We report carbon (δ13C) and oxygen (δ18O) isotope records of two modern giant clam (Tridacna squamosa) shells from two sites (Ishigaki-jima and Okinoerabu-jima) at different latitudes in the Ryukyu Islands, Japan. The δ13C profiles of samples from the inner shell layer on cross-sections along the maximum growth axis display no ontogenetic trends or seasonal variations. This finding suggests that the calcification site is likely to be unaffected by CO2 uptake and release resulting from the metabolic activity of the molluscan host and algal symbionts. The δ18O profiles show distinct seasonal cycles. After accounting for the influence of seawater δ18O, the time-series variations are consistent with variations in sea surface temperature, and the temperature dependency of oxygen isotope fractionation is nearly identical to previously published δ18O–temperature relationships for biogenic and synthetic aragonite. We conclude that δ18O records from pristine fossils of this species will enable accurate paleoenvironmental reconstructions at high temporal resolution.
Introduction
The tridacnines (Subfamily Tridacninae Lamarck, 1819) are a variety of bivalve that form large shells up to 1 m long. These bivalves appeared in the Eocene (Harzhauser et al., 2008) and at the present day are an important component of Indo-Pacific coral reef communities (Rosewater, 1965). The process of light-enhanced calcification (Goreau, 1961; Simkiss, 1964), caused by the association of all tridacnines with unicellular algal symbionts (zooxanthellae), results in unusually high calcification rates (Rossbach et al., 2019). The dense aragonitic shells formed by tridacnines exhibit distinct annual and daily growth bands in their inner shell layer (Aharon and Chappel, 1986; Watanabe and Oba, 1999), which provide a robust chronological constraint for high-resolution reconstruction of paleoenvironmental changes (Sano et al., 2012).
The oxygen isotope composition (δ18O values) of biogenic carbonates has been used in numerous paleoclimate reconstruction studies, owing to its relationship with the temperature and δ18O values of the ambient seawater in which the carbonates were precipitated (Epstein et al., 1953; Grossman and Ku, 1986; Kim and O'Neil, 1997; Bemis et al., 1998). These biogenic carbonates include those formed by foraminifers (Emiliani, 1955; Shackleton, 1967; Lea et al., 2000; Lisiecki and Raymo, 2005), corals (Linsley et al., 2004; Asami et al., 2005; Calvo et al., 2007; Felis et al., 2009; Gagan et al., 2012), mollusks (Schöne et al., 2004a, b; Carre et al., 2005; Chauvaud et al., 2005; Gillikin et al., 2005), and brachiopods (Parkinson et al., 2005; Yamamoto et al., 2011; Cusack et al., 2012; Takayanagi et al., 2013, 2015).
The δ13C and δ18O values of coral skeletons differ from those of equilibrium aragonite due to the effects of kinetic and metabolic isotope fractionation (McConnaughey et al., 1997). Although the shell calcite of the secondary layer of brachiopods was believed to be precipitated in carbon and oxygen isotope equilibrium with ambient seawater, Auclair et al. (2003) showed that the δ13C and δ18O values of modern brachiopod shells do not necessarily fall within the range of the values for equilibrium calcite and that intra-shell variability can be significant. This disequilibrium calcite precipitation is attributed, at least in part, to kinetic and metabolic isotope fractionation effects (Yamamoto et al., 2010a, b, 2011, 2013; Takayanagi et al., 2012, 2013, 2015) or unidentified chemical conditions at the calcification sites (Takayanagi et al., 2013).
Figure 1.
Location and map of the study site. Bathymetric data are from the ETOPO1 Global Model ( https://www.ngdc.noaa.gov/mgg/global/ [Cited 21 February 2020]). White circles in the subaerial photographs of Ishigaki-jima and Okinoerabu-jima indicate the sampling sites.

The δ18O values of modern tridacnine shells have been found to be in isotopic equilibrium with ambient seawater (e.g. Aharon and Chappell, 1986). However, although there are currently ten described species within the genus Tridacna Bruguière, 1797 and two within the genus Hippopus Lamarck, 1799 (WoRMS Editorial Board, 2018), most previous studies have examined only three species: T. gigas (Aharon, 1991; Patzöld et al., 1991; Elliot et al., 2009; Welsh et al., 2011; Yan et al., 2013); T. maxima (Jones et al., 1986; Komagoe et al., 2018); and H. hippopus (Watanabe and Oba, 1999; Aubert et al., 2009). There have been few studies of other species (e.g. T. derasa, Yamanashi et al., 2016).
Tridacna squamosa Lamarck, 1819 is one of the tridacnines that has not been well investigated. This species is common in upper Cenozoic coral-reef deposits (e.g. MacNeil, 1960) and is found in many archaeological sites (e.g. Asami et al., 2015; Radclyffe, 2015). As such, its skeletal records could constitute valuable paleoclimatic and paleoceanographic archives. In this study, we analyze the carbon (δ13CS) and oxygen isotope (δ18OS) records of T. squamosa shells collected from two different latitudes in the Ryukyu Islands (Ishigaki-jima, 24°19′–37′N; Okinoerabu-jima, 27°19′–26′N) and assess their potential as paleoenvironmental indicators.
Materials and methods
Study site and climate regime
The Ryukyu Islands, which are composed of several tens of islands and islets, are located to the southwest of mainland Japan (Figure 1). These islands are arranged in a curved row, called the Ryukyu Island Arc. They are divided into three segments, North, Central, and South Ryukyus, by two tectonic boundaries, the Tokara Strait in the north and the Kerama Gap in the south. Most of the islands are rimmed by well-developed fringing coral reefs, extending at relatively high latitudes in the coral reef province in present-day oceans (Iryu et al., 2006). The northern limit of reef formation in the northwestern Pacific Ocean is located at Tane-ga-shima, North Ryukyus (Ikeda et al., 2006).
Climate data for Ishigaki-jima and Okinoerabu-jima were obtained from the Japan Meteorological Agency (2018). Ishigaki-jima (24°19′–37′N, 124°04′–21′E; Figure 1) is the second largest island in the South Ryukyus. The climate of the island is subtropical, with a monthly mean atmospheric temperature ranging from 18.6°C (January) to 29.5°C (July) and an annual mean value of 24.3°C (Table 1). Semidiurnal tides dominate throughout the islands, with a maximal range of 1.9 m at spring tide and 1.0 m at neap tide. The annual rainfall is approximately 2100 mm, with rainy months in May–June and August–October. The prevailing wind is SSE in summer and NNW in winter.
Okinoerabu-jima is located in the Central Ryukyus (27°19′–26′N, 128°31′–43′E; Figure 1). The climate on the island is subtropical (Table 1). The monthly mean atmospheric temperature ranges from 16.2°C (January) to 28.4°C (July and August), with an annual mean value of 22.4°C. Semidiurnal tides in the northern Central Ryukyus have a maximal range of 1.9 m at spring tide and 1.3 m at neap tide in Amami-o-shima. The annual rainfall reaches 1800 mm, with rainy months in May–June and October. The prevailing wind is SSE in summer and NNW in winter.
The Integrated Global Ocean Services System (IGOSS) data (1° resolution gridded data; Reynolds et al., 2002) indicate that the sea surface temperature (SST) around Ishigaki-jima from 2009 to 2013 varied between 23.1°C and 30.0°C, with an annual mean of 26.4°C. The highest mean monthly SST was 29.1°C in June and the lowest was 23.8°C in February. During the same period, SST around Okinoerabu-jima ranged from 21.3°C to 29.9°C, average 25.2°C. The highest mean monthly SST was 29.1°C in August and the lowest was 21.8°C in February. The sea surface salinity (SSS) at Ishigaki Port during 1998–2004 ranged from 33.6 to 35.0, average 34.3, with short-term decreases of > 2 caused by heavy rainfall (Abe et al., 2009). SSS off Okinoerabu-jima (27–28°N, 128–129°E) ranged from 34.1 to 35.0, average 34.7, during 1906–2003 (Japan Oceanographic Data Center, 2018).
Table 1.
Climatology at Ishigaki-jima and Okinoerabu-jima. 1, Data for 1981–2010 are from the Japan Meteorological Agency (2018); 2, Integrated Global Ocean Services System data (1° resolution gridded data, centered at 24°530′N and 124°30′E [Ishigaki-jima] and 27°30′N, 128°30′E [Okinoerabu-jima] for 2009–2013; Reynolds et al., 2002).

Tridacna squamosa
Tridacnines are one of the most conspicuous marine invertebrates in coral reefs due to their large size and brilliantly colored mantle that contains photosynthesizing symbionts. Shells of T. squamosa, one of the eight species within the genus Tridacna, were examined in this study. Five species (T. derasa, T. gigas, T. crocea, T. squamosa, and T. maxima) are abundant in the Indo-Pacific coral reef province, with the latter two extending their distribution into the Red Sea (Huelsken et al., 2013). The maximum shell length of T. squamosa is 40 cm. This species lives byssally anchored to hard substrates. From phylogenetic analyses, the divergence time between T. squamosa and T. crocea was estimated to be ∼5 Mya (Schneider and Foighil, 1999) or ∼9 Mya (Huelsken et al., 2013). Based on the occurrence of T. squamosa shells on Kita-daito-jima (Nomura and Zinbo, 1935), Schneider and Foighil (1999) stated that the oldest occurrence of this species was in the late Pliocene or early Pleistocene, which was supported by Harzhauser et al. (2008). Although Nomura and Zinbo (1935) did not provide information on the sampling locality within the island, the shells' mineralogical composition and the stratigraphic setting of the island indicate that the shells were collected from unit 2 of the Daito Formation (Nambu et al., 2003; Suzuki et al., 2006). Based on strontium isotope ages from the island's surface dolomites (4.9–2.1 Mya; Takayangi et al., 2010) and the reconstructed geologic history of Kita-daito-jima (Iryu et al., 2010), the oldest occurrence of this species dates back to at least the late Miocene. This age is in accordance with molecular phylogenetic data.
Figure 2.
Studied Tridacna squamosa shells. Each shell was cut vertically (red line) along its maximum growth axis. A, Shell (left valve) from Ishigaki-jima (TSI); B, Shell (right valve) from Okinoerabu-jima (TSO); C, Cross-section through the TSO shell; D, Cross-section through the TSI shell. The red line indicates the sampling transect. Scale bar = 10 cm.

Materials
A Tridacna squamosa shell, designated TSI (Figure 2), was collected on March 7, 2013 at 6 m water depth in Sekisei Lagoon (24°34′N, 124°03′E) near Ishigaki-jima (Figure 1). The shell height was 15 cm and the shell length was 25 cm. The other T. squamosa shell, designated TSO (Figure 2), was collected at 12 m water depth off Okidomari, western Okinoerabu-jima (27°24′N, 128°33′E) on September 20, 2013. The shell height was 17.5 cm and the length was 29.0 cm. The studied material is deposited at the Institute of Geology and Paleontology, Graduate School of Science, Tohoku University, Sendai.
After the soft tissue was removed, a ∼1 cm-thick slab was cut vertically from each shell along the maximum growth axis (cross-section; Figure 2). The inner and outer shell layers were clearly discernible on each slab. Patzöld et al. (1991) showed that the biogenic (daily growth banding) and geochemical (δ18O values) records in the inner shell layer are more suitable for paleoenvironmental reconstruction than those of the outer shell layers or hinge, and many paleoenvironmental studies have analyzed this region of the shell (Watanabe and Oba, 1999; Aubert et al., 2009; Elliot et al., 2009; Welsh et al., 2011; Yan et al., 2013; Asami et al., 2015; Yamanashi et al., 2016).
Carbonate samples for isotope analysis were manually retrieved along a roughly median line on the inner shell layer at ∼0.4 mm intervals using a 0.5 mm diameter drill bit (Figure 2). TSI yielded 51 carbonate samples, and 62 samples were obtained from TSO.
Isotope analysis
Stable carbon and oxygen isotope analyses of carbonate powder samples were performed using a Thermo Fisher Delta V isotope ratio mass spectrometer (Bremen, Germany) coupled to a ThermoQuest Kiel-III (Bremen, Germany) automated carbonate device at the Institute of Geology and Paleontology, Tohoku University, Japan. The samples (∼0.1 mg) were reacted with 100% phosphoric acid at ∼72°C. The isotope ratios are expressed in conventional (δ‰) notation and calibrated to the NBS-19 international standard relative to the Vienna Pee Dee Belemnite. A value of 1.01025 was used for the oxygen isotope fractionation factor during phosphoric acid digestion (Sharma and Clayton, 1965). The external precision (1σ) based on replicate measurements (n = 91) of the laboratory reference material (JCt-1; Okai et al., 2004) was 0.04‰ for carbon isotope analyses and 0.06‰ for oxygen isotope analyses.
Following Yamanashi et al. (2016), the distance-domain δ18OS profiles, which clearly showed seasonal cycles (Figure 3), were converted to time series for better interpretation and comparison with those of SST. We converted the distance-domain δ18OS profiles to time series via peak-to-peak matching (i.e. annual maximum and minimum values in a year) with the SST profiles using AnalySeries software (Paillard et al., 1996). Assuming constant growth between each δ18OS peak, monthly resolved δ18OS profiles were determined. The δ13CS time series were generated simultaneously.
Results
Values of δ13CS and δ18OS
The δ13CS profiles of TSO and TSI did not show ontogenetic trends, regular cycles, or short-term fluctuations (Figure 3). The δ13CS values for TSI ranged from 0.89‰to 1.50‰ (average = 1.19‰, σ [standard deviation] = 0.14‰) and for TSO were 0.79‰–1.41‰ (average = 1.12‰, σ = 0.14‰; Table 2).
The δ18OS profiles were characterized by a series of regular cycles of varying amplitudes and frequencies. The δ18O profile for TSI exhibited three distinct cycles, with four cycles evident in the TSO profile (Figure 3). The δ18OS values for TSI ranged from – 1.92‰ to – 0.07‰(average = – 1.03‰, σ = 0.52‰) and for TSO from – 1.55‰ to 0.52‰ (average = – 0.72‰, σ = 0.55‰) (Table 2).
Equilibrium δ13C values of aragonite
Following Yamanashi et al. (2016), we estimated the δ13C values for aragonite precipitated in isotopic equilibrium with ambient seawater (equilibrium aragonite; δ13CEA) using previously published δ13C values for dissolved inorganic carbon (δ13CDIC) of seawater samples collected near Okinawa-jima (Takayanagi et al., 2013). As the pH of the surface seawater at Ishigaki-jima ranged from 7.9 to 8.0, the δ13CHCO3 – values for this seawater were assumed to be ∼0.2‰ greater than the δ13CDIC val- ues (Grossman, 1984; Romanek et al., 1992; Zhang et al., 1995). The δ13CEA values calculated using the δ13CDIC values of 1.1‰–1.6‰ and the aragonite-HCO3 – enrichment factor of 2.7‰ ± 0.6‰ (Romanek et al., 1992) range from 3.4‰ to 5.1‰ at the TSI and TSO growth sites.
Temperature dependency of δ18OS values
The two δ18OS profiles show a series of regular cycles of varying amplitudes and frequencies (Figure 3). As discussed by Yamanashi et al. (2016), these cycles correspond predominantly to the seasonal variation in seawater temperature. Similar distinct seasonal δ18O cycles have been detected in the inner shell layers of other tridacnines (Jones et al., 1986; Romanek et al., 1987; Romanek and Grossman, 1989; Patzöld et al., 1991; Elliot et al., 2009; Yan et al., 2013; Yamanashi et al., 2016).
The oxygen isotope composition of any carbonate is controlled by the temperature and oxygen isotope composition of the ambient water (δ18OW) in which it is formed. As such, adjusting the values of δ18OS to account for the contribution of δ18OW will enable a more accu- rate temperature calculation. The time-series variation for (δ18OS – δ18OW) of TSI and TSO during the period from 2009 to 2013 are shown in Figure 4A and B. The IGOSS SST recorded at sites at 24.5°N, 124.5°E and 27.5°N, 128.5°E were used for temperatures at the growth sites of TSI and TSO, respectively. To obtain the temperature dependency (δ18OS – δ18OW) for TSI, the monthly average δ18OW values at Ishigaki Port from December 1997 to May 2004, based on the study by Abe et al. (2009), were used. The monthly average δ18OW values for the TSO growth site were calculated based on salinity data at a water depth of 10 m off Okinoerabu-jima (27°–28°N, 128°–129°E), provided by Japan Oceanographic Data Center (2018), and the salinity–δ18OW relation proposed by Takayanagi et al. (2013). The δ18Os values for both TSI and TSO demonstrated distinct seasonal cycles and significant negative relationships with SST (Figure 4C and D). Because there was no significant difference between the TSI and TSO regressions, we combined the results for use in the discussion.
Figure 4.
Time-series variation and temperature dependency of δ18OS for shells TSI and TSO; δ18Os and δ18OW represent the δ18O values of T. squamosa shells and seawater, respectively. Left panels: Time-series variation of (δ18OS – δ18OW) and SST at Ishigaki-jima (A) and Okinoerabu-jima (B). IGOSS SST with 1 × 1 resolution gridded data are used, centered at 24.5°N and 124.5°E for Ishigaki-jima and centered at 27.5°N and 128.5°E for Okinoerabu-jima. Right panels: Relationship between (δ18OS – δ18OW) and SST for TSI (C) and TSO (D).

Discussion
Values of δ13CS
The δ13CS values show no seasonal cycles (Figure 3), as is common for most modern tridacnines: T. gigas (Aharon, 1991; Elliot et al., 2009); T. maxima (Jones et al., 1986) T. squamosa (Betenburg et al., 2011); and T. derasa (Yamanashi et al., 2016). The δ13C profiles of some tridacnines, including the studied shells, show irregular fluctuations with large amplitudes, contrasting well with the narrow range of variation of other tridacnines (Elliot et al., 2009; Yamanashi et al., 2016). In contrast, the δ13C of T. gigas shells from Kume-jima (Ryukyu Islands) have been shown to display distinct seasonal cycles (Watanabe et al., 2004).
Figure 5.
Cross-plots of δ13CS versus δ18OS for TSI and TSO. The δ13CS and δ18OS values show no significant correlations.

No statistically significant ontogenetic trend was recognized in the δ13CS profiles of the shells in this study, or in those of other tridacnines (Aharon, 1991; Watanabe et al., 2004; Elliot et al., 2009; Betenburg et al., 2011; Asami et al., 2015). However, it is known that some mollusks show decreases (Kennedy et al., 2001; Keller et al., 2002; Lorrain et al., 2004) or increases (Jones et al., 1986) in δ13C values during growth. The relatively constant δ13CS values, characterized by the absence of ontogenetic trends or seasonal cycles, suggest that the calcification site of T. squamosa is not affected by changes in CO2 uptake and release resulting from the metabolic activity (photosynthesis and respiration) of the molluscan host and algal symbionts. This is also common in other tridacnine species (e.g. Elliot et al., 2009; Yamanashi et al., 2016).
The δ13CS and δ18OS values show no statistically significant correlation (Figure 5), indicating no or a very weak kinetic effect (McConnaughey et al., 1997) on isotope fractionation during the secretion of T. squamosa shells. This is also common in other tridacnine species (Jones et al., 1986; Romanek and Grossman, 1989; Yamanashi et al., 2016).
The δ13CS values are ∼1.9‰–2.6‰ lower than the lowest δ13CEA values (3.4‰). The reason for the δ13CS values being so much lower than the δ13CEA values is not known. This discrepancy is also seen in the shell δ13C values of T. derasa and T. gigas. The offsets between shell δ13C and δ13CEA are not consistent among various tridacnine species (Yamanashi et al., 2016; this study), and we conclude that tridacnine shell δ13C profiles are of limited use for paleoenvironmental analysis.
Equilibrium oxygen isotope fractionation between aragonite and water
The aragonitic shell of tridacnines has been reported to be precipitated in oxygen isotope equilibrium with ambient seawater (e.g. Aharon, 1991; Watanabe and Oba, 1999). However, in their study of T. derasa, Yamanashi et al. (2016) found notable Δδ18O values (i.e. shell δ18O values minus equilibrium δ18O values), indicating disequilibrium precipitation.
The temperature dependency of the oxygen isotope fractionation factor between T. squamosa shell aragonite and seawater, as calculated from the δ18OS and δ18OW data, is compared with previously published δ18OS–temperature relationships for biogenic and synthetic aragonites (Figure 6). Patterson et al. (1993) investigated the δ18O values of various freshwater fish otoliths collected from six temperate lakes as well as temperature-controlled aquariums and obtained a temperature-fractionation factor (103lnα) for the temperature range 3.2°C to 30.3°C. Thorrold et al. (1997) investigated the δ18O values of otoliths of Atlantic croaker (Micropogonias undulatus) grown in temperature-controlled aquariums for a temperature range of 18.2°C to 25.0°C. White et al. (1999) determined the temperature–103lnα relationship for the shell of the land snail Lymnaea peregra grown in a controlled terrarium for the temperature range of 8–24°C. Watanabe and Oba (1999) analyzed the variation in δ18O values for a 1-year period in the shell of a giant clam (H. hippopus) shell collected at Ishigaki-jima, and determined the SST vs (shell δ18O – δ18Ow) relationship for a temperature range of 20.7–32.3°C. As Watanabe and Oba (1999) did not pro- vide the seasonal variation of δ18Ow values, we used a constant value of 0.203 to convert to 103lnα, calculated from the median salinity value of 34.3 and the regional SSS–δ18Ow relation they used (δ18Ow = 0.203SSS – 6.76; Oba, 1988). Böhm et al. (2000) measured δ18O values for tropical sclerosponges collected from Jamaica and the Great Barrier Reef, Australia, and calculated the temperature–103lnα relationship for biogenic aragonite between 3°C and 28°C, using data from their own results and studies by Tarutani et al. (1969), Grossman and Ku (1986), and Rahimpour-Bonab et al. (1997). Aubert et al. (2009) measured the variation in δ18O values during a 4-year period for an H. hippopus shell and determined the SST vs. (δ18Oshell – δ18OW) relationship for the temperature range 20.5–29.4°C. The specimen was collected from Ducos Island, New Caledonia and then grown in a non-temperature-controlled aquarium tank at Noumea. Similarly to Watanabe and Oba (1999), we used a constant value of 0.466 to convert to 103lnα, which was calculated from the average salinity value of 35.3 and the regional SSS–δ18Ow relationship (δ18Ow = 0.255SSS – 8.535). For abiogenic aragonites, Kim et al. (2007b) studied the temperature–103lnα relationships for synthetic aragonites over a temperature range of 0–40°C. In their study, they used a unique factor for oxygen isotope fractionation during phosphoric acid digestion of aragonites of 1.01063 (Kim et al., 2007a); therefore, we modified their equation by applying the conventional factor of 1.01025 instead of 1.01063 (Table 3).
Figure 6.
Temperature dependencies of oxygen isotope fractionation between aragonite and water. For the analyzed materials, see Table 3. *The original results for synthetic aragonites (Kim et al., 2007b) used an isotope fractionation factor of 1.01063 for phosphoric acid digestion, different from the value applied in other studies; standardization with the conventional value of 1.01025 was applied.

All equations, except for those of Watanabe and Oba (1999) and Aubert et al. (2009), exhibit similar regression slope values (Table 3; Figure 6). The average value for the slopes of these equations is 18.02, with 1σ = 0.70. The average is not significantly different from our result (slope = 17.90). This small variation in the slopes would induce a maximum difference of 0.21‰ within the temperature range 20–30°C. The difference would reduce to 0.1‰ if the calibration based on a relatively low temperature limit (up to 24°C) by White et al. (1999) was excluded. The value of 0.1‰ reflects an uncertainty in reconstructed temperature as small as ∼0.5°C. We thus conclude that the temperature sensitivity of the oxygen isotope composition within aragonite is robust between 20°C and 30°C. Our results extend this range up to 30°C for marine biogenic aragonites.
Given the slight inconsistencies between the various δ18O–temperature relationships, Kitamura (2018) argued that the reliability of tridacnine and other marine calcifier-based paleoenvironmental analysis is limited. However, it is notable that the two exceptional slopes depicted in Figure 6 were both derived for the tropical giant clam H. hippopus. The slope obtained from our results is consistent with other equations, although our study was based on the shells of another tridacnine genus, Tridacna. Further rigorous evaluation of the temperature–103lnα relationship within tridacnine shells is required. Long-term field observations of δ18OW and corresponding analyses of long time-series variation in shell δ18O, as well as developing an improved understanding of the physiological mechanisms of calcification for additional tridacnine taxa, would be particularly useful.
Table 3.
Comparison of published calibrations of the oxygen isotope fractionation between aragonite and water.

The intercepts of each calibration exhibit greater variation than the slopes. The average intercept value at 25°C is 29.42, with 1σ = 0.36. This includes the two above-mentioned H. hippopus results and extrapolated values for the equation of White et al. (1999). Our result (29.58) is consistent with the average value at 25°C. Kim et al. (2007b) noted that their calibration for inorganic aragonite was statistically indistinguishable from the equations of Patterson et al. (1993), Thorrold et al. (1997), White et al. (1999), and Böhm et al. (2000), by considering their standard error of 0.46, although other biological calibrations, except for Patterson et al. (1993), show systematically higher values than the abiological calibration of Kim et al. (2007b). In this context, we conclude that the aragonitic shell of Tridacna squamosa is formed in oxygen isotope equilibrium with ambient water.
The reason for the relatively wide uncertainty of intercepts is unclear. It is unlikely that differences between marine and freshwater precipitation regimes play a role, as the marine fish otolith and the land snail equations plot close to each other (Figure 6). This is consistent with the conclusion of Thorrold et al. (1997), who noted that the bias of their result with respect to those of Patterson et al. (1993) was not caused by the difference between marine and freshwater taxa. If the Patterson et al. (1993) equation is excluded, the average and 1σ values at 25°C would reduce to 29.51 and 0.26, respectively, bringing the average value much closer to our result.
Tridacnine shell aragonite is highly susceptible to meteoric diagenesis; as such, in most cases, fossil shells have been altered and do not retain their initial isotopic and chemical composition. Fossil tridacnine shells rarely occur in Pleistocene coral reef deposits in the Ryukyu Islands (Ryukyu Group; Iryu et al., 2006). To date, no aragonitic shell has yet been collected, even from younger members of the group (e.g. unit 13 on Irabu-jima, Sagawa et al., 2001). The use of tridacnine shells will be limited mostly to Holocene specimens, especially from archeological sites (e.g. Asami et al., 2015; Radclyffe, 2015). Compiling tridacnine-based climate records from multiple archeological sites could potentially enable high-resolution reconstructions of Holocene climate variability in the tropics.
Conclusions
We investigated the δ13CS and δ18OS records of two modern T. squamosa shells collected from Ishigaki-jima and Okinoerabu-jima, southwestern Japan. These records were obtained from cross-sections of the inner shell layer. The δ13CS profiles display no ontogenetic trends or seasonal variations. This finding suggests that the calcification site is not likely to be affected by CO2 uptake and release resulting from metabolic activities (photosynthesis and/or respiration) of the molluscan host and algal symbionts.
The δ18O profiles show distinct seasonal cycles. After removal of the δ18OW component, time-series variations are consistent with those of SST. Although there are differences in the published temperature–103lnα relationships for other biogenic and inorganic carbonates and further field observations and/or physiological investigations may be needed, Tridacna squamosa shell precipitates in oxygen isotope equilibrium with ambient seawater. The δ18O records of pristine T. squamosa specimens will be useful for accurate paleoenvironmental reconstructions at high temporal resolution.
Acknowledgments
We are grateful to K. Takamiyagi and S. Higashi (Seadream Okierabu) for collecting the T. squamosa shells. The manuscript was significantly improved by comments and suggestions from A. Kitamura and an anonymous reviewer.
References
Appendices
Author contributions
We declare that none of the material in this manuscript has been published or is under consideration for publication elsewhere. SK and YI conceptualized and designed this study. RA, OA, and YI conducted the fieldwork and sampling. SK, HT, and TTNH carried out the chemical and isotope analysis. SK, HT, and KY collaborated with the corresponding author in the interpretation of the data and the construction of the manuscript. All authors read and approved the final manuscript.