Evidence of methane hydrate instability has been reported from marine and lake sediments of the Quaternary [e.g., Wefer et al., 1994], the Late Paleocene [e.g., Dickens et al., 1995, 1997], Early Cretaceous [Jahren et al., 2001], and the Jurassic [Hasselbo et al., 2000]. Their evidence for Late Quaternary hydrate instability is reported in marginal sea sediments such as the California Margin [Hinrichs et al., 2003; Kennett et al., 2000], Gulf of California [Keigwin, 2002], Papua New Guinea Margin [de Garidel-Thoron et al., 2004], Amazon Fan [Maslin et al., 1998], and Greenland Sea Margin [Millo et al., 2005a; Smith et al., 2001], and in lake sediments in Lake Baikal [Prokopenko and Williams, 2004, 2005]. Kennett et al.  suggested, on the basis of foraminiferal δ13C in the Santa Barbara basin (SBB), the “clathrate gun” hypothesis, which highlights the potential role of methane hydrate in Quaternary variations in atmospheric methane concentrations. Moreover, biomarker records supporting methane hydrate instability were documented in the SBB [Hinrichs et al., 2003] and off the Shimokita Peninsula, Honshu island, Japan [Uchida et al., 2004]. These biomarker records provided aerobic methanotrophic molecular fossil evidence for periodic methane release, which may also support the clathrate gun hypothesis. More recently climate-sensitive hydrocarbon seepage of oil and gas, including large amounts of methane, has also been reported from the SBB during the last deglaciation [Hill et al., 2006]. The extent of support that foraminiferal δ13C records from sediments lend to the clathrate gun hypothesis are mainly dependent, however, on whether the planktonic δ13C excursions represent primary seawater δ13C values or secondary, postdepositional diagenetic δ13C values. Regarding interpretation of these foraminiferal δ13C records, some studies seem to be questionable, especially about the extent of methane oxidation that occurs in the water column. Recent studies have proposed that contamination by authigenic overgrowths or calcite replacement may be difficult to detect through microscopy [Hill et al., 2004a, 2004b]. Acid incremental leaching experiments for 13C-depleted foraminifera have been examined in detail to quantify the extent of authigenic overgrowth [Millo et al., 2005b; Torres et al., 2005]. Additionally, organic carbon in situ could contaminate the foraminiferal isotopic records, leading to a more depleted 13C composition, because total organic carbon (TOC) δ13C values (∼−23‰ to −22‰) are much lower than foraminiferal δ13C values (0‰ to −1‰). It is a matter of ongoing debate whether such extremely negative foraminiferal δ13C values are derived from dissolved inorganic carbon (DIC) at primary calcification or from subsequent authigenesis [Cannariato and Stott, 2004]. To address this controversy, several researchers have conducted extensive studies that have reached conflicting and complex conclusions. Thus we need to know the correct mechanism explaining the extremely 13C-depleted foraminifera found in glacial/interglacial sediments.
Ten or more regions from the northwest Pacific along the Japanese islands (including the Oyashio current region, the Sea of Okhotsk, the Nankai Trough, and the Japan Sea) have been identified as immense methane hydrate reservoirs [Center for Deep Earth Exploration (CDEX), 2002; Kvenvolden and Lorenson, 2001; Satoh, 1994, 2002] (Figure 1). Despite the potential for large episodic releases of methane in the western North Pacific, the relationship between climate and hydrate instability in the region has not been studied sufficiently. In a previous study that used the same core as was used in this study (core PC6, off Tokachi, Hokkaido, northeast Japan), many extremely 13C-depleted foraminiferal signals were found from sediment layers ranging in age from 21,000 to 17,800 cal years B.P. [Ohkushi et al., 2005]. Recent observations have shown the existence of a bottom-simulating reflector (BSR) over a large area that includes this study site [Satoh, 2002; TuZino and Noda, 2007].
Figure 1. (a) Location map of core sites PC6 (this study site), PC4/5, and CK. The red dot indicates the locations of the epicenters of the Tokachi-oki interplate earthquakes of 1952 and 2003 (41.780°N, 144.079°E; depth, 42 km; moment magnitude 8.0; 26 September 2003). (b) Detailed location map of core site PC6. The black dots indicate the locations of the epicenters of the Tokachi-oki interplate earthquakes of 1952 and 2003 (41.780°N, 144.079°E; depth, 42 km; moment magnitude 8.0; 26 September 2003). The yellow shaded areas show the locations of anomalous BSRs, which suggest the presence of methane gas hydrates [TuZino and Noda, 2007]. Each location has been described in detail previously by Kvenvolden and Lorenson , Satoh [1994, 2002], and Satoh et al. . The 13C-depleted planktonic and benthic foraminifera for several horizons in the PC4/5 cores were reported by Uchida et al.  and Hoshiba et al. . CK C9001 sites (41.10°N, 142.12°E; depth, 1180 m) represent the location at which methane hydrate was recently collected from the core depth of 189 m under the seafloor at a water depth of 1180 m during expeditions by the CK06-06 leg1 cruise by R/V Chikyu (information available at http://www.jamstec.go.jp/jamstec-e/PR/0608/0824_2/) [Taira and Curewitz, 2005].
Download figure to PowerPoint
In this study, we performed high-resolution natural radiocarbon measurements of planktonic and benthic foraminifera and organic carbon to decipher the carbon source of foraminiferal carbon isotope anomalies. 14C measurements of fossil foraminiferal carbonates were an essential tool for correctly determining the ages of the sediment horizons, because its 5730-year half-life of 14C makes it a convenient tracer of processes that occur on a millennial timescale (less than about 50,000 years). Using obtained natural radiocarbon data of planktonic and benthic foraminifera, we quantitatively evaluated the carbon source of foraminiferal isotopic anomalies by a coupled isotopic mass balance model (14C/C and 13C/12C).
Foraminifera use the DIC in the surrounding seawater when making their carbonate skeletons; thus the 14C content of the foraminiferal carbonate should reflect the 14C content of the DIC of the seawater when the foraminifera were alive. Methane hydrates buried under the deep seafloor have been known as fossil carbon (14C-free) because the methane hydrate was biogenically formed using ancient organic carbon or originated through thermodecomposition of fossil organic carbon [Grabowski et al., 2004; Kessler et al., 2005, 2006; Winckler et al., 2002]. Thus, if methane hydrate dissociation occurred as a result of any environmental changes and was accompanied by the oxidation of methane, the 14C content of seawater would be highly contaminated with the 14C-free DIC derived from methane. In this study, we took advantage of the complete absence of 14C in methane hydrates to quantify the carbon source derived from methane hydrate dissociation recorded in the foraminiferal δ13C anomalies. We also used analyses of trace metals, such as the Mg/Ca ratio, as an index of authigenic overgrowth and a phylogenetic type analysis to evaluate the past activity of methanotrophs and methanogens associated with past drastic environmental changes.