Laser ablation–accelerator mass spectrometry reveals complete bomb ¹⁴C signal in an otolith with confirmation of 60-year longevity for red snapper (Lutjanus campechanus)

Additional keywords : age validation, carbon-14, Gulf of Mexico, Lutjanidae, radiocarbon.

Bomb-produced ¹⁴C has been used to make valid estimates of age, growth and longevity for various marine organisms for 25 years (Kalish 1993). A long history of success with this approach began with fish otoliths and the validation of purported annual growth zones (Kalish 1995; Campana 1999). Early studies were crude and involved instrumentation that is currently outdated. Not only have the precision and accuracy of the instrumentation and analytical approaches used to measure ¹⁴C increased, but the amount of material required for an otolith measurement has also decreased considerably and continues to improve (Andrews et al. 2015; Grammer et al. 2015). In concert with these improvements, extraction techniques are more precise – most studies now use a micromilling machine – and there is a more thorough understanding of bomb-produced ¹⁴C in the marine environment (e.g. Grottoli and Eakin 2007; Andrews et al. 2016a, 2016c; Druffel et al. 2016). As a result, questions of age and longevity have been answered for fishes throughout the world (e.g. Kalish et al. 2001; Andrews et al. 2011, 2012, 2019; Kastelle et al. 2016; Campana et al. 2016) and, in some cases, the limits were pushed for species with very small otoliths and geographical origins that were not well constrained (Ishihara et al. 2017; Andrews et al. 2018a).

Bomb ¹⁴C dating is a modern form of radiocarbon dating (Libby 1955) that relies on the rise of ¹⁴C because of atmospheric testing of thermonuclear devices in the 1950s and 1960s as a temporal reference (Reimer et al. 2004; Druffel et al. 2016). For marine fishes, the approach typically relies on the extraction and ¹⁴C analysis of calcium carbonate from the earliest otolith growth (core material), for which an alignment is made with a bomb ¹⁴C reference from the marine environment (e.g. Campana 1997; Andrews et al. 2011, 2016b). Sampling beyond the core is rarely done because the temporal specificity of the extraction rapidly becomes more difficult to establish as otolith growth layers become thinner with increasing age (e.g. Cook et al. 2009). In most fish age validation studies, it was the period of the increase in ¹⁴C (from c. 1958 to c. 1970) that functioned as a diagnostic reference for determining a birth year from otolith ¹⁴C measurements. Recent advances include more extensive coral ¹⁴C records that have extended the utility of the technique in some tropical regions to more recent periods (c. 1980 to present) using the post-peak ¹⁴C decline to age younger fish (Andrews et al. 2013, 2016b, 2018a; Ishihara et al. 2017; Barnett et al. 2018; DeMartini et al. 2018). However, fish ages that lead to birth years earlier than the ¹⁴C rise period (prebomb) lose the ability to determine a specific date of formation because ¹⁴C levels in the marine surface layer plateau in the decades before the ¹⁴C rise. Hence, bomb ¹⁴C dating is limited by prebomb birth years in terms of establishing longevity (Baker and Wilson 2001; Cook et al. 2009; Andrews et al. 2013) – only a minimum age can be validated from the last year of prebomb levels, leaving longevity in question.

Although the age and growth of red snapper was verified or validated to a limited extent using otolith margin analyses, tag–recapture data, lead–radium dating and bomb ¹⁴C, some scepticism remains about longevity estimates that exceed 30 years (e.g. Szedlmayer and Beyer 2011). Otolith margin or tagging studies have led to some observations of red snapper age but are limited to early growth and short time spans (Patterson et al. 2001; Wilson and Nieland 2001). Lead–radium dating can be useful for age estimates on the order of a decade to ~100 years (Andrews et al. 2002, 2009; Andrews 2016; Tracey et al. 2017), but the method is typically limited to providing approximate ages for groups of fish otoliths (studies are subject to minimum mass and lead–radium activity requirements; Andrews et al. 1999a, 1999b). This approach has led to support for red snapper longevity exceeding 30 years (Baker et al. 1999). In contrast, bomb ¹⁴C dating has provided valid estimates of age for individual red snapper up to ~38 years (Baker and Wilson 2001). This method has evolved considerably since its inception, with recent indications that red snapper can live to at least 44 years (Barnett et al. 2018); the upper limit uncertainty for this age is based on a lack of time specificity for prebomb ¹⁴C measurements from modern otoliths.

In this study, red snapper otoliths were analysed for ¹⁴C using a novel laser ablation–accelerator mass spectrometry (LA-AMS) technique to provide a continuous record of ¹⁴C uptake (Welte et al. 2016) and a basis for age validation that may extend beyond the normal limits of bomb ¹⁴C dating. This approach may reveal the location in the otolith cross-section where bomb-produced ¹⁴C rises in a single continuous measurement (youngest to oldest material and covering the lifespan of the fish). The LA-AMS ¹⁴C time series from an otolith cross-section can be aligned with estimated years of formation from age reading of growth zones. The goal would be to assess the alignment or misalignment of the initial ¹⁴C rise (c. 1958) with the regional ¹⁴C reference to potentially refine age estimates that may not be accurate. In some scenarios, validated ages within the prebomb period are possible by locating the initial ¹⁴C rise at advanced ages away from the earliest otolith growth. The approach may also differentiate ages from ¹⁴C values that lead to ambiguous birth years (potentially attributed to either the ¹⁴C rise or decline period), a discrepancy that can be on the order of several decades.

The red snapper otoliths selected for LA-AMS were from fish specimens with age estimates that provided lifespans covering the bomb ¹⁴C signal. Great otolith mass (approaching or exceeding 4 g) was also considered to increase the chances of using the oldest fish. Based on these observations, three specimens were selected for the study (Table 1). The study began with a series of preliminary assays on the first otolith (RS07), followed by refined assays on two additional otoliths (RS04, RS17). These fish were among the largest for this species, exceeding 80-cm total length (TL), with whole otolith masses approaching 4 g (Fig. 1). Each fish was collected from the Gulf of Mexico by port samplers associated with the Panama City Laboratory (PCL) of the Southeast Fisheries Science Center, National Oceanic and Atmospheric Administration (NOAA) Fisheries. Estimated ages for the three selected specimens ranged from 33 to >50 years, with some level of variability in estimated age for each specimen from various age reader interpretations. These age estimates were either historical (anonymous) from age readers at PCL or from age reading during the present study (by A. H. Andrews). Similar estimates were made for two of the specimens in a parallel study (Barnett et al. 2018) and were within the range provided for the present study. Collection dates with age estimates revealed potential birth years in the 1950s for each specimen, but two fish may have been younger with birth years as early as 1964 and 1974, given age reading is providing an accurate age within the proposed age ranges. Two fish probably extend into the prebomb period, whereas one fish may have started life during the bomb ¹⁴C rise or peak period (Table 1).

Table 1.  Data for the red snapper (Lutjanus campechanus) specimens and otoliths that were selected for ¹⁴C assays
Fish length, otolith mass and estimated age indicated these fish were good candidates for capturing the full bomb ¹⁴C signal because birth years may predate the rise of bomb-produced ¹⁴C in c. 1958 for the Gulf of Mexico. Estimated ages covered a range of interpretations from various age readers (historical records from Panama City Laboratory, estimates from Barnett et al. (2018) and the present study). The potential birth years led to a range of hypothetical bomb ¹⁴C scenarios that can be explored with laser ablation–accelerator mass spectrometry. TL, total length

Fig. 1.  Illustration of red snapper (Lutjanus campechanus) and of one of the largest otoliths at more than 4 g. The bottom image is a transverse thin section of the otolith from specimen RS17. The series of small white dots on the left side of the section mark the growth zones counted for an age estimate of 43 years (dorsal).

Two of the three otoliths were sampled for the earliest otolith growth (core material) using a New Wave Research (ESI-NWR Division; Fremont, CA, USA) micromill and were analysed in the standard graphitisation AMS manner (Table 2; for details, see Andrews et al. 2015). These ¹⁴C values were used to generally confirm the predicted bomb ¹⁴C scenario and exemplify the limitations of single-sample bomb ¹⁴C dating (Barnett et al. 2018).

Table 2.  Radiocarbon data for the red snapper (Lutjanus campechanus) otoliths where the core (birth year) material of the adult otolith was extracted with a micromilling machine
The single measurements allowed refinement of the age from the original range of estimates to a minimum age. No initial core measurement was made for RS17. Fraction modern (F¹⁴C) data are given as the mean ± 2 s.d.

The first otolith analysed (RS07) was sampled beyond the initial milled core ¹⁴C measurement and into the growth zone sequence using the micromill. This series of samples was analysed by gas AMS (Rosenheim et al. 2008; Wacker et al. 2013). In this time-efficient online approach, the carbonate samples are placed in septum-sealed vials under a helium atmosphere. CO₂ is released from the carbonates by the addition of phosphoric acid. In contrast with conventional graphite AMS analysis where the liberated CO₂ is reduced to graphite and measurements are performed on solid targets, here the CO₂ gas is concentrated by means of a zeolite trap. In a final step, the CO₂ is transferred with He carrier gas into the syringe of the gas interface system, where it is further diluted with helium and fed into the gas ion source of a Mini Carbon Dating System (MICADAS) AMS system (Ionplus, Dietikon, Switzerland) for gas ¹⁴C analysis. The goal was to locate the position in the otolith section where the bomb ¹⁴C rise occurred using the radial sampling and to corroborate the findings from exploratory LA-AMS assays (Table 3; Fig. 2).

Table 3.  Radiocarbon data for the otolith core and radial sample series taken from sample RS07
Prebomb ¹⁴C levels covered a time span greater than expected (dating back to 1944 and an age of 60 years). Growth year was the year of formation for the respective portion of the otolith (see Fig. 2). Ages and years were rounded to the nearest whole number. Post-corrected Δ¹⁴C values are provided for comparison with other records. Fraction modern (F¹⁴C) and Δ¹⁴C data are given as the mean ± 2 s.d.

Fig. 2.  Transverse section of the red snapper otolith from specimen RS07 with the series of samples taken using both conventional micromilling extraction (right side) and laser ablation–accelerator mass spectrometry (LA-AMS; left side). Counts of well-defined growth zones were made before the sampling and are denoted as black dots radiating out from the core (earliest growth) to 10 years of age. Extractions with the micromill were estimated to be annual out to 10 years and then composed of 2–3 years of material out to an estimated age of 23 years. Analysis of the milled samples for ¹⁴C revealed the bomb ¹⁴C rise at 14–17 years (1958–61). Measurements of ¹⁴C from LA-AMS revealed the rise in bomb ¹⁴C in a similar structural location to the micromilling series for all LA-AMS scans (grey boxes). Scans 1 and 2 were early feasibility and instrument optimisation runs that provided a first look at the region of ¹⁴C rise (see Fig. S1, S2). The zigzag scan was focused on providing a stronger and longer-lasting signal, before development of the parallelogram zigzag or precision scan (Fig. 3). The earliest core extraction with the micromill was performed on the other otolith of the pair and is not shown here (Table 2).

The initial LA-AMS test scans were performed on an otolith section (RS07) originally prepared for making estimates of age from growth zone counting. The section was very thin (~0.3 mm) and previously mounted to a glass microscope slide with an unknown mounting medium. This sample was used opportunistically in a first attempt at locating the bomb ¹⁴C rise by LA-AMS within an otolith. The set-up used for these initial measurements is described in Welte et al. (2016); the sample was ablated by a 193-nm laser (ArF excimer laser: Ex5, GAM LASER, Orlando, FL, USA) with a rectangular spot measuring 110 × 680 µm in its focus. A helium carrier gas flow rate of ~1.5 mL min^–1 was used to transport the gases formed during laser ablation (including CO and CO₂) into the gas ion source of the MICADAS for ¹⁴C analysis. An exploratory scan (Fig. 2, Scan 1) revealed a ¹⁴C inflection that could be attributed to the rise of bomb-produced ¹⁴C, and the difference was measurable across the otolith section. This finding led to a more refined linear scan (Fig. 2, Scan 2) that further revealed the ¹⁴C inflection in greater detail. This exploratory series led ultimately to a ‘rectangular’ zigzag scan (precision scan; Welte et al. 2016) where the zigzag-shaped moving pattern of the sample under the laser was performed perpendicular to its continuous axial moving direction, ultimately resulting in a rectangular sampling area (Fig. 2). A more thorough sampling of the otolith material could be achieved with the zigzag scanning and allowed documentation of the precise location of the ¹⁴C rise in the otolith section (for these exploratory scan lines, see Fig. 2).

Based on the initial results, a refined LA-AMS analysis procedure was used for the next two otoliths (RS04, RS17) including modifications of: (1) the sample preparation procedure; (2) the LA-AMS set-up; and (3) the LA-based sampling strategy (see below).

The small otoliths required low-loss preparation to allow for precise positioning of the laser ablation scan lines relative to the growth zone structure in the otolith. Consequently, thicker (~1 mm) transverse sections were prepared and mounted in an epoxy (Epo-Tek 301-2/1LB/A with hardener 301-2/1LB/B; Epoxy Technology, Billerica, MA, USA) using disposable plastic embedding moulds (Peel-A-Way, R-40 22 × 40-mm rectangular, 20 mm deep; Catalogue number 18646C; Polysciences Inc., Warrington, PA, USA). Sections were placed directly on the bottom of the embedding moulds after minor polishing to make sure the cut surface was flat. Epoxy was added to the top of the otolith section to a depth of a few millimetres to create a wafer that allowed easy handling. The mounted sections were polished with 1200-grit silicon carbide wet–dry sandpaper on a lapidary wheel to remove epoxy and other surface contamination from the analysis surface. The embedded sample was cut for fitting and proper orientation within the LA-AMS sample carriage.

LA-AMS analysis of the two otoliths (RS04, RS17) was performed using a refined set-up with a smaller laser spot of 75 × 140 µm² on the sample allowing for improved spatial resolution.

The rectangular zigzag scanning pattern used for the LA-AMS test scan on RS07 proved to be insufficient for revealing the imprinted ¹⁴C-bomb signature within this sample (compare Fig. 3 and the Supplementary material) because it crossed multiple growth zones and consequently different years of growth during one zigzag step. In order to provide greater spatiotemporal specificity, the zigzag scans performed on RS04 and RS17 were improved by positioning the zigzag movement and the continuous axial displacement at an optimised angle. This procedure ensured sampling within the same growth zone structure during single zigzag steps and yielded an overall sampling area of a parallelogram (compare Fig. 4 and 5).

Fig. 3.  Comparison of ¹⁴C measurements from both micromilling and the laser ablation–accelerator mass spectrometry (LA-AMS) zigzag scan (Fig. 2) with the regional coral bomb ¹⁴C reference series for the Gulf of Mexico (GoM; Andrews et al. 2013). The initial core measurement indicated the age was >46 years but could have been up to 55 years given the high count scenario (Table 2). The micromilled sample series pushed the age to 60 years with a well-matched time series to the coral ¹⁴C record (±1–2 years). The zigzag LA-AMS data were in approximate agreement with the coral ¹⁴C record considering the cross-growth zone pattern of this initial assessment. Regardless of the low resolution, the goal of detecting a complete bomb ¹⁴C signal within an otolith was achieved. The attenuated peak may be attributed to an influence of deep water with a slightly depleted ¹⁴C record later in adult life. Horizontal error bars for LA-AMS were derived from the age range of the scan path width for each sample block and vertical error bars were 2 s.d. of the block mean.

Fig. 4.  Cross-sectioned otolith of specimen RS04 with the micromilled core extraction visible as a notch in the upper right and the laser ablation–accelerator mass spectrometry precision scan (zigzag parallelogram) progressing diagonally from near the core to near the outer otolith edge. Delineated are the scan analysis blocks (600 µm) that provided mean ¹⁴C values for the respective regions of the otolith. Mean age estimates from growth zone counting are enumerated along the centre of the scan. The original zone counting age estimate was shifted by 6 years from the maximum counted age of 48 years for an alignment with the ¹⁴C rise and an age of 54 years (Fig. 6). Zone counts for this section could not be clearly imaged and are not specifically marked.

Fig. 5.  Laser ablation accelerator mass spectrometry (LA-AMS) precision scans for both sides of the otolith section from specimen RS17 where growth zone structure was well defined: (a) long scan (dorsal), (b) short scan (ventral). Age was estimated as 43 years with original estimates that ranged from 36 to 56 years using different interpretations. Zone counts from the 43-year scenario were used to estimate mean age for the LA-AMS block means enumerated within the scans. The resulting ¹⁴C time series from each scan was well aligned with the coral ¹⁴C reference (Fig. 6) and proved useful in validating an age from near peak values; normally, a single measurement near the ¹⁴C peak is not diagnostic for age. No initial core ¹⁴C measurement was made for this specimen.

On each of the two otoliths (RS04, RS17), two LA-AMS sampling areas across the dorsal and ventral portions were selected. Because for the refined LA-AMS analysis thicker otolith sections were prepared, four stacked scans were performed at each sampling location, yielding eight scans for each otolith. On the RS04 otolith (Fig. 4), one LA-AMS sampling area may have been compromised by inadvertent inclusion of ¹⁴C-depleted epoxy (Welte et al. 2016) and was thus not considered. The overall analysis time per otolith sample was on the order of 2 h, whereas ~6–8 mg of material was consumed. The resulting precision using the refined setup and sampling method is in the order of 1% for a single LA-AMS data point on a modern sample.

Because the ¹⁴C data from LA-AMS is provided quasi-continuously (in 10-s intervals), it must be processed and recombined offline to reach the anticipated spatial resolution. For each scan, the number of ¹⁴C data points per unit time, and thus scan distance, were chosen in a way that provided enough spatial resolution while minimising instrument measurement variability (e.g. a wider range of years leads to lower instrument measurement error but can compromise the resolution needed to resolve the initial ¹⁴C rise). Therefore, a dedicated data reduction strategy was applied (Yeman et al. 2017). Briefly, first, single-cycle (10 s) AMS data were blank subtracted, isotopic fractionation corrected using concurrently acquired δ¹³C data and normalised using standard data processing routines (Wacker et al. 2010). Because the focus was on relative changes in ¹⁴C within the scans, the systematic error from standard normalisation (0.5%) was not considered. Second, single-cycle data points were subdivided and subsequently recombined (using a weighted average) so that for each resulting data point a spatial resolution (along the otolith growth axis) of 500–600 µm was reached. The final data are reported as F¹⁴C (fraction modern), which corresponds to the activity ratio of the sample relative to the primary ¹⁴C standard (Oxalic acid I, SRM 4990B) and consequently does not depend on the year of sample formation (Reimer et al. 2004). Because the main goal of this study was to determine the age of formation of the otolith layers, this is the most appropriate ¹⁴C unit. Calculated Δ (often also used as Δ¹⁴C) values are provided for comparison with previous studies, but are not primarily used because of the circularity in correcting to the time of formation (Stuiver and Polach 1977).

By comparing the section images used for growth layer counting with the corresponding laser tracks, ages were assigned to the LA-AMS data and then correlated with a regional coral ¹⁴C reference record (Andrews et al. 2013). Final adjustments to the age estimates and consequent year of formation because of small initial misalignments observed for the ¹⁴C rise in the otolith data were made to achieve congruence with the regional coral ¹⁴C reference. The refined ages were compared to growth zone counts and, in some cases, to the images of well-defined otolith sections to cross-check the age model.

The first transverse otolith section used in exploring the feasibility of LA-AMS to measure the bomb ¹⁴C signal through the ontogeny of an adult red snapper (RS07) revealed a consistent location in the cross-section and year of formation for the rise of bomb ¹⁴C between the methods (Fig. 2, 3). An original low estimate of 40 years was invalid based on a single core extraction that revealed the fish must be older than 46 years from prebomb levels (Table 2; Fig. 3). As a result, age estimates were reassessed in the sectioned otolith by counting finer growth zone structure that led to ages up to 55 years and were consistently quantified. Hence, it was hypothesised that the bomb ¹⁴C rise would occur between the core and the 10th year of growth in the otolith section. The subsequent extraction and conventional gas AMS ¹⁴C analysis of 15 micromilled radial samples revealed the bomb ¹⁴C rise was in a more recently formed part of the otolith section, corresponding to mean estimated fish ages of 14–17 years (Table 3; Fig. 2).

Measurements of ¹⁴C in specimen RS07 using LA-AMS revealed that the bomb ¹⁴C rise could be detected and was in a similar location and year of formation relative to a conventional AMS analysis of the micromilled series (Fig. 2, 3, S1, S2). The proof-of-concept for use of LA-AMS to detect the bomb ¹⁴C rise in an otolith section was addressed initially with two exploratory scans: Scan 1 revealed an estimate of the bomb ¹⁴C signal and Scan 2 was used to further investigate the ¹⁴C signal and to optimise the instrument, which was followed by a first trial of the zigzag scan on an otolith (Fig. 2). This approach was used to increase the sampling time within single growth layers and to provide a stronger measure of the bomb ¹⁴C rise, especially considering this otolith section was very thin (0.3 mm) and could not be analysed for an extended period in one location without reaching the mounting medium. Although the zigzag approach led to a longer signal collection time along the sample, the results suffered from a mix of ¹⁴C levels from various formation years through the scanning analysis timeline (offset from growth direction and layering; Table 4, Fig. 2). Specifically, as the scan entered the bomb ¹⁴C rise region of the otolith, the zigzag pattern of the laser criss-crossed otolith growth zones that were formed in different years, some of which would have had markedly different ¹⁴C levels. The result was a muted bomb ¹⁴C record across the otolith; however, the goal of this specimen analysis was a proof-of-concept for moving forward with analysis of additional fish specimens. LA-AMS on otolith RS07 achieved this goal by tracing a complete bomb ¹⁴C signal covering the prebomb, peak and decline periods in an otolith section on three separate occasions (Fig. 2, 3, S1, S2). These findings led to the development of a parallelogram zigzag pattern (precision scan) to maintain greater time specificity on the two following red snapper otolith specimens.

Table 4.  Radiocarbon and age data as block means from the laser ablation–accelerator mass spectrometry zigzag scan and the growth zone counting for specimen RS07
Ages and years were rounded to the nearest whole number. Post-corrected Δ¹⁴C values are provided for comparison with other records. Fraction modern (F¹⁴C) and Δ¹⁴C data are given as the mean ± 2 s.d.

The findings from both conventional gas AMS of micromilled samples and the LA-AMS scans supported an estimated age for RS07 that was slightly older than the maximum estimated age of 55 years from growth zone counting (observed by A. H. Andrews). Given the rise of bomb ¹⁴C within the section was located at counts near the mid-teens and the minimum age of 46 years was established with the single core ¹⁴C measurement, the birth year would be 1946 for an age of 60 years and a potential uncertainty of just a 1–2 years from the milled extraction series width (Table 3; Fig. 3).

The otolith sections from RS04 and RS17 were used to refine the application of LA-AMS to measure bomb ¹⁴C with greater precision through the ontogeny of adult red snapper. The original lowest estimates of age for RS04 were eliminated below 44 years by the single core extraction ¹⁴C measurement (Table 1); a prebomb birth year earlier than 1958 was the only possible scenario (Table 2). No core measurement was made for RS17 and age could have been 36–56 years with prebomb to peak birth years (Table 1). LA-AMS revealed two different bomb ¹⁴C patterns from which the prebomb scenario for RS04 was confirmed and a younger age scenario was discovered for RS17. Although both had the potential to be fish with prebomb birth years (Table 1), RS17 was restricted to near peak ¹⁴C levels followed by a ¹⁴C decline and RS04 was similar to RS07 with a full bomb ¹⁴C signal and greater precision (Tables 5, 6).

Table 5.  Radiocarbon data for the laser ablation–accelerator mass spectrometry precision scan taken from specimen RS04
The timeline determined from growth zone counting to 54 years provided a well-matched ¹⁴C rise time, extending the minimum age from the single core ¹⁴C measurement by 10 years and older than what growth zone counting covered (Table 1). Ages and years were rounded to the nearest whole number. Post-corrected Δ¹⁴C values are provided for comparison with other records. Fraction modern (F¹⁴C) and Δ¹⁴C data are given as the mean ± 2 s.d.

Table 6.  Radiocarbon data for the long and short laser ablation–accelerator mass spectrometry precision scans taken from both sides of the otolith section of specimen RS17
The ¹⁴C timelines agreed with the estimated ages relative to the coral ¹⁴C reference. Each RS17 series precluded alignment with a younger age associated with a strictly ¹⁴C decline scenario. Post-corrected Δ¹⁴C values are provided for comparison with other records. Fraction modern (F¹⁴C) and Δ¹⁴C data are given as the mean ± 2 s.d.

The time-specific parallelogram zigzag design (precision scan) led to circumstances that were easier to align (relative to RS07) with the otolith age estimates through ontogeny (Fig. 4, 5). Age estimates were reassessed for RS04 and RS17 in the sectioned otoliths by counting fine growth zone structure (observed by A. H. Andrews), which is similar to the growth zone structure that was used for RS07 and is exemplified by the RS17 section (Fig. 1). These counts led to consistent ages of 48 and 43 years for RS07 and RS17 respectively (Fig. 4, 5). Although the ages and dates of formation from the LA-AMS data series for RS17 were in alignment with the coral ¹⁴C reference record, RS04 required an additional 6 years to the maximum growth zone-derived age of 48 years to align properly (Fig. 6). Furthermore, although the single core ¹⁴C value led to a minimum age of 44 years (Table 2), the LA-AMS data series indicated RS04 was 54 years old because of the extent of prebomb levels in the earliest growth. This is similar to what was found for RS07, where the age was extended 5 years beyond the growth zone counting maximum (55 years). The minor misalignment for the more recent adult samples within the series, attenuated and phase lagged relative to the coral ¹⁴C record (Fig. 6), may be attributed to either uptake from deeper ¹⁴C-depleted waters or the laser sampling more years of growth than expected, leading to a dampened signal. Either way, the results were consistent with the findings for RS07 (Fig. 3). The replicated sample series from RS17 provided a consistent alignment with the bomb ¹⁴C peak and decline periods (Fig. 6) and indicated the age of this fish was 43 years; younger age estimates would not provide a proper alignment with the coral ¹⁴C peak and decline periods.

Fig. 6.  Comparison of ¹⁴C measurements from laser ablation–accelerator mass spectrometry (LA-AMS) precision scanning of two red snapper otoliths (Fig. 4, 5) with the regional coral bomb ¹⁴C reference series for the Gulf of Mexico (GoM; Andrews et al. 2013). The initial core measurement of specimen RS04 indicated the minimum age was >44 years, but an alignment of the LA-AMS series indicated the fish was 54 years old. The precision scan for RS04 agreed with the coral ¹⁴C record with regard to the time of ¹⁴C rise and was slightly attenuated relative to the ¹⁴C reference, similar to the findings for specimen RS07. Both precision scans on specimen RS017 agreed with the coral ¹⁴C record from the upper rise through the peak period into the decline. This alignment illustrates the potential utility of LA-AMS over single core ¹⁴C analyses because of the potentially ambiguous alignment with either the rise or decline portion of the reference record. Horizontal error bars for LA-AMS were derived from the age range of the scan path width for each sample block and vertical error bars were 2 s.d. of the block mean. F¹⁴C, fraction modern.

A 60-year longevity was strongly supported for red snapper in the Gulf of Mexico from both the conventional micromilled and LA-AMS ¹⁴C analyses. The radial micromilled samples placed the bomb ¹⁴C rise time at an estimated age of 14–17 years and well away from the earliest otolith growth for the opportunistic otolith section (RS07). In this case, the minimum age determined from a single ¹⁴C measurement in the otolith core was 46 years based on the last year of prebomb levels. Thus, it was necessary to rely on other studies that have validated early growth for the addition of 14 years to attain the 60-year estimate. Because biannual growth zone deposition during the early life history of red snapper (Szedlmayer and Beyer 2011) was not supported in a recent study using the post-peak decline period (Barnett et al. 2018), an age of 60 years for this 83-cm-TL red snapper was most probable and greater than previous estimates from growth zone counting by at least 5 years.

The utility of LA-AMS in revealing valid estimates of age by locating the rise of bomb-produced ¹⁴C was confirmed as feasible with two exploratory linear scans and a zigzag scan in the first otolith section (RS07). Even though the otolith section thickness was not optimal for LA-AMS, each scan revealed a complete bomb ¹⁴C signal and provided a time-specific location for the bomb ¹⁴C rise in the otolith that correlated with the results from micromilling. Although the radial micromilling provided what may have been greater precision in determining the age at which the bomb ¹⁴C rise occurred, these preliminary LA-AMS results indicated a more concerted effort on thicker samples prepared for LA-AMS analyses would yield a complete bomb ¹⁴C signal for fish with prebomb birth years and valid estimates of age.

Valid ages for two additional adult red snapper (RS04, RS17) were determined with the development and application of the LA-AMS precision scan, a zigzag pattern within a parallelogram that had greater time specificity for the laser track over the otolith surface. The thicker otolith section preparation allowed longer scan times and replicate runs across the same transect. In each case, the full length of the scan path (core to edge) was replicated four times and, as a result, precision increased. The rectangular laser spot sample size of 75 × 140 µm is estimated to have sampled close to 1 year of accreted growth based on otolith increment widths (narrowest zones) and may have sampled as little as 2 years of growth simultaneously as the zigzag pattern crossed the most compressed growth zones – these narrow zone widths appeared to begin after the first 10 years of growth and progress to the otolith edge. The finalised data points correspond to the average of numerous 10-s intervals (i.e. 36 measurements across 0.6 mm of the scan) to improve counting statistics and reduce noise. For the measurements closer to the core, where growth zones are thickest, the integration time corresponded to 1–2 years of growth. Hence, the greatest precision for calculated ¹⁴C mean values from the precision scans was within the region of earliest and most rapid otolith growth. Translating these measurements to a date and fish age is consequently related to the position on the coral bomb ¹⁴C reference curve, and temporal resolution is necessarily dependent upon the position within the otolith.

The second oldest fish of this study (RS04) was validated to be at least 44 years with the single core ¹⁴C measurement and was assumed to be 48 years based strictly on growth zone counting. This estimate was increased further by the LA-AMS precision scans to 54 years. Similar to the results from RS07, the timeline established for RS04 relied on support from early otolith age reading because of prebomb ¹⁴C levels in the first few years of growth. In this case, the ¹⁴C rise time was closer to the core and between estimated ages of 8 and 12 years, which led to an age of 54 years, greater than the highest estimate for this red snapper from various age readers (33–48 years) and similar in outcome to the underestimated age for RS07. It is most probable that this 85.3-cm-TL red snapper was 54 years old.

The third and youngest red snapper studied (RS17) provided an example of another kind of result that can be obtained from LA-AMS analyses on otoliths. This fish was estimated to be between 36 and 56 years from anonymous historical and recent age readers, and there was no initial core ¹⁴C value to narrow down birth year assignment. However, the age reading performed in this study provided an age estimate that was well defined at 43 years (Fig. 1). The series of LA-AMS precision scans on this specimen resulted in the best-case scenario for obtaining otolith ¹⁴C measurements in that both dorsal and ventral sides of the otolith section were scanned successfully. The results were similar for each side and provided two well-matched time series for both the most recent age reading (43 years) and agreement with the contour of the coral ¹⁴C reference. In this case, if a single core measurement had been made, the alignment of this value to the bomb ¹⁴C reference would have been ambiguous with potential birth years that span the peak (1967–82) and ages of 28–43 years. The advantage with the continuous ¹⁴C profile from LA-AMS was the elimination of the younger age scenario because of an obvious offset of the entire data series from the coral record when using younger age scenarios. Hence, the LA-AMS works well to resolve issues associated with core ¹⁴C values that can be ambiguous and may only be resolved with well-defined growth zone counting or the use of other otolith proxies for age, such as otolith mass (e.g. Andrews et al. 2013, 2016b, 2019). The age of this 88.0-cm-TL red snapper was well constrained with a valid estimate of 43 years.

A potential complication arises from the use of a continuous bomb ¹⁴C record within an otolith because of possible ontogenetic changes in habitat. The LA-AMS scans for the two oldest red snapper in this study provided information that is consistent with a deeper dwelling part of its life history – each fish exhibited an attenuated bomb ¹⁴C peak. This would be expected to some extent if the fish was living near or within the thermocline, a depth at which the well-mixed surface layer begins to give way to or mixes with deeper ¹⁴C-depleted waters (e.g. Grammer et al. 2015; Campana et al. 2016; Andrews et al. 2018b). A similar scenario was exhibited in the Gulf of Mexico from a study of yellowedge grouper (Epinephelus flavolimbatus), where a series of extractions from younger to older otolith material indicated there may have been a deep-water effect on otolith ¹⁴C because of ontogenetic habitat changes (Cook et al. 2009). However, even if the bomb ¹⁴C signal is attenuated to some extent from the mixing of surface waters with deeper waters near the thermocline, the timing of the initial ¹⁴C rise would not be expected to be phase lagged to a significant extent in tropical waters (likely at most 1–2 years and more related to the attenuation of the signal because of dilution, as opposed to a late arrival). This factor would be a concern for organisms that live well below the thermocline or in waters that are strongly affected by upwelling. In general, success with bomb ¹⁴C dating of organisms with this kind of habitat or life history is best addressed with a series of otolith core measurements in concert with known-age reference material, such as juvenile otoliths, to provide some ground truthing. Examples of a series of successes in this regard can be illustrated with a long series of studies on rockfishes (Family Sebastidae) of the north-eastern Pacific Ocean (e.g. Kerr et al. 2004; Piner et al. 2005; Andrews et al. 2007; Kastelle et al. 2008).

Otoliths of red snapper were selected for this pioneering study because the age estimates led to birth years in the prebomb or bomb ¹⁴C rise periods and the Gulf of Mexico provides the broadest, regionally consistent coral ¹⁴C reference records (Andrews et al. 2013; Barnett et al. 2018). Because the selected red snapper otoliths could span all or most of the bomb-produced ¹⁴C signal of this marine environment, it was hypothesised that an entire bomb ¹⁴C signal could be traced with LA-AMS based on previous results from other carbonates (Welte et al. 2016). A factor in selecting red snapper for this study was that the otoliths are massive and can reach or exceed 4 g. Hence, the high mass was a best-case scenario for a feasibility study using the new LA-AMS technology because it provided the greatest amount of carbonate through this 50- to 60-year period of formation that exhibits Δ¹⁴C changes in the marine environment on the order of 100–200‰ (Grottoli and Eakin 2007). The first specimen (RS07) was opportunistic because it had been prepared as a thin section for age reading and was only 0.3 mm thick. From previous analyses of another marine carbonate (black-lip pearl oyster Pinctada margaritifera; Welte et al. 2016), it was known that the laser could easily pass through this thin section and may be influenced by contamination from ¹⁴C-depleted mounting medium. In some circumstances with the black-lip pearl oyster, it was clear from the LA-AMS signal that the carbonate ¹⁴C levels were being diluted with sources that were well below what was expected from the marine environment. This was likely the reason for the low levels observed on one side of the RS04 otolith that led to it being dismissed from further analysis.

This research addresses a frontier in bomb ¹⁴C dating by providing avenues that may lead to the determination of age for marine organisms that have lived through all or part of the bomb ¹⁴C period when few other options exist. An example of a circumstance where this approach could be useful is with the otoliths of blue marlin (Makaira nigricans). A recent study revealed that the age of a large adult blue marlin (3.7-m fork length and 565 kg) was 20 years using a series of deductions from different otolith sample types (Andrews et al. 2018a). The initial measurement of the otolith core revealed the fish could have been born during either the bomb ¹⁴C rise or decline period, a discrepancy of up to several decades. It was the use of the whole lapillus and other otolith samples where a series of deductions was made to refine the age of this fish to 20 years. Assuming LA-AMS technology continues to improve well enough to deal with the very small otoliths from this species (<10 mg for full-sized adults), the whole otolith could be scanned from core to edge in a single continuous measurement within 30 min, thereby alleviating the need to make other ¹⁴C assays.

The authors declare that they have no conflicts of interest.

Christane Yeman is funded by Swiss National Science Foundation (project number 160064).