Environmental selection in marine organisms: spatial scaling considerations
Jerome Pinti, University of Delaware, jpinti@udel.edu (Presenter)
Matthew Oliver, University of Delaware, moliver@udel.edu (Presenter)
Aaron Carlisle, University Of Delaware, carlisle@udel.edu
Helga Huntley, Rowan College, helgah@udel.edu
Matthew Shatley, University Of Delaware, mshatley@udel.edu
Understanding the selection of environmental conditions by marine predators requires knowledge of the positions of the animals, and of the environmental conditions that are available to these animals. Both these data sources are hard to estimate accurately and/or are patchy.
Tracking marine organisms is challenging because animals can only be detected when they are at the surface. Further, the accuracy of such detections depends on the tag used and on the time that the animal spends at the surface, with uncertainties ranging as large as several tens of kilometers for ARGOS tags.
Mapping environmental conditions systematically at an ocean basin scale and at a fine resolution is difficult because of the large amount of data it represents. Remote sensing allows to monitor these conditions on a broad scale, but satellites are vulnerable to cloud cover and may not scout the entire ocean regularly. Using model outputs for environmental conditions allows to have a complete spatial coverage, but potentially at the expense of accuracy. Further, different environmental products vary over different scales. Consequently, there might be complex interactions between the scale at which we detect animal movement and the scale over which these environmental products vary.
Here, we explore the interplay between location accuracy and the scales at which environmental variables vary. We focus on three different environmental variables (sea surface temperature, chlorophyll concentration, and a measure of Lagrangian Coherent Structures, finite-time Lyapunov exponent). For each of these variables, we create synthetic animal tracks selecting for specific environmental conditions with different strengths. Then, we assess how well we can detect selection depending on the assumed accuracy of these tracks. The lower the accuracy, the harder it is to detect selection – especially when the environmental variable varies on scales comparable to or smaller than the accuracy of the animal tracks.
Associated Project(s):
Poster Location ID: 2-44
Presentation Type: Poster
Session: Poster Session 2
Session Date: Wed (May 10) 5:15-7:15 PM
CCE Program: BDEC