Abundance of the introduced seastar, Asterias amurensis, and spatial variability in soft sediment assemblages in SE Tasmania: Clear correlations but complex interpretation

Abstract

The northern Pacific seastar, Asterias amurensis, was first collected in southeast Tasmania in 1986. Mistaken for the endemic asteroid Uniophora granifera, its true identity was not realised until 1992. It is now a conspicuous predator in soft sediment habitats in this region, and is considered a major threat to native assemblages and commercial species. We examined the structure of soft sediment assemblages at different spatial scales in southeast Tasmania, and correlated spatial variation in community composition with seastar abundances. We found that the structure of soft sediment assemblages is highly variable at a range of spatial scales from metres to tens of kilometres. Clear differences in the composition of assemblages and abundances of major taxa were detected between areas with and without seastars and between areas with low and high seastar densities. However, the nature of these patterns suggests that they are more likely due to differences in sediment characteristics than due to impacts of the seastar. Thus, spatial differences in soft sediment assemblages might have been erroneously attributed to seastars without detailed information on important physical factors such as sediment characteristics. A second survey, using larger sampling units (1 m2) but across a more limited spatial extent, targeted bivalves and heart urchins that were identified as important prey of the seastar in observations of feeding and in experimental studies. Large-scale patterns of abundance and size structure were consistent with seastar effects anticipated from small-scale experimental and feeding studies for some, but not all, species. While the field survey ultimately provided evidence about the presence or absence of seastar impacts at large-scales, the identification of key ecological variables in experimental and feeding studies proved crucial to both the design and interpretation of patterns observed in the large-scale surveys. Overall, this work highlighted the necessity to consider multiple lines of evidence rather than relying on a single 'inferential' test, in the absence of pre-impact data.