Descriptive core logs, high-resolution images and derived data for Holocene reef cores collected from 1976 to 2017 along the Florida Keys reef tract

Collection of the Holocene reef cores: The USGS Core Archive (Reich and others 2012; https://usgs.maps.arcgis.com/apps/webappviewer/index.html?id=2b7d5dbc0be340b9b17f6d94aac5b713) housed at the St. Petersburg Coastal and Marine Science Center (SPCMSC) contains an extensive collection of Holocene reef cores from throughout the FKRT from 1976 to present. This archive represents the legacy of more than half a century of geological and ecological research programs in the region (Shinn and Lidz, 2018). The majority of the cores were collected using the USGS hydraulic wireline drilling system (Shinn and others, 1977). The coring system is positioned over the reef by suspending the hydraulic drill from a cable attached to an aluminum tripod. Core barrels and a water hose are attached to the drill and a hydraulic water pump is used to force seawater down the borehole to flush the drill cuttings and facilitate coring. Cores are collected using a double-barrel wireline system, in which successive 5-ft (~1.5 m) sections of reef framework are cored and then recovered by removing the inner barrel, while the outer barrel remains in the reef. Three of the cores (LK-SK-6, UK-CF-1, and UK-CF-4) were collected using the SCARID hydraulic drilling system developed by Hubbard (see Hubbard, 2014). The general concept of the SCARID system is the same as the USGS coring system, but instead of being suspended by a cable from a tripod, the drill is fixed to a rigid frame, which allows for more stability and more precise determination of the depths of recovered material (Hubbard, 2014).

Generation of core photograph and descriptive core logs: In order to capture high-resolution images of the core records, three, close-up photographs of each core were taken using a Canon EOS Rebel digital camera. Intervals between consecutive core barrels are marked in the core boxes were indicated by printed labels in the photographs. Unless otherwise indicated, the labels indicate depth in feet in the core because the USGS coring system is designed to use Imperial measurements. A ruler was also placed within the core photographs for scale and was used to generate scale markers at the bottom of each photo mosaic. Photographic core logs were generated by creating high-resolution photo-mosaics of the three core photographs using the Photomerge tool in Adobe® Photoshop® CC 2017. Using the cores for reference, all recovered reef material visible in the photographic core logs was identified. Coral skeletons were generally identified to the species level; however, some species of coral such as: Orbicella spp. (O. faveolata, O. annularis, and O. franksii), branching Porites spp. (P. divaricata, P. furcata, and P. porites), Pseudodiploria spp. (P. strigose and P. clivosa), and Millepora spp. (M. alcicornis, M. complanata, M. squarosa, and M. striata) were pooled to the genus level because they could not be reliably identified to the species level. Carbonate reef rocks and unconsolidated sediments were identified categorically.
The composition of each core was quantified by tracing the projected surface area of each core component on the digital core photographs using the Area Analysis Tool in the program Coral Point Count with Excel extensions (CPCe; Kohler and Gill, 2006). The data collected in CPCe were used to determine percentage recovery and percentage composition of coral taxa within each recovered interval of the cores. First, the theoretical projected surface area was calculated by multiplying the diameter of the core as measured in CPCe by the actual penetration depth of that core interval. Percentage recovery was then determined by dividing the total measured area of each interval by the theoretical projected surface area of that interval. Similarly, the absolute percentage composition of coral taxa was calculated. These data were simplified into five categories: (1) Acropora palmata, (1) Orbicella spp., (3) Pseudodiploria spp., (4) carbonate reef rock and (5) "Other corals", which included any corals not included in the other categories. The names of the species grouped under the "Other corals" category are listed in the Abbreviations section of Core log key.pdf, which can be found in FKRT_Descriptive_Core_Logs.zip. These data were used to generate simplified core logs for each core using RockWare© LogPlot™ 7 software. The symbology of the core logs represents the dominant coral taxon or reef carbonate in an interval. The description of the interval includes any taxa that represented at least 5% of that interval by area. The total percent recovery of each interval is represented by the width of the bars on the right side of the core log and the coloration of those bars represents the percentage contribution of the four categories of core constituents described above. The tops of the core logs also contain descriptive information about the core locations (summarized in FKRT_Core_Data), who collected the cores and when (from Reich and others, 2012), and who compiled the core log (L.T. Toth). Please note that the core logs are generally based on the information that could be gathered from the core boxes. Where core logs from the field were available, USGS staff also incorporated information about the position of likely voids in the cores, but this information was not available for cores collected before 2006. No attempt was made to recreate the detailed core logs for the cores from Sand Key (LK-SK-2, LK-SK-5, LK-SK-6, LK-SK-7) and Carysfort Reefs (UK-CF-1, UK-CF-2, UK-CF-3, UK-CF-4, UK-CF-5) generated by Toscano (1996) because the core boxes did not contain adequate information to cross-reference with the published core logs. Additionally, a core log for core DT-LB-2 was not generated because the recovery of this core was so poor.
Also included are calibrated ages and the 95% confidence intervals (CI) of those calibrated ages on the core logs (see Radiometric dating of Holocene reef cores process step). Only ages that were significantly different from adjacent ages in the cores logs were included. Researchers determined whether the differences between any pair of sequential dates with conventional 14C ages or calibrated U-series ages within 500 years (yrs) of one another were significant by calculating the standard error of the difference (SEdiff) and 95% CIs of two ages. If the CIs did not overlap, the ages were considered to be significantly different from one another and were included in the accretion calculations. In instances in which the 95% CI of two sequential dates in a core overlapped (were not significantly different), we omitted the age with a 1? uncertainty >50 yrs or the age that allowed for the most even spread of ages within the core.

Radiometric dating: Coral samples collected for radiometric dating were cut from the cores using a tile saw at USGS-SPCMSC, which is dedicated to that purpose. The samples were generally sonicated for 15 minutes in a bath of warm, deionized water and dried at 60 degrees Celsius prior to shipping to the laboratories. Radiocarbon ages were generally determined using accelerator mass spectrometry (AMS) at either the Lawrence Livermoore National Laboratory (processed at the USGS Radiocarbon Laboratory in Reston, VA) or the National Ocean Sciences AMS facility at Woods Hole Oceanographic Institution. Some samples were dated using standard radiometric dating at the University of Miami Radiocarbon Laboratory, Beta Analytic, Inc. or Geochron Laboratories. Conventional 14C ages were corrected for fractionation of 13C. The ?13C of those samples was either measured by University of California, Davis Stable Isotope Laboratory or, if not measured, were assumed to be 0±3‰ (Toth and others, 2018). The conventional radiocarbon ages were calibrated in Calib 7.0.2 (http://calib.org/calib/) using the time-varying estimates of the local reservoir age, ?R, for the nearshore and open-ocean environments of the FKRT developed by Toth and others (2017; available at https://coastal.er.usgs.gov/data-release/doi-F7P8492Q/). The full radiocarbon dataset is available in Toth and others (2018). Ages determined by U-series analysis were conducted using either multicollector inductively coupled plasma mass spectrometry at Xi’an Jiaotong University in China using standard methodologies (see Toth and others, 2018) or by Thermal Ionization Mass Spectrometry (see Toscano 1996; Toscano and Lundberg 1998, 1999).

Determining reef thickness: As a measure of the degree of reef development across the subregions of the FKRT, scientists quantified the thickness of the Holocene reef framework using the core records that reached the Pleistocene bedrock. Researchers only included records where they were able to confidently identify the Holocene-Pleistocene boundary on the basis of at least one of the following three criteria: 1) ages from samples on either side of the boundary, 2) the presence of a soilstone (or “caliche”) crust characteristic of the Holocene-Pleistocene boundary in south Florida, or 3) a clear distinction between the Holocene and Pleistocene based on diagenetic alteration to calcite or a shift from coral framework to carbonate grainstones or boundstones. In the 31 cores that met these criteria, the estimated depth of core penetration to the base of the Holocene reef framework was used to quantify reef thickness. In some of the cores, there was a sand layer between Holocene and Pleistocene facies. The thickness of the Holocene reef with and without these sediment layers are both included in the FKRT_Core_Data dataset. "Total Holocene thickness" includes the unconsolidated sediment layer and "Holocene reef thickness" does not include the sediment layer.

Reef accretion, initiation, and senescence: Rates of vertical reef accretion (referred to throughout as simply “reef accretion”), in meters per thousand years (equivalent to mm/yr), of dated intervals within individual cores were calculated by dividing the length of each interval by the time span of that interval (summarized in FKRT_Temporal_Core_Data). The ages used in the accretion calculation were the median probabilities of the radiocarbon calibrations. Temporal variability in overall FKRT reef accretion during the Holocene was evaluated by averaging reef accretion rates from all cores (±SE) within 500-yr bins from 8500 years before present (BP; where present is 1950) to present. Similarly, USGS researchers quantified spatial variability in temporal changes in reef accretion by calculating average accretion rates for each subregion (±SE) within 500-yr age bins from 7000 BP to present (FKRT_Subregion_Accretion_500yr_Bins). Researchers also calculated the average rates of reef accretion over the lifespan of the reef represented by each core record by dividing the depth of the deepest age in the core by that age. USGS researchers determined the average timing of reef initiation based on ages in the cores that were within 1 m of the Holocene-Pleistocene boundary. A reef was considered to be senescent when the rate of vertical reef accretion was no longer keeping pace with relative sea-level (RSL) rise. Researchers used the modeled rates of SL rise from Khan and other (2017) to approximate the rates of RSL rise through the Holocene and compared those rates to the rates of reef accretion in the cores (summarized in FKRT_Temporal_Core_Data). In most cores, reefs accreted rapidly for some time, but accretion declined significantly in recent millennia. For these cores, the uppermost measured age in the core before the reef began accreting more slowly than sea level was rising (within average uncertainty of the RSL rate reconstruction: ±1 m [2?]) was used to estimate the timing of reef senescence. In four cores accretion rates were never within 1 m of the modeled rate of RSL rise. These cores were excluded from the analysis of the timing of reef senescence. USGS researchers also did not determine the timing of reef senescene for cores that only started accreting on pace with the rate of RSL rise during the late Holocene (after 4200 years ago) because these records represented a separate, more recent period of reef growth. In cases where accretion rates measured in a core never fell below 1 meter per thousand years (m/ky), the age of the top surface of the core was used to estimate the timing of reef senescence. These samples were always from the first barrel (upper 5 ft, 1.5 m) of the core, and were generally within 0.5 m below the reef surface.

Criteria for excluding data from accretion analysis: To ensure that all the cores included Toth and others' (2018) study were from a similar environmental setting, USGS researchers used the model of Holocene sea-level variability developed by Khan and others (2017) to reconstruct the paleodepths of all dated intervals in the cores (summarized in FKRT_Temporal_Core_Data). Researchers excluded any intervals within the cores that were deposited in depths significantly >10 m (based on 95% confidence intervals of estimated paleodepths). We also excluded core records from the USGS coral archive that were taken from non-reef-building habitats (for example, reef grooves). Accretion rates can be artificially inflated when sequential dates within a core are similar enough that the entire layer could have been deposited simultaneously. To avoid this potential complication, USGS researchers determined whether the differences between any pair of sequential dates with conventional 14C ages or calibrated U-series ages within 500-yrs of one another were significant by calculating the standard error of the difference (SEdiff) and 95% CIs of two ages. If the CIs did not overlap, the ages were considered to be significantly different from one another and were included in the accretion calculations. In instances in which the 95% CI of two sequential dates in a core overlapped (were not significantly different), researchers omitted the age with a 1? uncertainty >50 yrs or the age that allowed for the most even spread of ages within the core.

The locations for cores collected prior to the mid 1990s are estimates. For the remaining cores, the NAD83 latitude/longitude coordinates were collected using a Garmin GPSMAP 78SC handheld Global Positioning System (GPS) and written into field notebooks. All of the high-resolution coral core photographs included in FKRT_Core_Photographs.zip were taken at the USGS SPCMSC, using a Canon EOS Rebel camera. EXchangeable Image File format (EXIF) headers were initially populated by the camera’s imaging software but were subsequently updated by USGS staff to include additional core-related details. A Python version 2.7.3 script (UpdatePhotoEXIFv3.py) was run to incorporate location information and auxiliary details into the appropriate locations in the EXIF header of each full-resolution JPEG image. The Python script used ExifTool (version 10.25) to write the information to the image headers. The following tags were populated in the JPEG image headers. Information is duplicated in some tags. This was done because different software packages access different tags.
GPS tags: The values populated are unique for each image and based on the information exported from the handheld GPS or estimated location.

GPSLatitudeRef
GPSLatitude
GPSLongitudeRef
GPSLongitude
GPSTimeStamp
GSPDateStamp

JPEG tags: The tag is listed along with the information used to populate it - which is the same for every image taken with a particular camera. The following information is based on the Canon EOS Rebel camera.

comment: Photo of cores collected from the Florida Keys Reef Tract (FKRT) to investigate Holocene coral-reef development. These data are associated with field activity number 2018-301-DD (https://cmgds.marine.usgs.gov/fan_info.php?fan=2018-301-DD). Published as USGS data release DOI:10.5066/F7NV9HJX.

EXIF tags: The tag is listed along with the information used to populate it - which is the same for every image.

ImageDescription: Photograph of coral core collected from the Florida Keys Reef Tract, Florida.

Artist: Lauren Toth
Copyright: Public Domain - please credit U.S. Geological Survey

IPTC tags: The tag is listed along with the information used to populate it - which is the same for every image.

Credit: U.S. Geological Survey
Contact: gs-g-spcmsc_data_inquiries@usgs.gov
Keywords: Dry Tortugas National Park, Florida Keys Reef Tract (FKRT), Florida, Florida Keys, coral, coral reef, Holocene, reef accretion, paleoecology, paleoenvironmental reconstruction, 2018-301-DD, USGS
CopyrightNotice: Public Domain - please credit U.S. Geological Survey
Caption-Abstract: Photograph of coral core collected from the Florida Keys Reef Tract, Florida.

XMP tags: The tag is listed along with the information used to populate it - which is the same for every image.

Caption: Photograph of coral core collected from the Florida Keys Reef Tract, Florida.

To extract the information from the image headers using ExifTool, the following command can be used (tested with ExifTool version 10.25):

exiftool.exe -csv -f -filename -GPSTimeStamp -GPSLongitude -GPSLatitude -n -Artist -Credit -comment -keywords -Caption -Copyright -CopyrightNotice -Caption-Abstract -ImageDescription photos/*.jpg > out.csv

The -csv flag writes the information out in a comma-delimited format. The -n option formats the latitude and longitude as signed decimal degrees.

Added keywords section with USGS persistent identifier as theme keyword.