Using a Bayesian modelling approach (INLA-SPDE) to predict the occurrence of the Spinetail Devil Ray (Mobular...

Area of Science:

• Marine biology
• Fisheries science
• Statistical ecology

Background:

• Understanding marine species' spatiotemporal distributions is crucial for conservation.
• Bayesian statistics are commonly used for bycatch species distribution modeling but can be computationally intensive.
• Existing habitat models may not identify all important bycatch areas.

Purpose of the Study:

• To evaluate the Integrated Nested Laplace Approximation with Stochastic Partial Differential Equation (INLA-SPDE) Bayesian approach for predicting Mobula mobular occurrence.
• To identify potential bycatch hotspots for Mobula mobular in the eastern Pacific Ocean.
• To compare the INLA-SPDE approach with traditional methods for bycatch data analysis.

Main Methods:

• Utilized a Generalized Additive Model (GAM) to analyze bycatch data from the Inter-American Tropical Tuna Commission's (IATTC) tropical tuna purse-seine fishery observer database (2005-2015).
• Applied the INLA-SPDE Bayesian approach to model Mobula mobular occurrence in the eastern Pacific Ocean.
• Assessed the computational efficiency and predictive capabilities of the INLA-SPDE method.

Main Results:

• The INLA-SPDE approach successfully predicted known areas of importance for Mobula mobular in the eastern Pacific Ocean.
• Identified previously unrecognized bycatch hotspots, including the Gulf of California and the Equatorial area.
• Demonstrated the potential of INLA-SPDE to reveal marginal hotspots missed by other habitat models.

Conclusions:

• The INLA-SPDE method is a computationally efficient and valuable alternative or complement to traditional approaches for bycatch data modeling.
• This approach can enhance the accuracy of management decisions by providing a more comprehensive understanding of species distributions.
• Challenges remain in handling categorical variables and spatial triangulation within the INLA-SPDE framework.