https://doi.org/10.5061/dryad.msbcc2g60

Description of the data and file structure

This repository is associated with McCraw et al., (2024) Ammonites as paleothermometers: Isotopically reconstructed temperatures of the Western Interior Seaway track global records. Paleogeography, Paleoclimatology, Paleoecology.

In this study, 113 ammonite fossils were sampled multiple times for stable oxygen isotopic analyses. The raw data are averaged into a specimen mean value, which can be averaged into a locality mean in which all specimens in a given locality are combined, and then an Ammonite Zone or Compiled Zone average, in which all specimens in a given time period are averaged. The dataset contains associated data described below.

This repository contains the scripts and raw data necessary to perform all analyses conducted in our paper, and produce all associated figures.

Files and variables

File: WIS ammonite raw master data.csv

This file contains the raw isotopic measurements for each sample run (multiple from each individual fossil specimen), and all associated information for each specimen. Every variable is listed usually in triplicate, so each isotopic measurement value has all the associated metadata.

Variables:

·      Loaning Institution- the museum abbreviation for the institution who loaned the specimen

·      Column B- the simplified numeric specimen ID for each sample

·      Column C- the Sample ID indicating the multiple samples taken from single individuals

·      Column D- morphology descriptor of each specimen: baculite, heteromorph, or scaphite

·      Column E- species name

·      Column F: Compiled Zone #, the CZ to which the specimen was assigned based on the species(see Figure 1)

·      Column G & H: raw stable isotope measurements of d13c (carbon) and d18O (oxygen)

·      Column I: Paleolatitude, calculated using the paleolatitude rotation website from the modern collection locality coordinates (columns J &K, latitude and longitude respectively)

·      Column L- assigned Preservation Index (PI) score for each specimen based on assessment of nacre microstructure

·      Column M: Ammonite bioZone (AZ)- the WI Ammonite Taxon Range Biozone (Cobban et al., 2006) each specimen belongs to based on the species

·      Column N: AZ midpoint- the median numerical age of each AZ based on the upper and lower ages of each zone, used for plotting isotopic values in the middle of their AZ for visualization purposes

·      Column O: # Localities represents a unique locality ID based on the latitude value for each specimen, to give a number of localities within AZs or CZ.

File: WIS ammonite specimen averages.csv

Here we provide the ammonite specimen data for n=113. The associated institutional data includes the museum the specimen was loaned from, the specific institutional ID number if applicable, the TT (Tom Tobin) ID number utilized in the study before application of institutional ID's, the taxonomic group or "morphotype" of each specimen (baculite, scaphite, or non-baculitid heteromorph) as a string for import into scripts, the latitude and longitude coordinates of the specimen collection locality, and the species name as described by the museum collection tag or identified by us. Information generated by our lab includes the Ammonite Zone (AZ) that each species belongs to as defined by the Cobban 2006 biostratigraphy of the Western Interior Seaway, and the median age of each AZ as defined by the boundaries of the biozones provided in the Geologic Time Scale (Gale et al., 2020) and refined boundaries from the work of Singer et al., 2024 and Sageman et al., 2014. We also defined Compiled Zones as 2-4 AZ's binned into one zone, so each AZ fits within a CZ. See Figure 1 in associated manuscript. We also provide the Preservation Index value assigned to each fossil specimen, as well as the estimated Late Cretaceous Paleolatitude as determined with the paleolatitude calculator derived from Vaes et al., 2023. Finally, the isotopic information presented is the mean specimen light stable isotope value for each specimen, calculated as the average of all 2-4 samples measured from each specimen, and presented with the 1 sigma Standard Deviation of said calculation of the mean for each specimen.

File: Locality Distance Processing.csv

This file contains all the same information as above, but the sheet is sorted by AZ to group each specimen together in time. The additional information provided is a measurement of the distance of specimen locality to the nearest paleoshoreline (column P) as identified by a given researcher (publication shoreline was taken from in Column Q). See supplemental figures with associated manuscript. This file is read into the associated MATLAB script.

File: morphotype data separation.xlsx

Each tab in this sheet was built from the results of the UN-filtered specimen level IsoCompiler script, sorted and separated by morphotype. All headings are the same as in previous sheets. 

File: Morphotype comparisons.csv

This file contains calculations done by hand with data derived from the MATLAB script, attempting to visualize the mean offset of oxygen isotope values between different morphotype groups of ammonites. This sheet takes the CZ average isotopic values after PI filtering from the morphotype data file (bac_data, het_data, scaph_data), and uses the heteromorph data combined with scaphite data to be a more comparable sample size to baculites. It displays the mean for each CZ if only baculites are included in the data (col. B), and the same for the non-baculitid heteromorphs (E). It also gives the 95% Confidence Interval calculated for each CZ average value (C & F) and displays the number of specimens of that morphotype in each zone (D & G). In column H, the offset between mean CZ for bacs and hets is calculated by subtracting the bac value from the het value for each CZ. There are only 4 CZs with both morphotype groups present, all other CZ have a NaN value as there is no offset to be calculated. Column I has the total number of PI filtered specimens as a check that all specimens are accounted for. We recalculate the mean CZ d18O with both groups included (J).  To "correct" heteromorph values into "baculite" values, we add the average offset of -0.3‰ (cell L7) to the CZ mean value if the zone has any heteromorphs present or just print the existing baculite value if there are no heteromorphs in that zone (K).

File: Obrien d18O data.csv

This file contains supplementary data from O’Brien et al. 2017 (see our manuscript). O'Brien et al. 2017 used the GTS 2012 to calibrate their age model, so to be consistent with our dataset we used their age markers and recalibrated the model using GTS 2020 (Gale et al., 2020), so the ages presented here are not what would be found in the reference citation. Age of samples (column A) are from our updated (recalibrated) age model. See O'Brien et al., 2017 publication and papers that data are sourced from (column C) for additional data about specimens. This data is used in our manuscript to provide a comparison global dataset of marine d18O values, see figure 6.

File: PI smooths.csv

In groups of 3 columns each, these data represent the lowess fit values of d18O and AZ age for different PI filter thresholds. The first column in every group of 3 is the sorted ages from the lowess fit model for the respective numerical PI value, the second column with the Uppercase X represents the d18O values of a 50% lowess span, and the lowercase x column is the d18O values of a 25% lowess span. For example, column A are the ages after lowess fitting the d18O data at a PI filter of 3.5, column B is the d18O lowess values for 50%, and column C is the span at 25%. Column D is the ages for a PI filter of 0, and so on. NaNs represent ages in which there are no specimens to fit at that PI threshold.

File: LIG Slope plot.csv

Contains the calculated slope values for each CZ for which we correlated the mean d18O value and latitudinal span, to determine if there was any correlation between the isotopic values and latitudinal position. Column A is the CZ, column B is the calculated slope extracted from the MATLAB script, Column C is the error associated with the correlation calculation, Column D lists the paleolatitudinal range of that CZ for printing on the figure, and Column E is the number of localities in that CZ on which the correlation was performed.

Code/software

Software used was MATLAB, for which the associated scripts are included.