TABLE 2.
Boosted regression tree model predictor variable contributions.
| Predictor | MARU1 | MARU2 | MARU3 | MARU4 | MARU5 |
|---|---|---|---|---|---|
| Song (full) | |||||
| logNPP | 11.24 | 8.50 | 11.28 | 6.39 | 8.19 |
| Month | 66.63 | 59.62 | 58.08 | 77.28 | 70.78 |
| SST | 6.16 | 13.59 | 15.73 | 9.20 | 7.79 |
| SST anomaly | 4.13 | 6.68 | 7.38 | 3.04 | 8.41 |
| Detection area | 11.53 | 11.60 | 7.50 | 4.07 | 4.81 |
| D calls (full) | |||||
| logNPP | 46.39 | 40.35 | 21.18 | 12.74 | 11.37 |
| Month | 33.82 | 45.78 | 47.55 | 44.83 | 64.26 |
| SST | 3.55 | 5.11 | 9.69 | 9.76 | 5.76 |
| SST anomaly | 3.75 | 5.71 | 15.42 | 13.69 | 7.23 |
| Detection area | 12.46 | 3.02 | 6.13 | 18.95 | 11.36 |
| D calls (spring) | |||||
| logNPP | 38.26 | 85.15 | 18.67 | 11.29 | 16.87 |
| SST | 13.43 | 5.42 | 35.78 | 26.21 | 44.08 |
| SST anomaly | 8.98 | 4.74 | 33.58 | 21.13 | 7.60 |
| Detection area | 39.31 | 4.67 | 11.95 | 41.35 | 31.43 |
| D calls (fall) | |||||
| logNPP | 48.34 | 37.66 | 37.53 | 16.88 | 25.94 |
| SST | 29.05 | 11.94 | 29.12 | 22.69 | 26.92 |
| SST anomaly | 13.90 | 38.71 | 21.05 | 36.34 | 32.68 |
| Detection area | 8.69 | 11.66 | 12.28 | 24.07 | 14.44 |
| Song (fall) | |||||
| logNPP | 24.10 | 39.11 | 29.79 | 29.73 | 73.00 |
| SST | 25.92 | 15.50 | 44.58 | 44.44 | 11.48 |
| SST anomaly | 19.78 | 15.13 | 13.14 | 13.83 | 5.11 |
| Detection area | 30.18 | 30.24 | 12.47 | 11.98 | 10.39 |
Note: Variable contribution for the boosted regression tree models for song and D calls at each of the hydrophones. The percent contribution in the model is listed for each predictor. The full models with all seasons included are reported first for both call types, followed by the within‐season peaks for each call type.