3D ecological niche models outperform 2D in predicting coelacanth (Latimeria spp.) habitat

Emmaline Sheahan ^1,2* Hannah Owens ^1,2,3 Robert Guralnick ^1,2 Gavin Naylor ^1,2

• ¹ University of Florida, Gainesville, United States
• ² Florida Museum of Natural History, Gainesville, Florida, United States
• ³ Center for Macroecology, Evolution and Climate, GLOBE Institute, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Capital Region of Denmark, Denmark

Discoveries of coelacanth populations off the East African coast and in the Indo-Pacific warrant an analysis of their potential distributions, but the necessary tools to model and project their distributions in 3 dimensions are lacking. Using occurrence records for the West Indian ocean coelacanth, Latimeria chalumnae, we produced 3D and 2D maximum entropy ecological niche models and projected them into the habitat of the Indonesian coelacanth, Latimeria menadoensis.We gauged each model's success by how well it could predict L. menadoensis presences recorded from submersible observations. While the 2D model omitted 33% of occurrences at the most forgiving threshold, the 3D model successfully predicted all occurrences, regardless of threshold level. Incorporating depth results in improved model accuracy when predicting coelacanth habitat, and projecting into 3 dimensions can give us insights as to where to target future sampling. This 3D modelling framework can help us better understand how marine species are distributed by depth and allow for more targeted conservation management.