Table 1.
Results of linear modelling with a random effect for year. Distance of sea-crossing over the Mediterranean Sea was investigated as a function of wind support, crosswind, temperature gradient and date of arrival at sea. AIC, Akaike information criterion.
| coefficient (s.e.) |
||||||||
|---|---|---|---|---|---|---|---|---|
| model | intercept | wind support | crosswind | ΔT | Julian date | random effects | AIC | adjusted R2 |
| A | 813.24 (36.43) | — | — | — | 62.65 (26.62) | yeara | 594.27 | 0.10 |
| B | 805.69 (33.69) | 2.21 (26.88) | 58.82 (26.01) | — | 51.04 (26.81) | yearb | 576.21 | 0.20 |
| C | 806.36 (31.96) | — | — | 71.61 (26.39) | 30.17 (27.32) | yearc | 580.98 | 0.23 |
| D | 803.58 (28.66) | — | 56.77 (23.30) | 69.22 (25.01) | 20.44 (25.96) | yeard | 569.20 | 0.33 |
aBetween-year standard deviation: 56.93; residual standard deviation: 168.51.
bBetween-year standard deviation: 49.82; residual standard deviation: 162.90.
cBetween-year standard deviation: 46.15; residual standard deviation: 158.50.
dBetween-year standard deviation: 37.75; residual standard deviation: 150.93.