Invasive species that form novel habitat often have large effects on ecosystems (Parker et al. 1999, Crooks 2002). These ecosystem engineers add physical structure, alter the abiotic environment, and change food webs (Jones et al. 1997, Hastings et al. 2007) and often these changes negatively affect native species (Crooks 2002, Levine et al. 2003). However, where the invasive species adds novel structure to ecosystems, positive effects on native species are frequently observed (Crooks 2002). In particular, invasive marine ecosystem engineers that colonize substrata with little aboveground structure, such as bare rock or sediment, often contain diverse epifaunal assemblages (Castilla et al. 2004, Byers et al. 2012), but there is little known about the relative importance of mechanisms that determine the facilitation.
Invasive ecosystem engineers can facilitate associated fauna via alterations to resource availability or by modifying abiotic conditions. Changes in resource availability occur when the invasive species provides a new food source for native species (Parker et al. 2006) including detritivores (Levin et al. 2006, Bradford et al. 2012). In the case of invasive seaweeds, small marine herbivores (mesograzers) are frequently a large component of invertebrate assemblages (Wikström and Kautsky 2004, Byers et al. 2012), but often consume less of the invasive species compared to native algae (Gollan and Wright 2006, Hammann et al. 2013). Although there are several engineering mechanisms by which invasive ecosystem engineers can facilitate native consumers (Crooks 2002), two are likely to be particularly important in the marine intertidal.
First, the structure provided by native intertidal engineers reduces abiotic stresses for associated species (disturbance by waves, air temperature, and evaporation rates), which facilitates plant (Bruno 2000) and animal (Bertness et al. 1999) communities. Second, engineered structure provides a refuge from predation for small marine invertebrates, such as amphipods (Stoner 1982, Duffy and Hay 1994). Invasive marine ecosystem engineers have similar effects (Gribben and Wright 2006, Neira et al. 2006). Overall, although a number of studies have identified trophic and engineering mechanisms by which invasive ecosystem engineers facilitate native fauna, few have explicitly distinguished between the different mechanisms in either marine or terrestrial systems.
The invasive Japanese seaweed Gracilaria vermiculophylla (hereafter Gracilaria), has established on both coasts of North America and northern Europe (Nyberg et al. 2009, Thomsen et al. 2009, Byers et al. 2012). In the southeastern United States (South Carolina and Georgia), Gracilaria is a novel addition to estuaries where the vegetation has historically been dominated by the salt marsh angiosperm Spartina alterniflora (hereafter Spartina) and its detrital input. With high turbidity and little substrate for seaweed attachment, these estuaries were mostly devoid of macroalgae prior to the invasion of Gracilaria. On mudflats in these estuaries, Gracilaria is now commonly found attached to the tube-building worm, Diopatra cuprea (Polychaeta: Diopatridae), and as drift, and its biomass reaches up to 1 kg/m2 (Byers et al. 2012). Prior to the Gracilaria invasion, the microphytobenthos was the only source of primary productivity on the mudflat.
The addition of this large amount of algal tissue to these homogenous mudflats appears to be having dramatic effects on the community. Small invertebrates including amphipods, gastropods, shrimp, and crabs are abundant on Gracilaria (Nyberg et al. 2009, Thomsen et al. 2009, Byers et al. 2012) but it remains unclear whether these invertebrates are responding to Gracilaria as a source of food, protective structure or an abiotic ameliorator. In Europe, mesograzers generally eat only small amounts of invasive Gracilaria (Nejrup et al. 2012, Hammann et al. 2013).
Our overall aim was to determine the relative importance of three main mechanisms by which an invasive ecosystem engineer facilitates its associated faunal community. Initially, we surveyed amphipods on Gracilaria vs. other benthic macrophytes on mudflats and in the salt marsh. Next we determined habitat selection by the numerically dominant mesograzer in the system (the omnivorous amphipod Gammarus mucronatus, family Gammaridae; hereafter Gammarus) for live Gracilaria vs. bare mudflat during high and low tide (submersion and emersion, respectively). Then, to understand the mechanisms promoting the high association of Gammarus with Gracilaria that we observed, we determined the use of Gracilaria as a food source by Gammarus using laboratory feeding and survivorship experiments and a stable isotope survey, and structural engineering effects of Gracilaria on amphipod survivorship at low tide via desiccation-resistance effects and at high tide via predation-refuge effects.
Results
Distribution patterns and habitat selection by Gammarus
Gammarus was significantly more abundant on the mudflat, where Gracilaria made up 97% of macrophyte biomass (pooled across locations) than on the marsh, where Spartina and Gracilaria made up 73% and 25% of the biomass, respectively (F1,44 = 6.470, P = 0.015; Fig. 1A and B). On the mudflat, Gammarus showed a strong habitat preference for Gracilaria over bare mud when surveyed at both high (F1,31 = 39.706, P < 0.001) and low (F1,31 = 76.446, P < 0.001) tide (Fig. 1C). Gammarus abundance also differed among locations at both high tide (F3,32 = 12.077, P < 0.001) and low tide (F3,32 = 12.201, P < 0.001; Appendix A: Table A1).
The use of Gracilaria as a food resource by Gammarus
In the no-choice experiment, Gammarus consumed more Spartina detritus than live Gracilaria but there was no significant difference in feeding rate (the diet × amphipod presence/absence interaction was marginally nonsignificant, F1,40 = 3.202, P = 0.081, Fig. 2A). In the choice experiment, Gammarus consumed significantly more Spartina detritus than live Gracilaria (t = −2.034, df = 34, P = 0.049).
Carbon and nitrogen isotope ratios indicated Gammarus does not utilize Gracilaria as a trophic resource (Fig. 2B). Trophic fractionation of 13C is generally minimal (∼0.4% [Post 2002]) and closely approximates the difference between Spartina and Gammarus (0.34%), whereas the difference between Gammarus and Gracilaria (4.66%) greatly exceeds this value, precluding Gracilaria as a major trophic resource. Additionally, δ15N values of Gammarus were indistinguishable from those of Gracilaria, making Spartina (3.14% enrichment) or sediment (2.18% enrichment) more likely resources for Gammarus.
Survivorship of Gammarus was lower on Gracilaria than on Spartina detritus or sediment by day 8 (χ2 = 17.86, df = 3, P < 0.001) and, by day 14, all Gammarus fed Gracilaria were dead (Fig. 2C). All Gammarus with no food were dead by day 8. In contrast, 40% of Gammarus with sediment and 20% of Gammarus fed Spartina detritus survived to day 28 and there was no significant difference in survivorship between those two food sources (Fig. 2C; χ2 = 3.43, df = 1, P = 0.064). Survivorship of Gammarus on Gracilaria was only higher than the no-food controls on day 5 (χ2 = 17.030, df = 1, P < 0.001).
Structural engineering effects of Gracilaria on amphipod survivorship
Amphipod survivorship at low tide was greater with Gracilaria than on bare mudflat (F1,18 = 42.205, P < 0.001; Appendix C: Table C1). After 1 hour exposure, amphipod survivorship in the presence of Gracilaria was 84.0% ± 7.5% (mean ± SE) compared to 24.0% ± 11.7% with no Gracilaria. After 3 hours exposure, amphipod survivorship in the presence of Gracilaria was 88.0% ± 8.0% compared to 4.0% ± 4.0% with no Gracilaria. There was no difference in amphipod survivorship between 1 and 3 hours exposure in the presence of Gracilaria (F1,17 = 0.985, P = 0.336; Appendix C: Table C1). During the experiment, amphipods on bare mudflats were exposed to sediments with less water and higher temperatures than amphipods with Gracilaria. After 1 hour, there was no difference in water loss from the sediment with Gracilaria (19.2% ± 1.3%) vs. without Gracilaria (18.3% ± 1.1%) but, after 3 hours, there was significantly greater loss of water without Gracilaria (37.2% ± 1.1%) vs. with Gracilaria (30.0% ± 2.0%; significant Gracilaria presence/absence × time interaction, F1,16 = 8.315, P = 0.011, Tukey's α < 0.05). Sediment temperature was slightly lower beneath Gracilaria (33.6° ± 0.2°C) compared to adjacent areas without Gracilaria (34.2° ± 0.4°C, N = 8, t = 3.416, P = 0.011, paired t test).
Gracilaria increased the survivorship of amphipods in the presence of the predatory shrimp Palaemonetes and Panopeid mud crabs. For Palaemonetes, there was a significant interaction between Gracilaria and predator presence (F1,14 = 9.789, P = 0.007; Appendix C: Table C1); after 72 hours there were no amphipods alive in tubs with Palaemonetes lacking Gracilaria compared to 53.8% ± 10.2% survivorship in tubs with Palaemonetes and Gracilaria (Tukey's α < 0.05). For Panopeid mud crabs, this interaction was marginally nonsignificant (F1,16 = 4.049, P = 0.061), although amphipod survivorship in the presence of Panopeid mud crabs was more than three times higher in the tubs with Gracilaria (52.4% ± 5.4%) compared to tubs lacking Gracilaria (14.0% ± 1.9%, Tukey's α < 0.05). In both experiments, the positive effect of Gracilaria only occurred when predators were present, such that there was no effect of Gracilaria on amphipod survival during submersion in the absence of predators (Tukey's α > 0.05, Palaemonetes experiment, with Gracilaria, 91.0% ± 4.6% vs. without Gracilaria, 85.0% ± 9.1%; Panopeid mud crab experiment, with Gracilaria, 76.0% ± 8.3% vs. without Gracilaria, 62.0% ± 6.4%).
Discussion
Although invasive ecosystem engineers often facilitate native species, the relative importance of mechanisms underpinning that facilitation is not well understood. In particular, small consumers could switch to invasive hosts because they are a novel food source; the host protects them from predators and harsh abiotic conditions, or both. We have shown that one of the most abundant benthic amphipods in the southeastern United States, Gammarus mucronatus, is facilitated by the invasive Gracilaria, with densities 5–100 times higher on mudflats with Gracilaria compared to uninvaded mudflats. Although Gracilaria potentially represents a novel food resource to native consumers, it is not an important food for Gammarus. Indeed, amphipods recruit to structural mimics of Gracilaria that have no nutritional value, albeit at lower levels than they recruit to Gracilaria itself (Byers et al. 2012). Instead, the high abundance of Gammarus on Gracilaria-dominated mudflats is likely maintained by reductions in predation mortality during high tide and abiotic stress during low tide.
Habitat structure is an important mediator of predator–prey interactions (Grabowski 2004) and the lower predation on Gammarus in the presence of Gracilaria is likely due to Gracilaria structure reducing predator–prey encounter rates or predator foraging efficiency. In the laboratory, we observed both Palaemonetes and Panopeid mud crabs actively pursuing and consuming Gammarus, particularly when Gracilaria was absent. Native habitat-forming seaweed and seagrass reduce predation by fish on herbivorous amphipods, particularly mobile species like Gammarus, which are at a higher risk of predation than more sedentary nest-building species (Van Dolah 1978, Stoner 1982, Duffy and Hay 1994). Moreover, Gracilaria's chemical defenses (Nylund et al. 2011) may deter omnivorous predators such as fish and reduce predation on amphipods in Gracilaria (Duffy and Hay 1994). In fact, lower predation rates on epifauna within the structure of chemically defended invasive seaweed are known across multiple taxa, including bivalve recruits and herbivorous isopods (Gribben and Wright 2006, Enge et al. 2013).
The role of canopy-forming seaweed in ameliorating harsh abiotic conditions and facilitating understory invertebrates on rocky shores is well established (Bertness et al. 1999). On the mudflats of coastal Georgia, temperatures regularly exceed 35°C and the Gracilaria canopy massively reduced mortality of Gammarus when exposed at low tide. Thus, in addition to the predator refuge effect at high tide, Gracilaria likely reduces desiccation and heat stress for amphipods at low tide, demonstrating a second mechanism of facilitation by this invasive seaweed that occurs at a different time in the tidal cycle. Although Gracilaria is typically only exposed for ∼3–4 hours at low tide, the strong survivorship benefit of Gracilaria after 3 hours exposure suggests survivorship over longer tidal periods is likely.
Given these survivorship results and that Gammarus were virtually absent from bare mudflat at low tide, it is likely they seek out structural refugia on the mudflat to avoid the harsh abiotic conditions. The low value of Gracilaria as a food emphasizes its importance as a structural, protective habitat for Gammarus. Indeed, Gracilaria provides virtually all aboveground habitat available for colonization by Gammarus on the mudflats with virtually no vegetative structure on these mudflats prior to Gracilaria invasion. Only small amounts of Spartina detritus occur on the mudflat because most Spartina wrack is deposited at the high-tide mark (where Spartina habitat occurs). Ulva is found in small quantities on the mudflat only during winter and few amphipods are present on Diopatra tubes themselves.
Previous feeding experiments showed Gammarus mucronatus consumed Ulva at higher rates compared to other seaweeds, including the native Gracilaria tikvahiae (Duffy and Hay 1994, Cruz-Rivera and Hay 2003), and consumed more seagrass detritus and epiphytic filamentous algae compared to other seaweeds (Zimmerman et al. 1979). Similarly, in northern Europe, Gammarus locusta eats relatively little invasive Gracilaria compared to native seaweeds, particularly Ulva (Nejrup et al. 2012). Our experiments indicate that Gammarus will eat some Gracilaria, but cannot survive when isolated with it.
Consistent with this, our isotope data show that natural populations of Gammarus consume little Gracilaria, but instead favor a combination of Spartina detritus, microalgae, or periphyton on the sediment surface or on Gracilaria. The low feeding of Gammarus on Gracilaria could reflect the alga's chemical defenses (Nylund et al. 2011). Our survivorship experiment showed that a diet comprised of either Spartina detritus or sediment alone increased survivorship relative to Gracilaria. At first glance, the low fitness on Gracilaria was surprising given that Spartina detritus is nutritionally poorer than Gracilaria, but combining Spartina detritus and microalgae in its diet may enhance fitness for Gammarus (Cruz-Rivera and Hay 2000, Parker et al. 2008). Moreover, fungi and bacteria may colonize Spartina detritus (Buchan et al. 2003) potentially enhancing its nutritional value for Gammarus. As Gammarus is relatively mobile (Duffy and Hay 1994), it will frequently come into contact with food of variable quality and it can maintain fitness as long as it consumes an adequate amount of high-quality food (Cruz-Rivera and Hay 2000).
The generally higher density of Gammarus on the mudflat compared to the saltmarsh and the near absence of Gammarus on bare mud compared to Gracilaria, suggests that Gracilaria's invasion onto mudflats may have allowed Gammarus to move out of the low marsh to utilize the mudflat. Gammarus was found in the marsh on Spartina wrack and Ulva, suggesting it occurred there prior to Gracilaria invasion although, we have no information on Gammarus abundance on these mudflats or in the saltmarsh before Gracilaria invasion. The substantially lower tidal elevation of the mudflat compared to the marsh may be advantageous for Gammarus by reducing the amount of time they are exposed to harsh abiotic conditions at low tide. In addition to structural effects, Gracilaria provides a massive increase in primary productivity on these mudflats. However, the lack of feeding on Gracilaria by the most abundant grazing epifaunal species associated with it suggests that secondary productivity is enhanced indirectly via structural provisioning. Our experiments indicate Gracilaria reduces Gammarus mortality from predation and harsh abiotic conditions but it is also possible that Gracilaria facilitates microalgae (either on the sediment or on Gracilaria itself), which Gammarus then feeds on.
Ecosystem engineers that create habitat are increasingly being considered as influential in determining community organization (Crooks 2002, Hastings et al. 2007). Gracilaria is making major changes to estuarine mudflats in southeastern USA that were historically devoid of macroalgae by providing a novel habitat that attracts and protects native epifauna. Although the abiotic changes caused by invasive ecosystem engineers can negatively impact associated species, in structurally depauperate systems, abiotic changes caused by the addition of novel structure may dampen harsh environment conditions allowing community wide facilitation and enhancing overall productivity.
November 2014
