Habitat selection models can provide insight into how animals respond to their environment at multiple scales. Here we investigate GPS tracks of the Mexican fish-eating bat as it forages in the open ocean. The dynamic marine environment provides a challenge for animals attempting to reliably find prey as often the same location between nights will yield very different results. Our previous research found that fish-eating bat foraging areas are on average several kilometers apart over consecutive nights, suggesting that bats cannot reliably predict prey location. Playback experiments and on-board audio recordings revealed that these bats rely on social information to find ephemeral prey patches, yet consistency in the direction of their foraging flights suggesting habitat selection at some level.
Here we ask if environmental variables measured by remote sensing are associated with foraging at two spatial scales, (nightly path, foraging patch) to identify predictors of habitat selection by fish-eating bats. Additionally, we examine if patterns change between years (2015, 2016, and 2018), spanning an oscillation between El Nino and La Nina. We determined foraging locations through evaluation of behavioral segmentation methods with on-board audio recordings of buzzes. Using the best performing method, hidden Markov model, we parse tracks into foraging and commuting behavioral states and then use randomization tests to determine associations with environmental variables. Our results show that environmental variation fails to predict foraging patches, but at a larger scales bats choose to forage in areas with higher chlorophyll and steeper ocean slope, likely increasing the general presence of prey. There do not appear to be significant yearly differences in habitat selection, suggesting that El Nino Southern Oscillation does not influence fine scale foraging patterns.
Habitat selection in a marine bat provides opportunities for comparison with other marine organisms foraging in the same unpredictable environment.
