Taphonomic and supplement data for: Quantification and mitigation of taphonomic biases in multi-taxon marine fossil assemblages using a mass-based, fragment-inclusive sampling approach
Data files
Sep 24, 2026 version files 1.03 MB
-
Appendix_1_Part_0_Custom_Helper_Functions.R
9.54 KB
-
Appendix_1_Part_1_Initial_Data_Processing.R
26.63 KB
-
Appendix_1_Part_2_Transect-Exclusionary_Data_Analyses_and_Figures.R
39.33 KB
-
Appendix_1_Part_3_Transect-Based_Analyses_and_Figures.R
71.37 KB
-
Appendix_2_Raw_Data.csv
20.42 KB
-
Appendix_3_Morphometric_Mass_and_Soft_Tissue_Data.csv
26.37 KB
-
Appendix_4_Photographed_Transect_Specimens.csv
79.37 KB
-
Appendix_5_Total_Corrected_Mass_Data_Transect-Inclusive.csv
11.22 KB
-
Appendix_6_Total_Corrected_Count_Data_Transect-Inclusive.csv
9.91 KB
-
Appendix_7_Total_Corrected_Mass_Data.csv
9.02 KB
-
Appendix_8_Total_Corrected_Count_Data.csv
7.72 KB
-
coast.bna
435.02 KB
-
coastCB.bna
69.19 KB
-
coastCK.bna
153.10 KB
-
coastCU.bna
21.62 KB
-
coastFK.bna
24.64 KB
-
README.md
19.19 KB
Abstract
Here we propose a fragment-inclusive mass-based (FIMB) approach to mitigate taphonomic biases and improve quantitative estimates of faunal composition in paleoecological studies. Using sympatric assemblages of irregular echinoids and shell-bearing mollusks collected from contemporary, shallow-water seafloors across four regions ([1] Carrabelle, Florida; [2] Cedar Key, Florida; [3] Curaçao, Netherlands Antilles; [4] Florida Keys, Florida), we tested the hypothesis that FIMB produces less biased estimates of faunal composition than traditional fragment-exclusionary (complete specimens only) mass or count-based approaches. We also tested two auxiliary taphonomic hypotheses, implied by the expected lower preservational potential of irregular echinoids compared to sympatric mollusks: (1) Irregular echinoid tests fragment more frequently than mollusk shells; (2) Whole specimen live-dead proportions are higher for irregular echinoids than for mollusks. Results support both taphonomic hypotheses, with irregular echinoids exhibiting higher fragmentation estimates and live-dead proportions than sympatric mollusks. However, both taxa are well represented by fragments. Consequently, the FIMB approach provides estimates of relative abundance of echinoids that better approximate their relative abundance in live assemblages. In contrast, traditional fragment-exclusionary approaches indicated that irregular echinoids were either absent or at most extremely rare in death assemblages. Our results demonstrate that although the FIMB approach offers lower taxonomic resolution than the methods based entirely on whole specimens, the fragment-inclusive method improves estimates of faunal composition by mitigating pronounced taphonomic biases that can affect multi-taxic assemblages.
Dataset DOI: 10.5061/dryad.g4f4qrg4f
Description of the data and file structure
Found here are the appendices and supplementary figures and tables from the manuscript "Quantification and mitigation of taphonomic biases in multi-taxic marine fossil assemblages using a fragment-inclusive, mass-based approach". These files are based on skeletal materials of irregular echinoids and shelled mollusks gathered from shallow coastal ecosystems in: (1) Carrabelle, Florida; (2) Cedar Key, Florida; (3) Florida Keys, Florida; and (4) Curaçao, The Netherlands Antilles. In this study, we demonstrate a fragment-inclusive, mass-based (FIMB) approach to quantifying relative abundances of multiple higher taxa, while comparing it to more traditional fragment-exclusionary mass and count-based approaches.
Files and variables
File: Appendix_1_Part_0_Custom_Helper_Functions.R
Description: The first section of a sequential R-code that primarily functions to define custom helper functions. Within this script the following functions are defined: (1) Image-Mass Model ["mass_model"]; (2) Cumulative Image-Mass Modeling Application ["mass.F"]; (3) Specimen-Level Image-Mass Modeling Application ["mass.P"]; (4) Image-Mass Modeling Application by Taxa ["predict_mass_base"]; (5) Transect Sample Mass & Count Estimation ["summarize_sam"]; (6) Boot Confidence Intervals ["CI_boot"]; (7) Sample Mass Standardization ["per.g.function"]; (8) Computing Variable-Level Mass Means + Proportions ["compute_means_props"]; (9) Higher-Taxonomic Composition Matrix ["compute_composition"];(10) Death Assemblage Fragment Proportion ["fragmentation"]; (11) Taxon-Selective Death Assemblage Fragment Proportion ["get_frag_prop"]; (12) PERMANOVA With Fragmentation Estimate Summaries ["run_permanova"]; (13) Live-Dead Proportion Calculation ["live.dead"]; (14) Live-Dead Dataframe Formation ["compute_ld_dataset"]; (15) Form Wilcoxon Signed-Rank Summary Datafragme ["wilcox_summary"]; (16) Assemblage Means Differences For Bootstrap ["twomeans.F"]; (17) Assemblage Confidence Intervals For Bootstrap ["boot.F"].
File: Appendix_1_Part_1_Initial_Data_Processing.R
Description: The first true section of a sequential R-code that primarily functions as pre-processing of data for further analyses in later sections. Within this script the following tasks are performed: (1) Image-based morphometric-mass modeling; (2) Sediment density calculation and transect mass estimation; (3) Data cleaning and bivalve count correction; (4) Soft tissue mass correction; (5) Subsample mass/count corrections; (6) Sample patchiness corrections for live specimens (averaged by site); (7) Production of transect-inclusive, corrected mass [Appendix 5] and count [Appendix 6] datasets; (8) Sampling method [sample type] contribution evaluation (9) Production of transect-exclusionary mass [Appendix 7] and count [Appendix 8] datasets; and (10) Producing study area map [Figure 1].
File: Appendix_1_Part_2_Transect-Exclusionary_Data_Analyses_and_Figures.R
Description: The second true section of a sequential R-code that primarily functions to process and analyze the transect-exclusionary dataset. Within this script the following tasks are performed: (1) Computing standardized mass, proportions and compositions; (2) Fragment death assemblage proportion calculations and analyses; (3) Sediment depth death assemblage composition calculation and analyses; (4) Live-dead proportion calculations and analyses; (5) Mass vs count pattern analyses; (6) Comparing methods of relative abundance estimation. Various figures are produced to illustrate the results.
File: Appendix_1_Part_3_Transect-Based_Analyses_and_Figures.R
Description: The third true section of a sequential R-code that primarily functions to process and compare transect-inclusive and transect-exclusionary datasets. Within this script the following tasks are performed: (1) Computing standardized mass, proportions and compositions; (2) Fragment death assemblage proportion calculations and analyses; (3) Live-dead proportion calculations and analyses; (4) Mass vs count pattern analyses; (5) Comparing methods of relative abundance estimation. Various figures are produced to illustrate the results.
File: Appendix_2_Raw_Data.csv
Description: Base sample-level dataset of raw mass estimates and counts of different specimen types (live mollusks, whole-dead [complete] mollusks, mollusk fragments, live echinoids, whole-dead [complete] echinoids, echinoid fragments). Also includes miscellaneous information regarding sampling conditions. Includes all processed samples.
Variables
- SID: unique sample identifier
- Region: sample region of collected sample
- Site: unique sample locality at sample region
- Sample: unique sample at sample locality
- SamType: sample collection method (surface bulk [bulk], HVAC bulk [hvac], transect-survey [transect])
- SamDepth: sediment depth of sample collection (surface = 0-10cm; deep = 30-40cm)
- SamDist/Surface Area: distance of samples from start point [m] for samples with bulk SamType/sampling dimensions [m] for samples with hvac or transect
- Sediment Volume Live (cm3): estimated volume [cm^3^] of sediment sampled for live specimens within the collected sample
- Sediment Volume Dead (cm3): estimated volume [cm^3^] of sediment samples for dead and fragmentary specimens within the collected sample
- Lat: latitude (WGS84)
- Long: longitude (WGS84)
- Depth: water depth of sample site [m]
- Temp: water temperature at sample site (from dive computer) [°C]
- Year: year of sampling
- Month: month of sampling
- Day: day in month of sampling
- Time: start of sampling dive
- BottomTime: dive time [min]
- CtLiveM(WS): raw count of live mollusks from the whole sample
- WgLiveM(WS): raw dry weight of live mollusks [g] from the whole sample
- CtDeadM(SS): raw count of dead complete mollusks from the subsample
- WgDeadM(SS): raw dry weight of dead complete mollusks [g] from the subsample
- CtFragM(SS): raw count of mollusk fragments from the subsample
- WgFragM(SS): raw dry weight of mollusk fragments [g] from the subsample
- CtLiveE(WS): raw count of live echinoids from the whole sample
- WgLiveE(WS): raw dry weight of live echinoids [g] from the whole sample
- CtDeadE(WS): raw count of dead complete echinoids from the whole sample
- WgDeadE(WS): raw dry weight of dead complete echinoids [g] from the whole sample
- CtFragE(SS): raw count of echinoid fragments from the subsample
- WgFragE(SS): raw dry weight of echinoid fragments [g] from the subsample
- WgTotal(WS): dry weight of the entire sample [g]
- WgSS: dry weight of subsample [g]
- BivCt(SS): counts of articulated dead bivalve mollusks from subsample
- GasCt(SS): count of dead complete gastropod and scaphopod mollusk shells from subsample
- Notes: beneficial information regarding data processing
File: Appendix_3_Morphometric_Mass_and_Soft_Tissue_Data.csv
Description: Dataset of various exemplar specimens used to produce image-mass modeling and soft-tissue mass estimation data.
Variables
- SID: unique specimen/group identifier
- SAMPLESID: sample of origin
- SPECIMEN: specimen ID number of sample
- CLADE: taxonomic classification by higher taxon
- GENUS: taxonomic classification by genus
- SPECIES: taxonomic classification by species
- ID CONFIDENCE: ordinal scale of species identification (1 = not very confident, 2 = somewhat confident, 3 = very confident, 4 = positive).
- STATUS: live (experimental) or dead (control) for soft tissue mass correction
- INDIVIDUAL: status as an individual (I) or a group (G) for mass measurements
- LENGTH: longest anterior-posterior distance [mm]
- WIDTH: longest lateral distance [mm]
- PRE-BLEACH MASS: dry weight w/ soft tissues [g]
- POST-BLEACH MASS: dry weight w/o soft tissues [g]
- POST-PRE MASS RATIO: post-bleach mass divided by pre-bleach mass
- PERCENT LOSS: percentage of mass lost during bleaching process
- LAT: latitude (WGS84)
- LONG: longitude (WGS84)
- NOTES: miscellaneous notes about the specimen/group
File: Appendix_4_Photographed_Transect_Specimens.csv
Description: Dataset of various released [non-exemplar] live specimens whose masses were estimated using image-mass modeling.
Variables
- SID: unique specimen/group identifier
- SAMPLESID: sample of origin
- SPECIMEN: specimen ID number of sample
- IDENTIFYING#: sample ID to link to Appendix 2
- CLADE: taxonomic classification by higher taxon
- GENUS: taxonomic classification by genus
- SPECIES: taxonomic classification by species
- LENGTH: longest anterior-posterior distance [mm]
- WIDTH: longest lateral distance [mm]
- MORPHOMETRIC MEAN: square root of length*width measurements [mm]
- PREDICTED MASS: dry weight [g] to be predicted in image-mass model
- LATITUDE: latitude (WGS84)
- LONGITUDE: longitude (WGS84)
- NOTES: miscellaneous notes about the specimen/group
File: Appendix_5_Total_Corrected_Mass_Data_Transect-Inclusive.csv
Description: Sample-level dataset of weights of different specimen types (live mollusks, whole-dead [complete] mollusks, mollusk fragments, live echinoids, whole-dead [complete] echinoids, echinoid fragments) standardized to per mass unit (g) of sediment.
Variables
- SID: unique sample identifier
- Region: sample region of collected sample
- Site: unique sample locality at sample region
- Sample: unique sample at sample locality
- SamType: sample collection method (surface bulk [bulk], HVAC bulk [hvac], transect-survey [transect])
- SamDepth: sediment depth of collected sample (surface = 0-10cm; deep = 30-40cm)
- LiveMol/g: dry weight of live mollusks per mass unit of sediment [g/g] averaged by site
- DeadMol/g: dry weight of dead complete mollusks per mass unit of sediment [g/g] unique to sample
- FragMol/g: dry weight of mollusk fragments per mass unit of sediment [g/g] unique to sample
- LiveEch/g: dry weight of live echinoids per mass unit of sediment [g/g] averaged by site
- DeadEch/g: dry weight of dead complete echinoids per mass unit of sediment [g/g] unique to sample
- FragEch/g: dry weight of echinoid fragments per mass unit of sediment [g/g] unique to sample
File: Appendix_6_Total_Corrected_Count_Data_Transect-Inclusive.csv
Description: Sample-level dataset of counts of different specimen types (live mollusks, whole-dead [complete] mollusks, mollusk fragments, live echinoids, whole-dead [complete] echinoids, echinoid fragments) standardized to per mass unit (g) of sediment.
Variables
- SID: unique sample identifier
- Region: sample region of collected sample
- Site: unique sample locality at sample region
- Sample: unique sample at sample locality
- SamType: sample collection method (surface bulk [bulk], HVAC bulk [hvac], transect-survey [transect])
- SamDepth: sediment depth of collected sample (surface = 0-10cm; deep = 30-40cm)
- CtLiveMol.g: count (n) of live mollusks per mass unit of sediment [n/g] averaged by site
- CtDeadMol.g: count (n) of dead complete mollusks per mass unit of sediment [n/g] unique to sample
- CtFragMol.g: count (n) of mollusk fragments per mass unit of sediment [n/g] unique to sample
- CtLiveEch.g: count (n) of live echinoids per mass unit of sediment [n/g] averaged by site
- CtDeadEch.g: count (n) of dead complete echinoids per mass unit of sediment [n/g] unique to sample
- CtFragEch.g: count (n) of echinoid fragments per mass unit of sediment [n/g] unique to sample
File: Appendix_7_Total_Corrected_Mass_Data.csv
Description: Sample-level dataset of weights of different specimen types (live mollusks, whole-dead [complete] mollusks, mollusk fragments, live echinoids, whole-dead [complete] echinoids, echinoid fragments) standardized to per mass unit (g) of sediment. Includes only bulk samples (hand-bulk/HVAC).
Variables
- SID: unique sample identifier
- Region: sample region of collected sample
- Site: unique sample locality at sample region
- Sample: unique sample at sample locality
- SamType: sample collection method (surface bulk [bulk], HVAC bulk [hvac], transect-survey [transect])
- SamDepth: sediment depth of collected sample (surface = 0-10cm; deep = 30-40cm)
- LiveMol/g: dry weight of live mollusks per mass unit of sediment [g/g] averaged by site
- DeadMol/g: dry weight of dead complete mollusks per mass unit of sediment [g/g] unique to sample
- FragMol/g: dry weight of mollusk fragments per mass unit of sediment [g/g] unique to sample
- LiveEch/g: dry weight of live echinoids per mass unit of sediment [g/g] averaged by site
- DeadEch/g: dry weight of dead complete echinoids per mass unit of sediment [g/g] unique to sample
- FragEch/g: dry weight of echinoid fragments per mass unit of sediment [g/g] unique to sample
File: Appendix_8_Total_Corrected_Count_Data.csv
Description: Sample-level dataset of counts of different specimen types (live mollusks, whole-dead [complete] mollusks, mollusk fragments, live echinoids, whole-dead [complete] echinoids, echinoid fragments) standardized to per mass unit (g) of sediment. Includes only bulk samples (hand-bulk/HVAC).
Variables
- SID: unique sample identifier
- Region: sample region of collected sample
- Site: unique sample locality at sample region
- Sample: unique sample at sample locality
- SamType: sample collection method (surface bulk [bulk], HVAC bulk [hvac], transect-survey [transect])
- SamDepth: sediment depth of collected sample (surface = 0-10cm; deep = 30-40cm)
- CtLiveMol.g: count (n) of live mollusks per mass unit of sediment [n/g] averaged by site
- CtDeadMol.g: count (n) of dead complete mollusks per mass unit of sediment [n/g] unique to sample
- CtFragMol.g: count (n) of mollusk fragments per mass unit of sediment [n/g] unique to sample
- CtLiveEch.g: count (n) of live echinoids per mass unit of sediment [n/g] averaged by site
- CtDeadEch.g: count (n) of dead complete echinoids per mass unit of sediment [n/g] unique to sample
- CtFragEch.g: count (n) of echinoid fragments per mass unit of sediment [n/g] unique to sample
File: coast.bna
Description: GIS file used in conjunction with the "mapmaker" function (Kowalewski 2025) in the code file “Appendix_1_Part 1_Initial_Data_Processing.R” to create our study area map (Figure 1) in the main manuscript.
File: coastCK.bna
Description: GIS file used in conjunction with the "mapmaker" function (Kowalewski 2025) in the code file “Appendix_1_Part 1_Initial_Data_Processing.R” to create our study area map (Figure 1) in the main manuscript. Specifically provides coordinates for the Cedar Key (CK) study region.
File: coastCB.bna
Description: GIS file used in conjunction with the "mapmaker" function (Kowalewski 2025) in the code file “Appendix_1_Part 1_Initial_Data_Processing.R” to create our study area map (Figure 1) in the main manuscript. Specifically provides coordinates for the Carrabelle (CB) study region.
File: coastCU.bna
Description: GIS file used in conjunction with the "mapmaker" function (Kowalewski 2025) in the code file “Appendix_1_Part 1_Initial_Data_Processing.R” to create our study area map (Figure 1) in the main manuscript. Specifically provides coordinates for the Curaçao (CU) study region.
File: coastFK.bna
Description: GIS file used in conjunction with the "mapmaker" function (Kowalewski 2025) in the code file “Appendix_1_Part 1_Initial_Data_Processing.R” to create our study area map (Figure 1) in the main manuscript. Specifically provides coordinates for the Florida Keys (FK) study region.
Code/software
This dataset utilizes R (R-4.5.3) along with a sequential four-part code (Appendix 1) to process and analyze data.
Appendix 1 Part 0 Custom Helper Functions - Defines custom helper functions to streamline the rest of the code.
Appendix 1 Part 1 Initial Data Processing - Clean and processes data with some data analyses and figure creation.
Appendix 1 Part 2 Transect-Exclusionary Data Analyses - The workhorse of the hypothesis testing of the manuscript. Performs most of the analyses and forms most of the figures directly related to the conclusions of the work.
Appendix 1 Part 3 Transect-Inclusive Data Analyses - Performs analyses and forms figures directly related to the transect-inclusive dataset.
Packages Used:
Canty, A., and B. Ripley. 2025. Package “boot”. R package version for boot: Bootstrap Functions v1.3-32. https://CRAN.R-project.org/package=boot. Maintainer: Alessandra R. Brazzale. https://doi.org/10.32614/CRAN.package.boot
Csárdi, G., J. Hester, H. Wickham, W. Chang, M. Morgan, D. Tenenbaum. 2024. Package “remotes”. R package version for**** remotes: R Package Installation from Remote Repositories, Including ‘GitHub’ v2.5.0. https://CRAN.R-project.org/package=remotes. Maintainer: Gábor Csárdi. https://doi.org/10.32614/CRAN.package.remotes
Kowalewski, M. 2025. Package “mapmaker”, R package version for mapmaker: Create maps using GSHHS-NOAA data files v0.1.0, https://github.com/MJKowalewski/mapmaker. Maintainer: Michal Kowalewski.
Oksanen, J., G. Simpson, F. Blanchet, R. Kindt, P. Legendre, P. Minchin, R. O'Hara, P. Solymos, M. Stevens, E. Szoecs, H. Wagner, M. Barbour, M. Bedward, B. Bolker, D. Borcard, T. Borman, G. Carvalho, M. Chirico, M. de Cáceres, S. Durand, H. Evangelista, R. FitzJohn, M. Friendly, B. Furneaux, G. Hannigan, M. Hill, L. Lahti, C. Martino, D. McGlinn, M. Ouellette, E. Ribeiro Cunha, T. Smith, A. Stier, C. ter Braak, and J. Weedon. 2025. Package “vegan”, R package version for vegan: community ecology package. v2.7-1, https://CRAN.R-project.org/package=vegan. Maintainer: Jari Oksanen. https://doi.org/10.32614/CRAN.package.vegan
Thulin, M. 2025. Package “boot.pval”, R package version for boot.pval: Bootstrap p-values v0.7.0, https://CRAN.R-project.org/package=boot.pval. Maintainer: Måns Thulin. https://doi.org/10.32614/CRAN.package.boot.pval
Wickham, H., R. François, L. Henry, K. Müller, and D. Vaughan. 2023. Package “dplyr”, R package version for dplyr: a grammar of data manipulation v1.1.4, https://CRAN.R-project.org/package=dplyr. Maintainer: Hadley Wickham. https://doi.org/10.32614/CRAN.package.dplyr
Access information
Other publicly accessible locations of the data:
- N/A
Data was derived from the following sources:
- N/A
