New frontiers in dinosaur exploration

200 years after the naming of the first dinosaur, taxonomic studies remain an important component of dinosaur research. Around 50 new dinosaurs are named each year, and are discovered from across the globe. The rate of new dinosaur discovery shows no signs of slowing, but not all geographic areas and temporal windows have been equally investigated. The potential for new dinosaur discoveries in India and Africa seems particularly high, while the Carnian, when dinosaurs probably originated, and the Middle Jurassic, when the major clades diversified, offer the best opportunities to make discoveries that will fundamentally change our understanding of dinosaur evolution. A major challenge to the discovery of new dinosaurs is funding. Frontier fieldwork is sometimes viewed as too risky to fund, while basic taxonomic work is considered to lack impact. As a consequence, we risk an ‘extinction of experience’, where researchers have limited training in the basic field and specimen-based research that underpins our discipline. Going forward, new remote sensing techniques may help to find prospective areas, while 3D scanning apps on smartphones will allow us to quickly record field data. Artificial intelligence is likely to be used increasingly for CT segmentation and identification of problematic fossils.

Methods
Collector curves–All dinosaur regular genera and species, both valid and invalid, were downloaded from the Paleobiology Database (PBDB; paleobiodb.org) on 17th December 2024. The data were cleaned to remove Avialae, ichnotaxa, and ootaxa. Taxa that were listed as invalid due to misspellings, obsolete variates, or that were renamed for grammatical or linguistic reasons were removed. Nomina dubia, nomina nuda, objective and subjective synonyms, and recombinations were retained. Collector curves (Fig. 1) were built in R 3.4.0 [124]. Code and raw data are available in the Supplementary Material. 

Time-calibrated phylogeny–A consensus dinosaur phylogeny was manually produced in Mesquite [125]. First and last appearance data were collected for all taxa in the phylogeny and are listed in the data file provided in the Supplementary Material. First and last appearances generally correspond to the earliest and latest dates of the Stage from which the taxon is known, unless more accurate information was available in the literature. Ages were derived from the International Chronostratigraphic Chart version 2024/12. The consensus phylogeny was time-calibrated using the TimePalaeoPhy function in the R package Paleotree [126] with minimum branch lengths of 0.5 million years. Code and raw data are available in the Supplementary Material.

Occurrences through time histogram–All dinosaur body fossil occurrences were downloaded from the PBDB on 22nd January 2025 and were manually cleansed to remove ichnotaxa and ootaxa. The midpoint of the first appearance datum and last appearance datum was taken for each occurrence, and these midpoints were plotted in 1 million year bins in R 3.4.0 [124]. Code and raw data are available in the Supplementary Material.

Occurrences map–Non-avialan dinosaur occurrences, excluding form taxa, were downloaded from the PBDB (27th January 2025). These were manipulated in R 4.4.2. The data set was pruned to limit it to species type occurrences only. Triassic, Jurassic and Cretaceous occurrences were plotted onto modern day maps using the packages maps and ggplot. Code and raw data are available in the Supplementary Material.

124. R Core Team. 2023. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

125. Maddison WP, Maddison DR. 2023. Mesquite: a modular system for evolutionary analysis. Version 3.81. http://www.mesquiteproject.org.

126. Bapst DW. 2012. Paleotree: an R package for palaeontological and phylogenetic analyses of evolution. Methods in Ecology and Evolution 3: 803–807.