Table S5.
Stepwise multiple regressions between the three phenology groups and local oceanic conditions
| Phenology group | Regression equation | df | AIC | F |
| Species with earlier phenology | ||||
| Full dataset | Y = 0.0436 + 0.9120XSST + 0.2355XZoo − 0.3823XAR1 | 86 | 75.2 | 26.6*** |
| Bias-corrected dataset | Y = -0.0250 + 1.2574XSST + 0.3690XZoo | 87 | 116.6 | 10.2*** |
| Species with no long-term, linear trend in phenology | ||||
| Full dataset | Y = 0.0121 | 105 | N/A | N/A |
| Bias-corrected dataset | Y = -0.0093 − 0.2898XAR1 | 104 | 152.7 | 9.3** |
| Species with later phenology | ||||
| Full dataset | Y = 0.0518 − 0.4556XSST + 1.3534XUpw | 40 | 33.8 | 7.5** |
| Bias-corrected dataset | Y = 0.0268 − 1.2713XSST + 1.8838XUpw | 40 | 44.3 | 11.0*** |
In the regression equations, Y refers to CT anomalies of species in each phenology group, whereas XSST, XZoo, and XUpw are the CT of CalCOFI SST, CalCOFI mesozooplankton volume, and the Bakun upwelling index, respectively. XAR1 is a first-order autoregressive term. Only the final regression model with the lowest AIC is shown. The bias-corrected dataset removes all data from May, September, and December, because missing data during these months in the 2000s could potentially affect results of this analysis. N/A, statistic could not be estimated because the null model provided the best fit to the data. Significance levels of regressions are indicated as follows: **P < 0.01; ***P < 0.001.