Previous   |    Next

doi:10.2204/iodp.proc.304305.204.2009

Results

Measurements were obtained from 109 samples (Tables T1, T2, T3; Plate P1) taken from the entire core in Hole U1309D, which include 7 oxide gabbros (oxide content > 2%), 72 gabbros and olivine gabbros, 14 troctolitic gabbros and troctolites, and 16 olivine-rich troctolites (olivine > ~70%). See the “Site U1309” chapter for further descriptions of the lithologies recovered from Hole U1309D.

Data quality is variable and was visually assessed on a case by case basis using an empirical scale for fit quality (from 1 = good to 3 = bad) (Fig. F1; Table T3). In 25 cases, the lowest salinity point tends to be low enough that it causes a deflection of the Revil and Glover (1998) model fit curve toward lower C_S values (Fig. F1C). This behavior was observed by Revil et al. (2002) in volcaniclastic material and was ascribed to the presence of zeolite in the samples. It is observed together with unusually high surface conductivities, which is also the case in our sample set. The 25 samples that show a deflection in the fit curve have high surface conductivities of >2.5 mS/m (mean = 19.7 mS/m), whereas the others have much lower surface conductivities of <2.6 mS/m (mean = 0.57 mS/m). Zeolite was described in Hole U1309D rocks (see the “Site U1309” chapter); hence, a similar behavior may be seen here. However, zeolite was only documented in the lower half of the hole, below ~700 mbsf (see the “Site U1309” chapter), whereas this electrical behavior is observed in 14 samples evenly distributed above 700 mbsf. Assessing the potential role of zeolite in the measured samples requires characterizing their alteration mineralogy, which is beyond the objective of this data report. An alternative or complementary contributor to these seemingly anomalous results could be serpentinite and/or associated minerals (e.g., brucite, talc, or magnetite), as many of these samples have a primary olivine-rich composition. We could not, however, identify a simple relationship between this peculiar electrical behavior and a compositional or textural parameter such as the amount of serpentinite (several heavily serpentinized samples do not show this behavior) or fracture density. One additional complexity to the problem of understanding this behavior is that in several cases, the low-salinity deflection of the Revil and Glover (1998) model fit curve is not observed, whereas the lowest salinity measurement clearly lies below the fit curve (e.g., Fig. F1B). Some poor-quality measurements in the higher salinity range, for example, may prevent the model from properly fitting the whole data set. When removing these samples from the data set, the remaining samples that show the standard curve fit as shown in Figure F1A correspond to samples with lowest degrees (<10%) of serpentinization.

In 21 samples, one or two points in the series of conductivity measurements were removed from the analyses because they were several orders of magnitude out of range when compared to the other points (Fig. F1C). In the majority of these 21 cases, the “bad” points were the lowest salinity measurements, as illustrated in Figure F1C, and we cannot exclude that these correspond to end-member versions of the zeolite-like behavior described above. Except for one oxide gabbro, these contain significant amounts of more or less altered olivine (olivine gabbro to olivine-rich troctolite).

The standard errors of F and C_S obtained from the analysis detailed above are given in Table T2 and shown in downhole distributions of F and C_S for both the Waxman and Smits (1968) and Revil and Glover (1998) models (Fig. F2). In most cases, the largest errors are smaller than the small-scale trends observed in these distributions, which then must correspond to real changes in some first-order controlling parameter(s) such as porosity or alteration. Downhole distributions of electrical properties are also shown in Figure F3, together with porosity and density and compared to the downhole distribution of lithology, variation of hand sample alteration, and borehole electrical resistivity.

As classically done in studies of electrical properties, the measurements are shown as a function of sample porosity, ϕ (Figure F4). Correlations in gabbros (Ildefonse and Pezard, 2001; Einaudi et al., 2005) and granites (Belghoul, 2007), such as a linear decrease of F and increases in C_S, τ, and m with ϕ, are still present but somehow disturbed by samples that tend to have higher C_S and lower F at fairly constant ϕ (~1%). As a result, τ and m tend to be more scattered, especially when abundant olivine/serpentine is present. Figure F5 shows the variations of ϕ, τ, and m as a function of the degree of fracturation (intensity of cataclastic fabric, as characterized onboard on hand samples) and the degree of serpentinization. Note that higher degrees of serpentinization do not significantly change porosity and tend to create a simpler geometry of the porous network (slightly lower τ and m). Not surprisingly, ϕ, τ, and m tend to increase in more fractured rocks.

Top of page   |    Previous   |    Next