Effects of antagonistic ecosystem engineers on macrofauna communities in a patchy, intertidal mudflat landscape

• Few experiments test if contrasting ecosystem engineers facilitate different organisms.
• We tested if seagrass removal and lugworm addition facilitate different macrofauna.
• Seagrass removal decimated density of attached epifauna, reduced overall diversity.
• Lugworm addition had weak effect, which contrast to strong effects in other studies.
• Different engineers facilitate different organisms, but effects are context-dependent.

Ecosystem engineers are organisms that strongly modify abiotic conditions and in the process alter associated communities. Different types of benthic ecosystem engineers have been suggested to facilitate different communities in otherwise similar marine environments, partly because they alter sediment conditions in contrasting ways. However, most studies testing this hypothesis have either not manipulated the presence of engineers, or have transplanted engineers into areas already dominated by other engineers, which limits the ability to assess the relative engineering effects. Here we combined a field survey and a field experiment to investigate if two contrasting ecosystem engineers – the sediment-stabilizing seagrass Zostera noltei and the bioturbating lugworm Arenicola marina – facilitate different macrofauna communities. The study was performed in a sheltered mudflat area of the eastern Dutch Wadden Sea, where seagrasses and lugworms form a mosaic of spatially alternating seagrass-dominated elevations (hummocks) and lugworm-dominated depressions (hollows). Results showed that seagrasses facilitated some organisms (mainly attached epifauna) while lugworms facilitated others (primarily burrowing infauna), generating distinctly different macrofauna communities in hummocks and hollows. However, seagrasses had a much stronger effect on the macrofauna communities than lugworms, and competitively excluded lugworms. This contrasts with results from similar studies in hydrodynamically more exposed sand flats, where lugworms instead dominate communities and exclude seagrass. We therefore propose that effects of ecosystem engineering (acting primarily on a local scale) and variation in abiotic conditions (acting on larger scales, e.g., hydrodynamic gradients along the Dutch coastline) strongly interact to dictate the distribution and fitness of engineering species, and indirectly, the diversity and structure of associated benthic communities.