Two ESA-listed corals around American Samoa occupy different environmental niches—and the models are only moderately predictive
WHAT THE EVIDENCE SAYS
Researchers built ensemble MaxEnt models for Acropora globiceps and Isopora crateriformis using multi-source occurrence records from 2007–2025, open environmental data and predictors including thermal stress, water clarity, wind variability, seafloor structure and land-based stressors. Models trained through 2024 were evaluated against independent 2025 observations. The two listed corals showed distinct habitat associations, demonstrating that closely related recovery targets should not be assumed to share the same environmental niche.
WHY IT MATTERS / OUR READ
For recovery planning, the value is in separating species-specific habitat hypotheses instead of managing ‘listed coral habitat’ as a single category. Open biodiversity and satellite datasets can help prioritize survey effort, but the model should be used to ask where to verify habitat—not to replace field observation.
CAVEATS / UNCERTAINTY
These are presence-only relative-suitability models in a data-limited setting, not abundance, demographic or survival models. Forecast discrimination was moderate rather than definitive, occurrence records came from multiple survey sources, and open environmental layers carry their own spatial and temporal uncertainty.
SOURCE
Scientific Reports — primary study
View original evidence ↗
Presence-only models separated the habitat associations of Acropora globiceps and Isopora crateriformis and were tested against independent 2025 observations to support monitoring and recovery planning.
Read the Wild Lands assessment →