Data and code from: Connectivity of proetid trilobites among geographic regions across the Late Devonian and Carboniferous
Data files
Jul 29, 2026 version files 19.72 KB
-
Brez_Bash_Mosc.csv
999 B
-
Brez_EVisean.csv
1.83 KB
-
Brez_Kas_Ghez.csv
299 B
-
Brez_LVisean.csv
1.24 KB
-
Brez_MidtoLate_Tour.csv
2.08 KB
-
Brez_Serp.csv
592 B
-
Brezinski_ETour_Occurrences.csv
1.10 KB
-
Fam_Occurrences.csv
1.07 KB
-
Fras_occurrences.csv
522 B
-
network_and_cooccur_code.R
6.16 KB
-
README.md
3.83 KB
Abstract
The Carboniferous world was a time of transition in terms of climate, geography, and its characteristic faunas. A holdover from the Devonian, proetid trilobites were the final representatives of the Class Trilobita and maintained diversity throughout the remainder of the Paleozoic. The biogeographic patterns of these trilobites in the Late Paleozoic have been studied to a certain degree, but not in totality. This study uses a network analysis and corresponding metrics to assess the connectivity of regions where proetids were found beginning in the Famennian until the end of the Carboniferous Period. Two time bins, the latter part of the Tournaisian and the Serpukhovian, display networks that are unique in comparison to the other time bins. The mid–late Tournaisian shows a network with strong connections and therefore more open dispersal pathways, while the Serpukhovian shows three distinct regional clusters. The closing of the Rheic Ocean most likely reduced connectivity from the Serpukhovian onwards, and strong connectivity between regions never fully returned.
Dataset DOI: 10.5061/dryad.kh18932q6
Description of the data and file structure
Principal Investigator Contact Information
Name: Katherine Jordan-Burmeister
Institution: University of Tennessee
Email: kjorda36@utk.edu
Dataset Overview
This dataset contains the data and code required to replicate analyses in Jordan-Burmeister (2027). These analyses include detecting network connections across geographic sites from the Late Devonian through the Carboniferous and calculating co-occurrence networks of genera per time bin.
Data files are coded presence-absence occurrences of genera by site. A presence is coded by a "1" while an absence is coded as a "0". Sites are determined by those described in Brezinski (2023).
The "Brezinski" or "Brez" at the beginning of the file names indicate these occurrences were taken directly from Brezinski (2023). The other two files with "Fras" and "Fam" represent the Devonian intervals (Frasnian and Famennian, respectively) and were coded using PBDB data (accessed January 2026). These data files can be read in using the R code provided (network_and_cooccur_code.R).
Files and variables
The data files have the following names:
Fras_occurrences.csv
Fam_Occurrences.csv
Brezinski_ETour_Occurrences.csv
Brez_MidtoLate_Tour.csv
Brez_EVisean.csv
Brez_LVisean.csv
Brez_Serp.csv
Brez_Bash_Mosc.csv
Brez_Kas_Ghez.csv
All cells should contain either a "1" for presence or a "0" for absence of a genus in each site.
Sites are as follows:
EastUS- Eastern USA
WestUS- Western USA
UK- United Kingdom
WestEur- Western Europe
CenEur-Central Europe
EastEur- Eastern Europe
Austria/ATA- Austria and Armorican Terrane Assemblage
Samerica- South America
Nafrica- North Africa
Russia- Russia
China- China
Japan- Japan
SEAsia- Southeast Asia
Aus- Australia
Alaska- Alsa
Code is named as such:
network_and_cooccur_code.R
Code/software
The R code provided allows a user to rerun all network analyses. The first part of the code allows one to load each file in separately as a matrix ("mat"). I ran these one by one, but these can be run altogether as long as each matrix is labelled individually (e.g., "mat" to "mat_Fras" to load in Fras_occurrences.csv).
These matrices are then transposed ("t(mat)") and co-occurrence for each matrix is calculated using the package "cooccur". Networks can then be created using "graph_from_adjacency_matrix". Other metrics such as degree (de), strength (st), and betweenness (be) are calculated after using "graph_from_adjacency_matrix". Networks can be plotted using the "plot" function. Additional plotting functions are provided which include both aesthetic choices and weighting of the network edges. Lastly, the functions " hclust()" and "cutree()" can be used to cluster sites based on distances (based on the transposed matrices) using different clustering methods (e.g., Ward's D).
Any questions about methods and code can be emailed to Katherine Jordan-Burmeister at kjorda36@utk.edu or katjordan90@gmail.com.
Access information
Other publicly accessible locations of the data:
- None
Data were derived from the following sources:
- Brezinski, D. K. (2023). Biogeographic patterns in Late Paleozoic trilobites. Palaeogeography, Palaeoclimatology, Palaeoecology, 609, 111319. https://doi.org/10.1016/j.palaeo.2022.111319.
- The Paleobiology Database ([https://paleobiodb.org/#/])
