Skip to main content

Table 1 Results of model selection after 5-times-repeated fivefold cross-validation

From: Influence of thermal stratification and storms on acoustic telemetry detection efficiency: a year-long test in the US Southern Mid-Atlantic Bight

  

Root-mean-square error of repeated folds

     

Mean ± standard deviation (%)

   

Model

Description

Overall

250 m

550 m

800 m

Adj. R^2

Dev. Expl

AIC

ΔAIC

wAIC

 ~ Distance + 

\(f\)(noise) + \(f\) (ΔT)

Nonlinear response to noise and stratification

17.1 ± 0.5

9.9 ± 0.9

21.2 ± 1.1

18.7 ± 1.1

0.811

76.8%

1026

0

0.99

\(f\)(noise, ΔT)

Nonlinear response to noise, modulated by stratification

17.1 ± 0.5

9.8 ± 0.9

21.2 ± 1.1

18.6 ± 1.1

0.811

76.8

1035

9

0.01

\(f\)(noise) + \(f\)(BWT)

Nonlinear response to noise and near-receiver temperature

19.0 ± 0.5

9.3 ± 0.9

23.5 ± 0.9

21.7 ± 1.0

0.752

72.1

1279

253

0.00

\(f\)(noise, BWT)

Nonlinear response to noise, modulated by near-receiver temperature

18.7 ± 0.5

9.2 ± 0.9

23.3 ± 0.9

21.1 ± 1.2

0.752

72.1

1280

254

0.00

\(f\)(noise)

Nonlinear response to noise

21.3 ± 0.6

10.1 ± 0.9

25.0 ± 1.0

25.7 ± 1.0

0.689

66.0

1580

554

0.00

\(f\)(ΔT)

Nonlinear response to stratification

22.3 ± 0.8

15.8 ± 1.9

28.0 ± 1.2

21.6 ± 1.2

0.678

62.9

1598

572

0.00

\(f\)(BWT)

Nonlinear response to near-receiver temperature

25.7 ± 0.6

15.5 ± 1.7

31.3 ± 1.2

28.4 ± 1.0

0.56

52.7

2049

1023

0.00

0

(null)

Linear response to distance

28.7 ± 0.6

16.0 ± 2.0

34.1 ± 1.0

33.1 ± 0.9

0.454

42.3

2502

1476

0.00

  1. Overall model error and error at distances of 250, 550, and 800 m away from the transmission are shown along with adjusted R2, percent deviance explained, AIC, delta-AIC, and AIC weights