Stephen R. Durham, Gregory P. Dietl, Quan Hua, John C. Handley, Darrell Kaufman, Cheryl P. Clark
{"title":"牡蛎礁死亡组合的年龄变异性和年代际时间平均","authors":"Stephen R. Durham, Gregory P. Dietl, Quan Hua, John C. Handley, Darrell Kaufman, Cheryl P. Clark","doi":"10.1130/g50778.1","DOIUrl":null,"url":null,"abstract":"Using paleoecological data to inform resource management decisions is challenging without an understanding of the ages and degrees of time-averaging in molluscan death assemblage (DA) samples. We illustrate this challenge by documenting the spatial and stratigraphic variability in age and time-averaging of oyster reef DAs. By radiocarbon dating a total of 630 oyster shells from samples at two burial depths on 31 oyster reefs around Florida, southeastern United States, we found that (1) spatial and stratigraphic variability in DA sample ages and time-averaging is of similar magnitude, and (2) the shallow oyster reef DAs are among the youngest and highest-resolution molluscan DAs documented to date, with most having decadal-scale time-averaging estimates, and sometimes less. This information increases the potential utility of the DAs for habitat management because DA data can be placed in a more specific temporal context relative to real-time monitoring data. More broadly, the results highlight the potential to obtain decadal-scale resolution from oyster bioherms in the fossil record.Decades of work on molluscan death assemblages (DAs) have successfully documented temporal changes in community composition or species attributes from direct assessments of the remains themselves (e.g., Kowalewski et al., 2000; Kidwell, 2007; Dietl and Durham, 2016; Albano et al., 2021) or from proxy information derived from them (e.g., Gillikin et al., 2019). Despite the promise of such geohistorical records for conservation paleobiology, examples of their use by resource managers are still uncommon (Groff et al., 2023). One reason is the difficulty of putting DA data in temporal context. Geochronological analyses (e.g., radiocarbon dating) are expensive and difficult to interpret, leading many conservation paleobiological studies to work around age-related uncertainties by citing general assumptions and/or studies from similar depositional settings (e.g., Dietl and Durham, 2016).However, assemblage- or specimen-level chronological control is often required to meaningfully compare DA data with the annual or subannual real-time monitoring data typically used for resource management. This was the case for the Historical Oyster Body Size (HOBS) project in Florida, southeastern United States—codeveloped by the Florida Department of Environmental Protection (FDEP) Office of Resilience and Coastal Protection (ORCP) and the Paleontological Research Institution (PRI; Dietl et al., 2023)—which aimed to use oyster reef DA samples to supplement real-time monitoring data on oyster body sizes for ORCP’s Statewide Ecosystem Assessment of Coastal and Aquatic Resources (SEACAR) project (www.floridadep.gov/SEACAR).The aquatic preserves managed by ORCP were established between 1966 and 2020 to be maintained “in an essentially natural or existing condition” (Florida Administrative Code R.18-20.001[2]; Florida Department of State, 1997). Thus, management of each preserve is often focused on its relative condition since establishment, meaning the ultimate utility of the DA approach for SEACAR would be influenced by the specific age and time-averaging properties of the oyster reef DAs. We hypothesized that oyster reef structure might limit postburial stratigraphic mixing enough such that samples from the DAs could yield data at a high enough temporal resolution to be integrated with real-time monitoring data from living oyster populations. To test this assumption and develop an understanding of both oyster reef taphonomy and the potential utility of DA data for FDEP, we produced a geochronological data set to quantify the absolute ages and temporal resolutions of oyster reef DAs from around the state.Here, we describe this investigation and show that oyster reef DAs preserve reliably recent and high-resolution stratigraphic records relative to most other molluscan DAs documented to date, suggesting these records are often appropriate for decadal-scale conservation paleobiological investigations. We also highlight the geographic variability in our data set and its implications for the importance of location-specific geochronological information for increasing the salience of paleoecological data for the resource management community.In order to build a geochronological data set to evaluate the utility of oyster DA samples for documenting trends over recent decades, we randomly selected 630 Crassostrea virginica left-valve specimens from oyster DA samples representing two stratigraphic intervals (15–25 cm and 25–35 cm) collected from up to three sample holes positioned across the densest living portion of each of 31 natural, intertidal oyster reefs in 11 locations around Florida (see the Supplemental Material1), i.e., between 2 and 7 specimens from each DA sample (Fig. 1). The selected specimens were dated by radiocarbon analysis of powdered carbonate targets (Bush et al., 2013; Hua et al., 2019)—a less expensive method with lower precision than the standard analysis of graphite targets, but one that yields similar ages (Bright et al., 2021)—to achieve a higher sample size (see the Supplemental Material for details on specimen selection for radiocarbon analysis as well as a sample size validation using 80 additional randomly selected specimens from four of the DA samples). Specimens were prepared at Northern Arizona University (NAU; Flagstaff, Arizona, USA) and analyzed at either the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry facility at the University of California, Irvine, or NAU’s own Arizona Climate and Ecosystems (ACE) Isotope Laboratory. Local corrections for the hardwater effect (e.g., Spennemann and Head, 1998) and/or estuarine influences (e.g., Ulm et al., 2009), in terms of dead carbon contribution, were developed using additional radiocarbon analyses of two live-collected oyster specimens from each sampling area (see the Supplemental Material).Age calibration was performed using OxCal v4.4 software (Bronk Ramsey, 2009) and the Marine20 calibration curve (Heaton et al., 2020) with a constant regional marine reservoir correction, ΔR = −134 ± 26 yr, which is equivalent to 5 ± 32 yr (Kowalewski et al., 2018) relative to Marine13 (Reimer et al., 2013), extended to 2022 using a regional marine bomb radiocarbon curve based on our data as well as 665 other radiocarbon results from the Gulf of Mexico, western Atlantic Ocean, and Caribbean Sea from 24 additional studies (see the Supplemental Material). Following Kowalewski et al. (2018), we used empirical posterior distributions of age probabilities for the specimens in each DA sample to generate estimates of (1) DA sample ages (we use the terms “specimen age” and “sample age” to refer to radiocarbon results for an individual oyster shell and all oyster shells from a given DA sample, respectively), and (2) time-averaging. Due to recently published concerns about the corrected posterior age estimate (CPE; sensu Kowalewski et al., 2018; also known as residual time-averaging in some studies), however, we used the interquartile range (IQR) of the average sample age probability distribution, with the quartiles weighted by the age probabilities—the total age variability (IQRTAV)—alone to estimate time-averaging instead of the IQRTAV and CPE (Ritter et al., 2023; see the Supplemental Material).Finally, to compare the contributions of location and burial depth to overall variation in DA sample median age and IQRTAV, we fit a hierarchical Bayesian model to the data for each burial depth as well as the burial depth difference for each DA sample hole (see the Supplemental Material). All data analyses were conducted using R statistical software v4.3.0 (R Core Team, 2023) and RStudio (RStudio Team, 2023).The radiocarbon results indicated that oyster reef DAs are high-resolution archives with abundant shells from the recent past and minimal time-averaging in comparison to other molluscan DAs. Among the 126 dated oyster DA samples, median calibrated ages ranged from 1567 to 2012 CE, but 91% were post-1950 (Fig. 2), and 6.4% of the DA samples had subdecadal-scale IQRTAV (0–10 yr), 72.8% had decadal-scale IQRTAV (11–100 yr), and 20.8% had centennial-scale IQRTAV (101–1000 yr) (Fig. 3; see Appendix S1 in the Supplemental Material for DA sample-level results). Moreover, collocated samples from different burial depths showed the expected temporal order (i.e., deeper = older) in most cases: Out of the 53 sample holes for which both depth intervals were processed and dated, 12 had median DA sample ages for the 15–25 cm burial depth that were older than those of material from the 25–35 cm burial depth, and five of those cases were from a single locality (Lone Cabbage; Fig. 2). The results also showed that the age and time-averaging of a given burial depth can vary substantially over small spatial scales (i.e., both intrareef and interreef assemblage variation; Fig. 2). In fact, the modeled standard deviations (SDs) for spatial variability in median age and IQRTAV (e.g., DA sample-hole-level median SDs were 12.8 and 20.9 yr for median age and IQRTAV, respectively, for the 15–25-cm-depth samples) were of similar magnitude to those for the difference between burial depths (e.g., DA sample-hole-level median depth difference SDs were 29.7 and 37.0 yr for median age and IQRTAV, respectively; Table S5; Figs. S5–S10).To our knowledge, this is the largest study of age-depth relationships and the first study to use radiocarbon to document time-averaging in oyster reef DAs. We found that, relative to other molluscan DAs, the oyster DA samples were younger, were less time-averaged, and had less spatial variability in both calibrated age and time-averaging estimates (Flessa et al., 1993; Meldahl et al., 1997; Kowalewski et al., 1998, 2018; Kosnik et al., 2009, 2015; Krause et al., 2010; Dexter et al., 2014; Dominguez et al., 2016; Ritter et al., 2017; Tomašových et al., 2019; Albano et al., 2020; see additional studies summarized by Kidwell [2013, their table 1]; but see also Tomašových et al. [2018] for an example of a non-reef DA with decadal-scale resolution). This result agrees well with that of a preliminary investigation of time-averaging on two southwest Florida oyster reefs by Lindland et al. (2001) that used amino acid racemization geochronology.Among the few recent studies that (1) focused on mollusks, (2) estimated time-averaging and sample ages in similar ways to our study, and (3) reported their unsummarized, sample-level results, the C. virginica DA samples typically had younger median ages by ~100 yr, and over half of our samples also had lower IQRTAV, some by an order of magnitude or more (Fig. 3). For instance, Dominguez et al. (2016) sampled the upper 20 cm of sediment (medium–fine sand, <2% mud) at six sites with ~9 m water depth in Sydney Harbor, Australia, and found time-averaging (IQRTAV) of ~84–>2000 yr in DAs of the bivalve Fulvia tenuicostata, but the median ages of the samples were ~150 yr. In contrast, IQRTAV across all of the C. virginica DA samples in our study ranged from 6 to 532 yr with a median of ~25 yr. The medians of the median calibrated ages across all of the DA samples from 15–25 cm and 25–35 cm burial depths were 23 yr and 29 yr, respectively. The SDs for both median age and IQRTAV among the locations sampled by Dominguez et al. (2016) were higher than the respective modeled locality-level SDs for the oyster reefs we sampled, despite the much greater geographic area covered by our study (Figs. S5–S10). A similar pattern is evident for DAs from other depositional settings and locations (e.g., southern Brazilian shelf—Ritter et al., 2017; subtidal sand flat, Fernandez Bay, San Salvador Island, Bahamas—Kowalewski et al., 2018; Fig. 3).One exception to this pattern is the study by Tomašových et al. (2018), which found comparable age and time-averaging estimates to ours in Corbula gibba DAs from cores of the Po and Isonzo prodeltas, northern Adriatic Sea (Fig. 3). However, the authors stated that the two deltas have some of the highest sedimentation rates in the northern Adriatic Sea, and median ages and time-averaging estimates for C. gibba DAs from the eastern Gulf of Trieste—across the gulf from the Isonzo River and characterized by low sedimentation rates—were older and more time-averaged than the prodelta samples by nearly two orders of magnitude (Tomašových et al., 2019). In contrast to these large differences in age and time-averaging of C. gibba DAs between depositional settings, decadal-scale resolution appears to be a common feature of DAs from intertidal C. virginica reefs in multiple estuaries across Florida.Overall, our results suggest that oyster reefs have a relatively high shell burial rate and less stratigraphic mixing relative to nonreef molluscan DAs, consistent with the hypothesis that the physical structure of oyster reefs reduces the susceptibility of DAs to some taphonomic processes. Despite their higher temporal resolution than other types of molluscan DAs in most cases, there is still considerable variation in the oyster DA median ages and IQRTAV values (Fig. 2), precluding useful regional or statewide generalizations of age versus burial depth relationships or scales of time-averaging (see the Supplemental Material for an example).This variability highlights a need for additional work to further refine the spatial and temporal specificity of dead C reference information for calibrating the fraction modern carbon (F14C) of estuarine carbonates to improve their accuracy and precision as much as possible (see Supplemental Material for further discussion of this point). It also illustrates why specific geochronological information will be important for many conservation paleobiological contributions to oyster management. Exactly how necessary they are for any given project will depend on the questions investigated, but trends in many indicators of oyster population condition, such as live-oyster size-frequency, are typically tracked at annual or subannual intervals by oyster monitoring programs. To integrate measurements from DA samples with such high-resolution records for trend analyses, it would likely be necessary to know, for instance, whether the median calibrated age and IQRTAV of a DA sample are 2011 and 53 yr, respectively, or 1979 and 16.5 yr—as was the case for two of the DA samples at 15–25 cm burial depth from our Big Hickory locality.Once these data are obtained, comparisons between the DA data and monitoring data that were impractical without them can become feasible—such as the HOBS project’s focus on integrating DA and real-time oyster size data into a single trend analysis, accounting for uncertainty in both oyster size and sample ages—instead of only focusing on more general “before/after” comparisons (e.g., Dietl and Durham, 2016). Further, our study demonstrated that most of the DA samples from oyster reefs represent a relevant time period for ORCP management (i.e., late 1960s to mid-2000s) and can yield decadal-scale (and sometimes subdecadal-scale) retrospective information from ORCP-managed areas where no long-term contemporaneous oyster monitoring took place.Lastly, the apparently limited stratigraphic mobility of shells in recent oyster DAs suggests the intriguing possibility that the degree of time-averaging in an in situ fossil oyster reef (bioherm) is not dramatically greater than that in the DA of a living oyster reef. In this case, bioherms might preserve decadal-scale records from time periods when information at such a fine temporal resolution is exceptionally rare, making them suitable records for otherwise impossible studies of short-term ecological processes in the deep past (e.g., Kowalewski et al., 1998; Kidwell and Tomašových, 2013).We thank Jordon Bright (Amino Acid Geochronology Laboratory, Northern Arizona University) and the staffs of the Keck Accelerator Mass Spectrometer Facility (University of California, Irvine) and the ACE Isotope Laboratory (Northern Arizona University) for 14C analyses, and Matthew Kosnik (Macquarie University, Australia) for help in adapting the Kowalewski et al. (2018) R script. We also thank colleagues and volunteers from the Florida Department of Environmental Protection (FDEP), the Paleontological Research Institution (PRI), and the Florida Department of Agriculture and Consumer Services (FDACS) who helped with field or laboratory work (*FDEP, †PRI, ‡FDACS, §Pelican Island National Wildlife Refuge). Staff: M. Anderson*, P. Benjasirichai‡, E. Bourque*, M. Brown*, C. Brunk*, C. Clark‡, S. Cofone‡, R. Cray*, E. Dark*, M. DeHaven‡, N. Dix*, S. Erickson*, J. Fleiger‡, J. Garwood*, T. Green*, B. Hamill*, K. Harshaw‡, D. Hersl*, T. Jones*, K. Lang*, P. Marcum*, M. McMurray*, B. Mowbray*, R. Noyes*, J. Pier†, R. Prado*, M. Pruden†. Volunteers: B. Alexander*, A. Bishop*, S. Gavirneni†, C. Hormuth*, A. McNeil†, M. Melekos†, R. Mondazzi*, D. Philipp*, E. Prest†, M. Raymond*, M. Schilling*, Z. Siper†, B. Skoblick†, A. Thorsness*, D. Thorsness*, J. Valentine§. We also gratefully acknowledge K. Flessa, M. Savarese, and an anonymous reviewer, whose comments helped to improve the manuscript, and funding from the National Oceanic and Atmospheric Administration (NOAA) Office for Coastal Management under the Coastal Zone Management Act of 1972, as amended, to the Florida Coastal Management Program (NOAA awards NA18NOS4190080 and NA19NOS4190064). The views, statements, findings, conclusions, and recommendations expressed herein are those of the author(s) and do not necessarily reflect the views of the State of Florida, NOAA, or any of their subagencies.","PeriodicalId":12642,"journal":{"name":"Geology","volume":"57 46","pages":""},"PeriodicalIF":4.8000,"publicationDate":"2023-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Age variability and decadal time-averaging in oyster reef death assemblages\",\"authors\":\"Stephen R. Durham, Gregory P. Dietl, Quan Hua, John C. Handley, Darrell Kaufman, Cheryl P. Clark\",\"doi\":\"10.1130/g50778.1\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Using paleoecological data to inform resource management decisions is challenging without an understanding of the ages and degrees of time-averaging in molluscan death assemblage (DA) samples. We illustrate this challenge by documenting the spatial and stratigraphic variability in age and time-averaging of oyster reef DAs. By radiocarbon dating a total of 630 oyster shells from samples at two burial depths on 31 oyster reefs around Florida, southeastern United States, we found that (1) spatial and stratigraphic variability in DA sample ages and time-averaging is of similar magnitude, and (2) the shallow oyster reef DAs are among the youngest and highest-resolution molluscan DAs documented to date, with most having decadal-scale time-averaging estimates, and sometimes less. This information increases the potential utility of the DAs for habitat management because DA data can be placed in a more specific temporal context relative to real-time monitoring data. More broadly, the results highlight the potential to obtain decadal-scale resolution from oyster bioherms in the fossil record.Decades of work on molluscan death assemblages (DAs) have successfully documented temporal changes in community composition or species attributes from direct assessments of the remains themselves (e.g., Kowalewski et al., 2000; Kidwell, 2007; Dietl and Durham, 2016; Albano et al., 2021) or from proxy information derived from them (e.g., Gillikin et al., 2019). Despite the promise of such geohistorical records for conservation paleobiology, examples of their use by resource managers are still uncommon (Groff et al., 2023). One reason is the difficulty of putting DA data in temporal context. Geochronological analyses (e.g., radiocarbon dating) are expensive and difficult to interpret, leading many conservation paleobiological studies to work around age-related uncertainties by citing general assumptions and/or studies from similar depositional settings (e.g., Dietl and Durham, 2016).However, assemblage- or specimen-level chronological control is often required to meaningfully compare DA data with the annual or subannual real-time monitoring data typically used for resource management. This was the case for the Historical Oyster Body Size (HOBS) project in Florida, southeastern United States—codeveloped by the Florida Department of Environmental Protection (FDEP) Office of Resilience and Coastal Protection (ORCP) and the Paleontological Research Institution (PRI; Dietl et al., 2023)—which aimed to use oyster reef DA samples to supplement real-time monitoring data on oyster body sizes for ORCP’s Statewide Ecosystem Assessment of Coastal and Aquatic Resources (SEACAR) project (www.floridadep.gov/SEACAR).The aquatic preserves managed by ORCP were established between 1966 and 2020 to be maintained “in an essentially natural or existing condition” (Florida Administrative Code R.18-20.001[2]; Florida Department of State, 1997). Thus, management of each preserve is often focused on its relative condition since establishment, meaning the ultimate utility of the DA approach for SEACAR would be influenced by the specific age and time-averaging properties of the oyster reef DAs. We hypothesized that oyster reef structure might limit postburial stratigraphic mixing enough such that samples from the DAs could yield data at a high enough temporal resolution to be integrated with real-time monitoring data from living oyster populations. To test this assumption and develop an understanding of both oyster reef taphonomy and the potential utility of DA data for FDEP, we produced a geochronological data set to quantify the absolute ages and temporal resolutions of oyster reef DAs from around the state.Here, we describe this investigation and show that oyster reef DAs preserve reliably recent and high-resolution stratigraphic records relative to most other molluscan DAs documented to date, suggesting these records are often appropriate for decadal-scale conservation paleobiological investigations. We also highlight the geographic variability in our data set and its implications for the importance of location-specific geochronological information for increasing the salience of paleoecological data for the resource management community.In order to build a geochronological data set to evaluate the utility of oyster DA samples for documenting trends over recent decades, we randomly selected 630 Crassostrea virginica left-valve specimens from oyster DA samples representing two stratigraphic intervals (15–25 cm and 25–35 cm) collected from up to three sample holes positioned across the densest living portion of each of 31 natural, intertidal oyster reefs in 11 locations around Florida (see the Supplemental Material1), i.e., between 2 and 7 specimens from each DA sample (Fig. 1). The selected specimens were dated by radiocarbon analysis of powdered carbonate targets (Bush et al., 2013; Hua et al., 2019)—a less expensive method with lower precision than the standard analysis of graphite targets, but one that yields similar ages (Bright et al., 2021)—to achieve a higher sample size (see the Supplemental Material for details on specimen selection for radiocarbon analysis as well as a sample size validation using 80 additional randomly selected specimens from four of the DA samples). Specimens were prepared at Northern Arizona University (NAU; Flagstaff, Arizona, USA) and analyzed at either the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry facility at the University of California, Irvine, or NAU’s own Arizona Climate and Ecosystems (ACE) Isotope Laboratory. Local corrections for the hardwater effect (e.g., Spennemann and Head, 1998) and/or estuarine influences (e.g., Ulm et al., 2009), in terms of dead carbon contribution, were developed using additional radiocarbon analyses of two live-collected oyster specimens from each sampling area (see the Supplemental Material).Age calibration was performed using OxCal v4.4 software (Bronk Ramsey, 2009) and the Marine20 calibration curve (Heaton et al., 2020) with a constant regional marine reservoir correction, ΔR = −134 ± 26 yr, which is equivalent to 5 ± 32 yr (Kowalewski et al., 2018) relative to Marine13 (Reimer et al., 2013), extended to 2022 using a regional marine bomb radiocarbon curve based on our data as well as 665 other radiocarbon results from the Gulf of Mexico, western Atlantic Ocean, and Caribbean Sea from 24 additional studies (see the Supplemental Material). Following Kowalewski et al. (2018), we used empirical posterior distributions of age probabilities for the specimens in each DA sample to generate estimates of (1) DA sample ages (we use the terms “specimen age” and “sample age” to refer to radiocarbon results for an individual oyster shell and all oyster shells from a given DA sample, respectively), and (2) time-averaging. Due to recently published concerns about the corrected posterior age estimate (CPE; sensu Kowalewski et al., 2018; also known as residual time-averaging in some studies), however, we used the interquartile range (IQR) of the average sample age probability distribution, with the quartiles weighted by the age probabilities—the total age variability (IQRTAV)—alone to estimate time-averaging instead of the IQRTAV and CPE (Ritter et al., 2023; see the Supplemental Material).Finally, to compare the contributions of location and burial depth to overall variation in DA sample median age and IQRTAV, we fit a hierarchical Bayesian model to the data for each burial depth as well as the burial depth difference for each DA sample hole (see the Supplemental Material). All data analyses were conducted using R statistical software v4.3.0 (R Core Team, 2023) and RStudio (RStudio Team, 2023).The radiocarbon results indicated that oyster reef DAs are high-resolution archives with abundant shells from the recent past and minimal time-averaging in comparison to other molluscan DAs. Among the 126 dated oyster DA samples, median calibrated ages ranged from 1567 to 2012 CE, but 91% were post-1950 (Fig. 2), and 6.4% of the DA samples had subdecadal-scale IQRTAV (0–10 yr), 72.8% had decadal-scale IQRTAV (11–100 yr), and 20.8% had centennial-scale IQRTAV (101–1000 yr) (Fig. 3; see Appendix S1 in the Supplemental Material for DA sample-level results). Moreover, collocated samples from different burial depths showed the expected temporal order (i.e., deeper = older) in most cases: Out of the 53 sample holes for which both depth intervals were processed and dated, 12 had median DA sample ages for the 15–25 cm burial depth that were older than those of material from the 25–35 cm burial depth, and five of those cases were from a single locality (Lone Cabbage; Fig. 2). The results also showed that the age and time-averaging of a given burial depth can vary substantially over small spatial scales (i.e., both intrareef and interreef assemblage variation; Fig. 2). In fact, the modeled standard deviations (SDs) for spatial variability in median age and IQRTAV (e.g., DA sample-hole-level median SDs were 12.8 and 20.9 yr for median age and IQRTAV, respectively, for the 15–25-cm-depth samples) were of similar magnitude to those for the difference between burial depths (e.g., DA sample-hole-level median depth difference SDs were 29.7 and 37.0 yr for median age and IQRTAV, respectively; Table S5; Figs. S5–S10).To our knowledge, this is the largest study of age-depth relationships and the first study to use radiocarbon to document time-averaging in oyster reef DAs. We found that, relative to other molluscan DAs, the oyster DA samples were younger, were less time-averaged, and had less spatial variability in both calibrated age and time-averaging estimates (Flessa et al., 1993; Meldahl et al., 1997; Kowalewski et al., 1998, 2018; Kosnik et al., 2009, 2015; Krause et al., 2010; Dexter et al., 2014; Dominguez et al., 2016; Ritter et al., 2017; Tomašových et al., 2019; Albano et al., 2020; see additional studies summarized by Kidwell [2013, their table 1]; but see also Tomašových et al. [2018] for an example of a non-reef DA with decadal-scale resolution). This result agrees well with that of a preliminary investigation of time-averaging on two southwest Florida oyster reefs by Lindland et al. (2001) that used amino acid racemization geochronology.Among the few recent studies that (1) focused on mollusks, (2) estimated time-averaging and sample ages in similar ways to our study, and (3) reported their unsummarized, sample-level results, the C. virginica DA samples typically had younger median ages by ~100 yr, and over half of our samples also had lower IQRTAV, some by an order of magnitude or more (Fig. 3). For instance, Dominguez et al. (2016) sampled the upper 20 cm of sediment (medium–fine sand, <2% mud) at six sites with ~9 m water depth in Sydney Harbor, Australia, and found time-averaging (IQRTAV) of ~84–>2000 yr in DAs of the bivalve Fulvia tenuicostata, but the median ages of the samples were ~150 yr. In contrast, IQRTAV across all of the C. virginica DA samples in our study ranged from 6 to 532 yr with a median of ~25 yr. The medians of the median calibrated ages across all of the DA samples from 15–25 cm and 25–35 cm burial depths were 23 yr and 29 yr, respectively. The SDs for both median age and IQRTAV among the locations sampled by Dominguez et al. (2016) were higher than the respective modeled locality-level SDs for the oyster reefs we sampled, despite the much greater geographic area covered by our study (Figs. S5–S10). A similar pattern is evident for DAs from other depositional settings and locations (e.g., southern Brazilian shelf—Ritter et al., 2017; subtidal sand flat, Fernandez Bay, San Salvador Island, Bahamas—Kowalewski et al., 2018; Fig. 3).One exception to this pattern is the study by Tomašových et al. (2018), which found comparable age and time-averaging estimates to ours in Corbula gibba DAs from cores of the Po and Isonzo prodeltas, northern Adriatic Sea (Fig. 3). However, the authors stated that the two deltas have some of the highest sedimentation rates in the northern Adriatic Sea, and median ages and time-averaging estimates for C. gibba DAs from the eastern Gulf of Trieste—across the gulf from the Isonzo River and characterized by low sedimentation rates—were older and more time-averaged than the prodelta samples by nearly two orders of magnitude (Tomašových et al., 2019). In contrast to these large differences in age and time-averaging of C. gibba DAs between depositional settings, decadal-scale resolution appears to be a common feature of DAs from intertidal C. virginica reefs in multiple estuaries across Florida.Overall, our results suggest that oyster reefs have a relatively high shell burial rate and less stratigraphic mixing relative to nonreef molluscan DAs, consistent with the hypothesis that the physical structure of oyster reefs reduces the susceptibility of DAs to some taphonomic processes. Despite their higher temporal resolution than other types of molluscan DAs in most cases, there is still considerable variation in the oyster DA median ages and IQRTAV values (Fig. 2), precluding useful regional or statewide generalizations of age versus burial depth relationships or scales of time-averaging (see the Supplemental Material for an example).This variability highlights a need for additional work to further refine the spatial and temporal specificity of dead C reference information for calibrating the fraction modern carbon (F14C) of estuarine carbonates to improve their accuracy and precision as much as possible (see Supplemental Material for further discussion of this point). It also illustrates why specific geochronological information will be important for many conservation paleobiological contributions to oyster management. Exactly how necessary they are for any given project will depend on the questions investigated, but trends in many indicators of oyster population condition, such as live-oyster size-frequency, are typically tracked at annual or subannual intervals by oyster monitoring programs. To integrate measurements from DA samples with such high-resolution records for trend analyses, it would likely be necessary to know, for instance, whether the median calibrated age and IQRTAV of a DA sample are 2011 and 53 yr, respectively, or 1979 and 16.5 yr—as was the case for two of the DA samples at 15–25 cm burial depth from our Big Hickory locality.Once these data are obtained, comparisons between the DA data and monitoring data that were impractical without them can become feasible—such as the HOBS project’s focus on integrating DA and real-time oyster size data into a single trend analysis, accounting for uncertainty in both oyster size and sample ages—instead of only focusing on more general “before/after” comparisons (e.g., Dietl and Durham, 2016). Further, our study demonstrated that most of the DA samples from oyster reefs represent a relevant time period for ORCP management (i.e., late 1960s to mid-2000s) and can yield decadal-scale (and sometimes subdecadal-scale) retrospective information from ORCP-managed areas where no long-term contemporaneous oyster monitoring took place.Lastly, the apparently limited stratigraphic mobility of shells in recent oyster DAs suggests the intriguing possibility that the degree of time-averaging in an in situ fossil oyster reef (bioherm) is not dramatically greater than that in the DA of a living oyster reef. In this case, bioherms might preserve decadal-scale records from time periods when information at such a fine temporal resolution is exceptionally rare, making them suitable records for otherwise impossible studies of short-term ecological processes in the deep past (e.g., Kowalewski et al., 1998; Kidwell and Tomašových, 2013).We thank Jordon Bright (Amino Acid Geochronology Laboratory, Northern Arizona University) and the staffs of the Keck Accelerator Mass Spectrometer Facility (University of California, Irvine) and the ACE Isotope Laboratory (Northern Arizona University) for 14C analyses, and Matthew Kosnik (Macquarie University, Australia) for help in adapting the Kowalewski et al. (2018) R script. We also thank colleagues and volunteers from the Florida Department of Environmental Protection (FDEP), the Paleontological Research Institution (PRI), and the Florida Department of Agriculture and Consumer Services (FDACS) who helped with field or laboratory work (*FDEP, †PRI, ‡FDACS, §Pelican Island National Wildlife Refuge). Staff: M. Anderson*, P. Benjasirichai‡, E. Bourque*, M. Brown*, C. Brunk*, C. Clark‡, S. Cofone‡, R. Cray*, E. Dark*, M. DeHaven‡, N. Dix*, S. Erickson*, J. Fleiger‡, J. Garwood*, T. Green*, B. Hamill*, K. Harshaw‡, D. Hersl*, T. Jones*, K. Lang*, P. Marcum*, M. McMurray*, B. Mowbray*, R. Noyes*, J. Pier†, R. Prado*, M. Pruden†. Volunteers: B. Alexander*, A. Bishop*, S. Gavirneni†, C. Hormuth*, A. McNeil†, M. Melekos†, R. Mondazzi*, D. Philipp*, E. Prest†, M. Raymond*, M. Schilling*, Z. Siper†, B. Skoblick†, A. Thorsness*, D. Thorsness*, J. Valentine§. We also gratefully acknowledge K. Flessa, M. Savarese, and an anonymous reviewer, whose comments helped to improve the manuscript, and funding from the National Oceanic and Atmospheric Administration (NOAA) Office for Coastal Management under the Coastal Zone Management Act of 1972, as amended, to the Florida Coastal Management Program (NOAA awards NA18NOS4190080 and NA19NOS4190064). The views, statements, findings, conclusions, and recommendations expressed herein are those of the author(s) and do not necessarily reflect the views of the State of Florida, NOAA, or any of their subagencies.\",\"PeriodicalId\":12642,\"journal\":{\"name\":\"Geology\",\"volume\":\"57 46\",\"pages\":\"\"},\"PeriodicalIF\":4.8000,\"publicationDate\":\"2023-11-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Geology\",\"FirstCategoryId\":\"89\",\"ListUrlMain\":\"https://doi.org/10.1130/g50778.1\",\"RegionNum\":1,\"RegionCategory\":\"地球科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"GEOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Geology","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1130/g50778.1","RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"GEOLOGY","Score":null,"Total":0}
引用次数: 0
摘要
在不了解软体动物死亡组合(DA)样本的年龄和时间平均程度的情况下,使用古生态数据为资源管理决策提供信息是一项挑战。我们通过记录牡蛎礁DA的年龄和时间平均值的空间和地层变化来说明这一挑战。通过对美国东南部佛罗里达州周围31个牡蛎礁两个埋藏深度的样本中的630个牡蛎壳进行放射性碳年代测定,我们发现(1)DA样本年龄和时间平均值的空间和地层变化幅度相似,(2)浅层牡蛎礁DA是迄今为止记录的最年轻、分辨率最高的软体动物DA之一,大多数具有十年尺度的时间平均估计,有时更少。这些信息增加了DA对栖息地管理的潜在效用,因为相对于实时监测数据,DA数据可以放在更具体的时间上下文中。更广泛地说,这些结果突出了从化石记录中的牡蛎生物礁获得十年尺度分辨率的潜力。几十年来对软体动物死亡组合(DA)的研究已经成功地记录了群落组成或物种属性的时间变化,这些变化来自对遗骸本身的直接评估(例如,Kowalewski等人,2000;Kidwell,2007;Dietl和Durham,2016;Albano等人,2021)或从中获得的代理信息(例如,Gillikin等人,2019)。尽管这种地质历史记录有望用于保护古生物学,但资源管理者使用它们的例子仍然很少(Groff等人,2023)。一个原因是将DA数据放在时间上下文中很困难。地质年代分析(例如,放射性碳定年)成本高昂,难以解释,导致许多保护性古生物学研究通过引用一般假设和/或类似沉积环境的研究来解决与年龄相关的不确定性(例如,Dietl和Durham,2016)。然而,为了将DA数据与通常用于资源管理的年度或亚年度实时监测数据进行有意义的比较,通常需要组合或样本级别的时间控制。佛罗里达州的历史牡蛎体型(HOBS)项目就是这样,美国东南部——由佛罗里达州环境保护部(FDEP)复原力和海岸保护办公室(ORCP)和古生物研究所(PRI;Dietl等人,2023)共同开发——旨在利用牡蛎礁DA样本补充牡蛎体型的实时监测数据,用于ORCP的全州海岸和水生生态系统评估资源(SEACAR)项目(www.floridadep.gov/SEACAR)。ORCP管理的水生保护区成立于1966年至2020年,旨在“保持基本上自然或现有的状态”(佛罗里达州行政法规R.18-20.001[2];佛罗里达州国务院,1997年)。因此,每个保护区的管理通常侧重于其自建立以来的相对条件,这意味着DA方法对SEACAR的最终效用将受到牡蛎礁DA的特定年龄和时间平均特性的影响。我们假设,牡蛎礁结构可能会充分限制洪后地层混合,从而使DA的样本可以产生足够高的时间分辨率的数据,与活牡蛎种群的实时监测数据相结合。为了检验这一假设,并了解牡蛎礁的埋藏和DA数据在FDEP中的潜在效用,我们制作了一个地质年代数据集,以量化该州牡蛎礁DA的绝对年龄和时间分辨率。在这里,我们描述了这项调查,并表明与迄今为止记录的大多数其他软体动物DA相比,牡蛎礁DA保存了可靠的近期和高分辨率地层记录,这表明这些记录通常适用于十年尺度的保护古生物学调查。我们还强调了我们数据集中的地理变异性,以及它对特定位置的地质年代信息的重要性的影响,以提高古生态数据对资源管理界的重要性。为了建立一个地质年代数据集来评估牡蛎DA样本在记录近几十年趋势方面的效用,我们从代表两个地层间隔(15–25 cm和25–35 cm)的牡蛎DA样本中随机选择了630个弗吉尼亚牡蛎左瓣标本,这些样本是从佛罗里达州周围11个位置的31个自然潮间带牡蛎礁中最密集的生活区的多达三个样本孔中收集的(见补充材料1),即。,每个DA样本中有2到7个样本(图1)。通过对粉末状碳酸盐靶进行放射性碳分析,确定了选定样本的年代(Bush等人,2013;华等人。 ,2019)——这是一种成本较低、精度低于石墨靶标准分析的方法,但产生的年龄相似(Bright et al.,2021)——以实现更高的样本量(有关放射性碳分析样本选择的详细信息,以及使用从四个DA样本中随机选择的80个额外样本进行样本量验证的补充材料)。样本在北亚利桑那大学(NAU;美国亚利桑那州弗拉格斯塔夫)制备,并在加州大学欧文分校的W.M.Keck碳循环加速器质谱设施或NAU自己的亚利桑那州气候与生态系统(ACE)同位素实验室进行分析。硬水效应的局部校正(例如,Spennemann和Head,1998)和/或河口影响(例如,Ulm等人,2009),就死碳贡献而言,对每个采样区采集的两个活牡蛎样本进行了额外的放射性碳分析(见补充材料)。使用OxCal v4.4软件(Bronk Ramsey,2009)和Marine20校准曲线(Heaton等人,2020)进行了年龄校准,并进行了恒定的区域海洋储层校正,ΔR=-134±26年,相对于Marine13(Reimer et al.,2013),这相当于5±32年(Kowalewski et al.,2018),使用基于我们的数据的区域海洋炸弹放射性碳曲线以及来自墨西哥湾、西大西洋和加勒比海的24项额外研究的665个其他放射性碳结果,将其延长至2022年(见补充材料)。根据Kowalewski等人(2018),我们使用每个DA样本中样本的年龄概率的经验后验分布来生成(1)DA样本年龄的估计值(我们使用术语“样本年龄”和“样品年龄”分别指给定DA样本中单个牡蛎壳和所有牡蛎壳的放射性碳结果),以及(2)时间平均值。然而,由于最近发表的对校正后验年龄估计的担忧(CPE;sensu-Kowalewski等人,2018;在一些研究中也称为残差时间平均),我们使用了平均样本年龄概率分布的四分位间距(IQR),用年龄概率加权的四分位数——总年龄变异性(IQRTAV)——单独估计时间平均值,而不是IQRTAV和CPE(Ritter et al.,2023;见补充材料)。最后,比较位置和埋深对DA样本中位年龄和IQRTAV总体变化的贡献,我们将分层贝叶斯模型拟合到每个埋深的数据以及每个DA样品孔的埋深差(见补充材料)。所有数据分析都是使用R统计软件v4.3.0(R Core Team,2023)和RStudio(RStudioTeam,2022)进行的。放射性碳结果表明,牡蛎礁DA是高分辨率的档案,与其他软体动物DA相比,具有丰富的近期贝壳和最短的时间平均值。在126个测年的牡蛎DA样本中,中位校准年龄范围为1567年至2012年CE,但91%的DA样本在1950年后(图2),6.4%的DA样本具有亚十年尺度IQRTAV(0–10年),72.8%的DA样本有十年尺度IQ RTAV(11–100年),20.8%的样本具有百年尺度IQRTV(101–1000年)(图3;DA样本水平结果见补充材料中的附录S1)。此外,在大多数情况下,来自不同埋深的并置样本显示出预期的时间顺序(即,更深=更老):在处理和确定两个深度间隔的53个样本孔中,有12个具有15-25厘米埋深的中值DA样本年龄,该年龄比25-35厘米埋深材料的年龄更老,其中5例来自同一地区(Lone卷心菜;图2)。结果还表明,给定埋深的年龄和时间平均值在小的空间尺度上可能会有很大的变化(即礁内和礁间组合的变化;图2)。事实上中位年龄和IQRTAV空间变异性的建模标准差(SD)(例如,对于15-25 cm深度的样本,DA样本孔水平的中位年龄为12.8年,IQRTAV为20.9年)与埋深差异的标准差大小相似(例如,中位年龄和IQRTAV的DA样品孔水平中位深度差SD分别为29.7年和37.0年;表S5;图S5–S10)。据我们所知,这是对年龄-深度关系的最大研究,也是首次使用放射性碳记录牡蛎礁DA的时间平均值的研究。我们发现,与其他软体动物DA相比,牡蛎DA样本更年轻,时间平均值更低,并且在校准的年龄和时间平均估计值中具有较小的空间变异性(Flessa等人,1993;Meldahl等人,1997;Kowalewski等人,1998、2018;Kosnik等人,2009、2015;Krause等人,2010;Dexter等人,2014;Dominguez等人,2016;Ritter等人。 ,2017;Tomašových等人,2019;Albano等人,2020;见Kidwell总结的其他研究【2013年,表1】;但也可参见Tomašových等人[2018],以获得具有十年尺度分辨率的非珊瑚礁DA的示例)。这一结果与Lindland等人(2001)使用氨基酸外消旋化地质年代学对佛罗里达州西南部两个牡蛎礁的时间平均进行的初步调查结果一致。在最近为数不多的研究中,(1)专注于软体动物,(2)以与我们的研究类似的方式估计时间平均值和样本年龄,以及(3)报告了其未汇总的样本水平结果,弗吉尼亚C.virginica DA样本的中位年龄通常较年轻约100岁,我们一半以上的样本IQRTAV也较低,有些样本的IQRTAV较低一个数量级或更高(图3)。例如,Dominguez等人(2016)在双壳类Fulvia tenuicostata的DA中对上部20厘米的沉积物(中细砂,2000年)进行了采样,但样本的中位年龄约为150年。相比之下,我们研究中所有弗吉尼亚C.virginica DA样本的IQRTAV在6至532年之间,中位年龄为约25年。15–25 cm和25–35 cm埋深的所有DA样本的中位校准年龄的中位数分别为23年和29年。Dominguez等人(2016)采样的地点的中位年龄和IQRTAV的SDs均高于我们采样的牡蛎礁的各自建模的地点级SDs,尽管我们的研究覆盖了更大的地理区域(图S5–S10)。其他沉积环境和位置的DA也有类似的模式(例如,巴西南部陆架——Ritter等人,2017;巴哈马圣萨尔瓦多岛费尔南德斯湾潮下砂滩——Kowalewski等人,2018;图3)。这种模式的一个例外是Tomašových等人的研究。(2018),该研究发现,亚得里亚海北部Po和Isonzo前三角洲核心的Corbula gibba DA的年龄和时间平均估计值与我们的相当(图3)。然而,作者指出,这两个三角洲在亚得里亚海北部的沉积速率最高,来自的里雅斯特湾东部(与Isonzo河隔海湾相望,以低沉积率为特征)的长臂猿DA的中位年龄和时间平均估计值比前三角洲样本更古老,时间平均值也更多,接近两个数量级(Tomašových et al.,2019)。与不同沉积环境中长臂猿DA的年龄和时间平均值的巨大差异相反,十年尺度分辨率似乎是佛罗里达州多个河口潮间带弗吉尼亚长臂猿礁DA的共同特征。总的来说,我们的研究结果表明,与非参考软体动物DA相比,牡蛎礁的外壳埋藏率相对较高,地层混合较少,这与牡蛎礁的物理结构降低了DA对某些陆生过程的易感性的假设一致。尽管在大多数情况下,牡蛎DA的时间分辨率高于其他类型的软体动物DA,但牡蛎DA的中位年龄和IQRTAV值仍存在相当大的差异(图2),排除了对年龄与埋深关系或时间平均尺度的有用的区域或全州概括(示例见补充材料)。这种可变性突出表明,需要进一步完善死碳参考信息的空间和时间特异性,以校准河口碳酸盐岩的现代碳(F14C)分数,从而尽可能提高其准确性和精度(关于这一点的进一步讨论,请参阅补充材料)。这也说明了为什么特定的地质年代信息对牡蛎管理的许多保护古生物学贡献都很重要。它们对任何特定项目的确切必要性将取决于所调查的问题,但牡蛎监测项目通常每年或每半年跟踪牡蛎种群状况的许多指标的趋势,如活牡蛎大小的频率。为了将DA样本的测量结果与这种高分辨率记录相结合进行趋势分析,可能有必要知道,例如,DA样本的中位校准年龄和IQRTAV是分别为2011年和53年,还是1979年和16.5年——就像我们大希克里地区15-25厘米埋深的两个DA样本的情况一样。一旦获得了这些数据,DA数据和没有它们就不切实际的监测数据之间的比较就变得可行了——比如HOBS项目专注于将DA和实时牡蛎大小数据集成到单一趋势分析中,考虑到牡蛎大小和样本年龄的不确定性,而不是只关注更普遍的“前后”比较(例如,Dietl和Durham,2016)。此外,我们的研究表明,来自牡蛎礁的大多数DA样本代表了ORCP管理的相关时间段(即。 20世纪60年代末至2000年代中期),并且可以从ORCP管理的没有进行长期同期牡蛎监测的地区产生十年尺度(有时是十年以下尺度)的回顾性信息。最后,在最近的牡蛎DA中,贝壳的地层流动性明显有限,这表明了一种有趣的可能性,即原位化石牡蛎礁(生物礁)的时间平均程度并不显著大于活牡蛎礁的DA。在这种情况下,生物礁可能会保存这样精细的时间分辨率的信息异常罕见的时间段的十年尺度记录,使它们成为过去不可能进行的短期生态过程研究的合适记录(例如,Kowalewski等人,1998;Kidwell和Tomašových,2013)。我们感谢Jordon Bright(北亚利桑那大学氨基酸地质年代实验室)和Keck加速器质谱仪设施(加州大学欧文分校)和ACE同位素的工作人员实验室(北亚利桑那大学)进行14C分析,Matthew Kosnik(澳大利亚麦考瑞大学)帮助改编Kowalewski等人(2018)R脚本。我们还感谢佛罗里达州环境保护部(FDEP)、古生物研究所(PRI)和佛罗里达州农业和消费者服务部(FDACS)的同事和志愿者,他们帮助了实地或实验室工作(*FDEP,†PRI,†FDACS,§Pelican Island National Wildlife Refuge)。工作人员:M.Anderson*、P.Benjasirichai⏴、E.Bourque*、M.Brown*、C.Brunk*、C.Clark⏴、S.Cofone⏴、R.Cray*、E.Dark*、M.DeHaven⏴、N.Dix*、S.Erickson*、J.Fleiger⏴、J.Garwood*、T.Green*、B.Hamill*、K.Harshaw⏴、D.Hersl*、T.Jones*、K.Lang*、P.Marcum*、M.McMurray*、B.Mowbray*、R.Noyes*、J.Pier†、R.Prado*、M.Pruden†。志愿者:B.Alexander*、A.Bishop*、S.Gavirneni†、C.Hormuth*、A.McNeil†、M.Melekos†、R.Mondazzi*、D.Philipp*、E.Prest†、M.Raymond*、M.Schilling*、Z.Siper†、B.Skoblick†、A.Thornness*、D.Thorness*、J.Valentine§。我们还感谢K.Flessa、M.Savarese和一位匿名评审员,他们的评论有助于改进手稿,并感谢美国国家海洋和大气管理局(NOAA)海岸管理办公室根据1972年《海岸带管理法》(经修订)为佛罗里达海岸管理计划提供的资金(NOAA授予NA18NOS419080和NA19NOS419064)。本文所表达的观点、陈述、调查结果、结论和建议均为作者的观点、声明、调查结果和建议,不一定反映佛罗里达州、美国国家海洋和大气管理局或其任何子代的观点。
Age variability and decadal time-averaging in oyster reef death assemblages
Using paleoecological data to inform resource management decisions is challenging without an understanding of the ages and degrees of time-averaging in molluscan death assemblage (DA) samples. We illustrate this challenge by documenting the spatial and stratigraphic variability in age and time-averaging of oyster reef DAs. By radiocarbon dating a total of 630 oyster shells from samples at two burial depths on 31 oyster reefs around Florida, southeastern United States, we found that (1) spatial and stratigraphic variability in DA sample ages and time-averaging is of similar magnitude, and (2) the shallow oyster reef DAs are among the youngest and highest-resolution molluscan DAs documented to date, with most having decadal-scale time-averaging estimates, and sometimes less. This information increases the potential utility of the DAs for habitat management because DA data can be placed in a more specific temporal context relative to real-time monitoring data. More broadly, the results highlight the potential to obtain decadal-scale resolution from oyster bioherms in the fossil record.Decades of work on molluscan death assemblages (DAs) have successfully documented temporal changes in community composition or species attributes from direct assessments of the remains themselves (e.g., Kowalewski et al., 2000; Kidwell, 2007; Dietl and Durham, 2016; Albano et al., 2021) or from proxy information derived from them (e.g., Gillikin et al., 2019). Despite the promise of such geohistorical records for conservation paleobiology, examples of their use by resource managers are still uncommon (Groff et al., 2023). One reason is the difficulty of putting DA data in temporal context. Geochronological analyses (e.g., radiocarbon dating) are expensive and difficult to interpret, leading many conservation paleobiological studies to work around age-related uncertainties by citing general assumptions and/or studies from similar depositional settings (e.g., Dietl and Durham, 2016).However, assemblage- or specimen-level chronological control is often required to meaningfully compare DA data with the annual or subannual real-time monitoring data typically used for resource management. This was the case for the Historical Oyster Body Size (HOBS) project in Florida, southeastern United States—codeveloped by the Florida Department of Environmental Protection (FDEP) Office of Resilience and Coastal Protection (ORCP) and the Paleontological Research Institution (PRI; Dietl et al., 2023)—which aimed to use oyster reef DA samples to supplement real-time monitoring data on oyster body sizes for ORCP’s Statewide Ecosystem Assessment of Coastal and Aquatic Resources (SEACAR) project (www.floridadep.gov/SEACAR).The aquatic preserves managed by ORCP were established between 1966 and 2020 to be maintained “in an essentially natural or existing condition” (Florida Administrative Code R.18-20.001[2]; Florida Department of State, 1997). Thus, management of each preserve is often focused on its relative condition since establishment, meaning the ultimate utility of the DA approach for SEACAR would be influenced by the specific age and time-averaging properties of the oyster reef DAs. We hypothesized that oyster reef structure might limit postburial stratigraphic mixing enough such that samples from the DAs could yield data at a high enough temporal resolution to be integrated with real-time monitoring data from living oyster populations. To test this assumption and develop an understanding of both oyster reef taphonomy and the potential utility of DA data for FDEP, we produced a geochronological data set to quantify the absolute ages and temporal resolutions of oyster reef DAs from around the state.Here, we describe this investigation and show that oyster reef DAs preserve reliably recent and high-resolution stratigraphic records relative to most other molluscan DAs documented to date, suggesting these records are often appropriate for decadal-scale conservation paleobiological investigations. We also highlight the geographic variability in our data set and its implications for the importance of location-specific geochronological information for increasing the salience of paleoecological data for the resource management community.In order to build a geochronological data set to evaluate the utility of oyster DA samples for documenting trends over recent decades, we randomly selected 630 Crassostrea virginica left-valve specimens from oyster DA samples representing two stratigraphic intervals (15–25 cm and 25–35 cm) collected from up to three sample holes positioned across the densest living portion of each of 31 natural, intertidal oyster reefs in 11 locations around Florida (see the Supplemental Material1), i.e., between 2 and 7 specimens from each DA sample (Fig. 1). The selected specimens were dated by radiocarbon analysis of powdered carbonate targets (Bush et al., 2013; Hua et al., 2019)—a less expensive method with lower precision than the standard analysis of graphite targets, but one that yields similar ages (Bright et al., 2021)—to achieve a higher sample size (see the Supplemental Material for details on specimen selection for radiocarbon analysis as well as a sample size validation using 80 additional randomly selected specimens from four of the DA samples). Specimens were prepared at Northern Arizona University (NAU; Flagstaff, Arizona, USA) and analyzed at either the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry facility at the University of California, Irvine, or NAU’s own Arizona Climate and Ecosystems (ACE) Isotope Laboratory. Local corrections for the hardwater effect (e.g., Spennemann and Head, 1998) and/or estuarine influences (e.g., Ulm et al., 2009), in terms of dead carbon contribution, were developed using additional radiocarbon analyses of two live-collected oyster specimens from each sampling area (see the Supplemental Material).Age calibration was performed using OxCal v4.4 software (Bronk Ramsey, 2009) and the Marine20 calibration curve (Heaton et al., 2020) with a constant regional marine reservoir correction, ΔR = −134 ± 26 yr, which is equivalent to 5 ± 32 yr (Kowalewski et al., 2018) relative to Marine13 (Reimer et al., 2013), extended to 2022 using a regional marine bomb radiocarbon curve based on our data as well as 665 other radiocarbon results from the Gulf of Mexico, western Atlantic Ocean, and Caribbean Sea from 24 additional studies (see the Supplemental Material). Following Kowalewski et al. (2018), we used empirical posterior distributions of age probabilities for the specimens in each DA sample to generate estimates of (1) DA sample ages (we use the terms “specimen age” and “sample age” to refer to radiocarbon results for an individual oyster shell and all oyster shells from a given DA sample, respectively), and (2) time-averaging. Due to recently published concerns about the corrected posterior age estimate (CPE; sensu Kowalewski et al., 2018; also known as residual time-averaging in some studies), however, we used the interquartile range (IQR) of the average sample age probability distribution, with the quartiles weighted by the age probabilities—the total age variability (IQRTAV)—alone to estimate time-averaging instead of the IQRTAV and CPE (Ritter et al., 2023; see the Supplemental Material).Finally, to compare the contributions of location and burial depth to overall variation in DA sample median age and IQRTAV, we fit a hierarchical Bayesian model to the data for each burial depth as well as the burial depth difference for each DA sample hole (see the Supplemental Material). All data analyses were conducted using R statistical software v4.3.0 (R Core Team, 2023) and RStudio (RStudio Team, 2023).The radiocarbon results indicated that oyster reef DAs are high-resolution archives with abundant shells from the recent past and minimal time-averaging in comparison to other molluscan DAs. Among the 126 dated oyster DA samples, median calibrated ages ranged from 1567 to 2012 CE, but 91% were post-1950 (Fig. 2), and 6.4% of the DA samples had subdecadal-scale IQRTAV (0–10 yr), 72.8% had decadal-scale IQRTAV (11–100 yr), and 20.8% had centennial-scale IQRTAV (101–1000 yr) (Fig. 3; see Appendix S1 in the Supplemental Material for DA sample-level results). Moreover, collocated samples from different burial depths showed the expected temporal order (i.e., deeper = older) in most cases: Out of the 53 sample holes for which both depth intervals were processed and dated, 12 had median DA sample ages for the 15–25 cm burial depth that were older than those of material from the 25–35 cm burial depth, and five of those cases were from a single locality (Lone Cabbage; Fig. 2). The results also showed that the age and time-averaging of a given burial depth can vary substantially over small spatial scales (i.e., both intrareef and interreef assemblage variation; Fig. 2). In fact, the modeled standard deviations (SDs) for spatial variability in median age and IQRTAV (e.g., DA sample-hole-level median SDs were 12.8 and 20.9 yr for median age and IQRTAV, respectively, for the 15–25-cm-depth samples) were of similar magnitude to those for the difference between burial depths (e.g., DA sample-hole-level median depth difference SDs were 29.7 and 37.0 yr for median age and IQRTAV, respectively; Table S5; Figs. S5–S10).To our knowledge, this is the largest study of age-depth relationships and the first study to use radiocarbon to document time-averaging in oyster reef DAs. We found that, relative to other molluscan DAs, the oyster DA samples were younger, were less time-averaged, and had less spatial variability in both calibrated age and time-averaging estimates (Flessa et al., 1993; Meldahl et al., 1997; Kowalewski et al., 1998, 2018; Kosnik et al., 2009, 2015; Krause et al., 2010; Dexter et al., 2014; Dominguez et al., 2016; Ritter et al., 2017; Tomašových et al., 2019; Albano et al., 2020; see additional studies summarized by Kidwell [2013, their table 1]; but see also Tomašových et al. [2018] for an example of a non-reef DA with decadal-scale resolution). This result agrees well with that of a preliminary investigation of time-averaging on two southwest Florida oyster reefs by Lindland et al. (2001) that used amino acid racemization geochronology.Among the few recent studies that (1) focused on mollusks, (2) estimated time-averaging and sample ages in similar ways to our study, and (3) reported their unsummarized, sample-level results, the C. virginica DA samples typically had younger median ages by ~100 yr, and over half of our samples also had lower IQRTAV, some by an order of magnitude or more (Fig. 3). For instance, Dominguez et al. (2016) sampled the upper 20 cm of sediment (medium–fine sand, <2% mud) at six sites with ~9 m water depth in Sydney Harbor, Australia, and found time-averaging (IQRTAV) of ~84–>2000 yr in DAs of the bivalve Fulvia tenuicostata, but the median ages of the samples were ~150 yr. In contrast, IQRTAV across all of the C. virginica DA samples in our study ranged from 6 to 532 yr with a median of ~25 yr. The medians of the median calibrated ages across all of the DA samples from 15–25 cm and 25–35 cm burial depths were 23 yr and 29 yr, respectively. The SDs for both median age and IQRTAV among the locations sampled by Dominguez et al. (2016) were higher than the respective modeled locality-level SDs for the oyster reefs we sampled, despite the much greater geographic area covered by our study (Figs. S5–S10). A similar pattern is evident for DAs from other depositional settings and locations (e.g., southern Brazilian shelf—Ritter et al., 2017; subtidal sand flat, Fernandez Bay, San Salvador Island, Bahamas—Kowalewski et al., 2018; Fig. 3).One exception to this pattern is the study by Tomašových et al. (2018), which found comparable age and time-averaging estimates to ours in Corbula gibba DAs from cores of the Po and Isonzo prodeltas, northern Adriatic Sea (Fig. 3). However, the authors stated that the two deltas have some of the highest sedimentation rates in the northern Adriatic Sea, and median ages and time-averaging estimates for C. gibba DAs from the eastern Gulf of Trieste—across the gulf from the Isonzo River and characterized by low sedimentation rates—were older and more time-averaged than the prodelta samples by nearly two orders of magnitude (Tomašových et al., 2019). In contrast to these large differences in age and time-averaging of C. gibba DAs between depositional settings, decadal-scale resolution appears to be a common feature of DAs from intertidal C. virginica reefs in multiple estuaries across Florida.Overall, our results suggest that oyster reefs have a relatively high shell burial rate and less stratigraphic mixing relative to nonreef molluscan DAs, consistent with the hypothesis that the physical structure of oyster reefs reduces the susceptibility of DAs to some taphonomic processes. Despite their higher temporal resolution than other types of molluscan DAs in most cases, there is still considerable variation in the oyster DA median ages and IQRTAV values (Fig. 2), precluding useful regional or statewide generalizations of age versus burial depth relationships or scales of time-averaging (see the Supplemental Material for an example).This variability highlights a need for additional work to further refine the spatial and temporal specificity of dead C reference information for calibrating the fraction modern carbon (F14C) of estuarine carbonates to improve their accuracy and precision as much as possible (see Supplemental Material for further discussion of this point). It also illustrates why specific geochronological information will be important for many conservation paleobiological contributions to oyster management. Exactly how necessary they are for any given project will depend on the questions investigated, but trends in many indicators of oyster population condition, such as live-oyster size-frequency, are typically tracked at annual or subannual intervals by oyster monitoring programs. To integrate measurements from DA samples with such high-resolution records for trend analyses, it would likely be necessary to know, for instance, whether the median calibrated age and IQRTAV of a DA sample are 2011 and 53 yr, respectively, or 1979 and 16.5 yr—as was the case for two of the DA samples at 15–25 cm burial depth from our Big Hickory locality.Once these data are obtained, comparisons between the DA data and monitoring data that were impractical without them can become feasible—such as the HOBS project’s focus on integrating DA and real-time oyster size data into a single trend analysis, accounting for uncertainty in both oyster size and sample ages—instead of only focusing on more general “before/after” comparisons (e.g., Dietl and Durham, 2016). Further, our study demonstrated that most of the DA samples from oyster reefs represent a relevant time period for ORCP management (i.e., late 1960s to mid-2000s) and can yield decadal-scale (and sometimes subdecadal-scale) retrospective information from ORCP-managed areas where no long-term contemporaneous oyster monitoring took place.Lastly, the apparently limited stratigraphic mobility of shells in recent oyster DAs suggests the intriguing possibility that the degree of time-averaging in an in situ fossil oyster reef (bioherm) is not dramatically greater than that in the DA of a living oyster reef. In this case, bioherms might preserve decadal-scale records from time periods when information at such a fine temporal resolution is exceptionally rare, making them suitable records for otherwise impossible studies of short-term ecological processes in the deep past (e.g., Kowalewski et al., 1998; Kidwell and Tomašových, 2013).We thank Jordon Bright (Amino Acid Geochronology Laboratory, Northern Arizona University) and the staffs of the Keck Accelerator Mass Spectrometer Facility (University of California, Irvine) and the ACE Isotope Laboratory (Northern Arizona University) for 14C analyses, and Matthew Kosnik (Macquarie University, Australia) for help in adapting the Kowalewski et al. (2018) R script. We also thank colleagues and volunteers from the Florida Department of Environmental Protection (FDEP), the Paleontological Research Institution (PRI), and the Florida Department of Agriculture and Consumer Services (FDACS) who helped with field or laboratory work (*FDEP, †PRI, ‡FDACS, §Pelican Island National Wildlife Refuge). Staff: M. Anderson*, P. Benjasirichai‡, E. Bourque*, M. Brown*, C. Brunk*, C. Clark‡, S. Cofone‡, R. Cray*, E. Dark*, M. DeHaven‡, N. Dix*, S. Erickson*, J. Fleiger‡, J. Garwood*, T. Green*, B. Hamill*, K. Harshaw‡, D. Hersl*, T. Jones*, K. Lang*, P. Marcum*, M. McMurray*, B. Mowbray*, R. Noyes*, J. Pier†, R. Prado*, M. Pruden†. Volunteers: B. Alexander*, A. Bishop*, S. Gavirneni†, C. Hormuth*, A. McNeil†, M. Melekos†, R. Mondazzi*, D. Philipp*, E. Prest†, M. Raymond*, M. Schilling*, Z. Siper†, B. Skoblick†, A. Thorsness*, D. Thorsness*, J. Valentine§. We also gratefully acknowledge K. Flessa, M. Savarese, and an anonymous reviewer, whose comments helped to improve the manuscript, and funding from the National Oceanic and Atmospheric Administration (NOAA) Office for Coastal Management under the Coastal Zone Management Act of 1972, as amended, to the Florida Coastal Management Program (NOAA awards NA18NOS4190080 and NA19NOS4190064). The views, statements, findings, conclusions, and recommendations expressed herein are those of the author(s) and do not necessarily reflect the views of the State of Florida, NOAA, or any of their subagencies.
期刊介绍:
Published since 1973, Geology features rapid publication of about 23 refereed short (four-page) papers each month. Articles cover all earth-science disciplines and include new investigations and provocative topics. Professional geologists and university-level students in the earth sciences use this widely read journal to keep up with scientific research trends. The online forum section facilitates author-reader dialog. Includes color and occasional large-format illustrations on oversized loose inserts.