APPLICATION OF AMINO ACID RACEMIZATION TO THE DATING OF FORAMINIFERS AND THE IDENTIFICATION OF REMANIÉ FOSSILS AND REWORKED SEDIMENTS

The reworking of older fossils and sedimentary particles into more recently deposited sedimentary successions has long been recognised as a challenge for defining the age of depositional events and the construction of high-resolution palaeoenvironmental records (Martin et al., 1995; Kidwell & Flessa, 1996; Martin, 1999). In many cases, reworked fossils might be indistinguishable from younger individuals based on their preservation states, representing a significant issue for their direct dating and the quantification of sedimentation rates. The cyclic nature of large-scale sedimentary processes, due to repeated climate and sea-level changes, in many sedimentary environments has amplified the degree of reworking of fossils into younger successions, such that it may represent the norm rather than an exception.

The early recognition that Orbital Forcing (Croll, 1875; Milankovitch, 1941) could explain long-term climate and glacio-eustatic sea-level changes (Shackleton & Opdyke, 1973; Hays et al., 1976) has provided a framework for identifying repeated environmental changes in marginal marine environments and an awareness of the increased probability of reworked fossils in the stratigraphical record. Repeated episodes of relative sea-level changes represent a key driver in the erosion, transportation, and recycling of sediment across continental shelves in both seaward and landward directions (Roy et al., 1994; James & Bone, 2021). These relationships are particularly evident for the last glacial cycle, the most recent example from the Quaternary, providing geomorphological evidence and sedimentary records of sediment fluxes and their exchanges from shelf to shallow water coastal environments. Following the end of sea level highstand conditions and glacial inception at the end of the Last Interglacial Maximum [Marine Isotope Stage (MIS) 5e] some 112 ka, mean sea level fell by about 20 m in 2000 years as revealed by marine and ice core oxygen isotope records (Lambeck & Chappell, 2001; Masson-Delmotte et al., 2010). Sea level remained at approximately 60–70 m below present sea level (BPSL) for much of the glacial cycle following the main glacial transition from the end of the Last Interglacial sensu lato (MIS 5) to the onset of the MIS 4 stadial, until ultimately falling to 125 m BPSL during the Last Glacial Maximum (LGM; MIS 2; Murray-Wallace & Woodroffe, 2014). With the demise of much of the continental ice sheets, the post-glacial sea-level rise following the LGM, inundated many coastal regions such that some Holocene coastal landscapes reoccupied areas of active deposition during the Last Interglacial (MIS 5e), resulting in coastal erosion and the supply of reworked Last Interglacial sediment to nourish modern coastal systems (Oliver et al., 2020a,b). Thus, the reworking of fossils and sedimentary materials may occur at a range of spatial and temporal scales, driven by relative sea-level changes, and may also be more significant than traditionally appreciated.

The tectonic setting of different coastal regions may also affect the magnitude of erosion and reworking of sediment, resulting from contrasting geomorphological processes. In many coastal sedimentary environments within tectonically highly stable settings, or in areas subject to slow rates of tectonic uplift or subsidence, older sedimentary successions and landforms may undergo significant erosion accompanying the return of similar sea level highstand conditions and the re-inundation of relict coastal landscapes, particularly in sedimentary carbonate settings. Sediment accumulation within such coastal environments will result in mixed-aged populations of sedimentary particles, rendering it difficult to derive sediment budgets and define the ‘true’ rates of bioclastic carbonate sediment production or dissolution of carbonate particles (Davies et al., 1989).

Recognition of the complexity of the stratigraphic record and the nuances of sediment reworking at a higher degree of resolution than previously possible has resulted from refinements in geochronological methods over the past thirty years. In particular, the ability to analyse samples of significantly smaller mass (microgram to milligram level) has enabled many questions to be resolved at a higher level of spatial and temporal resolution. Examples include uranium-series dating of discrete portions of fossils by laser-ablation mass spectrometry (Eggins et al., 2005), refinements in sample pre-treatment, and radiocarbon analysis by accelerator mass spectrometry (Taylor et al., 1992), with the possibility of analysing single foraminifers by amino acid racemization and radiocarbon methods (Hearty et al., 2004) as well as single grain luminescence dating of sediments (Jacobs, 2008). Questions such as the temporal homogeneity of fossil and modern foraminiferal assemblages, the duration of diastems defined by subtle down-profile changes in sediment colour, and the difference in age of fossils and their host sediments can now be addressed.

Within the perspective of higher temporal and spatial resolution in sediment analysis, this work examines the utility of amino acid racemization (AAR) in the dating of foraminifers and the application of the method for identifying remanié fossils and reworked sediments. Previous work on the analysis of fossil foraminifers is briefly reviewed to illustrate the historical development of the science and shows that technological developments in the measurement of fossil protein has led to the refinement in the scientific questions that now may be addressed by analysing smaller populations of fossils. This theme begins by examining in more detail the taphonomically active zone in a marine core from southern Gulf St. Vincent in southern Australia (Murray-Wallace et al., 2021). In addition, some of our previously unpublished results on AAR in fossil foraminifers from Tarawa Atoll from the western equatorial Pacific Ocean and from the Coorong Coastal Plain in southern Australia are also presented to examine sediment reworking over larger geographical scales than evident from a single marine core.

The application of amino acid racemization reactions to the dating of Quaternary marginal marine sedimentary successions is now well-established, having commenced with the pioneering work of Hare & Mitterer (1967). They reported a D/L value of 0.15 for the ratio of the protein amino acid L-isoleucine to its non-protein counterpart, D-alloisoleucne (D-aIle/L-Ile) in the marine bivalve mollusc Mercenaria sp., independently dated at 1000 years. A D-aIle/L-Ile value of 0.53 for an Upper Pleistocene Mercenaria shell with a minimum radiocarbon age of 40 ka was also reported. Hare & Abelson (1968) subsequently reported higher concentrations of amino acids in the D-configuration (right-handed molecule) for several amino acids with increasing fossil age, in specimens of Mercenaria sp., and negligible racemization in a modern specimen of the species. Collectively, these investigations highlighted the potential of amino acid racemization (epimerization) for dating fossil marine invertebrates with calcium carbonate matrices.

The α-amino carboxylic acids are non-volatile, organic molecules of low molecular mass (c. 100–200 Daltons) representing the structural components of proteins. Twenty protein amino acids commonly occur in living organisms, and many more are of diagenetic origin (non-protein amino acids). Amino acids are characterised by a centre of asymmetry at the α-carbon position, to which is attached an amino group (-NH₂), a carboxyl group (-COOH), a hydrogen atom (-H), and a hydrocarbon group (-R). Compositional variations between amino acids result from different R-group side chains. Structurally, all amino acids except glycine contain at least one asymmetric central tetrahedral carbon atom. Amino acids are present in all living organisms as polymers linked together through covalent peptide bonds to form enzymes and structural material such as protein (Schroeder & Bada, 1976; Kaufman & Miller, 2024).

During diagenesis, amino acids and their precursor peptides may undergo numerous reactions that include oxidation, decarboxylation, deamination, hydrolysis, and racemization (epimerization). In the protein of living organisms, amino acids are bound in peptides exclusively as left-handed molecules (L-amino acid: laevorotatory, in reference to rotating light in a left-handed direction in plane polarised light for a specific amino acid enantiomer). The exclusive presence of the L-form of amino acid in living organisms has been related to enzyme reactions inhibiting the formation of the non-superimposable mirror image, counterpart, D-enantiomer or diastereoisomer (Williams & Smith, 1977). Following the cessation of protein formation, or the death of an organism, racemization of amino acids commences. The cessation of protein formation and the death of an organism are not necessarily contemporaneous, as racemization may occur in non-metabolically active tissue and biominerals during life, such as older, formerly active growth layers of marine molluscs and corals (Goodfriend et al., 1992, 1995; Griffin et al., 2008). With the cessation of protein formation, the enzymic reactions that formerly maintained the disequilibrium condition cease (i.e., exclusively L-amino acids corresponding with an enantiomeric D/L value of D/L = 0), and amino acids then slowly and progressively interconvert from a left-handed molecule to a right-handed counterpart (D-amino acid: dextrorotatory). This process is termed racemization for amino acids with only one chiral carbon centre (i.e., enantiomers such as aspartic acid, glutamic acid, and leucine) and epimerization for amino acids with two carbon centres (e.g., isoleucine and threonine). Accordingly, the higher the D/L value for an amino acid, the older the age of the fossil in question.

The interconversion of L- to D-amino acids continues until equilibrium (D/L = 1 for enantiomers and approximately 1.3 for diastereoisomers, although a range of equilibrium values have been reported for the latter, between 1.25 and 1.4; Williams & Smith, 1977). Epimerization results in a chemically, not just optically different molecule, and is therefore readily distinguished on conventional amino acid analysers, accounting historically for the popularity of isoleucine for quantifying fossil age (Bada et al., 1986; Rutter & Blackwell, 1995). The term racemization is therefore restricted to a change in configuration only at the α-carbon of an amino acid with only one chiral carbon atom (Schroeder & Bada, 1976). The four commonly occurring diastereoisomers that undergo epimerization are isoleucine, threonine, hydroxyproline, and hydroxylysine.

In aragonite and calcite biominerals of marine invertebrates, the residual protein occurs between mineral crystals (inter-crystalline amino acids) and at boundaries between coherent blocks within crystals (intra-crystalline amino acids; Sykes et al., 1995). In the history of amino acid racemization studies, the total acid hydrolysable population of amino acids, including both the intra- and inter-crystalline fractions, have more commonly been analysed. More recently, however, the intra-crystalline amino acid fraction, although of lower concentration has also been analysed in view of the possible diffusive loss of a portion of the inter-crystalline fraction during diagenesis as well as potential sources of contamination (Sykes et al., 1995; Miller et al., 2000; Penkman et al., 2008). The intra-crystalline amino acids are isolated by removing the inter-crystalline fraction using an oxidation pre-treatment involving a concentrated solution of sodium hypochlorite or hydrogen peroxide. These analytical protocols, however, are constrained by having sufficient sample mass to ensure recovery of amino acids for analysis following pre-treatment processes. In some instances, bleaching has been shown to have little effect on reducing D/L variability in several types of carbonate biominerals including foraminifers (Millman et al., 2022).

Amino acid racemization analyses are most commonly reported for the total hydrolysable amino acids, in which the measured D/L value is based on the extent of racemization in the residual protein breakdown products for a range of molecular weights (peptides of varying mass and free amino acids). Fewer investigations have reported the extent of racemization for free amino acids (Murray-Wallace & Kimber, 1989). The measured D/L value tends to be higher for free amino acids than the total hydrolysable amino acids for a substantial portion of the reaction history. In the latter stages of the reaction history approaching equilibrium, however, the extent of racemization for both are similar. Although free amino acids show higher D/L values, their rate of racemization is slower than for the total hydrolysable amino acids. The analysis of free amino acids and total hydrolysable amino acids on the same fossil specimen provides a valuable framework for evaluating the diagenetic integrity of critically important fossils but can be time consuming.

Historically, the analysis of protein amino acids and their non-protein counterparts has involved ion-exchange chromatography for the diastereoisomer isoleucine and gas chromatography and high-performance liquid chromatography for a range of amino acids. Details of the analytical methods and sample preparation procedures for these methods are outlined elsewhere (Murray-Wallace, 1993; Kaufman & Manley, 1998). The enantiomeric amino acids commonly analysed include aspartic acid (Asp), leucine (Leu), glutamic acid (Glu), valine (Val), and the diastereoisomer isoleucine (Ile). Aspartic acid is one of the fastest racemizing amino acids offering the greatest potential for higher age resolution in determining fossil age over the past 1000 years, but its time range in application extends back to the Middle Pleistocene. Leucine, glutamic acid, valine, and isoleucine change configuration at slower rates and accordingly provide the opportunity to date significantly older materials, and depending on the diagenetic temperature history, have the potential of dating fossils older than 1–2 Ma and up to 10 Ma in Arctic settings (Wehmiller & Miller, 2000).

Amino acid racemization dating of coastal and marine sedimentary successions has predominantly involved the analysis of fossil molluscs given their larger mass, with fewer studies based on fossil foraminifers. Some of the earliest investigations of marine sediments using amino acid racemization involved the analysis of the whole sediment, representing a composite analysis of foraminifers, bioclasts and fine-grained carbonate sediment (Bada et al., 1970; Bada & Man, 1980). The stratigraphically concordant results showed an increase in the extent of amino acid racemization down-core consistent with increasing sediment age.

The separation of planktonic foraminifers from marine sediments in deep-sea cores represented the next phase of amino acid racemization investigations (Wehmiller & Hare, 1971; Bada & Man, 1980). Wehmiller & Hare (1971) reported the extent of isoleucine epimerization in the planktonic foraminifers Globorotalia multicamerata, Sphaeroidinella dehiscens, and the left coiling Pulleniantina primalis for the size fraction >74 µm (200 mg of foraminifers) from several deep-sea cores. The analytical results highlighted the potential of foraminifers to retain endogenous amino acids for prolonged periods and the possibility of assigning numeric ages for at least the past 400 ka.

Given the analytical limitations of chromatography in detecting protein residues, early investigations were based on the isolation of several hundred foraminifers to obtain the requisite sample mass for analysis by either ion-exchange or gas chromatography. Cann & Murray-Wallace (1986) reported the extent of aspartic acid, leucine, and proline racemization in the benthic foraminifer Massilina milletti from Holocene marine sediments in Core SG#198 from northern Spencer Gulf, South Australia (Fig. 1). The pilot investigation was undertaken to evaluate the potential of the species for amino acid racemization dating. The analyses by gas chromatography required up to 500 foraminifers for the requisite sample mass. In contrast, Sejrup et al. (1984) using ion-exchange chromatography were able to analyse between 20 to 100 individuals of the foraminifers Cibicides lobatulus and Elphidium excavatum (125–1000 µm fraction) from the continental shelf of Norway, but given the analytical methods used, were restricted to the measurement of the diastereoisomer isoleucine.

Amino acid racemization analyses on the large (c. up to 8 mm in diameter) benthic foraminifer Marginopora vertebralis from Wardang Island in Spencer Gulf, South Australia, required fewer individuals with up to 50 foraminifers for analysis by gas chromatography. The analyses were undertaken to evaluate the hypothesis that the species was extant on a reef flat between Wardang Island and Goose Island (Bone, 1984). The extent of alanine, proline, and aspartic acid racemization within specimens of M. vertebralis revealed they were of Last Interglacial (MIS 5e) age and were subsequently shown to have been eroded from coastal outcrops of the Last Interglacial Glanville Formation (Belperio & Murray-Wallace, 1984; Murray-Wallace & Belperio, 1994). The species is a Late Pleistocene index fossil in southern Australia.

Advances in chromatography, based on reverse phase, high-performance liquid chromatography (Kaufman & Manley, 1998) has enabled significantly smaller samples to be analysed. Hearty et al. (2004) reported results of amino acid racemization analyses on single tests of the large planktonic foraminifer Pulleniatina obliquiloculata from the Queensland continental margin. Based on an analysis of 462 Pulleniatina tests, they noted that the genus was relatively resistant to dissolution and that the tests reliably retained endogenous protein residues (2–4 nmol mg⁻¹ of test calcium carbonate) enabling age assessments back to at least 500 ka.

In view of the small sample mass, Kaufman et al. (2008) undertook amino acid racemization analyses on multiple individuals and up to 50 specimens of the planktonic foraminifer Neogloboquadrina pachyderma from Arctic Ocean deep-sea cores, reporting results for aspartic and glutamic acids. They found that by analysing up to 50 foraminifers, there was sufficient sample mass to analyse replicate subsamples, identify outliers, and confidently determine mean D/L values for a population of foraminifers. Subsequent investigations have reported results for the planktonic foraminifers Neogloboquadrina pachyderma, Globigerinoides obliquus, and G. sacculifer (Kaufman et al., 2013), as well as the benthic species Cassidulina neoteretis (West et al., 2019) and Cibicidoides wuellerstorfi (West et al., 2023) from Arctic Ocean deep-sea cores. Collectively, these works have illustrated the sophisticated application of amino acid racemization for developing geochronological frameworks for deep-sea sediments and have provided relatively continuous time-series to evaluate genus-specific rates of racemization in different foraminiferal taxa.

Blakemore et al. (2015) examined the extent of amino acid racemization in individual tests of the benthic foraminifer Elphidium crispum from a succession of aeolian dune facies of the Pleistocene Bridgewater Formation in the form of a series of coastal barriers between Mount Gambier and Port MacDonnell in the southern-most portion of the Coorong Coastal Plain, southern Australia. In a comparison of the extent of racemization in aeolian calcarenite samples (‘whole-rock’ analyses) with that in Elphidium crispum, the results showed that the foraminifers retained endogenous amino acids for longer periods of geological time than the calacarenite host sediments (Blakemore et al., 2015, their fig. 3). The calcareous sediments showed evidence of in situ leaching of residual protein residues, resulting in lower D/L values for the total hydrolysable amino acids. The most likely mechanism is the diffusive loss of the lower molecular weight peptide residues and free amino acids from the bioclastic sediment grains, resulting in lower amino acid D/L values based on the remaining higher molecular weight components, which are characterised by lower extents of racemization. Dismal Range, the oldest coastal barrier sampled in the transect, was dated at 933±145 ka based on the extent of amino acid racemization in E. crispum and correlated with MIS 23 (Blakemore et al., 2015).

Collectively, these works have highlighted the strong potential amino acid racemization for deriving chronologies using fossil foraminifers. A critical element in the use of foraminifers is selecting species that retain endogenous amino acids for geologically prolonged intervals, while not taking up amino acids from the surrounding micro-environment during diagenesis.

In principle, the amino acid racemization method can be applied to foraminifers in a similar manner to fossil molluscs. The method can be used for distinguishing modern from fossil foraminifers based on amino acid D/L values. Amino acid D/L values can be used to derive relative or numeric ages, the latter by calibration with an independent geochronological method, such as U-series or radiocarbon dating. These independent geochronological techniques enable calculation of racemization rates, and accordingly, amino acid racemization ages to be determined on additional fossils of unknown age. Foraminifers are particularly suited for the dating of Quaternary sediments, as their longevity (1–24 months; Murray, 1991) means that any age offset between fossils and the time of deposition is not resolvable by geochronological methods, unless the foraminifers are reworked. In this work, amino acid racemization has been applied to the dating of benthic foraminifers with relatively short life durations, compared with marine invertebrates such as molluscs, which may live for several years. Amino acid racemization does not, however, have the age resolving power to identify such subtle differences in age when applied to timescales of thousands of years. The longevity of the foraminifers referred to in this work include Elphidium macelliforme and E. crispum (12 months), Amphistegina sp. (4–12 months), Lamellodiscorbis dimidiatus (<24 months?), and Marginopora vertebralis (12–24 months; Hallock, 1985; Murray, 1991).

In this work, amino acid racemization analyses were undertaken largely following the analytical protocols established by Kaufman & Manley (1998). Amino acid D/L values are reported for aspartic acid (ASP) and glutamic acid (GLU), two of the more reliably resolved amino acids in view of their generally higher concentrations in fossils. Aspartic acid is one of the fastest racemizing amino acids providing the potential for a higher level of age resolution. Glutamic acid racemizes at an intermediate rate, so comparisons of the two acids assist in evaluating results. Individual or multiple tests (≤10 foraminiferal tests) were initially pre-treated with dilute H₂O₂ to remove surface contaminants followed by a brief (≤1 minute) ultrasonic clean in high purity water and subsequently digested in 6 M HCl. In view of the small sample mass (e.g., 0.028 mg per test for Elphidium macelliforme), foraminiferal tests were not given a 2 M HCl acid etch before complete digestion in HCl. Analyses were undertaken on the total hydrolysable amino acids involving heating for 22 hours at 110°C in 6 M HCl. Following hydrolysis, the amino acids were derivatized using 0-phthaldialehyde (OPA) together with the chiral thiol, N-isobutyryl-L-cysteine (IBLC) to yield fluorescent diastereomeric derivatives of the chiral primary amino acids. The determination of amino acid D/L values was based upon reverse phase, high-performance liquid chromatography (RP-HPLC) using an Agilent 1100 liquid chromatograph with a Hypersil C-18 column and auto-injector.

The basis for the uncertainties associated with the amino acid D/L values are outlined in each of the reported case studies. Outliers of D/L values were not rejected, as the principal objective of the case studies reported herein, is to evaluate the magnitude of reworking within the sedimentary successions. In addition to the AAR analyses on foraminifers, results are also reported for International Laboratory Comparison (ILC) samples to evaluate the integrity of the AAR data reported herein, which are based on the analysis of powders of fossil molluscs (Wehmiller, 2013; footer of Table 1).

The results of some earlier gas chromatography analyses are also reported in this work relating to a marine vibracore from southern Gulf St. Vincent, southern South Australia (next section). The analytical methods are set out in more detail elsewhere (Murray-Wallace, 1993). In summary, following the hydrolysis conditions set out above, the analyses of the protein residues were based on gas chromatography of N-pentafluoroproprionyl-D, L-amino acid isopropyl esters using a Hewlett-Packard model 5890A gas chromatograph with a flame ionisation detector and coiled, fused silica capillary column (25 m length) with the stationary phase Chirasil-L-Val.

A potential complication in deriving records of geologically recent environmental change that spans, approximately, the past 1000 years, is the role of the taphonomically-active zone (TAZ) in modifying sediment profiles, particularly the physical disturbance of sedimentary particles resulting in the down-profile loss of chronological fidelity (Kidwell & Flessa, 1996; Kidwell, 2013). The limitations of applying the radiocarbon method to this time interval has compounded this problem. Pronounced fluctuations in radiocarbon production in the upper atmosphere due to cosmic ray flux, has made the calibration of radiocarbon ages more challenging for this time interval, as well as the difficulty of distinguishing finite ages from ‘modern’ (200 years before 1950 AD; Sternberg, 1992; Bradley, 2015; Fallon & Murray-Wallace, 2020). The incorporation of remanié fossils into sediments within the TAZ presents a further difficulty when radiocarbon dating of fossil remains is contemplated, as some fossil remains may appear indistinguishable from modern individuals. With the passage of time and the upward movement of the TAZ with continued deposition, these chronological discordances become less evident, in view of the reduced age resolution of the geochronological methods over timescales of ka to Ma. Depending on the ages of the successions in question, it may not be possible to resolve the nuances of mixed-age populations.

The taphonomically-active zone (TAZ) refers to the uppermost portion of the sediment column, from the sediment-water interface downwards to the lower limit of sediment disturbance. Below the base of the TAZ, the sediment has a significantly lower likelihood of being disturbed and early diagenesis will begin to modify the sediment, unless the sediment is disturbed by exceptional events. The depth of the TAZ varies with different sedimentary environments and the intensity of the physical processes responsible for sediment disturbance and redeposition. In marine environments, the reworking of shallow-water continental shelf sediments in part relates to the wave energy and the capacity of infaunal biota to disturb sediment. As more sediment is deposited on the sea floor (sediment-water interface), the base of the TAZ progressively moves up through the sediment profile, isolating underlying sediment from disturbance, and thus becoming part of the fossil record.

Investigations of the submarine Quaternary stratigraphy of Spencer Gulf and Gulf St. Vincent (Fig. 1), two triangular shaped, fault-bound shallow marine basins in semi-arid, southern South Australia, involved the collection of several hundred marine vibracores (location of the submarine cores are shown in Gostin et al., 1984; Cann et al., 1993; Murray-Wallace et al., 2021). The vibracores commonly obtained undisturbed profiles of the seafloor sediment less than 5 m thick. Depending on their modern physical setting, the cores either chronicled a record of sedimentation during the last glacial cycle from present day submarine contexts of the gulfs (Hails et al., 1984) or Holocene peritidal sedimentation along the margins of the gulfs (Belperio et al., 1984). The TAZ was identified in all the cores examined with a thickness of generally 30–50 cm (Nicholas, 2012). In the initial description of cores, the sediment was referred to as “modern sea floor sediments”. Commonly, primary bedding within these sediments was indistinct or destroyed by bioturbation; traces of seagrass roots were evident in the shallower water facies, and a subtle colour change signified the boundary between the base of the TAZ and the underlying “fossil” sediment. The physical reworking of modern subtidal sediments within the South Australian gulfs is a function of turbulent water during storms and the action of prawns, crabs, and holothurians (Shepherd & Sprigg, 1976).

Core SV#23 from southeastern Gulf St. Vincent (35°18.10S, 138°14.15E; Fig. 1), collected from a present water depth of 40 m, is discussed here to highlight issues related to the TAZ in shallow marine sediments in general. The core was investigated as part of two larger studies examining post-glacial and Holocene sea-level rise following the Last Glacial Maximum (Cann et al., 2006) and Late Pleistocene interstadial sedimentation and relative sea-level changes within Gulf St. Vincent (Cann et al., 1993; Murray-Wallace et al., 2021). As the taphonomically active zone was not the primary focus of these works, the topic is explored in more detail herein, as it relates to wider issues of the reworking of fossils in shallow marine environments. In the initial description of Core SV#23 immediately following collection, the base of the TAZ was defined at 30 cm below the modern sea floor. The TAZ sediment comprised light olive grey (5Y6/1), poorly sorted muddy sand. Entire bivalves, including the beaked mussel Brachidontes sp. and the stepped Venerid cockle Katelysia scalarina, abundant foraminifers, and encrusting algae are common within the sediment. Radiocarbon dating of molluscs from two depth intervals within the modern seafloor sediments illustrates the effects of reworking within the TAZ. Specimens of the cockle Katelysia scalarina at 18–20 cm (unavailable for AAR analysis owing to use in conventional, liquid scintillation radiocarbon dating) from the top of the core (four disarticulated valves with a mass of 19.3 g following HCl dissolution for radiocarbon pretreatment for liquid scintillation counting) were dated at 5040±270 yr cal BP (SUA-2710), while an articulated specimen of Bassina sp. at 22–24 cm was dated at 6870±250 yr cal BP (OZF-049; Murray-Wallace et al., 2021; Table 1). The shell radiocarbon ages are in sidereal years with 2-sigma uncertainties and include a correction for the marine reservoir effect of 450±35 years for southern Australian ocean surface waters (Gillespie & Polach, 1979). The ages relate to the growth of the subfossil molluscs during the early post-glacial Holocene sea-level highstand, but their incorporation within TAZ sediments is most likely from a lateral source within the shallow gulf environment.

The extent of amino acid racemization measured collectively in 5–10 tests of the foraminifer Elphidium macelliforme (to obtain the requisite sample mass for reverse phase, high performance liquid chromatography; RP-HPLC) was determined based on the sampling of 2-cm thick sediment slices down the entire core barrel of Core SV#23 (Murray-Wallace et al., 2021; Fig. 2). Within the vertical profile of the TAZ in Core SV#23 (top 30 cm), a consistent down-core trend in the extent of amino acid racemization is evident with an average glutamic acid D/L value of 0.067±0.009 and 0.183±0.025 for the faster racemizing aspartic acid (n = 15 analyses; Fig. 2; Table 1; Appendix 1). The 1-sigma uncertainty in D/L values is based on up to five subsamples (A–E in Table 1) of 5–10 foraminifers from the total population of foraminifers within each 2-cm slice of sediment down-core. The down-core variability in D/L values is not significantly different to that noted for replicate fossils of the same age from a range of sedimentary deposits of Holocene to Late Pleistocene age (Murray-Wallace, 1995). The D/L values for modern Elphidium crispum for glutamic and aspartic acid are 0.045±0.011 and 0.080±0.015, respectively (Blakemore et al., 2015), and are significantly different to the average D/L values determined for the TAZ in Core SV#23. Thus, the Elphidium macelliforme within the TAZ of Core SV#23 are at most only a few hundred years older than modern but are of a common age within the 30-cm thick sediment profile. In contrast, as noted above, based on the radiocarbon dating of a specimen of Bassina sp., at 22–24 cm, the shell is significantly older and dated at 6870±250 yr cal BP (OZF-049). Glutamic acid and aspartic acid D/L values of 0.154±0.016 and 0.381±0.008, respectively, for the Bassina sp. by RP-HPLC reaffirm the radiocarbon age showing that the shell is significantly older than the E. macelliforme specimens within the TAZ. The D/L value for aspartic acid in the Bassina sp. by RP-HPLC also accords with the value determined by gas chromatography (Table 1).

The amino acid racemization data reveal that the D/L values for the upper-most 30 cm of Core SV#23 extend down-core for a further 18 cm, indicating that the TAZ extends to 48 cm, within sediment logged as “Restrictive transgressive marine/lagoon facies” and originally considered to represent a distinctively different lithofacies (Fig. 2). Although an indistinct lithological change was noted in the core at 30–35 cm, the AAR data reveal that at 48–50 cm down-core, a stepwise increase in extent of racemization is evident. Accordingly, the TAZ extends to 48–50 cm rather than at a boundary of 30–35 cm beneath the seafloor. These results highlight the potential of AAR for discriminating age differences in sediments where subtle diastems are present but not evident in conventional lithostratigraphical descriptions of sediment packages.

The extent of aspartic acid racemization in other shell specimens previously analysed by gas chromatography GC) from the TAZ in Core SV#23 reveal that several are significantly older than the foraminifers within the host sediments (Cann et al., 2006; Table 1). The differences in the extent of racemization far exceed potential differences related to a genus-effect on racemization rate. A confounding point is that the taphonomic state of shell preservation is not necessarily a reliable indicator of fossil age. For example, the articulated specimen of Bassina sp. at 12–14 cm has a D/L aspartic acid value of 0.466±0.046 (UWGA-670) consistent with a Late Pleistocene age (Last Interglacial; MIS 5e), while a disarticulated specimen of the same genus has a similar D/L value of 0.443±0.037, revealing that articulated specimens do not necessarily indicate that they are in situ. Both specimens were reworked into the TAZ with the most likely source being a submarine outcrop of the Last Interglacial (MIS 5e) Glanville Formation (Belperio et al., 1995). Collectively, the results from Core SV#23 highlight that the physical reworking of larger fossil shells is likely to be more common than for foraminifers, as their physical size and hydrodynamic character renders them more easily transported.

The lower portion of Core SV#23 reveals two older packages of sediment within the core, of Late Pleistocene age and beyond the range of radiocarbon dating (>50 ka). Based on the average extent of racemization down-profile, a step change is evident in amino acid D/L values indicating two chronostratigraphically distinct sediment packages.

Situated in the western equatorial Pacific Ocean, Tarawa Atoll comprises a series of connected reef islands forming a triangular atoll up to 40 km long and 25 km wide (Fig. 3). On the western side of the atoll, a shallow lagoon with a maximum water depth of 24 m is partly protected from the open ocean, by a series of small reef islands. Tarawa island is part of the Republic of Kiribati, a country consisting of 33 islands (coral atolls) forming three principal groups: Gilbert, Phoenix, and Line Islands.

Besides a longer history of reef subsidence, modern Tarawa Atoll has formed since the culmination of the post-glacial marine transgression some 8 ka (Biribo, 2012). Shallow drilling to 30 m below the ground surface identified four lithological units, which in ascending order include a basal, leached limestone of last interglacial age (MIS 5e) dated at 125±9 ka (Marshall & Jacobson, 1985) overlain by a poorly consolidated coral unit, and in turn overlain by unconsolidated sediment and cemented conglomerate (cay rock). The upper leached portion of the limestone unit represents a subaerial exposure surface, which would have formed during the Last Glacial Cycle at a time of lower sea levels between MIS 5d stadial lowstand and the LGM. The subaerial exposure surface characteristically occurs at 11–30 m below the modern land surface.

Modern sediments of Tarawa Atoll are derived from several sources that include the ocean reef flat, ocean beaches, channels, lagoon reef flat, and the lagoon floor and beaches. As with the older sediment, the modern sediments comprise bioclastic, skeletal carbonate sand. Sediment grains coarser than 0.25 mm in diameter include coral, foraminifers, marine molluscs, the octocoral Halimeda, and algal debris (Ebrahim, 1999). Foraminifers are a major contributor to the sediment bioclasts of many Pacific reef islands with up to 50% of sediment on Tarawa Atoll represented by the genera Amphistegina, Baculogypsina, and Calcarina.

Given the concern about the environmental effects of sea-level rise on Tarawa Atoll, a program of research was undertaken to examine the rate of sedimentary carbonate production and how the islands are currently responding to coastal landscape change (Biribo, 2012). In particular, the susceptibility of different sub-environments to coastal erosion is central to the understanding of how the Tarawa reef complex may change in the future.

One of the challenges to quantifying sedimentation rates on the Tarawa reef islands is isolating a sediment component for dating that may reliably define the age of the sediment production and accordingly, provide insights on whether the reef islands are increasing in area or experiencing erosion. The absence of quartz precludes the application of luminescence methods for defining the time of sedimentation within the reef islands. Given that the reef island sediments are dominated by bioclastic skeletal carbonate sands, ‘whole-rock’ sediment analyses by either radiocarbon or amino acid racemization may yield ages that pre-date the time of deposition. ‘Whole-rock’ analyses provide an ‘average’ age based on the integrated effect of the different residual radiocarbon activities or the differential extents of amino acid racemization for individual bioclasts within a sediment. Thus, depending on the residence time of individual sediment particles (bioclasts), the derived ages may significantly pre-date the time of sedimentation. To overcome this difficulty and improve the age resolution of these analyses, in this work an assessment of sediment age is based on the analysis of individual specimens of the foraminifer Amphistegina within the sediment samples. This analytical strategy involved the radiocarbon analysis by accelerator mass spectrometry of individual specimens of Amphistegina and the subsequent amino acid racemization analysis of the phosphoric acid residue from CO₂ evolution for each radiocarbon sample. This analytical protocol (Murray-Wallace & Bourman, 1990) permitted the direct radiocarbon calibration of the extent of amino acid racemization in the same fossil Amphistegina, providing a basis to assign numeric ages to Amphistegina specimens of unknown age based on a knowledge of the kinetics of amino acid racemization. Radiocarbon measurements were undertaken at the Australian Nuclear Science and Technology Organisation (ANSTO), using the Australian National Tandem Research Accelerator (ANTARES) facility. Amino acid racemization analyses were completed at the Amino Acid Geochronology Laboratory at the University of Wollongong and follow the analytical protocol set out elsewhere (Kaufman & Manley, 1998). In view of the difficulty of re-concentrating the acid residue from the radiocarbon sample preparation of single foraminifers, and the subsequent hydrolysis of the residual protein in 6 M HCl, in this suite of analyses, only one injection was undertaken in each of the RP-HPLC AAR analyses. Current mean annual air temperature for Tarawa Atoll, a relevant consideration for the interpretation of the amino acid racemization data, is 28.3°C, based on unpublished data from Kiribati Meteorological Services.

Seventeen individual specimens of Amphistegina underwent paired radiocarbon and amino acid racemization measurements (Table 2). The radiocarbon ages include a correction for the marine reservoir effect (ΔR = 39±56 years) based on a value for Christmas Island, Kiribati. The ages expressed in sidereal years with 2-sigma uncertainties for the individual Amphistegina sp., range from 5125±210 yr cal BP (OZL964) to Modern (post-1950 AD: OZL963, but the modern individual analysed was not alive at the time of collection; Table 2). The extent of aspartic acid racemization ranges from a D/L value of 0.12 in the radiocarbon-determined ‘Modern’ (post-1950 AD) Amphistegina to a D/L value of 0.44 in the second oldest specimen dated by radiocarbon at 4505±315 yr cal BP (OZN137), while the extent of racemization in the slower racemizing glutamic acid ranges from a D/L value of 0.05 in the Modern specimen to a D/L value of 0.24 in the second oldest specimen. A high level of association is noted in the measured extent of amino acid racemization and the radiocarbon ages for the individual Amphistegina sp., specimens with r²-values of 0.8409 and 0.79194 for aspartic acid and glutamic acid respectively (Fig. 4).

An auger transect across the reef island of Notoue (Fig. 3) was undertaken to evaluate the age of the bioclastic sediment within the reef island and determine if the sediment of the lagoon beach (western side of the atoll) and the ocean beach (eastern side of the reef island) comprised modern or relict sediment and accordingly, if these beaches are accreting or eroding. The modern beach sediments were collected from the upper-most 2 cm of beach face sediment. Aspartic acid D/L values for single tests of Amphistegina sp., from the ‘modern’ ocean beach sediments of Notoue range from 0.06 to 0.29 with measured aspartic acid D/L values of 0.06, 0.07, 0.09, 0.1 and 0.29, and a mean D/L value of 0.12±0.01 (1-sigma uncertainty). The two lowest D/L values represent modern individuals (alive at the time of collection) while the highest aspartic acid D/L value of 0.29 indicates at face value, an age >1000 years, with a numeric age of 1700±260 yr, based on a model of apparent parabolic racemization kinetics (Mitterer & Kriausakul, 1989), an aspartic acid D/L value of 0.06 for a Modern Amphistegina sp., and a D/L value of 0.44 in an Amphistegina test with a radiocarbon age of 4505±315 yr cal BP (Table 2). The uncertainty in the AAR age accounts for a 1°C uncertainty in diagenetic temperature. Thus, the beach sediment comprises a mixed age population of Amphistegina sp. but is dominated by younger foraminifers.

Aspartic acid D/L values for Amphistegina sp. from the ‘modern’ lagoon beach sediments of Notoue range from 0.13 to 0.35 (measured D/L values of 0.13, 0.14, 0.15, 0.21, 0.20, 0.22, 0.24, 0.25, 0.27, 0.28, 0.29, 0.31, and 0.35; with a mean D/L value of 0.23±0.07). These D/L values are consistently higher than for the modern ocean beach and reveal an older population of bioclasts in the lagoon beach sediments at Notoue. Collectively, these results imply that the modern ocean beach is receiving younger sediments and foraminifers and is likely to be accreting, while the lagoon side of Notoue is receiving less modern sediment.

The auger transect across Notoue reveals that the bioclastic sediments sampled from the base of each auger hole (1 m below the ground surface) all show a similar extent of aspartic acid racemization with a mean D/L value of 0.33±0.02 (n = 4), indicating that the foraminifers analysed, and by inference the host sediments, are of a common age and older than modern ocean and lagoon beach sediments (Fig. 5). The uncertainties for the aspartic acid D/L values from each site are based on 6–9 AAR analyses on individual tests of Amphistegina sp., from each pit (Table 3; Appendix 2). Although these D/L values are in accord with the uniformity of the independently derived radiocarbon ages on single Amphistegina sp. from each pit, the amino acid D/L values imply significantly older numeric ages than the radiocarbon analyses (Fig. 5; Table 3). We have not been able to resolve this discord in ages and note that these observations are at variance with the wider data set for paired radiocarbon and amino acid racemization results (Table 2, Fig. 4). The individual Amphistegina sp. analysed from the Notoue reef island transect were not analysed by both methods, and different foraminifer specimens were analysed by each geochronological method. The aspartic acid D/L values for the individual Amphistegina sp., from within each pit and between the four pits are highly concordant (Table 3, Appendix 2). We have reported these results, however, to emphasize the need for extreme caution in sampling from open pits. The possibility that the radiocarbon-dated foraminifers represent intrusive contaminants during sampling cannot be excluded.

The Coorong Coastal Plain is part of a major cool-water, temperate carbonate sediment province in southern Australia (James & Bone, 2011; Murray-Wallace, 2018; Fig. 6). The coastal plain which extends from near the River Murray mouth, southeast to the western-most coastline of Victoria, is a low-gradient coastal plain that has been differentially uplifted throughout the Pleistocene (Fig. 6). Up to 21 coastal barriers occur across the coastal plain but are not all encountered in a single line of section. The coastal plain extends up to 100 km landwards from the present coastline. Luminescence, amino acid racemization, electron spin resonance and uranium-series dating, have established that the individual coastal barriers formed predominantly during Pleistocene interglacials, with some barriers such as Robe Range having formed during the warm interstadials of the Last Glacial Cycle (MIS 5c, 105 ka and MIS 5a, 85 ka). An overview of numerous geochronological investigations that have provided a timeframe for the Quaternary evolution of the coastal plain is presented elsewhere (Murray-Wallace, 2018). The barrier sediments comprise mixed quartz-carbonate, calcarenites mapped as Bridgewater Formation (Belperio, 1995). The sand-sized bioclasts characteristically include abraded mollusc fragments, coralline algae, bryozoans, and foraminifers in association with subordinate quartz and heavy minerals (Murray-Wallace et al., 2001).

The Bridgewater Formation refers to predominantly windblown successions of Pleistocene calcarenite (aeolianite, also referred to as dune limestone) that occurs extensively along the high-energy open ocean coastlines of southern Australia (Belperio, 1995). The formation was originally defined at Bridgewater Bay in western Victoria in reference to three superposed units of aeolianite of presumed early to late Pleistocene age (Boutakoff, 1963). The sedimentary successions of the Bridgewater Formation characteristically comprise trough and planar cross-stratified aeolian dune facies, which interfinger with beach and back-barrier, estuarine-lagoon sediments, the latter used for defining palaeo-sea level (Murray-Wallace et al., 1999). Besides the Coorong Coastal Plain, the Bridgewater Formation occurs extensively along the open ocean coastlines of Kangaroo Island, southern Yorke Peninsula and Eyre Peninsula in southern Australia (Fig. 1).

As more geochronological results became available for the ages of the coastal barriers of the Bridgewater Formation, a question arose about the rate of bioclastic calcium carbonate formation and the extent of incorporation of older bioclasts into younger sediments. While the Lacepede and Bonney shelves (Fig. 1) are regions of known high rates of temperate carbonate sediment production (James & Bone, 2011, 2007), the question arose as to whether the magnitude of carbonate production was sufficient to form coastal barrier landforms contemporaneously with carbonate production on the adjacent shelves within each of the interglacials represented across the coastal plain. To address this question, a pilot investigation was undertaken to quantify the extent of amino acid racemization in single specimens of the benthic foraminifer Lamellodiscorbis dimidiatus in modern beach sediments of the Coorong Coastal Plain (Limestone Coast). The rationale for selecting the foraminifer Lamellodiscorbis dimidiatus is that it appears to be robust to mechanical abrasion with its tests surviving long-term cycles of sedimentation involving deposition and subsequent reworking of sediment and redeposition (Ryan et al., 2020). The species also appears to retain endogenous protein residues for protracted periods (≥400 ka; Lachlan, 2011) and for biocenose individuals, more accurately defines the time of sedimentation compared with calcareous bioclastic sand, which may comprise skeletal grains with a range of inherited ages.

In this pilot investigation, modern beach sediments were sampled from the beach face of four beaches in a southerly transect along the modern coastline of the coastal plain (Fig. 6). The sediment samples were collected from the swash zone of each beach, at a depth of 10–20 cm below the beach surface. The beaches sampled include The Granites (139°51′13.8″E, 36°39′35.0″S), adjacent southern Lacepede Bay, and three beaches adjacent to the Bonney Shelf in the southern portion of the coastal plain. The latter include sediment from West Beach at Robe (139°44′36.3″E, 37°10′1.1″S), Canunda Beach (140°13′17.1″E, 37°39′31.7″S), and Cape Douglas Beach (140°34′51.5″E, 38°1′45.5″S; Fig. 6). Single tests of Lamellodiscorbis dimidiatus, preferably ≥500 µm in size, were selected using a binocular microscope with 170 individuals analysed. Amino acid racemization analyses followed the analytical methods outlined earlier in this work. In view of the low recovery of amino acids in the analysis of single tests of Lamellodiscorbis dimidiatus, each plotted datum is based on only one injection, and accordingly, no uncertainties are plotted in Figure 7. Current mean annual air temperature along the seaward portions of the coastal plain range from 15.4°C at Meningie to 14.7°C at Robe and 13.9°C at Cape Northumberland.

Results for single specimens of Lamellodiscorbis dimidiatus reveal that there is a similar trend in the extent of aspartic and glutamic acid racemization in L. dimidiatus for the four modern beach sediment samples (Fig. 7). The raw data are provided in Appendix 3. The D/L values for the Holocene tests show an elongate pattern within the plots, reflecting the nature of racemization kinetics in early diagenesis. Following the death of an individual and the cessation of protein formation, racemization occurs at a rapid initial rate, which is particularly evident in Holocene fossils (Murray-Wallace & Kimber, 1987). With time, the rate of racemization slows to the point that the rate of racemization proceeds at approximately 10% of the rate evident within fossils of Holocene age, reflecting the parabolic nature of racemization kinetics (Clarke & Murray-Wallace, 2006). As a result, the age resolving power of AAR decreases for Pleistocene fossils with discrimination of ages at the scale of successive interglacials. All four modern beach sediment samples have populations of D/L values consistent with a late Pleistocene age and accordingly represent reworked tests. The reworked tests show a pattern of sediment recycling at a regional scale along the Limestone Coast. Many of the reworked tests are either brown-stained or corroded, while pristine or transparent tests are categorized as reworked based on the extent of amino acid racemization, indicating that the taphonomic state of foraminiferal tests is not a reliable indicator of fossil age.

In a previous investigation of the coastal barrier landscape evolution of the Late Pleistocene Robe Range in southern South Australia, Hidayat et al., (2023; Fig. 6) identified distinct groups of glutamic acid D/L values in L. dimidiatus corresponding with increasing fossil age. This observation is consistent with the earlier mapping of Schwebel (1978, 1984) who suggested that Robe Range represents a composite structure, comprising three components, two Late Pleistocene units corresponding with MIS 5c (105 ka) and 5a (85 ka) and capped by a Holocene succession of dunes younger than 8 ka associated with the culmination of post-glacial sea-level rise. Glutamic acid D/L values of 0.25–0.30 correspond with Late Pleistocene interstadials MIS 5c (105 ka) and MIS 5a (85 ka) based on independent optically stimulated luminescence ages of the aeolianites within Robe Range. It is not possible to distinguish these two Marine Isotope Stages based on the extent of amino acid racemization alone. Last interglacial L. dimidiatus showed glutamic acid D/L values of 0.31–0.40, which are in accord with values reported by Ryan et al. (2020) from calcarenites from the northern-most Coorong Coastal Plain.

In this work, the AAR results for single specimens of L. dimidiatus from the modern sediment at The Granites beach (Figs. 6, 7a) reveal that of the 45 individual foraminifers analysed, five (11%) are of last interglacial age (MIS 5e) based on previous analyses from the Woakwine Range I of 125 ka age (Hidayat, 2022; Hidayat et al., 2023). Eleven L. dimidiatus are of interstadial age (MIS 5c or MIS 5a), 28 of Holocene age (≤8 ka), and only one modern specimen is represented (Fig. 7a). In the West Beach sample near Robe (Fig. 6), of the 44 foraminifers analysed, five individuals are of MIS 5e age, 29 of interstadial age (MIS 5c or MIS 5a), five Holocene, and three modern specimens (Fig. 7b). Canunda Beach (Fig. 6) shows a similar pattern, with seven individuals of MIS 5e age, 30 of interstadial age (MIS 5c or 5a), five Holocene L. dimidiatus, and two modern individuals, based on the analysis of 44 individuals (Fig. 7c). Last Interglacial age L. dimidiatus are not represented at Cape Douglas beach, and of the population of 37 foraminifers analysed, 17 are of interstadial age (MIS 5c or 5a), 18 are Holocene, and two modern foraminifers.

The groupings of L. dimidiatus noted in the bivariate plots of glutamic acid and aspartic acid above can be directly related to the coastal landscape settings of the four modern beach sediment samples. As noted earlier, the Coorong Coastal Plain is a differentially uplifted coastal landscape with the elevation of Last Interglacial (MIS 5e) intertidal shelly facies progressively rising in a southeasterly direction over a horizontal distance of 300 km from 2 m to 18 m APSL, within a far-field geographical province that has experienced the same Late Quaternary glacio-eustatic sea-level history. The coastal plain shows the highest level of uplift in the south at a Pleistocene-Holocene volcanic centre near Mount Gambier with Last Interglacial shelly facies at 18 m above present sea level (APSL: Murray-Wallace et al., 1996; Fig. 6). In the north-western-most sector of the coastal plain, subsidence is evident in the River Murray mouth region due to basin subsidence of deltaic successions with Last Interglacial coastal sandflat facies occurring at 1 m APSL on northwestern Hindmarsh Island (Murray-Wallace, 2018). The differential uplift and subsidence of the coastal plain explains the present spatial distribution of the youngest two barriers of the coastal plain, the Woakwine Range I and II (predominantly MIS 5e with an inner core of MIS 7 age representing Woakwine II) and Robe Range III, II, and I (MIS 5c, 5a and 1, respectively). Accordingly, in the area of most pronounced uplift in the south, within the Mount Gambier region, Robe Range is exposed at the modern coastline while Woakwine Range I is between 5 to 8 km landwards from Robe Range. In contrast, farther north near Kingston, Robe Range does not occur along the modern coastline and only appears as a few subdued, elongate islands less than 1 km long within the Coorong Lagoon (Cann & Murray-Wallace, 2012; Fig. 6). In the northern sector of the coastal plain, the aeolianites of Woakwine Range I crop out along the landward side of the Coorong Lagoon. These contrasting spatial distributions explain the availability of older aeolianite sediment to nourish modern beaches through coastal erosion.

At The Granites, the dominance of Holocene tests of L. dimidiatus relates to the sediments of southern Younghusband Peninsula representing the principal sediment source. Younghusband Peninsula is a large-scale Holocene coastal barrier up to 194 km long and 1 to 3 km wide in cross-section that formed following the culmination of post-glacial sea-level rise some 7 ka. As Robe Range does not crop out in this region owing to the overall slower rate of coastal emergence in the area of the sample site, the MIS 5c and 5a foraminifers are most likely associated with either an earlier, post-Last Glacial Maximum phase of coastal erosion on the inner Lacepede Shelf, or more recent erosion of Robe Range to the south near Robe and the transport of a reworked sediment component, north by longshore drift.

At West Beach, Robe (Fig. 6), the dominance of interstadial sediment relating to MIS 5c and 5a are more likely to relate to the recent erosion of Robe Range along the modern rocky, cliffed coastline (Fig. 8). The high wave energy of the modern coastline is responsible for significant coastal erosion of the Bridgewater Formation, providing an immediate source of sand to nourish West Beach. The inclusion of Last Interglacial (MIS 5e) L. dimidiatus may have also been derived from the erosion of Robe Range and accordingly represents foraminifers that effectively have already been through a complete cycle of sediment erosion and re-deposition. In the Robe area, the former seaward side of the Last Interglacial Woakwine Range occurs up to 6 km inland from the modern coastline and is not actively eroding. The relict coastal landform is capped by a laterally persistent calcrete, representing an armoured landscape.

The dominance of interstadial age (MIS 5c and 5a) L. dimidiatus in the sediments from Canunda Beach is also directly associated with coastal outcrops of Robe Range, particularly between Cape Banks Port MacDonnell where the dune range is undergoing active erosion (Fig. 6). The L. dimidiatus of Last Interglacial age, as with West Beach, Robe, may also relate to reworked individuals derived from Robe Range.

Cape Douglas Beach (Fig. 6) is close to the area of maximum uplift of the Coorong Coastal Plain. Along this sector of coastline, vertical cliffs up to 20 m high and 53 near-shore islands representing erosional remnants of Robe Range highlight the vigorous coastal erosion responsible for the generation of ‘new’ sediment to nourish the modern beaches. Up to 2 km of coastal cliff retreat has occurred within the region over the past 7000 years, since the culmination of the post-glacial marine transgression, accounting for the erosional reworking of bioclastic sand and deposition on beach ridge plains along this coastline (Oliver et al., 2020b).

A surprising observation is that the four ‘modern’ beach sediment samples contain very few tests of modern L. dimidiatus. Only one test was noted at The Granites, three at West Beach, two at Canunda Beach, and two at Cape Douglas Beach. This may relate to a change in the nature of the modern offshore environments and a reduction in suitable habitats associated with the high degree of coastal erosion.

Historically, the application of amino acid racemization to foraminifers has involved establishing time frameworks for sedimentation in marine and coastal sedimentary environments, particularly in relative and numerical age assessments of the host sediments (Sejrup et al., 1984; Hearty et al., 2004; Kaufman et al., 2013). This paper, however, has focused on applying the method to identify reworked fossils and sediments, principally in sedimentary carbonate environments. Given their short lifespan (c. 2–24 months), foraminifers represent a valuable means of dating sedimentary deposits in contexts where they are not obviously reworked, as they represent a form of skeletal grain that more closely approximates the time of deposition, rather than bioclastic grains derived over longer periods from the physical breakdown of marine molluscs.

The reworking of fossils may occur over a wide range of spatial and temporal scales. At the smallest scale, the taphonomically active zone (TAZ) represents ‘nature’s’ first attempt at blurring the stratigraphical record. The TAZ is a geologically active sediment profile that may vary in thickness depending on the sedimentary environment but is commonly up to 30–50 cm. Bioturbation and disturbance of sediment by high-wave-energy storm events in shallow marine environments, such as storm-dominated continental shelves, are particularly susceptible to substrate sediment being disturbed, and in that process, the inclusion of remanié fossils into younger sediments. The Lacepede and Bonney Shelves in southern Australia are representative examples of wave-swept shelf environments, in which the entire continental shelf is disturbed by wave action in extreme storm events (James & Bone, 2011; Fig. 1). In terms of the field of observation in identifying remainé fossils, Core SV#23 in southern Gulf St. Vincent is 7 cm in cross-section with a TAZ of up to 50 cm below the seafloor. The recognition of foraminifers showing a high degree of mechanical resistance and reduced likelihood of assimilating younger sediment particles represents critical elements for defining the age of sediments within the taphonomically active zone.

At a larger scale of observation, the transect across the reef island of Notoue, Tarawa, illustrates the application of amino acid racemization in assessing the sediment dynamics of reef islands and whether such coastal landscapes are undergoing accretion or erosion. The preliminary results show that the modern ocean beach is receiving younger sediment from offshore while the lagoon side of Notoue is receiving less modern sediment.

The Coorong Coastal Plain in southern South Australia represents an even larger scale of observation of sediment reworking, resulting from erosion of the Late Pleistocene coastal barriers of Robe Range and the Woakwine Range, with the inclusion of older foraminifers into modern beach sediment. Modern sediments from the beaches reported in this work (The Granites, West Beach at Robe, Canunda Beach, and Cape Douglas Beach) occur over a 190-km sector of the coastal plain, yet they show broadly similar patterns of sediment reworking, as seen in the plots of aspartic acid and glutamic acid racemization (Fig. 7). The timescale involved in this example of reworking relates a substantial portion of the last glacial cycle, occurring over tens of thousands of years, and involves changes in relative sea-level, erosion of subaerially-exposed sediments, episodes of sediment movement across the shelf in both seaward and landward directions, and ultimately the inclusion of relict sediment particles and foraminifers within modern beach sediment. These observations affirm that the tests of the foraminifer Lamellodiscorbis dimidiatus are particularly robust and able to withstand several cycles of reworking and retain endogenous amino acids for geologically prolonged periods, as noted by Ryan et al. (2020) for a succession of calcarenites from the northern-most Coorong Coastal Plain. Collectively, Elphidium macelliforme, E. crispum, Lamellodiscorbis dimidiatus, and Amphistegina sp. are well suited for amino acid racemization geochronology and in the identification of reworked foraminifers.

Technological refinements in chromatography, particularly the ability to analyse samples of significantly smaller mass have enabled single or only a few foraminifers for a single analysis in investigations of amino acid racemization in fossils of Holocene and Late Pleistocene age. This has enhanced the application of higher spatial and temporal resolution palaeoenvironmental investigations, particularly in sampling from thinner sediment slices down-profile and for evaluating the chronological fidelity and the degree of reworking of fossils within strata. The nonlinear nature of racemization kinetics in carbonate biominerals, however, remains a limitation for age resolution given the significantly slower rates of racemization observed in Pleistocene fossils. Despite the dangers of prediction, it is unlikely that resolution higher than being able to discriminate different interglacials in Middle or Early Pleistocene time is likely to be achieved. Other applications of amino acid racemization in fossil foraminifers include relative age assessments, stratigraphical correlation, and the derivation of numeric ages when the rate of racemization has been determined by calibration with other geochronological methods.

In this work, the foraminifers Elphidium macelliforme, Amphistegina sp., and Lamellodiscorbis dimidiatus have been further shown as reliable genera for amino acid racemization analyses and for relative and numeric ages assessments as well as the identification of remanié fossils, with quantitative assessments of the magnitude of their incorporation into younger sediments. The case studies presented highlight the range of spatial scales of sediment reworking evident in the geological record and reveal that such processes most likely represent the norm and highlight the nuanced and complex nature of the stratigraphical record.

This research is a distillation of ideas about the nature of sedimentary successions that has arisen as a result of research funded over many years. An AINSE grant (06/134) covered the cost of some of the radiocarbon analyses and the analytical component of part of the research was also covered by ARC Large Grant (A10012002). We thank Peter Johnson and Tsun-You Pan for skilfully drafting the figures. We thank two anonymous reviewers and editors Marci Robinson and Robert Poirier for their very helpful and constructive reviews of the manuscript. The Appendices can be found linked to the online version of this article.

Appendix 1. Extent of glutamic acid, aspartic acid and serine racemization (THAA) in the foraminifer Elphidium macelliforme from Core SV#23, Gulf St. Vincent, southern Australia.

Appendix 2. Extent of aspartic acid racemization in individual Amphistegina sp., in reef island sediments at Notoue, northern Tarawa Atoll. Amino acid D/L values are based on peak area calculations.

Appendix 3. Extent of aspartic acid and glutamic acid racemization in individual tests of the benthic foraminifer Lamellodiscorbis dimidiatus from modern Beach sediments, Coorong Coastal Plain, southern South Australia.