Multi-Scale Approach for Predicting Fish Species Distributions across Coral Reef Seascapes
2011

Predicting Fish Distributions across Coral Reef Seascapes

Sample size: 1018 publication Evidence: high

Author Information

Author(s): Pittman Simon J., Brown Kerry A., Gratwicke Brian

Primary Institution: National Oceanic and Atmospheric Administration (NOAA)

Hypothesis

Can a multi-scale approach using seafloor morphology and location predict fish species distributions in coral reef ecosystems?

Conclusion

The study successfully models and maps fish species distributions using environmental predictors, demonstrating the importance of geographical location and topographic complexity.

Supporting Evidence

  • Model predictions were assessed using two machine-learning algorithms: Boosted Regression Trees (BRT) and Maximum Entropy (MaxEnt).
  • BRT provided outstanding predictions for three fish species, while MaxEnt was more accurate overall.
  • Geographical location and topographic complexity were identified as the most important predictors of fish distributions.
  • Map accuracy for MaxEnt models was consistently high, with up to 97% correct predictions for some species.

Takeaway

This study shows how scientists can use underwater maps to find out where different fish live in coral reefs, helping to protect their homes.

Methodology

The study used underwater visual surveys and advanced machine-learning algorithms to analyze fish distributions based on seafloor structure and geographical location.

Potential Biases

Potential biases may arise from using pseudo-absences in MaxEnt models.

Limitations

The study may not account for all environmental variables affecting fish distributions and relies on presence-only data.

Digital Object Identifier (DOI)

10.1371/journal.pone.0020583

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication