Skip to main content

Table 2 Performance output of the track- and dive-based metrics using GLMM in explaining and predicting nPEE at the scale of dives and days

From: Track and dive-based movement metrics do not predict the number of prey encountered by a marine predator

Model

R2F (%)

R2I (%)

R2S (%)

RMSE

mRMSE

sdRMSE

Corr

R2F (%)

R2I (%)

R2S (%)

RMSE

mRMSE

sdRMSE

Corr

Dive

Day

Descent rate

19

7

1

7.4 ± 2.2

0.78 ± 0.10

1.05 ± 0.12

0.31 ± 0.12

21

34

5

308.6 ± 134.6

0.49 ± 0.19

1.33 ± 0.48

0.41 ± 0.29

Ascent rate

33

8

1

7.2 ± 2.1

0.77 ± 0.14

1.02 ± 0.14

0.46 ± 0.08

37

37

2

313.9 ± 165.3

0.50 ± 0.24

1.36 ± 0.65

0.61 ± 0.17

Bottom duration

3

6

1

7.6 ± 2.3

0.81 ± 0.11

1.08 ± 0.08

0.08 ± 0.12

0

27

7

299.9 ± 136.1

0.46 ± 0.16

1.24 ± 0.24

0.03 ± 0.32

Surface duration

2

5

3

7.6 ± 2.2

0.81 ± 0.09

1.08 ± 0.07

0.11 ± 0.14

1

25

15

292.6 ± 134.6

0.45 ± 0.13

1.20 ± 0.18

0.13 ± 0.31

Efficiency

3

4

4

7.6 ± 2.1

0.81 ± 0.08

1.08 ± 0.06

0.17 ± 0.17

9

22

13

277.2 ± 112.2

0.43 ± 0.13

1.17 ± 0.21

0.37 ± 0.33

Sinuosity

11

7

2

7.5 ± 2.2

0.80 ± 0.11

1.07 ± 0.09

0.21 ± 0.17

18

36

10

325.8 ± 134.5

0.52 ± 0.19

1.43 ± 0.52

0.37 ± 0.35

Hunting time

28

15

3

10.1 ± 4.3

1.14 ± 0.68

1.52 ± 0.88

0.37 ± 0.11

28

45

18

861.6 ± 668.2

1.52 ± 1.53

4.45 ± 4.76

0.45 ± 0.32

Horizontal speed

5

5

2

7.8 ± 2.2

0.81 ± 0.10

1.07 ± 0.07

0.18 ± 0.11

21

29

5

300.2 ± 122.7

0.48 ± 0.22

1.26 ± 0.45

0.50 ± 0.24

Turning angle

0

5

0

7.9 ± 2.2

0.82 ± 0.10

1.08 ± 0.07

0.02 ± 0.04

17

27

9

389.1 ± 224.1

0.66 ± 0.61

1.69 ± 1.16

0.30 ± 0.17

FPT

4

5

1

7.7 ± 2.1

0.81 ± 0.10

1.07 ± 0.06

0.15 ± 0.09

15

28

5

322.4 ± 111.3

0.52 ± 0.21

1.36 ± 0.43

0.38 ± 0.19

Move persistence

9

15

3

8.0 ± 2.0

0.84 ± 0.10

1.10 ± 0.07

0.17 ± 0.11

22

31

7

313.8 ± 90.9

0.49 ± 0.15

1.32 ± 0.33

0.42 ± 0.24

Dive

45

13

5

7.1 ± 2.0

0.77 ± 0.13

1.03 ± 0.18

0.59 ± 0.07

46

27

17

282.0 ± 147.8

0.45 ± 0.24

1.25 ± 0.74

0.81 ± 0.11

Track

7

11

8

7.8 ± 2.1

0.82 ± 0.09

1.08 ± 0.07

0.20 ± 0.10

24

23

20

291.5 ± 109.7

0.46 ± 0.16

1.20 ± 0.32

0.53 ± 0.22

All

36

13

22

7.8 ± 2.1

0.82 ± 0.13

1.09 ± 0.21

0.58 ± 0.07

41

28

24

324.2 ± 158.6

0.50 ± 0.23

1.38 ± 0.74

0.81 ± 0.12

Descent rate

9

6

1

7.5 ± 2.4

0.79 ± 0.11

1.05 ± 0.10

0.22 ± 0.12

13

33

6

311.4 ± 137.1

0.49 ± 0.18

1.32 ± 0.35

0.31 ± 0.29

Ascent rate

10

6

3

7.5 ± 2.5

0.79 ± 0.12

1.05 ± 0.12

0.26 ± 0.14

14

32

6

316.0 ± 162.5

0.48 ± 0.17

1.30 ± 0.35

0.28 ± 0.27

Bottom duration

5

6

2

7.7 ± 2.3

0.81 ± 0.11

1.08 ± 0.08

0.10 ± 0.13

0

26

12

301.6 ± 138.8

0.46 ± 0.15

1.24 ± 0.22

−0.05 ± 0.35

Efficiency

0

6

3

7.7 ± 2.2

0.82 ± 0.09

1.09 ± 0.07

0.12 ± 0.15

8

24

12

279.9 ± 122.0

0.43 ± 0.13

1.17 ± 0.22

0.34 ± 0.35

Hunting time

2

5

1

7.6 ± 2.1

0.81 ± 0.09

1.07 ± 0.07

0.14 ± 0.11

0

27

10

299.9 ± 132.7

0.47 ± 0.17

1.25 ± 0.25

−0.00 ± 0.38

  1. The variance in nPEE explained by the GLMM (each row) is partitioned into the proportion of variance explained by the predictors via the fixed slope variance (\(R_{F}^{2}\)), by the individual-specific means via the random intercept variance (\(R_{I}^{2}\)), and by the predictors via the random slopes variance/covariance (\(R_{S}^{2}\)). The root-mean-square error (RMSE; mean \(\pm\) sd), the mean-based normalized RMSE (mRMSE), the standard-deviation-based normalized RMSE (sdRMSE), and the correlation coefficient (Corr.) are computed between the observed values of nPEE for each seal and the predicted values by the model fitted without the focal individual (i.e., leave-one-individual-out cross-validation). The dive model includes all the dive-based metrics, the track model includes all the track-based metrics, and the all model includes all the metrics. Dive metrics are calculated from the high-resolution dive profiles (i.e., at 1 Hz; upper section) or from the low-resolution dive profiles (lower section), i.e., simplified to five segments using the broken stick method similar to data transmitted by CTD-SRDL loggers.
Back to article page

Movement Ecology

ISSN: 2051-3933