Aquat Ecol (2007) 41:129–147 DOI 10.1007/s10452-006-9053-3 ORIGINAL PAPER An examination of the spatial and temporal generality of the influence of ecosystem engineers on the composition of associated assemblages Katrin Berkenbusch Æ Ashley A. Rowden Received: 29 June 2005 / Accepted: 8 May 2006 / Published online: 5 August 2006 Ó Springer Science+Business Media B.V. 2006 Abstract The present study evaluated the gen- between sites were best explained by sediment erality of ecosystem engineering processes by variables. In New Zealand, % fines and seagrass examining the influence of sympatric burrowing debris showed the highest correlation to differ- shrimps (Callianassidae) and intertidal seagrasses ences in assemblage composition, and in the (Zosteraceae) on benthic assemblage composi- U.S.A. % fines, % carbon and sediment turnover tion in two temperate regions, south-eastern New (by shrimp) appeared to be the most important Zealand and north-western U.S.A. In each re- environmental parameters measured. Four to six gion, intertidal macrofauna assemblage composi- taxa exhibited the greatest discriminating signifi- tion was determined at sites of different cance (including corophiid amphipods, spionid burrowing shrimp/seagrass density and where polychaetes and oligochaetes) for dissimilarities both species co-occured, in three different size in assemblage composition observed at the dif- estuaries/tidal inlets, on two occasions. Results ferent sites, with generally lower abundances at from both regions showed that the presence of shrimp than at seagrass sites. The present study shrimps and seagrasses consistently influenced the highlights the functional importance of seagrasses composition of the associated infaunal assem- and bioturbating shrimps as ecosystem engineers blages at all sites, in both summer and winter. in soft-sediment environments, and reveals the Macrofauna assemblages at shrimp sites were generality of their influence on associated macro- significantly different to those at seagrass-only invertebrate assemblages. The findings also allow and mixed sites, whereas the composition of the for further development of a heuristic model for latter sites was similar. The differences observed ecosystem engineering by shrimp and seagrass which indicate that numerical models that aim to explore the relationship between ecosystem K. Berkenbusch (&) Æ A. A. Rowden engineer populations and habitat modification National Institute of Water and Atmospheric should be expanded to capture the interaction of Research, Private Bag 14-901, Wellington, co-occurring engineers and be both spatially and New Zealand e-mail: [email protected] temporally explicit. Present Address: Keywords Callianassidae Æ Macro-invertebrate K. Berkenbusch assemblages Æ Ecosystem engineer Æ Portobello Marine Laboratory, Department of Marine Science, University of Otago, P.O. Box 8, Environmental variables Æ Intertidal sandflat Æ Portobello, Dunedin, New Zealand Temperate regions Æ Zosteraceae 123 130 Aquat Ecol (2007) 41:129–147 Introduction classic examples of ecosystem engineers (Levinton 1995), as they affect physical and biogeochemical Ecologists concerned with the organisation of properties of the sediment, such as near-surface natural communities have long recognised the sediment stability, sediment grain size, organic importance of species that have a disproportion- content and nutrient loading; environmental ate influence on assemblage composition. Some parameters which affect the habitat suitability organisms, such as so-called ‘keystone’ or ‘foun- for other species (de Wilde 1991). However, dation’ species, profoundly impact associated application of the formal assessment criteria for organisms in an assemblage through predation recognising and scaling the influence of ecosystem and competition; and ecological experiments and engineers (Jones et al. 1994) has been lacking for theory have predominantly focused on consider- marine bioturbators until very recently. ing how these biotic interactions structure ter- Callianassid shrimps (Thalassinidea), also restrial and aquatic assemblages (Paine 1966; referred to as ghost, burrowing or mud shrimps, Dayton 1972; Schoener 1983). However, rela- are a ubiquitous group of decapod crustaceans tively recent ecological studies have highlighted that have long been recognised as significant the importance of abiotic interactions between infaunal bioturbators of intertidal and shallow organisms and their environment, which can be subtidal soft sediments (Suchanek 1983; Swift equally dramatic to those of keystone species, but 1993; Cade´e 2001). Application of the formal are not included in the current ecological frame- ecosystem engineer assessment criteria to one work (Bertness 1985; Jones et al. 1994; Bruno such ghost shrimp, Callianassa filholi, identified et al. 2003). Jones et al. (1994) tried to address this species as an important allogenic ecosystem this shortfall by proposing the concept of organ- engineer, which influenced macrofauna assem- isms as ‘‘physical ecosystem engineers’’. This no- blage composition over a small spatial scale tion encompasses interactions between organisms through its large per capita bioturbation activity and their environment that are not directly tro- (Berkenbusch and Rowden 2003). Having phic or competitive, and which result in the established the engineering significance of this modification, maintenance and/or creation of species (endemic to New Zealand) in one habitats. That is, rather than providing resources intertidal habitat raises the question as to whe- directly, ecosystem engineers physically change ther ghost shrimp species are generally impor- their environment and impact the supply of re- tant in structuring assemblage composition sources for other species, thereby affecting spe- across similar habitats at different spatial scales. cies abundance and diversity, population, The study on the impact of Callianassa filholi community and ecosystem stability (Jones et al. bioturbation on associated community assem- 1994). Jones et al. (1994) distinguished between blages also indicated that the engineering influ- two kinds of engineers: ‘Autogenic engineers’ ence of the ghost shrimp was moderated by the impact the habitat in which they live through their presence of a small intertidal seagrass Zostera own physical structures and are an integral part of capricorni (previously Zostera novozelandica, the engineered ecosystem. ‘Allogenic engineers’, see Les et al. 2002) which buffered the effect of on the other hand, modulate resources from one the shrimp during summer, when seagrass bio- physical state to another through their behaviour mass was high. Zostera capricorni can be con- and activity. sidered an autogenic ecosystem engineer, as Following the initial proposal of the concept, a seagrasses, for example, provide living space for number of studies illustrated its application to others through their own physical structure terrestrial and aquatic habitats by providing (Berkenbusch et al. 2000, and see conceptual examples of plants and animals as autogenic or model in Berkenbusch and Rowden 2003). The allogenic ecosystem engineers; e.g. Sphagnum temporally manifested moderation of the influ- moss (van Breemen 1995) and detritivorous ence of one type of engineering species on that tropical fish (Flecker 1996). In the marine envi- of another raises the question as to whether such ronment, bioturbators have been presented as an interaction generally occurs between ghost 123 Aquat Ecol (2007) 41:129–147 131 shrimps and seagrasses, which frequently co- region, three estuaries/tidal inlet locations were occur in the same habitats. selected that were 10s of km apart and which Thus, the present study aimed to assess the contained both ghost shrimp and seagrass within generality of ecosystem engineering processes for the same intertidal area. Estuaries/inlets were sympatric allogenic engineering shrimps (Cal- selected to have different areas, but with similar lianassidae) and autogenic engineering seagrasses tidal regimes across regions. In New Zealand, (Zosteraceae). Specifically, we addressed the fol- study locations were Otago Harbour (46 km2), lowing questions: (1) Do ghost shrimps and Blueskin Bay (6.9 km2) and Papanui Inlet seagrasses consistently influence associated ben- (3.5 km2). In the U.S.A., the study locations were thic assemblages over different spatial scales and Tillamook Bay (33.5 km2), Yaquina Estuary time? and (2) What are the potential mechanisms (15.8 km2) and Netarts Bay (9.4 km2). All estu- by which these ecosystem engineers exert their aries/tidal inlets were characterised by semi- influence on associated macrofauna assemblages? diurnal tides with a similar tidal range (1.4–2.2 m Addressing such questions allows for the explo- in New Zealand, 1.3–2.9 m in the U.S.A.). Ghost ration of what Jones et al. (1997) term ‘multiple shrimps included in the study were Callianassa engineers and coupled and uncoupled interac- filholi in New Zealand, and Neotrypaea califor- tions’ and ‘coupled engineering and trophic cas- niensis in the U.S.A. Both callianassid species are cades’, the development of the local heuristic burrowing deposit-feeders, of similar size (12– ecosystem engineering model earlier proposed for 16 mm adult carapace length) and commonly shrimp and seagrass (Berkenbusch and Rowden occur in intertidal soft-sedimentary habitats in 2003), and the potential to develop or parame- their respective regions (MacGinitie 1934; Devine terize the models of