The mud volcanoes of the Mariana Trench region provide a unique look at rocks and fluids from a subduction zone. In this region, rock clasts and fluid-rich serpentine mud are brought from the forearc mantle wedge to the surface through faults in the overriding plate (Fryer, 1992, 1996). The densities and velocities of these rocks can be used in conjunction with seismic studies to constrain lithologies in subduction zones and the degree of alteration of the forearc mantle wedge.
Figure F2 shows the average compressional velocity of the 14 samples measured as a function of confining pressure. These samples show relatively little increase in velocity with pressure, particularly at low pressures. Many rocks show a large initial rise in velocity at pressures below 100 MPa due to closing of grain-boundary microcracks in the rock (Birch, 1961). The slow rise in velocities in the Leg 195 serpentinites indicates a low microcrack content in these rocks. Shipboard velocity measurements were made on different Leg 195 clasts at ambient pressure (Shipboard Scientific Party, 2002). These compressional wave velocities ranged from 3.80 to 5.48 km/s and are significantly lower than compressional wave velocities measured in this study. The shipboard measurements are likely to have been more affected by porosity and microcracks in the samples rather than the mineralogy and structure of the clasts.
Christensen (1966) established a velocity-density relationship for partially serpentinized rocks. He found that both density and velocity decrease with increasing serpentinization. The data demonstrating these relationships at a pressure of 200 MPa are shown in Figure F3. Samples in the 1966 study included serpentinite, partially serpentinized peridotites, peridotites, and dunites, most of which were from Burro Mountain, California (USA). Rocks with a density of 2500 kg/m3 are nearly 100% serpentinized, and rocks with a density of ~3300 kg/m3 are fresh, unserpentinized peridotite and dunite. The Leg 195 samples measured at 200 MPa were compared to samples from Christensen (1966) and show a narrow range of both velocity and density (Fig. F3). All of the Leg 195 samples fall in the range of rocks measured by Christensen (1966) that are 61%–95% serpentinized. The low velocities and densities of these clasts, along with this observation, indicate that the Leg 195 clasts measured here are all highly serpentinized in agreement with petrographic observations. The Leg 195 data agree well with the 1966 data. For the limited range of velocity and density, the velocity-density trend defined in Christensen (1966) is seen in the Leg 195 samples as well. They have slightly higher velocities for their densities compared to the 1966 trend, which may be due to variations in accessory mineralogy or phase of the serpentine.
Compressional velocity vs. shear velocity for Leg 195 rocks at a pressure of 200 MPa is shown in Figure F4. A linear trend line VP = 0.757 VS + 3.5375 is plotted to fit the data. Considering the narrow range of velocities, the scatter about the line is low, with a regression coefficient of r2 = 0.81. Dashed lines in the plot are lines of constant Poisson's ratio. Serpentinites characteristically have high Poisson's ratios (Christensen, 1996), and this is seen for these samples as well. Poisson's ratio increases with the degree of serpentinization, and is ~0.35 for 100% serpentine. As expected, samples with lower velocities, and therefore more serpentinization, have higher Poisson's ratios.
We used seismic studies of compressional wave velocities combined with knowledge of the geology of subduction zones to calculate the variation of serpentinization of the mantle rock in subduction zones. Assuming that the relationships from Christensen (1966) between velocity, density, and serpentinization apply at 200 MPa (~8 km depth), we can calculate percent serpentinization from a velocity. The percentage of serpentinite for the samples in Table T1 at a pressure of 200 MPa, corresponding to the mean velocity of 5.8 km/s using Christensen (1966), is ~66%. This values agrees reasonably well with the average degree of serpentinization of clasts (~70%) found by the Shipboard Scientific Party (2002).
For comparison, we calculated the percentage of serpentinite, assuming a lithostatic pressure of 200 MPa for the forearc mantle wedge, using velocities from a seismic study of the Izu-Bonin subduction zone (Kamimura et al., 2002). In that study, seismic surveys were conducted perpendicular to the Izu-Bonin Trench axis over a serpentine mud volcano, the Tori-shima serpentine forearc seamount, on a 130-km-long east-west line and parallel to the Izu-Bonin Trench axis on a 130-km-long north-south line. The Tori-shima mud volcano was studied during ODP Leg 125 and provided much of the data used to understand the South Chamorro serpentine mud volcano during Leg 195. Kamimura et al. (2002) identified nine layers in their east-west velocity model. Their Layer 6 corresponds to the forearc mantle wedge, the probable source of the muds for the serpentine mud volcanoes that line the Izu-Bonin and Marianna trench systems. In their velocity model of the forearc mantle wedge, the velocities decrease from 8.0 km/s at a distance of 162 km from the trench axis to 6.5 km/s at a distance of 62 km from the trench axis. This corresponds to an increase of percentage serpentinite from 4% at a distance of 162 km from the trench axis to 46% at a distance of 62 km from the trench axis. In a similar manner, the percentage of serpentinite in the layer corresponding to the forearc mantle wedge for the north-south seismic line of Kanimura et al. (2002) was also calculated. As noted in Kamimura et al. (2002), the area of lowest velocity on the north-south line in the forearc wedge is located nearest the Tori-shima mud volcano. This lowest velocity of 6.2 km/s corresponds to the highest percentage of serpentinization, equal to 54%. Based on these velocity models and assuming the relationships between velocity, density, and serpentinization from Christensen (1966) apply, the most altered rock in the forearc wedge is closest to the subducting slab and to the Tori-shima serpentine mud volcano. We also note that the average calculated percent serpentine of the Leg 195 clasts (66%) is more than that determined from the seismic velocity model (maximum = 54%). It is likely that these clasts were either taken from zones of increased serpentinization at depth or experienced additional serpentinization after the clasts were removed from their host rock.
The relationship between velocity and density might also aid in the interpretation of gravity measurements over the trench axis. Using the seismic model and the velocity-density curves, densities could be assigned to the altered forearc wedge that could provide an additional constraint to a gravity model of the subduction zone.