We examined the potential synchronising influence of climate on kelp by evaluating the spatial wavelet coherences between the MEI, PDO, and NPGO climate indices, and quarterly kelp density time-series drawn from the California coast. We applied a clustering algorithm to the kelp time-series to identify spatial patterns in their fluctuations. To determine how the statistical relationship between the long timescale fluctuations in kelp and the climate arises, and why synchrony differs between central and southern clusters, we examined the local surface nitrate (derived from sea surface temperature) and maximum wave height timeseries corresponding to the measured kelp sites. Multivariate wavelet models of kelp were constructed, using best-fit complex coefficients incorporating phase shifts between the variables. We applied a wavelet synchrony attribution theorem to decompose the contributions to synchrony made by the two drivers, and their interaction.
Results/Conclusions
The kelp time-series have a component of variability in common at long timescales, associated with the NPGO. Accordingly, some spatial synchrony is associated with the influence of climate. At long timescales 82 percent of the squared synchrony in Central sites is attributable to the component in common with the NPGO, and 84 percent in Southern sites. However, the central and southern clusters differ markedly in their synchrony at annual timescales, and this is not explained by their relationship with the NPGO.
Wave height fluctuations and kelp fluctuations are generally in antiphase: when wave action is high, kelp is destroyed. Phase shifts between nitrate fluctuations and kelp fluctuations are generally around zero: when nitrate is high, kelp is abundant. At the annual timescale (1-year periodic oscillation associated with the difference between seasons) the synchronising effects of wave action and nitrate fluctuations reinforce one another. At long timescales (slow, positively autocorrelated changes in kelp biomass) the effects of wave action and nitrate fluctuations partially cancel one another out. High annual synchrony in central California is associated with higher synchrony predicted by the model, attributable to these drivers.