Skip to main content

Table 2 Predicting changes in move-persistence as a function of environmental covariates for narwhals selecting offshore versus nearshore migration routes

From: Divergent migration routes reveal contrasting energy-minimization strategies to deal with differing resource predictability

 

Model

df

ΔAIC

Deviance

Offshore

 ~ ice.con + slope + dist + dist2 + (1 | id)

10

0

− 2460.9

 ~ ice.con + bathy + slope + dist + dist2 + (1 | id)

11

1.15

− 2461.7

 ~ ice.con + slope + sst + dist + dist2 + (1 | id)

11

1.79

− 2461.1

 ~ ice.con + bathy + slope + sst + dist + dist2 + (1 | id)

12

2.92

− 2461.9

-

–

–

–

 ~ dist + dist2 + (1 | id)

8

14.04

− 2442.8

 ~ sst + bathy + dist + dist2 + (1 | id)

10

15.06

− 2445.8

 ~ sst + dist + dist2 + (1 | id)

9

16.00

− 2442.9

Nearshore

 ~ ice.con + bathy + slope + sst + dist + dist2 + (1 | id)

12

0

− 5006.9

 ~ ice.con + bathy + slope + dist + dist2 + (1 | id)

11

1.64

− 5003.3

 ~ ice.con + slope + sst + dist + dist2 + (1 | id)

11

3.05

− 5001.9

–

–

–

 

 ~ dist + dist2 + (1 | id)

8

61.25

− 4937.7

 ~ sst + dist + dist2 + (1 | id)

9

63.06

− 4737.9

  1. Condensed model results table with top models are highlighted in bold. Models with ≤ 4 ΔAIC are highlighted as compared to the null model. Shown here are the model formula for environmental covariates, also indicating random slopes for each individual in brackets (id), the degrees of freedom, delta AIC, and model deviance. Predictors of move-persistence included ice concentration (ice.con), depth (bathy), slope, sea surface temperature (sst), and the mandatory variables of distance to shore as a linear (dist) and quadratic variable (dist2). For each route, the top model selected to generate spatial predictions is highlighted in bold