Modern samples

Sand samples for luminescence measurements were collected from the modern and Holocene beach and shoreface of the Kujukuri coast (Fig. 1B; Tables 1 and 2). A modern foreshore sample (gsj15160) was taken from the 5-cm-thick surface layer at Katakai, in the central part of the coast, using a light-tight plastic tube (Table 1). Four modern shoreface samples (gsj17282, 17283, 17284, and 17273) were scooped by the Smith–McIntyre sediment grab during a research cruise in August and September 2014 (Nishida et al., 2019). Plastic tubes >5 cm thick were hammered into the bulk sediment in the grab to collect light-unexposed sediments, which were transported to the laboratory.

Holocene samples

Two sediment cores (GS-QAS-2 and GS-QYH-1) were drilled in the northeastern part of the Kujukuri beach-ridge plain in 2015 (Table 2; Komatsubara, 2019). Cores GS-QAS-2 and GS-QYH-1 contained subsurface sediment successions representative of the landward and seaward parts of the prograded beach-ridge plains, respectively; both cores include the lower unit, consisting of the transgressive estuary and lagoon deposits in an incised valley, and the upper unit, containing regressive beach–shoreface deposits.

The upper unit of core GS-QAS-2 overlies and exhibits a gradual boundary with the incised-valley deposits at approximately 22.5 m depth and contains a succession of lower shoreface, upper shoreface, and foreshore and backshore facies, in ascending order. The lower shoreface facies occurs in the depth interval 8.2–22.5 m and consists of horizontally to low-angle cross-laminated very fine to fine sand. The upper shoreface facies is high-angle cross-laminated fine to very coarse sand in the depth range of 3.3–8.2 m. The foreshore and backshore facies is observed at 1.7–3.3 m depth and is characterized by horizontally laminated fine to medium sand with concentrations of heavy minerals. The upper unit of core GS-QYH-1, similarly to GS-QAS-2, consists of the lower shoreface (8.4–26.1 m depth), upper shoreface (3.0–8.4 m depth), and foreshore and backshore facies (0.5–3.0 m depth).

Six sand samples were taken from the upper units of each of the cores GS-QAS-2 and GS-QYH-1 (Table 2), yielding a total of 12 samples for OSL dating. Shell-rich layers were avoided in sampling to minimize the potential inaccuracy in dose-rate determination caused by the heterogeneous distribution of radionuclides in the sediment (e.g., Cunningham et al., 2019).

Radiocarbon dating

Accelerator mass spectrometry radiocarbon dating of shells and plant fragments from the sediment cores was performed by the Institute of Accelerator Analysis Ltd. (Kawasaki City, Japan) to establish an independent chronology for comparison with the OSL dating results (Table 3). Six intact shells and one plant fragment sample were dated for core GS-QAS-2, and six intact shells were analyzed for core GS-QYH-1. The ages obtained were converted into calendar ages by the program CALIB rev. 7.1.0 (Stuiver and Reimer, 1993) using the data set MARINE13 with a delta R of 0 years for the shell samples and INTCAL13 for the plant fragment sample (Reimer et al., 2013).

The radiocarbon ages are consistent with the stratigraphy in both cores (Figs. 2 and 3). In the lower unit of core GS-QAS-2, four samples were dated between 10,878–10,933 cal BP and 8,224–8,388 cal BP, supporting the interpretation of transgressive deposition. Likewise, two radiocarbon ages from the lower unit of GS-QYH-1 indicate the occurrence of transgressive deposition. Three ages from the upper unit of GS-QAS-2 range from 7,411–7,540 to 6,019–6,239 cal BP, corresponding to the initial regressive phase. The upper unit of GS-QYH-1 was dated as Late Holocene. Cores GS-QAS-2 and GS-QYH-1 thus represent the middle and late Holocene regressive succession of a prograding beach–shoreface deposit (Fig. 2).

Fig. 3

Fading-corrected final luminescence age estimates of samples taken from (A) core GS-QAS-2 and (B) core GS-QYH-1. Radiocarbon ages of molluscan shells are also shown for comparison. Radiocarbon ages at the ground surface were assumed to be 5.7 ka and 0 ka for cores GS-QAS-2 and GS-QYH-1, respectively, to define the expected ages (red lines).

Radiocarbon ages were used to define the expected ages of the Holocene samples taken from the cores (Table 2). The expected age of an OSL sample was obtained by linear interpolation of the intercepts of 2σ radiocarbon ages just above and below the sample and was associated with the mean uncertainties of the 2σ radio-carbon ages. For interpolation above the uppermost radiocarbon age, the dates of the tops of cores GS-QAS-2 and QYH-1 were assumed to be 5.7 ka and 0 ka, respectively.

Luminescence dating

Sample preparation and luminescence measurements were performed at the luminescence laboratory of the Geological Survey of Japan. Samples taken from drill cores were split into two, with one part used for luminescence measurements and the other for dosimetry; the modern samples were exclusively used for luminescence measurements. Samples for luminescence measurements were processed under subdued red light to avoid depletion of the luminescence signal. They were dried, sieved to extract grains 180–250 μm in diameter, and then treated with hydrochloric acid and hydrogen peroxide to remove carbonate and organic matter. Quartz and feldspar grains were then separated using sodium polytungstate liquids of densities 2.70, 2.58, and 2.53 g/cm³. The 2.53–2.58 g/cm³ fraction was taken as K-feldspar. Quartz samples were purified by etching the 2.58–2.70 g/cm³ fraction in 40% hydrofluoric acid for 1 h, followed by hydrochloric acid treatment for 1 h. Grains were mounted on stainless steel disks to form large (6 mm diameter) aliquots for luminescence measurement. Measurements were performed with a TL-DA-20 automated Risø TL/OSL reader equipped with blue and infrared LEDs centred on wavelengths of 470 and 970 nm, respectively, for stimulation and a ⁹⁰Sr/⁹⁰Y beta source for laboratory irradiation. For stimulation by blue and infrared LEDs, emitted luminescence through a Hoya U-340 filter and a combination of Schott BG3 (3 mm thick), BG39 (2 mm), and GG400 (3 mm) filters was measured with a photo-multiplier tube. Preliminary measurements showed that natural quartz OSL signals are dominated by the medium component and not appropriate for dating, as is also the case for other quartz sands in Japan (e.g., Tsukamoto et al., 2003; Tokuyasu et al., 2010; Tamura et al., 2015, 2017; Riedsel et al., 2018); therefore, only the IRSL and pIRIR signals of the K-feldspar grains were further investigated.

The modified single-aliquot regenerative dose protocols of pIRIR measured at 150°C and 290°C after a prior IRSL at 50°C (referred to here as pIRIR₁₅₀ and pIRIR₂₉₀, respectively; Table 4; Thomsen et al., 2008; Buylaert et al., 2009, 2012; Reimann and Tsukamoto, 2012) were applied to the K-feldspar grains. The prior IRSL signal in the pIRIR₁₅₀ protocol was also investigated and is referred to here as IR₅₀. In the pIRIR₁₅₀ protocol, the IR₅₀ and pIRIR₁₅₀ signals were sampled every 0.1 s for 100 s, and the net signal was derived from the integral of the first 2.0 s of the signal after subtracting the background estimated from the last 20 s of the signal. The pIRIR₂₉₀ signal was sampled every 0.1 s for 200 s with signal and subtraction integration intervals of the first 2.0 s and last 20 s, respectively. Preheat temperatures were fixed at 180°C for 60 s for pIRIR₁₅₀ and 320°C for 60 s for pIRIR₂₉₀. Bleaching experiments and dose recovery tests were practised for sample gsj15151. Six aliquots for each protocol were exposed to artificial sunlight for 4 h in a UVACUBE 400 chamber (Hönle) with an SOL 500 lamp module. After bleaching, three aliquots were dosed by 24 Gy for pIRIR₁₅₀ and by 72 Gy for pIRIR₂₉₀, and then used for a dose recovery test; the remaining aliquots were used to estimate the residual dose. Dose recovery was assessed after subtracting the residual dose from the recovered dose. IR₅₀, pIRIR₁₅₀, and pIRIR₂₉₀ yielded recovery ratios of 0.97±0.01, 0.96±0.02, and 1.17±0.08, respectively. The residual doses observed in the bleaching test were 0.31±0.09, 1.51±0.17, and 13.66±1.10 Gy for IR₅₀, pIRIR₁₅₀, and pIRIR₂₉₀, respectively. The mean value of the equivalent dose was determined from measurements of six replicates per sample by applying the Central Age Model of Galbraith et al. (1999). For all samples except gsj15151, 15155, 15157, and 15158, fading tests were also performed on aliquots after equivalent dose measurement to determine fading rates (expressed as the g_2days-value) following the method of Auclair et al. (2003).

For drill core samples, the environmental dose rate was determined using the DRAC program of Durcan et al. (2015) based on the contributions of both natural radioisotopes and cosmic radiation (Table 2). The concentrations of K and U were measured by inductively coupled plasma optical emission spectrometry and those of Th and Rb by inductively coupled plasma mass spectrometry; both sets of results were converted to dose rates by applying to the conversion factors of Adamiec and Aitken (1998). The attenuation factors used for beta and alpha rays were based on Mejdahl (1979) and Bell (1980), respectively. We used an a-value of 0.15±0.05 (Balescu and Lamothe, 1994). The a-values for IRSL and pIRIR signals are generally considered variable (e.g., Kreutzer et al., 2014); however, potential uncertainties in total dose rate related to the a-value are not significant (less than a few per cent), and thus we assume an identical a-value for IRSL and pIRIR. The internal K content of K-feldspar was assumed to be 12.5±0.5% (Huntley and Baril, 1997). The water content was also considered for the dose rate determination based on the measured value with an error of 5%. The cosmic dose rate was estimated based on Prescott and Hutton (1994) assuming overburden with a density of 1.8 g/cm³. The mean equivalent dose values were divided by the environmental dose rate to obtain fading-uncorrected ages. Mean g-values determined from six aliquots per sample and their standard errors for each signal were used for fading corrections of uncorrected ages based on Huntley and Lamothe (2001) and using the R Luminescence Package (Kreutzer et al., 2012; Fuchs et al., 2015). For correction of the four samples on which no fading test was performed, the average of the rest of the samples was applied. All dates are expressed relative to AD 2015.

Modern samples

In all modern samples, a bright natural pIRIR₂₉₀ signal was observed, whereas the IR₅₀ and pIRIR₁₅₀ signals were generally weak compared to the pIRIR₂₉₀ signal (Fig. 4A). Nevertheless, a dose–response curve was well defined for each signal, from which the equivalent dose was estimated. Equivalent doses of IR₅₀ were <0.1 Gy except for sample gsj17284, taken at a water depth of 23 m, which retains a residual dose of 0.18±0.01 Gy. The pIRIR₁₅₀ signal yielded higher residual doses than the IR₅₀, ranging from 0.93 to 3.25 Gy. The residual doses of pIRIR₂₉₀ were much higher than pIRIR₁₅₀: they were 11.93±1.66 Gy for the foreshore sample (gsj15160) at Katakai and 23–30 Gy for other subtidal samples.

Fig. 4

Decay curves and dose–response curves (inset) of IR₅₀, pIRIR₁₅₀, and pIRIR₂₉₀ measured on samples (A) gsj15259 and (B) gsj17282. The intersections of the natural OSL intensity with the dose–response curve and the corresponding equivalent dose are highlighted with red dashed lines.

Equivalent dose and burial age estimate of Holocene samples

Bright natural IR₅₀ and pIRIR₁₅₀ signals, as well as bright pIRIR₂₉₀ signal, were observed for Holocene samples (Fig. 4B). Equivalent doses were estimated from well-defined dose–response curves. Equivalent doses of IR₅₀ decrease upwards in both cores, ranging from 5.25 to 8.10 Gy in core GS-QAS-2 and from 0.28 to 2.49 Gy in GS-QYH-1 (Table 5). Equivalent doses of pIRIR₁₅₀ also decrease upwards, and are higher than the equivalent doses of IR₅₀ in all samples. In contrast, the pIRIR₂₉₀ signal yields much higher equivalent doses than pIRIR₁₅₀, and does not show systematic trends in either core. Average g-values of IR₅₀ were high (7.96±0.15%/decade), whereas those of pIRIR₁₅₀ and pIRIR₂₉₀ were as low as 1.21±0.11%/decade and 1.28±0.09%/decade, respectively (Fig. 5).

Fig. 5

Fading test results of IR₅₀, pIRIR₁₅₀, and pIRIR₂₉₀ signals measured for sample gsj15258. A regression line was defined for each signal based on the least-squares method and then applied to estimate the g-value.

In both cores, fading-corrected burial age estimates from IR₅₀ and pIRIR₁₅₀ display consistent trends with the stratigraphy, whereas those from pIRIR₂₉₀ exhibit remarkable reversals (Tables 5–7, Fig. 3). Estimated dose rates were consistent, characterized by a small range of 2.07–2.30 Gy/ka. Fading-corrected ages of IR₅₀ show vertical sequences from 7.7 to 5.8 ka in core GS-QAS and from 2.2 to 0.24 ka in core QYH-1, consistent with the stratigraphy (Fig. 3). The pIRIR₁₅₀ results also exhibited stratigraphically consistent trends of fading-corrected ages, but their ages are generally older than those of IR₅₀ by up to 1,200 years in core GS-QYH-1. In core GSQAS-2, the pIRIR₁₅₀ ages are, although overlapping within error, generally younger than the IR₅₀ ages. As expected from the equivalent dose, the fading-corrected ages of pIRIR₂₉₀ were inconsistent with the stratigraphy and older than 9 ka. However, it is noteworthy that the corrected pIRIR₂₉₀ ages in core GS-QAS-2 are generally older than those in core GS-QYH1 by several thousand years.

Residual dose characterization

As the five modern samples are supposed to have burial ages of 0 years, their equivalent doses are equal to the residual dose. The residual doses of IR₅₀ observed in the modern samples are less than 0.2 Gy, equivalent to only several decades for the given dose rate (Table 2). In contrast, pIRIR_50/150 retains residual doses of 1–3 Gy, equivalent to 400–1,300 years.

Characterization of residual doses for Holocene samples requires comparison with the expected ages obtained from radiocarbon dating. In core GS-QYH-2, fading-corrected IR₅₀ ages are consistent with radiocarbon ages. The radiocarbon ages of marine shells are likely older than the depositional ages as shells are not always buried immediately after their death; however, previous attempts confirmed that selective dating of intact shells results in consistent and precise chronology (Tamura et al., 2003, 2007) and thus minimizes the potential age overestimation. In core GS-QAS-2, which comprises early to middle Holocene deposits, the IR₅₀ ages appear to be generally older than the expected ages derived from radiocarbon dating. The g-values determined here are higher than those of the majority of K-feldspar samples from other regions (Li et al., 2014), which may have resulted in a slight inaccuracy of the ages in core GS-QAS-2. The opposite trend is identified for pIRIR₁₅₀ ages: the ages are older than the expected ages in core GS-QYH-1 and consistent with the expected ages in core GS-QAS-2. This finding suggests two possibilities: that either slight under-correction of the pIRIR₁₅₀ ages was caused by an underestimation of the g-value, which cancels residual doses, or the residual doses were negligible compared with the accuracy of ages determined for the early to middle Holocene deposits. Despite these slight discrepancies, these results show that the fading-corrected IR₅₀ and pIRIR₁₅₀ ages in the present study provide a consistent chronology that is at least as accurate as radiocarbon dating. This finding is regionally important because it demonstrates that for dating young sediments, the IR₅₀ and pIRIR₁₅₀ of K-feldspar can be an effective alternative for quartz OSL, which many studies have shown to be inappropriate for dating in the Japanese archipelago (e.g., Tsukamoto et al., 2003; Tokuyasu et al., 2010; Tamura et al., 2015, 2017; Riedsel et al., 2018).

Appreciable residual doses were observed for the pIRIR₂₉₀ signal of modern samples up to 30 Gy (Table 1). Holocene samples also show much older age estimates of pIRIR₂₉₀ than expected, which is attributed to the residual dose. The residual dose of Holocene samples, D_residual, is defined as:

Plots of fading-corrected equivalent doses D_e (= T_osl×Ḋ) against expected D_e (= T_expected×Ḋ) visualize the residual dose as a departure from the 1:1 line (Fig. 6). Although pIRIR₂₉₀ shows a larger scatter, its trend is distinct from that of the 1:1 line, and the y-intercept of the regression line of these plots indicates the average residual dose (e.g., Sohbati et al., 2011), which is 24.2 Gy (Fig. 6C). In contrast, as expected from the age results, the corrected D_e values of pIRIR₁₅₀ are consistent with the expected values, and only modern and late Holocene samples show identifiable departures from the 1:1 line (Fig. 6B). The departure of the IR₅₀ D_e from the 1:1 line is only identified for early Holocene samples and possibly results from over-correction (Fig. 6A).

Fig. 6

Plots of fading-corrected D_e derived from (A) IR₅₀, (B) pIRIR₁₅₀, and (C) pIRIR₂₉₀ against expected D_e. The y-intercept of the regression line for pIRIR₂₉₀ represents the average residual dose, which is estimated as 24.2 Gy. Dotted lines indicate the 95% confidence interval.

The average residual doses of pIRIR₂₉₀ are equivalent to an overestimate of 11,000 years, assuming an average dose rate of 2.2 Gy/ka. This overestimate leads to significant inaccuracy in dating Holocene sediments and possibly causes considerable overestimation even for deposits formed during the Last Interglacial Periods if the residual dose is not taken into account appropriately. No appreciable difference in the average g-value was identified between pIRIR₁₅₀ and pIRIR₂₉₀. If these g-values also apply for Late Pleistocene sediments, the lower residual dose of pIRIR₁₅₀ is an obvious advantage. However, an underestimated g-value of pIRIR₁₅₀ is also inferred from the apparent consistency of the pIRIR₁₅₀ ages with the expected ages for lower Holocene deposits in core GS-QAS-2, despite the influence of the residual dose. This possible inaccuracy in fading correction likely leads to large errors in dating older sediments. In selecting an optimal signal for feldspar dating, the g-value should also be carefully examined.

Correlation of residual dose with water depth

Saltwater absorbs the ultraviolet component of sunlight; thus, in shallow-marine environments, the bleachability of the luminescence signal is expected to rapidly drop as the water depth increases. However, the residual doses of IR₅₀ and pIRIR₁₅₀ in five modern samples show no obvious correlation with water depth. The depositional water depth of the Holocene samples is estimated from the difference between the bases of the foreshore and backshore facies, equivalent to the mean low-tide level at the time of deposition (Tamura et al., 2008a), and the sample level. The relationships of the residual dose of pIRIR₂₉₀ in modern and Holocene samples with water depth exhibit no obvious trend (Fig. 7), but a few samples deposited in water depths shallower than 5 m show obviously lower residual doses of pIRIR₂₉₀. Along the Kujukuri coast, the upper shoreface, i.e., the area shallower than 5–6 m, is characterized by frequent changes in topography caused by migration of longshore bars and troughs associated with an exchange of sand between the surf zone and the beachface (e.g., Lee et al., 1998; Kuriyama, 2002; Tamura et al., 2007). Samples from the foreshore and upper shoreface are likely to have been exposed to sunlight for a relatively long time, which may have bleached the difficult-to-bleach component of pIRIR₂₉₀. In the bleaching test on sample gsj15151, the residual dose of pIRIR₂₉₀ was depleted as low as 13.66±1.10 Gy after 4 h of exposure to artificial sunlight. Exposure to artificial sunlight for more than 4 h is known to further deplete pIRIR signals (e.g., Murray et al., 2012). Therefore, in the beach–shoreface system at Kujukuri, sunlight bleaching of pIRIR₂₉₀ was incomplete for most of the samples; the degree of bleaching varies from sample to sample but is likely minimized in water depths <5 m.

Fig. 7

Plots of residual dose estimated for pIRIR₂₉₀ against depositional water depth of the sample. The regression line exhibits only a very weak correlation; however, a few samples taken from less than 5 m water depth exhibit slightly lower residual doses than others. Dotted lines indicate the 95% confidence interval.

Comparisons with other coastal systems

Examination of the bleachability of luminescence signals has generally been limited to underwater environments of coastal systems, except for river deltas. Roberts and Plater (2007) quantified the residual dose of OSL in two quartz sand samples from the modern subtidal near-shore of the Dungeness Foreland, southern England, as equivalent to 15 and 40 years. Most efforts to quantify the residual dose of coastal sediments have been made in the intertidal zone and have generally reported residual doses of quartz OSL equivalent to less than a few tens of years (e.g., Banerjee et al., 2001; Ballarini et al., 2003; Madsen et al., 2005; Armitage et al., 2006; Madsen and Murray, 2009; Brill et al., 2017). Davids et al. (2010) reported that the K-feldspar IR₅₀ of modern beach sands in New England, USA, retains a residual dose corresponding to 40 years, and similar results were obtained for tidal-flat sediments on the western coast of the Korean Peninsula (Hong et al., 2003). Although the quartz OSL was not investigated here, the IR₅₀ of Kujukuri sands is shown to have a comparable residual dose to the quartz OSL, and feldspar IRSL observed in other coastal sands.

Applications of OSL dating to deltaic systems are not limited to intertidal and supratidal sediments (e.g., Giosan et al., 2006; Tamura et al., 2012a, 2012b; Chamberlain et al., 2018); the method can also be implemented in offshore and distributary sediments (e.g., Sugisaki et al., 2015; Chamberlain et al., 2017). Sugisaki et al. (2015) sampled suspended silts from turbid river water 300 km upstream of the Yangtze River estuary, southern China, and estimated the residual doses of quartz OSL, polymineral IR₅₀, and polymineral pIRIR₁₆₀ as 0.2, 1.4, and 7.4 Gy, respectively. In contrast, OSL and pIRIR₁₆₀ yielded almost identical equivalent doses for most of the samples in Holocene sediment cores collected in an offshore prodelta at 40 m water depth, which is a good indication of complete bleaching (Murray et al., 2012). Thus, Sugisaki et al. (2015) concluded that delta sediments are well bleached by sunlight before they reach the prodelta. Chamberlain et al. (2017) examined the bleachability of sediments in distributaries of the Ganges–Brahmaputra–Meghna delta, Bangladesh, and reported that the OSL of fine quartz grains is generally well-bleached, whereas the pIRIR of fine polymineralic sediments is poorly bleached. They additionally observed slightly lower residual doses in grains that had been tidally reworked upstream relative to grains obtained around the river mouth and attributed this to the longer transport path and additional sunlight exposure. Gao et al. (2017) found nearly consistent quartz OSL and polymineralic pIRIR ages for fine sediments in a 40-m-thick Holocene subaqueous delta succession in the northern Yangtze delta plain, giving a good indication that the bleaching was sufficient. Nian et al. (2018) also reported consistent and precise age estimates in the Holocene Yangtze delta sediments from OSL dating of coarse-grained quartz; results from quartz appeared to be much better than radiocarbon dating, which showed many age reversals. These studies show that quartz OSL in subaqueous deltaic environments is generally well bleached and retains residual doses that are negligible in terms of millennial-scale chronology. This conclusion is also supported in other deltaic systems (Kim et al., 2015; Li et al., 2018b). Although quartz can be well bleached in a deltaic system, quartz grains from the Kujukuri coast are not suitable for OSL dating due to their unfavourable OSL properties. For the feldspar signal, based on comparisons with quartz OSL, Li et al. (2018b) estimated the residual doses of pIRIR₁₅₀ and pIRIR₂₂₅ as <5.5 Gy and 2–11 Gy, respectively, for fine polymineralic sediments from a Holocene subaqueous delta in the Bohai Sea, northeastern China. As in the Kujukuri coast, such results can be applied to clarify the possible overestimate of the pIRIR ages of Pleistocene sedimentary sequences, providing practical grounds for selecting robust feldspar signals for dating.