume). The regional geometry of the Expedition 327 tracer experiment is illustrated in Figure F1, which shows the distribution of boreholes (and associated observatory systems), basement outcrops, and the inferred dominant fluid flow direction in basement. We discuss experimental procedures after introducing tracer testing concepts and describing systems prepared for use during Expedition 327.

Numerous oceanographic studies have used geochemical tracers as indicators of the direction and rate of fluid flow beneath the seafloor. In most cases, geochemical tracers are naturally occurring (e.g., chlorinity, methane, silica, or 14C) (Elderfield et al., 1999; Lilley et al., 1993; Mottl and Wheat, 1994; Wheat and McDuff, 1995). Additional studies have benefited from use of surface seawater as a drilling fluid, creating a geochemical perturbation when deep-sea boreholes are drilled and cored (e.g., Mottl and Gieskes, 1990; Wheat et al., 2003, 2004). Fewer studies of subseafloor fluids have used artificial tracers introduced as part of a flow experiment. Solomon et al. (2009) deployed an automated tracer injection and fluid sampling system in a CORK deployed within a sedimentary interval seaward of the Middle America Trench, offshore of Costa Rica. Tracer injected into the borehole at a constant rate was sampled to assess both the flow of formation fluid in and out of the borehole and the relative direction of fluid transport. Wheat et al. (2010) deployed a similar injection and sampling system in a CORK completed in basement at a younger site on the eastern flank of Juan de Fuca Ridge. Both of these experiments were completed in single holes, and in each case the apparent rate of borehole exchange is considered to be a proxy for lateral transport rates in the formation surrounding the borehole.

Single-hole tracer experiments have been conducted in numerous aquifer systems on land, generally as a means to assess the bulk rate of fluid and solute transport around a borehole (e.g., Altman et al., 2002; Haggerty et al., 2001; Leap and Kaplan, 1988; Novakowski et al., 1998). There also is a rich literature involving land-based multihole tracer experiments in aquifers to assess formation and fluid flow properties, including dispersivity, effective porosity, particle transport rates, and the extent of solute-formation interactions (e.g., Becker et al., 2003; Birk et al., 2005; Göppert and Goldscheider, 2008; Hall et al., 1991). In some cases, multihole pumping and tracer experiments have been combined to acquire an understanding of linked physical-chemical transport processes (Dann et al., 2008; Day-Lewis et al., 2006). Completing and interpreting the results of tracer experiments is inherently challenging, both because of technical difficulties in running the tests and controlling experimental conditions and because of complexities in formation property distributions and the scaling of hydrologic and solute transport processes (Brouyère et al., 2005; Khaleel, 1989; Neuman, 2005; Niemann and Rovey, 2009; Novakowski et al., 1995).

The conceptual basis for a hydrologic tracer injection experiment is illustrated schematically in Figure F2. Consider fluid flow through an aquifer system at a steady rate of Qin = Qout. Injectate containing a detectable tracer is added at a low rate (but with a high concentration) at an upgradient location, such that the initial tracer concentration in the aquifer is Cin. Fluid in the aquifer flows toward a downstream monitoring location while tracer mixing, spreading, and reaction occur simultaneously. In an aquifer containing complex and heterogeneous flow paths, there can also be exchange between multiple flow networks and between the primary flow paths and background (matrix) pores. If the injection lasts a relatively short time relative to the period of measurement, a plume of tracer will migrate through the aquifer, increasing in size and being reduced in mean concentration with time (Fig. F2B).

Tracer is detected at one or more downstream locations, generating a record of tracer transport versus time (Cout), known as a “breakthrough curve” (Fig. F2C). The timing of arrival of the downstream record of tracer concentration, as well as the shape of this record, provides information about physical (and potentially chemical and biological) processes in the aquifer. If there were no lateral spreading and mixing and monitoring sites were located immediately downstream from the injection site, all of the tracer mass would be represented in the breakthrough curve. In practice, flow patterns and processes tend to be complex, monitoring points are not located perfectly downstream, and tracer mass recovery is rarely complete. Many tracers behave nonconservatively during transport. In addition, a single period of tracer injection can result in generation of a breakthrough curve having multiple peaks separated in time (Fig. F2B). One interpretation for this kind of result is that tracer is partitioned and transported along distinct primary and secondary flow paths.

Little is known about the actual flow paths of large-scale fluid transport in the ocean crust. The crust is often idealized as being a single homogeneous and isotropic hydrologic layer, or one in which properties vary smoothly with depth, but examination of core samples, wireline logs from boreholes, and ocean crust in outcrop suggests that flow pathways in the ocean crust are complex (e.g., Bach et al., 2004; Bartetzko, 2005; Gillis and Robinson, 1990; Gillis et al., 2001; Karson, 1998). The Expedition 327 tracer experiment was developed to provide fundamental information on rates and directions of transport, relations between solute and particle transport, the effective porosity of the crust (fraction contributing to the majority of fluid and solute transport), and the nature of connectivity (vertical and horizontal) between boreholes drilled into the extrusive ocean crust. We injected a mixture of solute, gas, and particle tracers (conservative and nonconservative) to assess how transport of these phases might differ. Because this kind of experiment has never been attempted in the ocean crust, simply demonstrating that two sections of the crust are in hydrologic communication across a distance of hundreds of meters and quantifying the time required for molecules of tracer to be transported between the sites would be an important result. This experiment was also developed to take advantage of the operational capabilities of the R/V JOIDES Resolution in conducting a tracer experiment (using existing shipboard systems to the greatest extent possible) and to determine how these capabilities might be modified for future tracer transport studies.