The seasonal cycle revisited: interannual variation and ecosystem consequences

The annual seasonal cycle accounts for much of the total temporal variabili ty of mid- and high-latitude marine ecosystems. Although the general patter n of the seasons repeats each year, climatic variability of the atmosphere and the ocean produce detectable changes in intensity and onset timing. We use a combination of time series data from oceanographic, zooplankton and s eabird breeding data to ask if and how these variations in the timing of th e spring growing season affect marine populations. For the physical environ ment, we develop an annual index of spring timing by fitting a non-linear 2 -parameter periodic function to the average weekly SST data observed in Bri tish Columbia from I January to the end of August each year. For each year, the phase parameter describes the timing of seasonal warming (the timing i ndex) and the amplitude parameter describes the magnitude of the temperatur e increase between the fitted winter minimum and summer maximum. For the zo oplankton, which have annual and strongly synchronous cycles of biomass, pr oductivity, and developmental sequence, we use copepodite stage composition to index the timing of the annual maximum. For seabirds, we examine (1975- 1999) the timing of hatching, nestling growth performance, and diet for fou r species of alcids at Triangle Island, British Columbia's largest seabird colony and the world's largest population of the planktivorous Cassin's auk let. Temperature, zooplankton, and seabirds have all shown recent decadal t rends toward 'earlier spring', but the magnitudes of the timing perturbatio ns have differed from variable to variable and from year to year. Recent (1 996-1999) extreme interannual variation in spring timing and April SST help ed to facilitate a mechanistic investigation of oceanographic features that affect the reproductive performance of seabirds. Our results demonstrate a significant negative relationship between the annual spring timing index ( and April mean SST) and nestling growth rates for both Cassin's auklet and rhinoceros auklet. Nestling growth rates were significantly lower in early, warm years. We demonstrate that low growth rates of Cassin's auklet occurr ed when copepod composition in nestling diet was low overall and copepods w ere scarce or absent in samples collected later in the season. We propose t hat when spring is early and warm, the duration of overlap of seabird breed ing and copepod availability in surface waters becomes reduced, effectively creating a seasonal mismatch of prey and predator populations. Such a mism atch could explain the reduced reproductive performance of seabirds compare d to years when spring was later and colder. The relationships we develop h ere can be used as simple predictive models to examine the effects of ocean climate change on seabird reproductive performance within our region. (C) 2001 Elsevier Science Ltd. All rights reserved.