Journal article
No Model Fits All: Dynamic Ensemble Species Distribution Model Reveals Seasonal Patterns of Essential Habitat Use by Ocean Giants in the Southwest Pacific
Diversity and Distributions, Vol.32(4), pp.1-20
2026
Abstract
Aim: Essential habitats are areas that support biological and ecological functions critical for species' survival. For highly mobile and elusive marine species, aggregations in these habitats provide rare opportunities to study their ecology and inform conservation. We aimed to build a dynamic species distribution model (SDM) to predict essential habitats for migratory marine species. The model was tested on whale sharks (Rhincodon typus), a species with 30 documented aggregation sites across their global distribution, addressing a major knowledge gap in the southwest Pacific (SWP).
Location: Coral Sea, southwest Pacific.
Methods: High‐resolution, movement‐informed SDMs were built using behaviourally filtered juvenile whale shark satellite tracks (low move‐persistence locations) to quantify key environmental drivers of inferred foraging and predict dynamic suitability of essential habitat in the SWP. An ensemble modelling approach was applied to account for model uncertainty and improve model reliability by combining regression and machine learning algorithms.
Results: Model predictions indicated high suitability in the northern Great Barrier Reef during the monsoon season (November–April), shifting eastward into the Coral Sea and beyond during the dry season (May–October). Bathymetric variables (depth, distance to deepwater drop‐off) were key drivers of occurrence, while dynamic variables like sea surface temperature and productivity proxies also contributed largely to model predictions. Across algorithms, spatial block cross‐validation and external validation with independent sightings indicated moderate but consistent discriminatory ability. Habitat suitability predictions varied across algorithms, underscoring the advantages of integrating diverse modelling approaches.
Main Conclusions: This study presents the first movement‐informed predictions of essential habitat suitability for juvenile whale sharks in the SWP, providing a framework for improving population assessments and guiding research and management. The dynamic SDM approach is broadly applicable, facilitating essential habitat identification, research prioritisation in data‐limited regions, and targeted conservation in dynamic marine environments.
Details
- Title
- No Model Fits All: Dynamic Ensemble Species Distribution Model Reveals Seasonal Patterns of Essential Habitat Use by Ocean Giants in the Southwest Pacific
- Authors
- Ingo B. Miller (Corresponding Author) - AIMS@JCUYuri Niella (Author) - Integrated Marine Observing SystemVinay Udyawer (Author) - Sharks PacificMark V. Erdmann (Author) - Re:wild (United States)Kátya G. Abrantes (Author) - James Cook UniversitySimon J. Pierce (Author) - University of the Sunshine CoastRichard Fitzpatrick (Author) - Biopixel Oceans Foundation (Australia)Lisa A. Hoopes (Author) - Georgia Aquarium (United States)Alistair D. M. Dove (Author) - Georgia Aquarium (United States)Adam Barnett (Author) - Biopixel Oceans Foundation (Australia)
- Publication details
- Diversity and Distributions, Vol.32(4), pp.1-20
- Publisher
- Wiley-Blackwell Publishing Ltd.
- Date published
- 2026
- DOI
- 10.1111/ddi.70186
- ISSN
- 1472-4642
- Copyright note
- This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,provided the original work is properly cited. © 2026 The Author(s). Diversity and Distributions published by John Wiley & Sons Ltd.
- Data Availability
- Location data used for modelling are accessible in the Zenodo data repository: https://doi.org/10.5281/zenodo.17759819. R code are available in Zenodo: https://doi.org/10.5281/zenodo.17809753. Environmental predictor variables for habitat suitability modelling are available from the Copernicus Marine Data Store (https://data.marine.copernicus.eu/products) operated by the E.U. Copernicus Marine Environment Monitoring Service (CMEMS), and Bluelink Re-Analysis (BRAN) 2020 data is accessible through Australia's National Computational Infrastructure (https://doi.org/10.25914/6009627c7af03) and the ‘remora’ R package (Jaine et al. 2024). Bathymetry data is available from the General Bathymetric Chart of the Oceans (GEBCO) global grid (https://www.gebco.net), Geoscience Australia (https://doi.org/10.4225/25/5a207b36022d2) and figshare (https://doi.org/10.6084/m9.figshare.11986797) for Papua New Guinea. Additional References cited in Supporting Information: Gleiss et al. (2009); Hammerschlag et al. (2011); Hodson (2022); Willmott and Matsuura (2005).
- Grant note
- We acknowledge funding support by the Queensland Government's Threatened Species Research (round 1) grant (TSR069), the Sapphire Project, Blancpain Ocean Commitment, Georgia Aquarium, Sea World Foundation (SWR/9/2023), Conservation International, MAC3 Impact Philanthropies, the Slattery Family Trust, and 4planet.
- Organisation Unit
- School of Science, Technology and Engineering
- Language
- English
- Record Identifier
- 991228951602621
- Output Type
- Journal article
Metrics
1 Record Views