The working model for hydrothermal circulation used to plan Leg 168 is one in which permeable extrusive igneous rocks are between low-permeability deeper crustal rocks and low-permeability sediments. The physical nature of layering and basement exposure is readily apparent from extensive seismic surveying along the Leg 168 transect (Rosenberger et al., Chap. 2, this volume), as are age variations with distance from the spreading ridge (Shipboard Scientific Party, 1997; Su et al., Chap. 4, this volume). Cold, unaltered seawater enters the crust where basement is exposed at the seafloor and is heated and reacts as it flows through the rock. Buoyancy-driven fluid flow is confined within igneous basement, except where permeable rock is exposed at the seafloor and fluid exchange between the crust and the ocean can occur. This model served well, and both initial drilling results and subsequent studies provided excellent quantitative constraints on the nature of shallow, ridge flank circulation (e.g., Davis et al., 1999; Elderfield et al., 1999).
Perhaps the most surprising result of Leg 168 was the lateral distance over which the hydrothermal circulation in the igneous crust provides hydrologic communication beneath the relatively continuous sediment cover on the ridge flank. Thermal and geochemical evidence for large-scale transport was found at the time of drilling (Shipboard Scientific Party, 1997). Systematic increases in heat flow and basement temperatures with distance from basement outcrop (Pribnow et al., in press; Fig. 1) showed a clear influence of ventilated circulation as far as 20 km from the point of sediment onlap, and sulfate concentrations in basement waters indicated influence of seawater recharge over the full length of the drilling transect (Shipboard Scientific Party, 1997). Even at sites over 80 km from the area of extensive basement outcrop near the ridge, sulfate concentrations were found to have declined only to half seawater values, despite the diffusive loss into the sediment section where sulfate is consumed by bacteria. Shore-based studies confirmed this scale of influence and allowed estimates of lateral fluid flow on the order of meters per year (Davis et al., 1999; Elderfield et al., 1999; Wheat et al., 2000). Radiocarbon dating of large-volume water samples showed that nowhere along the transect was water >8 ka in age (Elderfield et al., 1999), even at the extreme eastern end of the transect, farthest from the spreading axis.