Predicting Pelagic Habitat with Presence-only Data using Maximum Entropy for Olive Ridley Sea Turtles in the Eastern Tropical Pacific
Repository Usage Stats
Little is known about the oceanic distribution of olive ridley sea turtles (Lepidochelys olivacea) in the eastern tropical Pacific (ETP), or what governs their offshore movements. In collaboration with NOAA’s Southwest Fisheries Science Center, I opportunistically sampled 350 olive ridley sea turtles in the ETP between August and December 2006. Using these presence-only observations and remotely-sensed oceanographic data, I developed a maximum entropy habitat model using the Maxent software package, and considered the influences of chlorophyll(a) concentrations, sea surface temperature, and bathymetry. I compared Maxent results with two more common approaches used to describe pelagic species distribution and habitat suitability: generalized linear and additive models (GLM and GAM). Statistically, the GLM model performed well (AUC > 0.78), whereas the GAM (highly variable AUC: 0.64 – 0.88) and Maxent (AUC < 0.68) models did not. However, based on expert knowledge the Maxent results were the most reasonable despite low AUC values. For the scope of this study, Maxent was determined to produce a viable species distribution model, although areas of improvement are recommended. Maxent is growing in popularity among marine habitat modelers, but I caution against using the method as a panacea for predictive habitat modeling for open-ocean, rare species, despite its accessible and easy-to-use software. Similar studies should be repeated for other oceanic species (e.g. migratory marine megafauna) and compared with richer datasets (e.g. line transect surveys and telemetry) to gain a better understanding of Maxent’s ability to accurately predict oceanic habitat.
CitationPeavey, Lindsey (2010). Predicting Pelagic Habitat with Presence-only Data using Maximum Entropy for Olive Ridley Sea Turtles in the Eastern Tropical Pacific. Master's project, Duke University. Retrieved from https://hdl.handle.net/10161/2247.
More InfoShow full item record
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.
Rights for Collection: Nicholas School of the Environment