hide
Free keywords:
-
Abstract:
Planktonic foraminiferal species identification is central to many paleoceanographicstudies, from selecting species for geochemical research to elucidating the biotic dynamics ofmicrofossil communities relevant to physical oceanographic processes and interconnected phenomenasuch as climate change. However, few resources exist to train students in the difficult task of discerningamongst closely related species, resulting in diverging taxonomic schools that differ in species conceptsand boundaries. This problem is exacerbated by the limited number of taxonomic experts. Here wedocument our initial progress toward removing these confounding and/or rate‐limiting factors bygenerating thefirst extensive image library of modern planktonic foraminifera, providing digitaltaxonomic training tools and resources, and automating species‐level taxonomic identification ofplanktonic foraminifera via machine learning using convolution neural networks. Experts identified34,640 images of modern (extant) planktonic foraminifera to the species level. These images are servedas species exemplars through the online portal Endless Forams (endlessforams.org) and a taxonomictraining portal hosted on the citizen science platform Zooniverse (zooniverse.org/projects/ahsiang/endless‐forams/). A supervised machine learning classifier was then trained with ~27,000 images ofthese identified planktonic foraminifera. The best‐performing model provided the correct species namefor an image in the validation set 87.4% of the time and included the correct name in its top threeguesses 97.7% of the time. Together, these resources provide a rigorous set of training tools in modernplanktonic foraminiferal taxonomy and a means of rapidly generating assemblage data via machinelearning in future studies for applications such as paleotemperature reconstruction..