Table 1.

Relevance of environmental indicators to sardine dynamics

VariableρmultiΔFP
SIO Pier temperature (surface)0.40850.08720.04
PDO0.39680.07550.04
SCB satellite SST0.36520.04390.12
NPGO0.34520.02390.19
SIO Pier temperature (bottom)0.3128−0.0085
MEI0.3034−0.0179
SOI0.2434−0.0779
NPI0.2152−0.1061
Newport Pier temperature (surface)0.2117−0.1096
  • Forecasts of sardine ichthyoplankton are compared between the multivariate embedding [X(t), X(t−1), X(t−2), Y(t)] and the univariate embedding [X(t), X(t−1), X(t−2)] for each candidate environmental indicator variable, Y. The forecast skill (correlation ρ between observations and predictions) of multivariate embeddings using each of the candidate variables are shown; the univariate forecast skill is ρuni = 0.3213. ΔF = ρmulti − ρuni > 0 indicates that the environmental variable influences sardine, whereas large negative ΔF indicates the variable is irrelevant. Significance of ΔF > 0 is tested using phase-randomized surrogate time series of the environmental variables. MEI, Multivariate El Niño Index.