Alteration of ecosystem structure and processes on SCI occurred not just because of the presence of non-native species that had ecosystem-level effects, but also as a result of their removal. Although the EE concept did not exist prior to the removal of sheep and cattle from SCI, it is instructive to ask to what extent could non-native species such as feral sheep, feral pigs, and cattle, as well as fennel and annual grasses and forbs, be considered ecosystem engineers.
It could be argued that the feral sheep, pigs, and cattle do not fit the definition of an EE species. One part of the definition is that the EE process is independent of assimilation (Jones and Gutiérrez, Ch 1), and the denuding of the landscape by the sheep and the extensive rooting by the pigs were, in large part, a result of their feeding behavior. Whether they precisely meet the definition of an EE species though may not be as important as their effects; severe erosion, soil compaction, changes in hydrology, fragmentation and alteration of vegetation community structure and function, and at least local extirpation of species are serious ecosystem consequences by any standard. Another part of the definition of an EE is that they can be considered species that change the abiotic environment with subsequent consequences to biotic components of an ecosystem (Jones and Gutiérrez, Ch 1). The sheep and pigs clearly fit this part of the definition; their effects had profound consequences for both abiotic and biotic parts of the environment on SCI. Following from this are two important considerations though: (1) Some non-native species on SCI with effects consistent with the EE concept were clearly suppressing other potential non-native EE species; and (2) the relative magnitude of the effects of the sheep, pigs, and cattle depended in large part on their abundance.
SCI had been invaded by many non-native animals and plants, and these species had interacted for decades. The ecosystem engineering effects of the sheep, cattle, and pigs were extensive and severe, but they also suppressed (or masked) effects of other species that could also be considered ecosystem engineers. It wasn't until the sheep and cattle were removed that fennel, as well as the non-native annual grasses and forbs, could be thought of in an EE context. Physical features such as ecosystem structure (vegetation height and patchiness), light environment, and moisture availability become much different in stands of fennel and in grasslands dominated by non-native annuals than in the communities they replace (Gordon and Rice 1992, Dyer and Rice 1999, Brown and Rice 2000, Brenton and Klinger 2002, Ogden and Rejmanek 2005). On SCI this resulted in effects on the biotic features of the ecosystems, such as species composition and richness (Klinger and Messer 2001, Brenton and Klinger 2002). In essence, after the sheep and cattle were removed from SCI effects of one group of EE species (sheep and cattle) were replaced by others (fennel and, collectively, non-native annual plants), while effects from another species (pigs) remained largely unaltered.
In contrast with some other EE species (Wright and Jones 2006), none of the non-native species on SCI that could potentially be categorized as ecosystem engineers had intrinsically strong per capita effects. Rooting by pigs probably had a far greater per capita effect on ecosystem processes than grazing and browsing by sheep. But the abundance of feral sheep was at least an order of magnitude greater than that of feral pigs, so on an absolute basis their ecosystem engineering effects were far more extensive. In either case, though, trophic and nontrophic effects of both species on the island's ecosystems were proportional to their abundance. Similarly, the ecosystem effects of fennel and non-native annual herbaceous species did not become serious until they increased in abundance after the sheep and cattle were removed.
If management of one EE species in systems with multiple invaders could result in release of another EE species, then determining when a non-native species is most likely to begin having ecosystem engineering effects becomes very important. Biological invasion can be thought of as a process occurring in phases of colonization, establishment, spread, and equilibrium (Ricklefs 2005, Salo 2005). In addition, invasion effects tend to increase the longer a non-native species has persisted in a region (Rejmanek et al. 2005a). These two aspects of biological invasions may explain a great deal of the complex interactions among the sheep, pigs, cattle, fennel, and annual grasses on SCI. They may also provide a crucial link for evaluating when a non-native species could begin to have significant ecosystem engineer effects, either as a result of their being introduced to an area or as an outcome of management actions.
Because their per capita effects were relatively small when they were not abundant, thinking of species such as sheep, pigs, cattle, fennel, and non-native annual plants as ecosystem engineers when they are in the colonization and establishment phases of invasion is not particularly informative. But as they enter the spread phase and become much more widespread and abundant it does start to be useful to begin thinking of them as potential EE species. Whether they actually start altering ecosystem properties as they enter the equilibrium phase will depend on their long-term mean abundance and interactions with other species. The sheep, pigs, cattle, fennel, and non-native annual plants had co-occurred on SCI for many decades and each was clearly in the equilibrium phase of the invasion process (albeit a dynamic equilibrium). Interaction strengths among them were well established; grazing by the sheep and cattle prevented the plants from having strong ecosystem engineering affects. Removal of the sheep and cattle created a non-equilibrium situation though. As a result, interaction strengths were changed and the fennel and non-native annual plants rapidly increased in abundance and began to have major EE effects.
Was this article helpful?