Previous   |    Next

doi:10.2204/iodp.proc.348.103.2015

Geochemistry

Hole C0002M

Interstitial water chemistry for core samples

One whole-round sample of 20 cm length was collected from each of the four Hole C0002M cores (348-C0002M-1R-1, 87–107 cm; 2R-2, 111–131 cm; 3R-1, 90–110 cm; and 4R-2, 113–133 cm). Interstitial water (IW) was obtained with a 55 mm diameter Manheim squeezer. The squeezing method used for this experiment differs from the standard squeezing method (see “Geochemistry” in the “Methods” chapter [Tobin et al., 2015]). Because the cores were not processed until 72 days after they were collected, the data are not useful for geochemical interpretations, but the cores were used to test the new squeezers to determine maximum squeezing pressure to use for future IW sampling. Increasing pressures up to 60,000 lb (corresponding to an applied pressure of 112 MPa) with different time steps were applied for assessing the effect of high pressure on the measured value of chlorinity, and IW was sampled at each different time interval. The squeezing sequences and their corresponding aliquots of IW and chlorinity results are listed in Table T16.

All samples were squeezed first according to standard hydraulic press Recipes 1–3 (Table T16; up to 20,000 lb [37 MPa] applied force), and Aliquot A was collected. The core samples were squeezed according to recipes 4–6 (up to 25,000 lb applied force [47 MPa]), and Aliquot B was collected. Finally, for samples from Sections 1R-1 and 2R-2, a force of 60,000 lb (112 MPa) was reached in four increments and Aliquot C was collected. For samples from Sections 3R-1 and 4R-2, an additional aliquot of fluid (Aliquot C′) was collected at 50,000 lb (94 MPa) just before ramping applied force up to 60,000 lb (112 MPa). This force was held for 12 h, and then Aliquot C” was collected.

Results of chlorinity measurements for different applied pressures are reported in Figure F52. Overall, results show a freshening of the fluid when the same sample was squeezed at higher pressure and for a longer time. With the exception of the highest pressure water sampled from Section 3R-1, chlorinity values are higher than those observed at the same depth interval (475–504.41 mbsf) during Expedition 338. This is likely due to contamination by higher chlorinity mud water while the cores were stored. During the pressure experiments, freshening increased with increasing pressure. The signal is strongest with core samples squeezed at pressures above 25,000 lb (47 MPa). In all four experiments, the first two squeezing steps yielded water with the same chlorinity value within error. Relative to the average of the first two steps, the long, highest pressure step (~60,000 lb [112 MPa] for 12 h) shows chlorinity decreases of 3% for samples from Sections 1R-1, 2R-2, and 4R-2 and 7% for Section 3R-1. The mechanism that induces the freshening of IW at high pressure is not yet understood, and later shore-based experiments will be conducted.

Organic geochemistry

Carbonates, organic carbon, and total nitrogen

Carbon values were analyzed for 15 samples from the four cores from Hole C0002M (Fig. F53; Table T17). Calcium carbonate content ranges from 4.26 to 13.67 wt% (median = 6.24 wt%). The highest point value of 13.67 wt% is from 478.70 mbsf (Sample 348-C0002M-1R-3, 87–89 cm), whereas Core 348-C0002M-4R has the highest average value. Carbonate content at this depth interval in Hole C0002M is higher than the highest and average values determined in the same interval in Hole C0002L obtained during Expedition 338, which were 6.1 and 1.4 wt%, respectively. Total organic carbon (TOC) is low, ranging from 0.46 to 0.82 wt%, and decreases slightly with depth. Hole C0002M TOC values are similar to those observed during Expedition 338 at the same depth interval. Total nitrogen (TN) is also low in Hole C0002M, similar to Expedition 338, and low TOC/TN values in Hole C0002M indicate a marine origin for the organic matter.

Comparison of XRD and coulometric data shows that the relative trends are the same, but XRD underestimated the abundance of carbonates (see “Lithology”).

Gas chemistry from cores

Headspace gas samples were taken from the four cores in Hole C0002M. Methane, ethane, and propane concentrations in the headspace gas were measured by a gas chromatograph–flame ionization detector (GC-FID) (see “Gas analysis in core samples” in the “Methods” chapter [Tobin et al., 2015]).

The concentrations of methane and ethane are shown in Table T18 and Figure F54. Propane was not detected in the core samples. The concentrations are shown in parts per million by volume (ppmv) and moles per kilogram, which is the concentration of gas molecules dissolved in 1 kg of interstitial water, as calculated by the equation in “Gas analysis in core samples” in the “Methods” chapter (Tobin et al., 2015). The concentrations of methane and ethane are 2532–7355 and 5.4–9.9 ppmv, respectively. Although variable, the concentrations generally increase with increasing depth.

Temperature estimation from headspace gas and total organic carbon

A cross plot of C₁/(C₂ + C₃) ratios and temperature is commonly used to get quick information about the origin of the hydrocarbons (i.e., to distinguish between biogenic gas and gas migrated from a deeper source of thermogenic hydrocarbon). A very high C₁/(C₂ + C₃) ratio indicates in situ gas formation by microbiological process. On the other hand, the occurrence of major amounts of C₂ + C₃ in shallow depths is associated with thermogenic hydrocarbon generation and migration. In general, the C₁/(C₂ + C₃) ratio consistently decreases with burial depth, resulting in a consistent (normal) decrease in the C₁/(C₂ + C₃) with increasing temperature. Anomalously low C₁/(C₂ + C₃) ratios suggest the presence of migrated thermogenic hydrocarbons. The separation pattern of normal versus anomalous zones was suggested by Stein et al. (1995) and JOIDES Pollution Prevention and Safety Program (PPSP; 1992) and is shown in Figure F55. Temperature estimations are based on heat probe and LWD data (JOIDES PPSP, 1992). Also shown by the solid lines is the approximate influence of different levels of organic carbon content (i.e., TOC). Sediment showing high C₁ concentrations is characterized by relatively high organic carbon contents and high sedimentation rates because the decomposition of the organic matter is dominated by anaerobic microbial processes under these conditions (Claypool and Kvenvolden, 1983). Although this diagram was initially established for safety purposes, it is used here to estimate approximate temperatures in the borehole using TOC in cuttings (>4 mm) and the C₁/(C₂ + C₃) ratio in drilling mud-gas data obtained during Expedition 348. Following this approach, it seems that the C₁/(C₂ + C₃) ratios are too low for the given TOC concentrations and thus would result in unrealistically high temperatures (Table T19). Most likely, hydrocarbon gas concentrations are not only generated in situ but are also influenced by gas migration. Consequently, the temperature-C₁/(C₂ + C₃)-TOC relationship is not suitable for estimating formation temperature (JOIDES PPSP, 1992).

Holes C0002N and C0002P

Interstitial water chemistry for core samples

Five whole-round samples varying from 10 to 41.5 cm were collected from Cores 348-C0002P-2R through 6R over the 2176.28–2211.31 mbsf depth interval (see “Geochemistry” in the “Methods” chapter [Tobin et al., 2015]). Recovery in Core 1R was insufficient for an IW sample. Traditional squeezing methods did not produce any IW because of sediment consolidation, so all samples were processed to obtain pore water (GW) using the ground rock interstitial normative determination (GRIND) method used during Expedition 338 (Wheat et al., 1994). Data are summarized in Table T20, after correction of dilution based on the water content determined at 60° and 105°C, with the exceptions of salinity (refractive index) and pH.

Salinity values obtained by the standard GRIND method, including determination of IW percentage by drying samples at 105°C, show a large range paralleled by chlorinity values, which range from 387 to 848 mM (Fig. F56). The lowest value of the chlorinity range is similar to values obtained in Kumano Basin sediment, either from the standard or GRIND method (Strasser et al., 2014a). Chlorinity in Section 348-C0002P-4R-2 samples is >50% higher than average seawater (551 mM). The standard drying temperature of 105°C to determine GW content could give erroneously high estimates of IW by also releasing some of the clay-bound water. If estimates of pore water obtained using a lower drying temperature of 60°C are considered; however, the chlorinity values are significantly higher for three of five samples (Sections 2R-3, 4R-2, and 6R-2).

Chlorinity values show a downhole variation of alternating high and low values paralleled by several other major and minor ions, including Br^–, NH₄⁺, Na⁺, K⁺, Mg²⁺, Ca²⁺, Li, Mn, Ba, Si, B and Sr (Figs. F56, F57). Although the possibility of localized brines in the formation cannot be excluded, it is unlikely that such large fluctuations in concentrations could occur in samples taken only 10 m apart with respect to depth. It is more likely that the variations in concentrations are related to lithology or mineralogy, core retrieval, and sample processing of the GRIND method. Crushing the samples may have damaged clay minerals and increased their capacity to absorb water. Also, addition of water in the sample produces a paste that is also subject to water loss during the transfer process. Both of these processing steps could induce artifacts in measured values of chlorinity.

Cross plots of Na⁺, K⁺, and Ca²⁺ versus chlorinity in GW, along with concentrations in mud water, show trends indicating that most of the variation in compositions is related to changes in salinity (Fig. F58). The plots show that mud contamination can be eliminated as the cause of this salinity increase, as no mixing trend between a lower saline fluid and the mud-water concentrations is observed except for Sample 2R-3, for which a slight trend toward mud potassium values exists. KCl was a significant additive to the mud fluids and K⁺ in mud-water samples is almost 1–2 orders of magnitude higher than in GW, whereas Cl in mud water is only ~4 times the GW values (Table T21). Moreover, the perfluorocarbon (PFC) tracer data indicate that mud-water contamination is possibly 5% in this sample (Table T22; Fig. F59).

The cause of the salinization trends in Hole C0002P cores is not clear. The increase in concentrations might be related to uptake of freshwater by hydrous clay minerals, occurring in response to unloading during the wireline trip of the core barrel back to the surface, which took ~2 h. Uptake could also occur during the milling process of preparing GRIND samples. The differences in salinity increases would then be related to differences in clay mineral abundance or clay mineralogy in the different cores. This will be investigated further with shore-based research.

To examine whether signals related to diagenesis or other geological processes are present in GW data from the cores, the concentrations were normalized to chlorinity and plotted relative to depth for selected ions (Fig. F60). This assumes that actual chlorinity values do not vary substantially over the cored interval or the processes effecting the concentration data was a simple addition or subtraction of water from the pore space that equally impacted all ions (e.g by simple hydration/dehydration of hydrous clays as one example). Clearly, chlorinity variations can occur, so the following observations can only be considered as approximations.

Normalization of concentrations to chlorinity significantly reduces the scatter in the data for many elements, although it does not completely eliminate the alternating pattern of high and low values observed in the concentration data (Table T23; Fig. F61). K and Ca show significant variation after normalization. The variations are in the range of those observed in shallower parts of Site C0002. The highest normalized K value could reflect some contamination by drilling mud, which has high K contents. However, the other alkali metals, Rb and Cs, also show high values in this core (not present in any quantity in the mud), and thus the large values are more likely to reflect diagenetic or other natural processes occurring at this depth. The range of normalized K and Rb suggests variations related to clay mineral diagenesis, such as might be expected for the conversion of smectite to illite. Normalized Ca and Sr values also show ranges that suggest variation related to carbonate diagenesis. Core 5R exhibits large carbonate veins associated with a fault zone in Section 5R-4.

The results of trace element concentrations extracted by the GRIND method are plotted in Figure F62 and listed in Table T20. Most of the minor and trace elements fall in the range of concentrations obtained from the standard squeezing method on cores from Holes C0002B and C0002J (lithologic Unit III/IV; Strasser et al., 2014b). Trends similar to salinity are observed for Rb and Cs, and presumably Zn, despite missing data for Section 5R-2 (not detected during analysis). Mo shows remarkably high concentrations that are 10 times greater than those from Expedition 338 (Strasser et al., 2014b). One explanation could be contamination by the grease used to connect drilling pipes, but core contamination is generally not indicated by the PFC test. Cs concentrations are also two times higher than in Holes C0002B and C0002J. Pb decreases with depth comparably to SO₄^2–. As previously observed in GRIND results from Expedition 338 (Strasser et al., 2014b), no Fe was detected with the inductively coupled plasma–mass spectrometry except in Section 6R-2, with a concentration of 2.04 mM after correction with the water content at 60°C. It is likely that Fe was oxidized by the grinding process and adsorbed on mineral surfaces.

Organic geochemistry

Carbonates, organic carbon, and total nitrogen

Carbonates as CaCO₃, TOC, and TN were determined from the 1–4 and >4 mm size fractions of the cuttings from Holes C0002N and C0002P and in Hole C0002P cores. TOC was calculated by the difference of total carbon (TC) and inorganic carbon (IC) (see “Geochemistry” in the “Methods” chapter [Tobin et al., 2015]. Results are plotted in Figure F63 and listed in Tables T24 and T25. Contamination by cement from drilling operations was observed in CaCO₃ and CaO profiles of XRF data (Fig. F13B) from 870.5 to 950.5 mbsf (Sample 348-C0002N-20-SMW) in Hole C0002N and from 1960.5 to 2040.5 mbsf (Sample 348-C0002P-32-SMW) in Hole C0002P. Those data were therefore not plotted in the figure and are not discussed in the following section.

CaCO₃ in cuttings varies from 0.46 to 9.61 wt% with a median of 3.2 wt%. Scattered data are mostly observed in cuttings from Hole C0002N, whereas those from Hole C0002P show a narrower range of values. A prominent peak is observed at ~1900 mbsf, where CaCO₃ is as high as 7.94 wt%. Plots of the data from the drilling mud seem to follow the same trend and could reflect an uptake of carbonates in the mud. The peak is also reflected by CaO in XRF data (see “Lithology”). A wider peak at ~2650 mbsf shows CaCO₃ values as high as 5.80 wt%. Despite occurrences of carbonate veins recovered in the fault zone at 2205 mbsf, no increase in CaCO₃ is visible at that depth in the core samples or the cuttings.

TOC varies between 0.47 and 2.07 wt%, with a median of 0.9 wt%. The trend shows a gradual downhole decrease with local minor variations along the profile (e.g., a small peak at ~1790.5 mbsf [Sample 348-C0002N-205-SMW]). The range of values at the bottom of Hole C0002N between 2040 and 2320 mbsf does not strictly match those at the beginning of Hole C0002P, even though data for CaCO₃ and TN match in the overlapping interval. Data from core samples in Hole C0002P agree with those of cuttings from Hole C0002P more than with cuttings from Hole C0002N.

TN varies between 0.024 and 0.064 wt% with a median of 0.053 wt%. TN does not show any particular trend in cuttings from Hole C0002N. A shift toward higher values is observed around 2600 mbsf in cuttings from Hole C0002P. Most of the core samples have higher TN values than cuttings of both holes, except for two core samples (Sections 2R-4 and 5R-5) that have lower values

The C/N ratio that depicts marine (4~10) or terrestrial (>10) origin of organic matter varies from 7.58 to 35.58 with a median of 17.53. Therefore, most of the organic matter in Holes C0002N and C0002P is considered to be of terrestrial origin. However, core samples seem to tend toward to marine origin except for two samples (Sections 2R-4 and 6R-4), highlighting possible contamination from mud water in cuttings. The decreasing trend of the C/N ratio and the variations within match those observed in the TOC profile.

Gas chemistry from cores

Concentrations of methane (CH₄), ethane (C₂H₆), and propane (C₃H₈) in the headspace gas collected from core in Hole C0002P (2160–2220 mbsf) are shown in Table T26 and Figure F64. Methane is the predominant hydrocarbon in all core samples, and ethane and propane were detected in all samples. The concentration of methane varies between 1,704 and 10,676 ppmv, except for two peaks at 2176.7 mbsf (20,183 ppmv) and 2183.4 mbsf (23,455 ppmv). The concentrations of ethane and propane are 57.5–346.9 and 24.6–351 ppmv, respectively, and increase slightly with increasing depth. The Bernard parameter ratio (i.e., C₁/(C₂ + C₃) is also shown in Table T26 and Figure F64. The C₁/(C₂ + C₃) ratio varies between 7 and 68, and its trend mostly resembles methane concentrations. At 2205.5 mbsf, where the fault zone was identified (see “Structural geology”), the concentrations of ethane and propane decrease slightly relative to methane, and the C₁/(C₂ + C₃) ratio increases.

Comparison between headspace gas data and drilling mud-gas data obtained by gas chromatograph–natural gas analyzer (GC-NGA) and GeoServices analysis shows that the concentrations of methane, ethane, and propane in the drilling mud are much lower than those in the headspace gas (Fig. F65). The highest methane concentration in the drilling mud is 696.7 ppmv at 2174 mbsf. The gas recovery rate by GC-NGA analysis is up to 27% at 2174 mbsf. Methane concentration estimated by GeoServices analysis is also lower (<750 ppmv) than headspace gas, with a recovery rate up to 6%. The highest ethane and propane concentrations are 7.1 ppmv (2180.4 mbsf) and 2.81 ppmv (2164.5 mbsf), respectively, and the recovery rate of these gases is 8% by GC-NGA analysis. This underestimate is likely caused by the high mud density and low rate of penetration during coring. Despite the different concentrations, relative changes in methane are similar, with the highest methane values in Cores 348-C0002P-2R and 3R and an overall decrease in concentrations with increasing depth. Within the cores, ethane and propane variations are similar; ethane and propane start to increase with Core 5R, but the drilling mud-gas concentrations decline.

The C₁/(C₂ + C₃) ratios calculated using GC-NGA and GeoServices data are higher than the headspace data (36–124 and 15–47, respectively) (Fig. F66). Figure F67 shows the Bernard diagram, which is usually used to distinguish between thermogenic and biogenic sources of hydrocarbon gas (Bernard et al., 1978). Microbial origin methane has δ¹³C-CH₄ values <55‰ Vienna Peedee belemnite (VPDB) (Rice and Claypool, 1981), and C₁/(C₂ + C₃) ratios as high as 1000 (Bernard et al., 1978). Thermogenic methane has δ¹³C-CH₄ values between –50‰ and –25‰ VPDB (Schoell, 1983) and C₁/(C₂ + C₃) ratios lower than 100. The headspace methane sampled from Hole C0002P core samples primarily falls in the region of thermogenic origin (Figs. F66, F67).

Comparison with other data sets

Plotting the C₂/C₁ ratios of headspace gas samples versus δ¹³C-CH₄ from the methane carbon isotope analyzer (MCIA) data set (Fig. F68) indicates normal maturation of organic matter rather than migration or diffusion of thermogenic gas. Figure F69 shows moisture and density (MAD)-derived porosity versus total headspace gas volume. In general, higher gas concentrations are expected for elevated porosities. Two outliers with exceptionally high gas concentrations at a porosity of ~0.2 are at a horizon where faults were identified (see “Structural geology”). Such high concentrations were not observed in the drilling mud-gas data, however. The significance of Rn data generated during coring and hole opening by underreaming are ambiguous. Although ²²²Rn in Hole C0002N is only present in traces, coring and drilling Hole C0002P revealed concentrations up to ~200 and 250 Bq/m³, respectively (Fig. F70).

Mud-gas chemistry

Overview of mud-gas composition

Continuous drilling mud-gas monitoring took place while drilling Holes C0002N and C0002P from 838 to 2330 mbsf and 1954 to 3058.5 mbsf, respectively. Two autonomous data sets were generated during the operation: a Rn data set available from the stand-alone Rn monitoring instrument (see “Online radon analysis” in the “Methods” chapter [Tobin et al., 2015]) and one from the SSX database, which was compiled with data from the GeoServices and scientific mud-gas monitoring systems on the Chikyu (hereafter referred to as “SciGas system”; see “Onboard mud-gas monitoring system” in the Methods chapter [Tobin et al., 2015]). The latter provided data from the GC-NGA, MCIA, and process gas mass spectrometer (PGMS). Gas concentrations determined by the MCIA, GC-NGA, and PGMS differ from those measured by GeoServices, most likely due to having a different degasser configuration of the onboard system during drilling of Hole C0002N (Strasser et al., 2014b). For Hole C0002P, the scientific degasser system and location were changed, which resulted in similar concentrations to the GeoServices’ system (Figs. F65, F71; see also “Geochemistry” in the “Methods” chapter [Tobin et al., 2015]). For the PGMS, this is only of secondary importance because the concentrations of different gases were normalized to 100%. Nevertheless, the individual data sets include data from different instruments, measurement techniques, and sampling intervals, and all data are influenced by drilling operations (e.g., rate of penetration, mud pump activity, and mud weight).

In the GeoServices data set, the total hydrocarbon gas volume content range from ≤0.0001% (i.e., below detection limit) to 8.2% at ~1305 mbsf. At a similar depth, methane has a maximum concentration of 8.1%. Overall, methane trends resemble the concentrations of total gas, and methane dominates the gas show. Ethane and propane are only present in minor concentrations, with the highest values at 979 (0.05% ethane) and 1302 mbsf (0.02% propane), respectively. Higher homologs (i.e., n-butane, i-butane, n-pentane, and i-pentane) typically stay below 0.01% and do not add significantly to the total gas composition. In Hole C0002N, absolute gas concentrations determined by the SciGas system are significantly lower than those from the GeoServices data set, except for ethane. However, relative changes in gas concentration in the SciGas data are similar to GeoServices with maxima between 1320 and 1405 mbsf (Figs. F71, F72, F73). In Hole C0002P, the new degasser configuration led to absolute gas concentrations for methane and propane on the same order of magnitude as the GeoServices instruments (Figs. F65, F74, F75). Ethane values from GeoServices remain higher by a factor of ~10.

For nonhydrocarbons, the PGMS data set is dominated by nitrogen (~77.6%) and oxygen (~20.7%), whereas the concentrations of the remaining nonhydrocarbons rang from 1.15 × 10^–6 percent for CO₂ to 0.99% for Ar (Fig. F76). Absolute concentrations determined with the PGMS are influenced by sensitivity checks and calibration, which lead to sudden concentration shifts, and therefore the data might be biased (see “Online gas analysis by process gas mass spectrometer” in the “Methods” chapter [Tobin et al., 2015]). In addition, depth intervals for which the ion current of the PGMS was unstable were removed from the data set. The concentrations of He, H₂, and Xe usually remain below the detection limit. H₂ was detected with the GC-NGA only in Hole C0002P (1954–3058 mbsf), with values up to 0.78% (Fig. F77). ²²²Rn concentrations usually stay below 450 Bq/m³ (Fig. F78).

Hydrocarbon gas distribution with depth

From 838 to 983 mbsf in Hole C0002N, total gas concentration stays below 2% and then abruptly increases to ~6.4% (Fig. F71). Gases in this interval are mainly hydrocarbons. Between 983 and 1500 mbsf, total gas concentration remains between 0.01% and 8.2%. Below 1500 mbsf, the values decrease sharply and stay less than 3.0%, except at 2145 mbsf, where the concentration increases to 3.2%. In Hole C0002P, total gas concentrations follow similar trends but are ~1 order of magnitude smaller compared to the Hole C0002N data set (Fig. F65). Starting at ~2170 mbsf, total gas concentrations in Hole C0002P decreases sharply to <0.2%, followed by an increase to ~0.9% at ~2184 mbsf. Farther downhole, no change in trend is visible.

As mentioned above, the gas show is dominated by methane (Figs. F71, F72). Data quality for ethane and propane from 838 to 1560 mbsf is poor, with most of the values below detection limit. Both ethane and propane show distinct peaks at ~980 mbsf with a sudden increase to 0.05% and 0.01%, respectively. Between 980 and 1560 mbsf, ethane concentrations remain less than 0.03%, whereas propane concentrations further increase with a maximum of 0.02% at 1320 mbsf. Below 1560 mbsf, data quality increases as ethane and propane concentrations rise above the detection limit. However, ethane values stay less than 0.03%, and propane decreases to concentrations less than 0.01%. Intervals relatively enriched in ethane and ethane + propane were found between 1825 and 1970 mbsf and between 2070 and 2220 mbsf, respectively. The sudden increase in propane from 0.001% to 0.01% between 1990 and 2000 mbsf is exceptional in the hydrocarbon gas show. Similar to the total gas/methane concentrations, ethane and propane in Hole C0002P are up to 2 orders and 1 order of magnitude smaller, respectively, compared to gas concentrations in Hole C0002N (Figs. F65, F74, F75). Relative changes, however, are similar. At ~2170 mbsf, ethane and propane decrease below detection limit for the GeoServices instruments, whereas the GC-NGA-reported concentrations are <10 ppm. Starting at 2200 mbsf, all data sets show a continuous increase of higher hydrocarbons (ethane and propane) with depth.

Distribution of nonhydrocarbon gas data with depth

²²²Rn values detected by the stand-alone Rn monitor show an increase with depth with some scatter (Fig. F78). In Hole C0002P between 1954 and 2330 mbsf, ²²²Rn values are almost twice as high as in Hole C0002N. Also, the significant decrease in Hole C0002N between 1835 and 2000 mbsf is absent in Hole C0002P. Overall, the Hole C0002P data follow the trend starting at 1835 mbsf in Hole C0002N, and concentrations increase to values >300 Bq/m³ at ~2500 mbsf. Below 2500 mbsf, ²²²Rn data show no trend but are highly scattered between 56 and 692 Bq/m³.

Nonhydrocarbons detected by the PGMS show major variations only at points when calibration or sensitivity checks were carried out (Fig. F76). With averages of 77.6%, 20.7%, and 0.96% for N₂, O₂, and Ar, respectively, concentrations are similar to atmospheric values. CO₂ shows a slightly negative trend downhole from ~838 to 1300 mbsf. With an average of 0.0005%, concentrations of CO₂ stay below atmospheric values, except for ~1677 mbsf, where a single peak of 0.16% occurs, and below 2600 mbsf, where values >0.2% were encountered. In Hole C0002N, He and Xe values are as high as 0.06% and 0.02%, respectively (3 orders of magnitude higher than atmospheric values). Starting at 2600 mbsf, He and Xe have even higher concentrations, to 3% and 2.9%, respectively. Overall, He and Xe show a slightly positive trend in Hole C0002P.

H₂ data detected by the GC-NGA are relatively constant downhole and stay below 0.075%. At two exceptional peaks at 1969 and 3043.5 mbsf, H₂ reaches values to 0.34% and 0.78%, respectively.

Comparison with data from Expedition 338, Hole C0002F

The gas data from 838 to 2007 mbsf in Hole C0002N and from 1954 to 2007 mbsf in Hole C0002P are compared with the gas show from Hole C0002F (Strasser et al., 2014b). Comparison between the three mud-gas data sets reveals the following (Figs. F71, F76, F78, F79, F80, F81, F82):

1. Gas show: in all holes, gas show is dominated by methane.
2. Total gas and methane: the prominent gas kick of total gas and methane in Hole C0002F between 918 and 1000 mbsf is absent in Hole C0002N. By contrast, a small gas kick occurs in Hole C0002N between 1300 and 1460 mbsf, whereas no excursion is visible in Hole C0002F. The overall total gas/methane concentrations in Hole C0002N remain slightly higher downhole. Between 1950 and 2010 mbsf, data from Hole C0002P became available. Overall concentrations are lower, and the peak at ~1993 mbsf is absent in the Hole C0002F data (Fig. F79).
3. Ethane: overall, ethane concentrations in all boreholes are less than 0.05%. Between ~900 and 1500 mbsf, the two data sets show no apparent correlation. The gas kick in Hole C0002N at ~980 mbsf correlates well with the lower boundary of the gas kick in methane from Hole C0002F but has no corresponding signal in ethane from Hole C0002F. Between 1250 and 1460 mbsf in Hole C0002N, there is no sudden shift to higher concentrations in ethane, contrary to Hole C0002F. Between 1500 and 1700 mbsf, ethane concentrations show no significant deviation. Between 1700 and 1800 mbsf, ethane concentrations in Hole C0002F increase, contrary to the concentrations in Hole C0002N. Below 1800 mbsf, ethane concentrations drop in Hole C0002F but rise in Hole C0002N. In Hole C0002P, the few data points show concentrations ~1 order of magnitude lower compared to data from Hole C0002F (Fig. F79).
4. Propane: overall, propane concentrations in both Holes C0002F and C0002N remain very low (<0.008%). The elevated concentrations starting at ~980 mbsf are absent in Hole C0002F. Between ~1250 and ~1500 mbsf, propane concentrations in Hole C0002N are high, whereas a decline is visible in Hole C0002F. Below ~1500 mbsf, propane concentrations in Hole C0002N are similar to those in Hole C0002F. In Hole C0002P, propane is only present in trace amounts (Fig. F79). Concentrations are similar to Hole C0002F.
5. δ¹³C-CH₄: except for some outliers, data from Holes C0002N and C0002P match data from Hole C0002F. At ~1700 mbsf, a shift to more thermogenic methane is indicated by values greater than or equal to –60‰ (Fig. F80).
6. Other nonhydrocarbon gases: comparison between the two mud-gas data sets obtained by the PGMS (both data sets are normalized to 100%) and the ²²²Rn data sets reveal the following (Figs. F78, F81, F82):
1. ²²²Rn: above 1835 mbsf, ²²²Rn concentrations in Hole C0002N are significantly lower than in Hole C0002F (Figs. F78). Except for the peak between 918 and 1000 mbsf in the Hole C0002F data, relative changes in concentrations are the same. Between 1835 and 2007 mbsf, both data sets correspond well in absolute values and relative changes in concentration. The same is true for data from Hole C0002P (Fig. F81).
2. Despite some problems with the ion current, data quality from the PGMS improved (Figs. F76, F82). Fewer calibrations were necessary, which decreased the number of data shifts in the Expedition 348 data set.
3. All data sets are dominated by atmospheric gases (O₂ + N₂ = 98.3%).
4. N₂ and O₂ concentrations in Hole C0002F are slightly higher and lower (Fig. F82), respectively, when compared to the values from Expedition 338. Ar values are higher in Hole C0002P, whereas Hole C0002N data resemble the Hole C0002F concentrations. CO₂ is highest in Hole C0002N and lowest in Hole C0002P. At ~1930 mbsf, a CO₂ peak in Hole C0002F corresponds to elevated values in Hole C0002N. A limited increase in CO₂ at ~1985 mbsf is reproduced in Hole C0002P.
5. Values of Xe are similar, whereas at depths shallower than 1200 mbsf, He concentrations are lower, and at depths deeper than 1600 mbsf, He concentrations are higher (Fig. F76). Between 1200 and 1600 mbsf, the apparently higher He concentrations are most likely caused by calibration of the PGMS during monitoring. Overall, concentrations in Holes C0002N and C0002P are higher compared to Hole C0002F (Figs. F76, F82).

Influence of drilling parameters and comparison with other data sets

The drilling mud-gas data can be easily influenced by drilling, or operational, parameters. For example, degassing is less effective when the mud level drops in the return mud trough, which can produce mud gas “artifacts” in gas data. The movement of the BHA up and down (“swabbing”) within already drilled sections can also produce mud-gas anomalies, which underlines the importance of correlating operational procedures with mud gas analyses. This movement of the BHA can produce an effect similar to “pumping,” as the movement of the BHA acts similarly to a piston in the borehole and can draw in mud gas from the formation, influencing mud-gas concentration analyses.

The influence of changes in drilling parameters becomes obvious when comparing the concentrations of hydrocarbon gases between drilling mud-gas data and in headspace gas data from the cored section of Hole C0002P; drilling mud-gas data from Hole C0002P were 1 order of magnitude lower. Propane in both Holes C0002N and C0002P are 1 order of magnitude lower compared to the headspace gas concentrations. It remains possible that the use of progressively higher mud weights in Hole C0002P compared to Hole C0002N (1.12 to 1.32 sg) could reduce the intake of higher hydrocarbons, and at the surface, hydrocarbons are likely to be retained in the cold, dense drilling mud (Ablard et al., 2012). As a consequence, the difference is most likely caused by underestimation of hydrocarbons in the drilling mud gas.

Two more examples of drilling operation artifacts occur at ~1320 and ~1954 mbsf. The prominent gas peaks correspond clearly to a time when the drill string was moved up and down several times and thus are interpreted as “trip gas” or “swab gas” rather than a true increase in formation gas concentration. An example of swab gas is visible between 30 December 2013 and 1 Jan 2014. When swabbing took place at ~4916.0 and 5010.7 m BRT, gas concentrations in drilling mud increased abruptly (total gas >3.5% and methane >28000 ppmv). The swabbing effect of upward movement of the drill string induced a temporary reduction of pressure in borehole. If the hydrostatic pressure at the bottom of the borehole decreases below the formation pressure of a gas-bearing formation, gas would be introduced into the borehole. Breakouts, possibly caused by temporary underpressure in the borehole, are also suggested by the presence of “pressure cuttings.” The first major total gas peak clearly corresponds to swabbing at ~4915 m BRT and is within the expected lag time (i.e., ~135 min at this depth) for a constant mudflow rate. The next major peak occurs after swabbing at ~5015 m BRT between 2200 and 2300 h (i.e., after 2345 h), with the bit moving up and down in smaller steps. At this time the role of the mudflow rate remains unclear. Although the time difference is understandable given the low mudflow rate during swabbing, it is still within the expected lag time. Large volumes of gas tend to move upward faster than the drilling mud, and this upward movement increases with decreasing depth. Alternatively, if changes in flow rate are taken into account, swabbing at more shallow depths could be responsible as well.

The influence of changes in drilling parameters becomes obvious when comparing the concentrations of hydrocarbon gases in the drilling mud-gas data with the concentrations found in headspace-gas data from Hole C0002P (Fig. F65). Although drilling mud-gas data from Hole C0002N showed concentrations for methane and ethane similar to those in headspace gas from Hole C0002P, drilling mud-gas data from Hole C0002P are 1 order of magnitude lower. Propane in both Holes C0002N and C0002P are 1 order of magnitude lower compared to the headspace gas concentrations. Higher hydrocarbons are present in headspace gases at up to 49 ppmv (for i-butane) (Table T26). Subsequently, the Bernard parameter based on headspace gases indicates a slightly more thermogenic regime (Fig. F67). Reasons for this discrepancy are manifold, including the use of a higher mud weight in Hole C0002P and changing degasser efficiency (i.e., how much gas could be extracted from the drilling mud). Higher mud density would reduce the intake of higher hydrocarbons in the borehole, and at the surface, hydrocarbons are likely to be retained in the cold, dense drilling mud (Ablard et al., 2012). As a consequence, the difference is most likely caused by underestimation of hydrocarbons in the drilling mud gas.

In addition, the distributions of hydrocarbon gas expelled from Hole C0002P (1966.5–2780.2 mbsf) during widening of Hole C0002P (reaming gas) is different compared to that obtained during drilling. Concentrations of methane and ethane in the reaming gas are similar to gas concentrations encountered during drilling at shallower depths. However, methane and ethane found during drilling increased more steeply with depth. Concentrations of propane in the reaming gas hardly change with depth profile. At the same time, propane increased with increasing depth during drilling and became higher than the reaming gas at ~2400 mbsf. To estimate the difference of those hydrocarbon gas characteristics, these data are plotted on the Bernard diagram. Whereas gas concentration ratios obtained from drilling plot in a mixing area, reaming gas concentration ratios indicate a strong thermogenic regime.

Based on the overall changes detected in hydrocarbons in the GeoServices and SSX data sets, 3 boundaries (4 gas packages) are defined (Fig. F83). The first boundary, separating log Unit III from IV, is set at 983 mbsf, where a minor gas increase exists. The overall drop in hydrocarbon gas concentrations at 1500 mbsf is considered a second boundary. The third boundary is set at ~2200 mbsf, where the total gas concentrations increase overall. Considering other shipboard data allows correlation of some gas geochemical data to log unit boundaries at 915 mbsf (Unit III/IV boundary), 1514 mbsf (Subunit IVd/IVe boundary), and 2191 mbsf (Subunit Vb/Va boundary). Between 860 and 2008 mbsf, 7 hydrocarbon gas packages were defined during Expedition 338 (Strasser et al., 2014a), of which the Package 1/2 boundary at 918 mbsf and Package 4/5 boundary are close to the gas package boundaries defined during Expedition 348. The fact that none of the gas-package boundaries precisely match the log unit boundaries might be related to hydrocarbon gas diffusion and the uncertainties related to drilling mud-gas monitoring. Later postcruise research might elucidate the cause of these differences.

Similar to the data from Expedition 338 (Strasser et al., 2014a), the causes of nonhydrocarbon-gas data shifts at ~900, 1190, 1650, 2008, and ~2200 mbsf are not fully resolved. Most of the anomalies in the PGMS data correlate well with periods in which sensitivity checks or calibrations were carried out (Fig. F76).

Nonhydrocarbon gases are clearly dominated by atmospheric components. Although N₂ can also originate from various sources, including clay-rich sedimentary rock (e.g., Krooss et al., 1995; Mingram et al., 2005) the N₂/Ar ratio < 100 (Jenden et al., 1988; Krooss et al., 1995), and the overall constant He/Ar and Xe/Ar support an atmospheric source (Fig. F84). Additionally, the stable O₂/Ar ratio rules out oxidation processes at the drill bit. It remains unclear if the overall data scatter is related to a real change in gas composition. Postcruise noble gas studies will shed light on this issue.

The presence of H₂ in Hole C0002P can have various sources (Fig. F77). The low concentrations of H₂ compared to hydrocarbon gas concentrations suggest organic sources (Wiersberg and Erzinger, 2008, and references therein). δD analyses could help with a clear distinction of source. Alternatively, H₂ can be a product of artificial processes, such as electrolytic products released by interaction of the drilling fluid products and the drilling equipment, or by processes related to metal-to-metal friction.

Classification of hydrocarbons

Clear classification of the hydrocarbon gases proved to be difficult, mainly due to the conflicting data sets (Fig. F80). Above 1000 mbsf, δ¹³C-CH₄ has values less than or equal to –60‰. Combined with a C₁/(C₂ + C₃) ratio of 1557 based on GC-NGA data, this indicates a bacterial source of methane. By contrast, the Bernard parameter based on the GeoServices data set indicates a thermogenic to mixed regime in this interval, with ratios between 75 and 377. Farther downhole, the Bernard parameter indicates the presence of mixed gases from bacterial and thermogenic regimes. Starting at 1500 mbsf, this pattern becomes more prominent as ethane and propane concentrations stay above the detection limit. At the same time, δ¹³C-CH₄ steadily declines. The δ¹³C-CH₄ ratio, Bernard parameter, and C₁/(C₂ + C₃) ratios from the GC-NGA point to an increase in thermogenically derived hydrocarbon gases below 1700 mbsf. This trend stops at ~2025 mbsf. Farther downhole, δ¹³C-CH₄ values stay almost constant at an average of –48‰, whereas the Bernard parameter averages 91.2 for the drilling mud-gas data and ~23.3 for the headspace gas. Between ~2325 and 2600 mbsf, δ¹³C-CH₄ declines again, and the Bernard parameter drops to values >50, indicating an early mature thermogenic regime (e.g., Whiticar, 1994). Below 2600 mbsf, the Bernard parameters from GeoServices and GC-NGA data remain at 21 and 40, respectively, and methane carbon isotope ratios remain at an average of –43‰.

By using a Bernard diagram (i.e., Bernard parameter versus δ¹³C-CH₄; see also Bernard et al., 1978; Whiticar et al., 1994), it is possible to further evaluate the hydrocarbon gas composition (Fig. F67). Plotting the data in the Bernard diagram indicates that the gas show in Hole C0002N is influenced by secondary effects like mixing, diffusion, and/or microbial oxidation (Whiticar et al., 1994). Following the procedure of Prinzhofer and Pernaton (1997), it is possible to qualitatively estimate if the gas show is affected by mixing or diffusion fractionation in Hole C0002N (Fig. F85). Hydrocarbon data from GeoServices, however, show no clear trend, and although the data from the GC-NGA might indicate mixing, the low number of data points precludes further interpretation. Below 2325 mbsf in Hole C0002P a clear change to thermogenic gas is evident (Figs. F67, F85). Later onshore analysis will focus on the δ¹³C-CH₄ of ethane and propane, so it will be possible to calculate a proper mixing model for the gases present in the accretionary prism (Whiticar et al., 1994).

The diagram of C₁/(C₂ + C₃) ratio and temperature relationship is also used to get quick information about the origin of the hydrocarbons (i.e., to distinguish between biogenic gas and gas migrated from a deeper source of thermogenic hydrocarbon). A very high C₁/(C₂ + C₃) ratio indicates in situ gas formation by microbiological process. On the other hand, the occurrence of high amounts of C₂ and C₃ at shallow depths is associated with thermogenic hydrocarbon generation. The separation pattern of normal versus anomalous zones was suggested by Stein et al. (1995) and Shipboard Scientific Party (1995), as shown in Figure F86. Stein et al. (1995) estimated the sediment temperature following temperature gradients given by JOIDES PPSP (1992). Also shown by the two solid lines is the approximate influence of different organic carbon content (i.e., TOC). Sediment showing high C₁ concentrations is characterized by relatively high organic carbon contents and high sedimentation rates because the decomposition of the organic matter is dominated by anaerobic microbial process under these conditions (Claypool and Kvenvolden, 1983).

We plot values for TOC in >4 mm cuttings (with some 1–4 mm cuttings data) and the C₁/(C₂ + C₃) ratios in drilling mud-gas data obtained during Expeditions 338 and 348 (Fig. F86; Table T27). The C₁/(C₂ + C₃) ratio is calculated using data measured by the GC-NGA. The C₁/(C₂ + C₃) ratio is high (>1000) at shallow levels and decreases normally with increasing depth. Below 2200 mbsf, the C₁/(C₂ + C₃) ratio decreases to 100, and the Bernard diagram plots indicate that the majority of methane is thermogenic (Fig. F67). Therefore, the diagram cannot be used to estimate the temperature in the borehole deeper than 2200 mbsf.

Top of page   |    Previous   |    Next