An examination of the spatial and temporal generality of the influence of ecosystem engineers on the composition of associated assemblages

The present study evaluated the generality of ecosystem engineering processes by examining the influence of sympatric burrowing shrimps (Callianassidae) and intertidal seagrasses (Zosteraceae) on benthic assemblage composition in two temperate regions, south-eastern New Zealand and north-western U.S.A. In each region, intertidal macrofauna assemblage composition was determined at sites of different burrowing shrimp/seagrass density and where both species co-occured, in three different size estuaries/tidal inlets, on two occasions. Results from both regions showed that the presence of shrimps and seagrasses consistently influenced the composition of the associated infaunal assemblages at all sites, in both summer and winter. Macrofauna assemblages at shrimp sites were significantly different to those at seagrass-only and mixed sites, whereas the composition of the latter sites was similar. The differences observed between sites were best explained by sediment variables. In New Zealand, % fines and seagrass debris showed the highest correlation to differences in assemblage composition, and in the U.S.A. % fines, % carbon and sediment turnover (by shrimp) appeared to be the most important environmental parameters measured. Four to six taxa exhibited the greatest discriminating significance (including corophiid amphipods, spionid polychaetes and oligochaetes) for dissimilarities in assemblage composition observed at the different sites, with generally lower abundances at shrimp than at seagrass sites. The present study highlights the functional importance of seagrasses and bioturbating shrimps as ecosystem engineers in soft-sediment environments, and reveals the generality of their influence on associated macro-invertebrate assemblages. The findings also allow for further development of a heuristic model for ecosystem engineering by shrimp and seagrass which indicate that numerical models that aim to explore the relationship between ecosystem engineer populations and habitat modification should be expanded to capture the interaction of co-occurring engineers and be both spatially and temporally explicit.