Abstract
Aim: Creating a network of marine protected areas in the Southern Ocean requires extensive knowledge on species’ abundances, distributions and population trends especially in the Weddell Sea where year-round pack ice makes most of the Weddell Sea inaccessible. We combine satellite images and citizen science to model habitat suitability for crabeater seals (Lobodon carcinophaga) throughout the Weddell Sea. Location: Weddell Sea, Antarctica. Methods: High-resolution satellite images covering 18,219 km2 of the Weddell Sea during crabeater seal breeding season (October—November) were hosted on the crowd-sourcing platform Tomnod (DigitalGlobe). Citizen scientists marked “maps” where seals were present/absent and these votes were compared with the votes of an experienced observer. Correction factors were used to correct votes to either a continuous probability of seal presence, or a binary seal presence/absence value. We modelled probability of seal presence using ensemble models of Random Forests (RF), Boosted Regression Trees (BRT) and Support Vector Machines (SVM), and used fitted Maxent models to model seal presence/absence data. Results: Model predictive power was low (RF: R2 = 0.076 ± 0.002: BRT: R2 = 0.086 ± 0.0008; SVM: R2 = 0.082 ± 0.003) to average (Maxent: AUC = 0.71 ± 0.004). Distance to the ice edge and bathymetry were the most important variables that influenced crabeater seal distribution. Main conclusions: Crabeater seals were more likely to be present over abyssal water, which coincides with typical adult Antarctic krill habitat — crabeater seal preferred prey. Where ice concentrations were more variable, that is more accessible, crabeater seals were also more likely to occur. Results agreed with the known ecology of crabeaters seals and the abundance, distribution and ecology of Antarctic krill. We were able to survey the largest area ever surveyed in the Weddell Sea and provide a model to assist furthering policy around the proposed protected area.
Original language | English (US) |
---|---|
Pages (from-to) | 1291-1304 |
Number of pages | 14 |
Journal | Diversity and Distributions |
Volume | 26 |
Issue number | 10 |
DOIs | |
State | Published - Oct 1 2020 |
Externally published | Yes |
Bibliographical note
Publisher Copyright:© 2020 The Authors. Diversity and Distributions published by John Wiley & Sons Ltd
Keywords
- Antarctic krill
- Maxent
- boosted regression tree
- citizen science
- ensemble modelling
- machine learning
- random forests
- support vector machines