Abstract
The Holocene sediments of the eastern margin of the Japan Sea are characterized by high total organic carbon (TOC) and total nitrogen (TN) contents, low TOC/TN and TS/TOC values with enriched signatures, as a result of high marine productivity during present oxic highstand. On the other hand, the LGM sediments are characterized by low TOC and TN contents, high TOC/TN and TS/TOC values with depleted signatures, characteristic of C3-derived terrestrial organic matter input during that anoxic lowstand. However, at the top of mounds at the seafloor, where gas hydrate and authigenic carbonate nodules occur, the host sediments have a mixture of both Holocene and LGM geochemical signatures. Both gas hydrate and authigenic carbonate, formed by the anaerobic oxidation of methane, increased the sedimentary volume and caused an uplift of older sediments, inducing mound formation. The thickness of the Holocene sediments over mounds is very small or absent exposing the last glacial maximum (LGM) sediments to the seafloor. The uplift of the LGM sediments within mounds is estimated to be >2 m. We conducted geochemical analysis to detect such sediment movement, using samples collected by shallow cores in the Joetsu Basin, eastern margin of the Japan Sea.
1. Introduction
Japan Sea is a typical back-arc basin formed behind the Japanese island-arc system initiated by the rifting of the eastern margin of the Eurasian Continent at around 25 Ma [1]. The opening was almost completed before 15 Ma [2, 3]. During the Middle Pliocene, the tectonic style changed from the extensional to compressive [4] and a series of NNE-SSW trending folds were formed along the eastern margin of the Japan Sea [5], where an incipient subduction zone extends throughout the western side of the Japanese island-arc system [6]. As a result, several potential hydrocarbon traps were formed during this period and continuous subsidence created kitchen areas with mature source rocks [5]. Joetsu Basin is located southwest of Sado Island (Figure 1) and was formed during the Miocene [5, 7]. A favorable source rock was developed by high production of organic matter under anoxic conditions in the Nanatani Fm. (<12.5 Ma) and Teradomari Fm. (12.5~5.5 Ma) [5]. Oil generation occurred in the Miocene and a 15 m oil column was confirmed in tuffaceous sandstones in the lower part of the Shiiya Fm. (5.5~3.5 Ma) [5]. The top of the Nishiyama Fm. (3.5~1.3 Ma) is characterized by a “domino style” with several horsts and grabens and both normal and reverse faults are observed, reflecting the complex stress field involved [8]. Some of these faults belong to the rifting time and were reactivated during the inversion tectonic process [9]. The Haizume formation has been deposited since the late Pleistocene, and it is dominated mainly by clayey sediments [10].
Umitaka Spur and Joetsu Knoll are two anticlines formed since the middle Pliocene with a regional NE-SW trend [11], located approximately 30 Km offshore Joetsu city (Figure 1). A complex NE-SW axial fault system, composed of both normal and reverse faults, is observed in the central parts of both anticlines [11, 13]. The combination of faults, carrier beds, depositional surfaces, thermogenic gas source location, and the geometry of the anticlines focuses gas migration towards the top of both structures and also provides gas to the gas hydrate stability zone (GHSZ). The axial fault system results in an inefficient seal and trap and high amounts of methane are delivered to the GHSZ, where gas hydrate is generated within gas chimneys (Figure 2) [11, 13].
Associated to the gas chimneys, Matsumoto [12, 15] observed several mounds and pockmarks at the seafloor. According to his studies, mounds were formed by the crystallization of gas hydrates, while pockmarks were the result of intense gas hydrate dissociation during the LGM, when the sea level dropped ca. 120 m below that of the present [16]. He proposed that pockmarks were the final stage after mound formation, when gas hydrate stability conditions were severely lost during the lowstand of the LGM.
The main propose of this study is to detail the influence of both gas hydrate and authigenic mineral formation for the growth of mounds related to gas seepages. Both processes within the porous space increase the sedimentary volume and induce the formation of mounds by an uplift of old sediments associated with high methane flux venting. To confirm this, we performed core descriptions and geochemical analysis in sediments collected by piston and push cores at the mounds and surrounding areas for comparing their geochemical signatures.
2. Methodology
Dozens of piston and push cores were collected for gas hydrate research from the Joetsu Knoll, Umitaka Spur, and surrounding areas since 2005 by the R/V Umitaka Maru of the Tokyo University of Marine Science and Technology and by the R/V Kaiyo of the Japan Agency for Marine-Earth Science and Technology (JAMSTEC). These studies have been conducted by the University of Tokyo and other institutions, providing a significant improvement in the geological knowledge of the eastern margin of the Japan Sea, in particular of the Joetsu Basin [11–15, 17–19].
A total of 22 piston and push cores, respectively, 6 to 9 m and 0.50 m long, were used for this study, recovered on both Umitaka Spur and Joetsu Knoll at Joetsu Basin (Figure 1). One piston core (PC701) was collected at Oki Trough (Figure 1), representing the reference characteristics for open sea conditions, where there is no indication of the occurrence of gas hydrates. From these cores, a total of 475 samples were analyzed for TC, TOC, TN, TS, and δ13Corg. A total of 67 samples belonging to the gas hydrate-bearing sediments (Figure 3) were collected over topographic mounds (Figures 4, 5, and 6).
(a)
(b)
(c)
(a)
(b)
For TOC, TN, and δ13Corg analysis, sediment samples were powdered and treated using 10% HCl solution to remove carbonate. An aliquot of each sample was preserved for analysis of TC, with no acid treatment, to calculate TIC (TC-TOC) and to control the quality of the acid treatment. Acidified samples were dried on a hot plate at 55°C for 1 day and later in an oven at the same temperature for 4 additional days. The weight of each dried sample was measured before and after acid treatment for later normalization, considering possible salt formation and weight increase [13, 14].
Ca. 20 mg of each sample were analyzed with a Thermo Finnigan Flash EA 1112 series CNS analyzer at the laboratory of the Department of Earth and Planetary Science of the University of Tokyo, using a retention time of 720 s. The analytical error was <0.2 wt% for TOC, <0.02 wt% for TN, and <0.1 wt% for TS, using sulfamethazine standard. The reproducibility error for duplicate analysis was <0.5 wt% for both TOC and TS, while TN had error <0.05 wt%.
A combination of Thermo Finnigan Flash EA 1112 series analyzers, CONFLO III, and Delta Plus mass spectrometer was used to analyze δ13Corg. The weight of each sample was ca. 1.5 mg, depending on the TOC content. The analytical error, using standard IAEA-C6 sucrose, was <0.1% relative to Peedee belemnite (PDB) and the reproducibility error of duplicate analysis was <0.1%.
3. Results
3.1. Lithology
Five lithologic units were identified from core descriptions from the bottom to the top (Figure 7). Unit 5 is at the bottom and is characterized by light gray bioturbated silty mud belonging to the early LGM sediments [13, 14]. Unit 4 (TL-2) is characterized by thinly laminated dark gray mud deposited under anoxic conditions during the LGM lowstand.
Unit 4 represents the TL-2 described by Tada et al. [20] and others, one of the most important and widespread layers of the late Quaternary in the Japan Sea. Unit 3 is a slightly bioturbated light gray silty mud and represents the LGM/Holocene transition [13, 14]. Unit 2 (TL-1) is a 5–20 cm dark gray thinly laminated mud layer, which is also common in the Japan Sea. TL-1 was deposited under anoxic/suboxic bottom water conditions as a result of the water stratification caused by the inflow of fresh water during the beginning of the Holocene [20]. Finally, Unit 1 is characterized by light gray bioturbated mud and represents the oxic bottom water conditions from the early-middle Holocene to the present [13, 14].
Tephra were identified in PC701 and their glass shards and composition were correlated [13, 17] with the Atlas of Tephra in and around Japan [21]. The upper tephra is a pumice type in Unit 1, ca. 50 cm above the top of Unit 2 (TL-1) at 1.88 m below the sea floor (mbsf-Figure 7). It was identified as Ulreung-Oki (U-Oki) tephra (10.7 ka cal BP) [21]. The lower tephra is a bubble wall glass type and is in the Unit 5, ca. 50 cm below the base of Unit 4 (TL-2) at 5.95 mbsf (Figure 7). Both the shape and the composition of the shards are well correlated with the Aira-Tanzawa (AT) tephra (28-29 ka cal BP) [21].
A total of four foraminifera samples were collected from PC701 for 14C dating: one sample of Neogloboquadrina dutertrei (warm water planktonic) at 0.80 mbsf in Unit 1 and three of Globigerina umbilicata (cold water planktonic) at 2.60 mbsf, in Unit 2 (TL-1) and at 3.63 and 4.34 mbsf respectively, both in Unit 4 (TL-2). Sedimentation rates were estimated for PC701 based on both tephra and foraminiferal data (Table 1) and inferred for the cores located in Joetsu Basin (Figure 7).
A depth-age conversion was made on the basis of tephra, 14C of foraminifera, and lithologic correlation [13, 14, 17]. Consequently, it was possible to obtain a good correlation between the reference core PC701 and those from the Joetsu Basin (Figure 7). The top of Unit 2 (TL-1) is inferred to occur at around 11.0 ka cal BP, while the top of Unit 3 is inferred to occur at 12.5 ka cal BP. The top of Unit 4 (TL-2) is placed around 18.0 ka cal BP and the bottom at 26.0 ka cal BP, which represents the top of Unit 5. The top of Unit 1 is the seafloor. High sedimentation rates were observed in PC701 (Figure 7), located at a favorable depositional site (Oki Trough depocenter). Samples below AT were estimated by extrapolation.
3.2. TOC, δ13Corg, TOC/TN, and TS/TOC Signatures
Graphs comparing age versus TOC (wt%), age versus δ13Corg (‰ PDB), age versus TOC/TN, and age versus TS/TOC were constructed to promote an age-based correlation between Joetsu Basin and Oki Trough (Figure 8). TOC varies from 0.5 to 2.0 wt% and δ13Corg varies from −22.5 to −26.0% in Unit 5 at both locations (open sea and enclosed bay). TOC/TN strongly oscillates from 9 to 58, while TS/TOC ranges from near zero to 2.5. Unit 4 (TL-2) is characterized by TOC ca. 0.5 to 1.5 wt% and δ13Corg varying from −23.0 to −26.0‰. TOC/TN strongly oscillates from 14 to 96 and TS/TOC varies from 0.5 to 2.0 (Figure 8).
TOC increases from 0.5 to 1.5 wt% and δ13Corg varies from −23.8% at the base to −21.7% at the top of Unit 3, representing the LGM/Holocene transition. TOC/TN decreases from 46 at the base of Unit 3 to around 10 at the top, while TS/TOC oscillates from 0.5 to 2.5, mainly in cores located at Joetsu Basin.
Unit 2 (TL-1) is characterized by TOC from 1.0 to 2.5 wt%, and δ13Corg varying from −21.0 to −24.0‰. TOC/TN is <20 and TS/TOC oscillates from near zero to 1.0. Unit 1 is characterized by TOC with a maximum value of 2.5 wt% at the base to a minimum of around 1.0 wt% at around 6.5 ka cal BP at open sea conditions (Figure 8). The δ13Corg varies from −22.4% at the base to −20.1% in the upper part, representing the present Holocene sedimentation pattern in the area. TOC/TN oscillates nearly 10 on average and TS/TOC ranges from near zero to 1.0 (Figure 8).
Gas hydrate fragments and carbonate nodules recovered from mounds obstructed the penetration of piston cores, causing low recovery of sediments. All the piston-cores in these areas recovered surface sediments shallower than 2.0 mbsf despite the 4 to 6 m piston corer tube length used. In the same way, push cores recovered the shallowest 50 cm of seafloor sediments.
At these sites, the lithologic correlation is inaccurate due to the disturbance of sediments during coring operation. However, it is reasonable to assume that coring recovered shallowest sediments present at the seafloor (Figure 9).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
4. Discussion
The shallowest Holocene sediments of Joetsu Basin, until ca. 2.0 mbsf, are characterized by high TOC and TN contents, followed by low TOC/TN values (Figure 9). Enriched δ13Corg signatures, combined with low TOC/TN values, indicate a predominance of marine organic matter deposited under oxic conditions, suggested by low TS/TOC values [13, 14]. On the other hand, the LGM sediments are characterized by low TOC and TN values, depleted δ13Corg signatures, and high TOC/TN values, indicating a predominance of C3-terrestrial organic matter deposited under anoxic conditions, suggested by high TS/TOC values [13, 14]. These values reflect that the terrestrial organic matter input was strongly controlled by the sea level changes that occurred since the LGM [14].
At mounds, however, shallow sediments from depths equivalent to those of the surrounding background Holocene sediments are strongly depleted in δ13Corg followed by high TOC/TN values, similar to those of the underlying LGM sediments (Figure 9).
Elevated TS/TOC values are observed in the near seafloor sediments at mounds. These values can be explained by the anaerobic oxidation of methane (AOM) caused by the shallow boundary between the sulfate reduction and the methanogenesis zones.
AOM occurs within the ocean floor sediments, where sea waters, enriched in sulfate, meet methane formed at both methanogenesis and/or thermogenesis zones. This interface is named sulfate-methane interface (SMI) or transition (SMT) [22]. At the SMT the precipitation of sulfide and carbonate is common, which can explain the higher TS content and the presence of authigenic carbonate nodules in shallow sediments of mounds. AOM can be represented as the following equation:
However, this reaction cannot explain the anomalous depleted values of δ13Corg (−26‰ to −31‰) and high TOC/TN signatures (almost >20), in an opposite trend of the neighboring Holocene sediments. Based on the age-based graphs (Figure 8) it is possible to infer that the gas hydrate-bearing sediments, located at the top of mounds, have similar geochemical signatures of those of the LGM sediments. On the other hand, samples collected in the mound flanks are apparently similar to the Holocene samples of Unit 1. Gas hydrate-bearing sediments are strongly disturbed but, in spite of this, Dr. Fumio Akiba (personal communication) was able to identify diatom species from the LGM in those sediments located on the top of mounds (Figure 10). On the other hand, he found Holocene species at similar depths in cores located in the flanks of the same mounds.
The diatom analysis of eight piston cores has revealed that they contained common to abundant diatom assemblages, and they can be clearly subdivided into four diatom zones as A, B, C, and D zones, in descending order, primarily based on the limited occurrences of two marker diatoms in the studied sediments: Fragilariopsis dolilus and Thalassiosira hyperborea. The former is a warm water species, and it is common to abundant in A Zone. The latter is a cold water species often associated with ice sheets and low-salinity water, and rare to common occurrence is recognized in C Zone. The B Zone is an interval between the A and C zones, and the D Zone is a horizon below the C Zone (Figure 10).
The characteristic occurrences of two diatoms in the late Quaternary sediments of the Japan Sea have been well noticed previously and were linked with dated paleoceanographic events of Japan Sea [23–25]. The diatom zones A, B, C, and D range from 0–8 ka, 8–15 ka, 15–24 ka, and >24 ka, respectively.
Crossplots of geochemical parameters are strong tools to evaluate the characteristics of both Holocene and LGM sediments. The crossplot TOC/TN versus TS/TOC (Figure 11(a)) illustrates that the sediments of Unit 1 and almost all of the sediments from Unit 2 (TL-1), both from the Holocene, plot in the field corresponding to marine organic matter deposited under oxic conditions. In the same graph it is possible to observe that sediments from Unit 3 (LGM-Holocene transition) partially plot in both marine/oxic and terrestrial/anoxic fields, while almost all samples from both LGM Unit 4 (TL-2) and Unit 5 plot in the terrestrial/anoxic field (Figure 11(a)). Gas hydrate-bearing sediments recovered from the top of mounds, however, have signatures partially compatible to those of LGM sediments, suggesting a mixture with Holocene sediments.
(a)
(b)
(c)
(d)
(e)
Excepted when related to foraminifera-enriched layers [14], the crossplot TOC versus TIC (Figure 11(b)) shows that Holocene sediments from both Unit 1 and Unit 2 (TL-1) have high TOC content and relatively low TIC content. However, superficial gas hydrate-bearing sediments are strongly enriched in TIC caused by authigenic carbonate precipitation as a result of AOM.
The crossplot TOC versus TS (Figure 11(c)) is useful to differentiate euxinic from normal marine (noneuxinic) sediments [20]. Sediments of Unit 1 and Unit 2 (TL-1) partially plot on both oxic- and suboxic type fields. Gas hydrate-bearing sediments, however, plot totally in both suboxic-type and euxinic-type fields, similar to those sediments of Unit 4 (TL-2).
The crossplot TOC/TN ratio versus δ13Corg (Figure 11(d)) suggests that geochemical signatures of gas hydrate-bearing sediments are partially similar to those of LGM sediments, suggesting the mixture of different origins. Such similarities are also confirmed by the graph TOC versus δ13Corg (Figure 11(e)). This graph shows that gas hydrate-bearing sediments are strongly depleted in δ13Corg in a similar trend of the LGM sediments of Unit 4 (TL-2).
Based on the geochemical signatures (Figure 11), supported by the disappearance of the Holocene diatom zones A and B in the shallow gas hydrate-bearing sediments at the top of the mounds (Figure 10), the following process is proposed here to explain these unusual characteristics. (a)The formation of gas hydrate nodules and blocks (Figure 3) causes an increase in the pore space of sediments located around the fault-conduits and within the fault planes themselves.(b)The increase of volume in both pore space and fault planes causes an increase in the volume of the sediments as a whole. Combined with this, an intense gas/water pore pressure toward the top induces a strong upward movement promoting the uplift of the sediments and the formation of the mounds at the seafloor. The pore pressure is created by the upward migration of both methane and water released from deeper gas hydrate dissociation at the base of the GHSZ, or direct from deeper hydrocarbon reservoirs below the BSR [11, 13].(c)On the seafloor, erosion or nondeposition of the Holocene sediments at the top of mounds can explain the absence of the Holocene sediments and the consequent exposure of deeper LGM sediments.
Another proof of the proposed uplift of sediments is the presence of carbonate nodules at the seafloor over mounds. Methane-related carbonate nodules were not formed at the seafloor. They were formed by AOM at the SMT, below the seafloor [22].
Carbonate nodules recovered in piston cores from mounds of both Joetsu Knoll and Umitaka Spur show 14C ages ranging from 26 cal ka BP to 47 cal ka BP at depths randomly varying from 0 to around 4 mbsf, indicating that they were formed during the LGM [26], in spite of the nodules, older than 35 cal ka, and should not be considered due to the potential presence of dead carbon.
On the other hand, carbonate nodules were observed in some Holocene sediment (Figure 11), indicating recent generation of authigenic minerals caused by active methane seep and gas hydrate dissociation.
This may indicates that carbonate nodules formed at different times are now exposed on the seafloor, and they need to be uplifted first, and then eroded to be spread out around the mounds (Figure 6). On the other hand, only erosion cannot explain the phenomena observed at both Umitaka Spur and Joetsu Knoll because all the seep sites are located at mound areas, not favorable for sediment deposition (Figure 6).
The combination of all these factors indicate that the uplift of pre-Holocene sediments to the surface of the seafloor, followed by the erosion or nondeposition of the Holocene sediments, is the better explanation for the growth of mounds of both Umitaka Spur and Joetsu Knoll.
The uplift >2 m, inferred by correlation using both geochemical and sedimentological parameters, was caused by the increase in the sedimentary pore space and the consequent increase of the volume of the whole sedimentary section creating sediment uplift. This process is still ongoing at the present time, and the gas/water migration from deeper zones through faults and fractures amplifies the uplift and creates the gas seeps and plumes observed over mounds.
A model of the uplift of solid materials caused by the increase in the volume of the sedimentary pore space, associated with a migration of gas and water released after gas hydrate dissociation is presented in Figure 12.
5. Summary
Based on this investigation, it is possible to conclude that the formation of gas hydrates and authigenic carbonate, associated with the pore-pressure growth caused by gas hydrate dissociation, promotes an increase in the volume of pore space inducing the formation of mounds at the seafloor. The uplift of older sediments is followed by nondeposition or erosion of the Holocene sediments on the top of the mounds.
Acknowledgments
The authors are thankful to R. O. Kowsmann for comments. Thanks go to the editor and anonymous reviewers for their comments and suggestions.
Supplementary Materials
Suplementary data 1: Piston- and push-cores location, objective, water depth and recovery.