Aim We tested whether a hybrid zone that has formed between an endemic and an invasive species of marine mussel has shifted poleward as expected under a general hypothesis of global warming or has responded instead to decadal climate oscillations.
Location We sampled 15 locations on the coast of California, USA, that span the distributions of the two species of marine mussels and their hybrids.
Methods Mussels were sampled in 2005–08 and analysed at three nuclear gene loci using methods identical to those used in a study a decade earlier in order to document the genetic architecture of this system. Change in the system was determined by comparing the frequency of species-specific alleles and multi-locus genotypes over the intervening decade. Climate variation over the same period was examined by comparing the Pacific Decadal Oscillation (PDO), El Niño/Southern Oscillation (ENSO), upwelling indices and sea surface temperature (SST) during and prior to the study period.
Results Contrary to the general expectations of global warming we show that the highly invasive warm-water mussel Mytilus galloprovincialis and the hybrid zone formed with the endemic species Mytilus trossulus has rapidly contracted southwards. Mytilus galloprovincialis declined in abundance over the northern third of its geographic range (c. 540 km) and has become rare or absent across the northern 200 km of the range it previously colonized during its initial invasion. The distribution of the native species M. trossulus has remained unchanged over the same time period.
Main conclusions The large-scale range shift in the warm-water invasive species M. galloprovincialis and the hybrid zone it forms with M. trossulus has been exceptionally rapid and is in the opposite direction to that predicted by the global warming hypotheses. This shift, however, is consistent with decadal climate variation associated with the ENSO and the PDO. Since the biogeography of this system was first described in 1999, the PDO has shifted from a warm phase, dominated by frequent and large El Niño events, to a cold-phase period, with minimal ENSO activity. Thus recent decadal climate variation can oppose global trends in average temperature and this study illustrates the need to integrate the effects of climate change across multiple time-scales.