Recent studies suggest that continental flood basalts can be emplaced as inflated compound pahoehoe flows with a three-part internal structure of an upper crust, a lava core, and a basal zone (Hon et al., 1994; Self et al., 1996; 1997; 1998). The upper crust is highly vesicular, with hypocrystalline texture (i.e., a high original glass content) and irregular jointing comprising typically 40%-50% of the total flow thickness. The lava core comprises 40%-60% of the total flow thickness. It is characterized by low vesicularity, regular jointing, and a holocrystalline texture, whereas the porosity is dominated by diktytaxitic voids between the crystals. The basal zone is thin (0.5-1 m), has a high original glass content, and has poorly developed jointing.
A three-part lava flow structure has been inferred from log and wireline data on volcanic margins (Planke, 1994), being similar to the upper crust, lava core, and basal zone internal structure seen in the field. However, the petrophysical data from Hole 990A suggest the upper crust may actually be divided into two zones (Fig. 4) (Bücker et al., in press). This subdivision is based on the physical property measurements, but it can also be related to the severity of the post-emplacement alteration. The high initial vesicularity and jointing in the upper crust provides pathways for fluids and facilitates complete alteration of the dominantly glassy crust. Studies suggest that the initial alteration of the flow top is rapid, soon after the emplacement in a subaerial environment, and that smectite- and iron oxide-rich paleosols develop (Desprairies et al., 1989; Singer et al., 1994). During subsequent submergence of the basalt pile, the paleosols may have become relatively impermeable, and hydrothermal fluid flow may be in fractures below the highly altered layer. Such late-stage hydrothermal alteration has been by proposed by Desprairies et al. (1989) and might explain the relatively high potassium content and magnetic susceptibility in the intermediately altered Zone II, compared to the highly altered Zone I and the fairly fresh Zone III (Fig. 4).
The thickness and degree of alteration appear to be related to the primary emplacement environment. The jointed, highly vesicular upper crust corresponds to the strongly and intermediately altered Zones I and II. The paleoclimate and time duration between subsequent eruptions are important parameters in determining the alteration stage and, thus, the physical properties of the lavas. The completely altered Zone I is not present in the highly vesicular Unit 9 (Fig. 4). We relate this to rapid emplacement of subsequent pahoehoe flows not providing sufficient time for a soil horizon to develop.
Goldberg (1997) reviewed seismic wave propagation in oceanic basalts. In particular, variations in total porosity and the pore aspect ratio spectrum are commonly regarded as the major cause of velocity variations, a hypothesis substantiated by theoretical considerations and numerical modeling (e.g., Wilkens et al., 1991). In addition, oxidation and hydration alteration processes are considered to be significant in modifying the physical properties of oceanic crust with age (Johnson and Semyan, 1994).
The crack concentration and aspect ratio distribution are considered to be the main causes of velocity variations in the massive parts of subaerial lava flows. Cerney and Carlson (Chap. 3, this volume) show that the variations in measured P- and S-wave velocities in minicores from the massive parts of lavas from Hole 990A can be modeled as primarily caused by variation in total porosity. The modeling also suggests that a common pore aspect ratio distribution normalized by porosity exists within the massive parts of the upper three aa flows from Hole 990A.
The effect of clay alteration on seismic properties of flood basalts is more difficult to quantify. First, it is difficult to recover the altered zones using rotary drilling techniques. Second, very altered samples are fragile and, thus, it is difficult to measure reliable velocities in the laboratory. Third, the velocities of altered samples are often severely reduced during storage because of clay swelling.
The Vp measured under atmospheric conditions during the shore-based studies of Hole 990A is 0.5-1 km/s lower than the corresponding shipboard velocities (Table 2). We relate this difference primarily to the swelling properties of smectite as microfissures develop during storage. The dominant alteration clay mineral in Hole 990A is smectite, which is a low-temperature alteration mineral that is physically unstable because of its swelling properties (Douglas et al., 1994; McGreevy, 1982). Microfissures are developed during cyclic wetting and drying of basalts with minor amounts of smectite, leading to a deterioration of the rock, which is associated with a significant decrease in seismic velocity (McGreevy, 1982). The microfissure development is nonreversible and will also affect anisotropic measurements as the cracks develop preferentially parallel to the rock foliation (Ruiz de Argandona et al., 1995). Confining pressures have to be increased to 200 MPa to match the original shipboard P-wave velocities (Cerney and Carlson, Chap. 3, this volume). Thus, to provide reliable data, velocity data from altered basalt cores should be measured on board ship as soon as possible after recovery or kept saturated under confining pressure.
The large mineralogical changes within Unit 2 and Subunit 3A (Fig. 4) correspond to changes in the elastic properties within the basalt units. The Vp in Zone I is very low (~2.5 km/s) and fairly constant. This zone is clay dominated (>50% smectite; Fig. 4). Velocity logs in smectite-rich intervals in the North Sea show similar low velocities (~2 km/s) and insignificant velocity changes with depth (Thyberg et al., unpubl. data).
Numerous rock-property theories have been developed to describe changes in elastic moduli in clastic rocks consisting of clay and sand mixtures (e.g., Hansen, 1997). The three sand/clay models shown in Figure 6 cannot be used directly to predict velocity variations in altered crystalline rocks such as basalts, but they provide insight into the significance of clay minerals on the elastic properties of basalts. All three theories predict a significant decrease in velocity when the clay proportion exceeds ~30%-40%. This is related to a pronounced softening of the rock when the clay minerals become a main part of the rigid rock framework. Zone II is characterized by a rapid increase in Vp. The modal clay proportion in this zone varies from 5% to 16% (Table 2). As clay minerals and iron hydroxides are part of the rigid framework, they may significantly soften the rock. But, as both the porosity and alteration mineral percentages vary in this zone, a reduction in either or both may locally contribute to the overall increase in the Vp.
Susceptibility and remanence values are lowest in Zone I (Fig. 4). These low magnetic values can be attributed to the hematite, a magnetic mineral with the lowest susceptibility and magnetic intensity values, which confirm the highly oxidized and altered zone (Table 2). Zone II can be subdivided into two parts characterized by the magnetic property behavior. In the upper part, the magnetic intensity is highest with moderate susceptibility values, and, in the lower part, the susceptibility is highest with lowest magnetic intensity values. This is possibly caused by two independent effects-smaller grain sizes (single domain particles) in the upper part, and larger grain sizes (multiple domain particles) in the lower part. The susceptibility behavior can be explained by a combined effect of grain size and oxidation state. The upper part with small grain sizes could be less oxidized (maghemite), and, thus, it shows lower susceptibility values than the higher oxidized lower part (magnetite) with large grain sizes.
The fresh and unoxidized Zone III shows the original magnetic properties with low magnetic intensity values and moderate susceptibility values, pointing to magnetite with a low titanium content and relatively large grain sizes. This dense part of the flow may serve as a seal and may partly protect the lowest part of the flows against intense alteration from the top. The lower part of Zone IV again shows high magnetic intensity values and moderate susceptibility values. Because of the oxidation scheme, the magnetic mineral may be (titano)maghemite with small grain sizes.
The most prominent feature of the magnetic properties susceptibility and remanence is the high values in Zone II (Fig. 2, Fig. 4). The question is whether it is caused by primary magnetic minerals or alteration. Ore microscopy of polished thin sections suggests that the dominant magnetic mineral is magnetite with differing amounts of titanium ([titano]magnetite) (Table 2). This observation is also confirmed by studies on volcanic rocks from Leg 152 (Fukuma, 1998). (Titano)magnetite undergoes alteration during low-temperature oxidation and hydrothermal alteration. In the upper part of the lava flows, the highly vesicular and fractured texture provides pathways for fluids. In the first step of low-temperature oxidation, the (titano)magnetite is changed to (titano)maghemite (Fig. 7). This is accompanied by decreasing susceptibility and magnetization intensity and increasing Curie temperatures. In the second step, the (titano)maghemite is changed into magnetite and then further into hematite. The oxidation to magnetite is accompanied by increasing Curie temperatures, susceptibility, and magnetization intensity, whereas the oxidation to hematite lowers the susceptibility and magnetic intensity but increases the Curie temperature. It is not necessary in each case during low-temperature oxidation that the (titano)magnetite is first changed into (titano)maghemite and then into magnetite or hematite. There might also be a direct way of oxidizing (titano)magnetite into magnetite and hematite (Fig. 7). In addition to the oxidation effects on the magnetic properties, susceptibility and magnetic intensity are also dependent on the abundance of the magnetic minerals, grain-size distribution, and shape.
In summary, we suggest from the magnetic property observations that the lower parts of the highly vesicular, scoriaceous, and fractured lava top and bottom are characterized by high magnetic susceptibility and intensity caused by different states of low-temperature oxidation. During these different states of low-temperature oxidation, susceptibility and magnetic intensity may be enhanced or reduced depending on the alteration to magnetite or hematite. A secondary effect on the magnetic properties is caused by the grain-size distribution.
Wireline logs are important to link downhole core data with surface geophysical measurements and to correctly interpolate missing sections in the core data (e.g., Goldberg, 1997). Detailed core-log integration is often difficult as the accurate depth location of the cores with respect to logging data frequently is impossible to obtain on the required centimeter or tens of centimeters scale. The high core recovery and densely sampled MST data and P-wave velocities from Hole 990A provide a unique data set in a flood basalt sequence. The Hole 990A core data can therefore be used to gain insight into interpretation of wireline logs recorded in flood basalt terrains. This is confirmed by downhole logging data from the nearby Hole 917A (Fig. 1) (Planke and Cambray, 1998) drilled into similar volcanic basement showing very similar intraflow variations in petrophysical logs as the Hole 990A core data. Figure 8 shows a sketch of the internal Vp variation in a thick flood basalt unit based on the Hole 990A data. The velocity data can be interpreted in terms of superimposed alteration and porosity variations and is typically represented by Unit 2 and Subunit 3A (Fig. 4).