FN Archimer Export Format
PT J
TI Geographical differences in habitat relationships of cetaceans across an ocean basin
BT 
AF Mannocci, Laura
   Roberts, Jason J.
   Pedersen, Eric J.
   Halpin, Patrick N.
AS 1:1,2;2:2;3:3;4:2;
FF 1:;2:;3:;4:;
C1 MARBEC, Univ. Montpellier, CNRS, Ifremer, IRD Sète France
   Marine Geospatial Ecology Laboratory, Nicholas School of the Environment, Duke Univ. Durham NC USA
   Dept of Biology, Concordia Univ. Montreal QC Canada
C2 IRD, FRANCE
   UNIV DUKE, USA
   UNIV CONCORDIA, CANADA
UM MARBEC
IN WOS Cotutelle UMR
   copubli-int-hors-europe
IF 5.992
TC 14
UR https://archimer.ifremer.fr/doc/00630/74163/73721.pdf
   https://archimer.ifremer.fr/doc/00630/74163/73722.zip
LA English
DT Article
DE ;environmental predictors;geographical variation;habitat relationships;highly mobile marine species;North Atlantic Ocean;species distribution modeling
AB The distributions of highly mobile marine species such as cetaceans are increasingly modeled at basin scale by combining data from multiple regions. However, these basin‐wide models often overlook geographical variations in species habitat relationships between regions. We tested for geographical variations in habitat relationships for a suite of cetacean taxa between the two sides of the North Atlantic basin. Using cetacean visual survey data and remote sensing data from the western and eastern basin in summer, we related the probability of presence of twelve cetacean taxa from three guilds to seafloor depth, sea surface temperature and primary productivity. In a generalized additive model framework, we fitted 1) basin‐wide (BW) models, assuming a single global relationship, 2) region‐specific intercepts (RI) models, assuming relationships with the same shape in both regions, but allowing a region‐specific intercept and 3) region‐specific shape (RS) models, assuming relationships with different shapes between regions. RS models mostly yielded significantly better fits than BW models, indicating cetacean occurrences were better modeled with region‐specific than with global relationships. The better fits of RS models over RI models further provided statistical evidence for differences in the shapes of region‐specific relationships. Baleen whales showed striking differences in both the shapes of relationships and their mean presence probabilities between regions. Deep diving whales and delphinoids showed contrasting relationships between regions with few exceptions (e.g. non‐statistically different shapes of region‐specific relationships for harbor porpoise and beaked whales with depth). Our findings stress the need to account for geographical differences in habitat relationships between regions when modeling species distributions from combined data at the basin scale. Our proposed hypotheses offer a roadmap for understanding why habitat relationships may geographically vary in cetaceans and other highly mobile marine species. 
PY 2020
PD AUG
SO Ecography
SN 0906-7590
PU Wiley
VL 43
IS 8
UT 000531110000001
BP 1250
EP 1259
DI 10.1111/ecog.04979
ID 74163
ER

EF