Disruption of essential habitat by a coastal invader: new evidence of the effects of green crabs on eelgrass beds

Abstract

Malyshev, A., and Quijón, P. A. 2011. Disruption of essential habitat by a coastal invader: new evidence of the effects of green crabs on eelgrass beds. – ICES Journal of Marine Science, 68: 1852–1856.

Eelgrass (Zostera marina) beds have been declining in Atlantic Canada and elsewhere, partly as a result of sediment disruption and direct feeding/cutting of basal meristems by the green crab (Carcinus maenas). Green crabs are detrimental to eelgrass beds, and field and laboratory experiments have confirmed that the deleterious role of this invasive species is mediated by at least two mechanisms, depending on the size/age of the crabs: uprooting by adults and grazing by juveniles. Eelgrass uprooting and grazing by green crabs are likely to contribute to further declines or a lack of recovery of eelgrass beds.

Carcinus maenas, eelgrass decline, grazing, uprooting, Zostera marina

Introduction

Zostera marina is a perennial seagrass that grows from horizontal rhizomes underground (Olesen and Sand-Jensen, 1994) and is currently distributed from Alaska to Baja California along the North American Pacific and Atlantic coasts (Meling-López and Ibarra-Obando, 1999). Its dense beds provide a structured nursery habitat for juvenile and adult fish, shrimp, lobsters, and other commercially important species (Heck et al., 2003; Short et al., 2006), so bed decline has broad ecological and economic implications. Global warming, eutrophication, decreased water clarity, and the introduction of invasive species have all been suggested to have a negative impact on eelgrass beds (Neckles and Short, 1999). In the past two decades, the area of all seagrasses has declined by 20%, reducing the oceans' ability to provide nursery habitats and to absorb carbon emissions (Duarte et al., 2004). Atlantic Canada has also seen a substantial reduction in eelgrass in the highly productive shorelines of the Gulf of St Lawrence. For instance, above ground biomass declined by a mean of 40% in 13 estuaries from 2001 to 2002 (Locke and Hanson, 2004). One of several potential causes of such losses is the arrival of the European green crab, Carcinus maenas, in the mid 1990s (Audet et al., 2003; Garbary et al., 2004).

The omnivorous diet of adult green crabs includes primarily crustaceans, molluscs, many polychaetes, and even green algae (Grosholz and Ruiz, 1996). Juvenile crabs prefer mussel and eelgrass beds as nursery habitat (Breteler, 1976; Moksnes, 2002), but do not consume eelgrass, so their negative effect on the beds is thought to be a result of sediment disruption as they dig for clams and other invertebrates. Green crabs rework the top few centimetres of sediment, exposing and uprooting eelgrass roots and rhizomes (Davis et al., 1998) or damaging eelgrass sheath bundles (Garbary et al., 2004). For instance, when the effects of green crab bioturbation were studied on eelgrass transplants, it was found that they destroyed up to 39% of the plants (Davis et al., 1998). However, no direct feeding on eelgrass has been reported, although other crab species do graze on eelgrass (Perkins-Visser et al., 1996).

In Tracadie Harbour, Nova Scotia, green crabs have been estimated to be able to remove up to 87 000 eelgrass shoots per day (Garbary et al., 2004). However, the exact mechanism by which they cause such severe impacts is not clear. Prince Edward Island (PEI) has a well-established population of green crabs as well as abundant eelgrass beds, so it represents an ideal area to conduct manipulative experiments in which to study their interactions. In this study, we used field and laboratory experiments to measure the extent to which sediment disruption and direct feeding on eelgrass by green crabs mediate their potential effects on the critical habitat.

Methods

During a spring low tide of July 2008, 20 galvanized wire mesh cages (0.4 m × 0.4 m × 0.75 m; 1 cm mesh) with open bottoms were placed in an eelgrass bed in Stewart Cove, PEI. The cages were anchored to the bottom by pushing them 10 cm into the sediment while trying to minimize the disruption of eelgrass shoots. They were arranged in pairs (distance between paired cages ∼3–4 m), with each pair separated from the next by 6–10 m. A window at the top of the cages allowed placement of an adult male green crab (50–60 mm carapace width, CW) in one randomly selected cage of each pair. These green crabs were all unharmed males that had been starved for 48 h before the start of the experiment, to standardize the hunger levels. Green crabs remained in the cages for 6 d, after which the cages were carefully opened to recover the green crabs (100% recovery) and to collect floating eelgrass shoots from inside the cages. The eelgrass shoots were packed, labelled, and brought to the laboratory for estimations of dry weight biomass. Subsequently, a large cylinder (0.03 m⁻²) was inserted ∼15 cm into the sediment (within the cage) to collect an eelgrass sample that was also packed and brought to the laboratory for estimation of above-ground dry biomass.

Crab-inclusion experiments also took place in the laboratory using fresh eelgrass shoots 15–25 cm tall that had been collected in Stewart Cove, Stratford. Each shoot was made up of several leaves naturally attached to a single basal meristem and a ∼2-cm long rhizome with roots. Roots and rhizomes were buried in a ∼4-cm bottom layer of sandy sediment from the same location inside glass tanks (10 cm × 40 cm × 20 cm). Shoots were planted randomly within the tank along with four soft-shell clams (Mya arenaria), a preferred food choice of green crabs (Davis et al., 1998). Soft-shell clams are common on PEI sandy bottoms and buried quickly into the sediment, mimicking to some extent the habitat and food of green crabs. Tanks were filled with (artificial, i.e. prepared) saltwater of 25 psu and kept at 18–20°C. After allowing 24 h for shoot and clam acclimatization, one adult male green crab (50–60 mm CW) was placed into each tank (n = 10) and matched with identical tanks without green crabs (controls; n = 10). These laboratory experiments lasted 6 d, after which the number of uprooted shoots in each tank was quantified. An identical experiment was run with juvenile green crabs (20–30 CW). After 6 d, the number of uprooted eelgrass shoots per tank was recorded and compared with control tanks.

The same laboratory setting was used to compare the quantity (dry biomass) of floating eelgrass shoots resulting from either uprooting or feeding/cutting by juvenile green crabs. Floating shoots with no roots attached were attributed to biomass loss through grazing, whereas shoots still attached to their roots were considered uprooted. As in the procedures described above, these replicated measurements (n = 10) were matched with those from control tanks without green crabs (n = 10). An additional (Supplementary material) laboratory experiment used bundles of four eelgrass shoots placed and held (weighed down) at the bottom of each tank, this time without sediment. The sole purpose of this additional experiment was to record on video whether juvenile green crabs did ingest eelgrass tissue or if eelgrass damage was solely attributable to physical damage while digging in the sediment. Preliminary observations had indicated that adult green crabs do not graze on the eelgrass, so only juveniles were used. Individual activities of three juvenile crabs were recorded by video over two nightly 8-h periods, after which the video was examined carefully for evidence of feeding behaviour.

The results of the field experiment were analysed using a two-way completely randomized analysis of variance (ANOVA) design that took into account treatment (presence or absence of green crabs) and cage location (cages were arranged in pairs and scattered along the eelgrass beds). The results of the laboratory experiments were analysed by one-way ANOVA that assessed the influence of treatment (tanks with and without green crabs) on each dependent variable (number of shoots uprooted by adult and juvenile green crabs, and the percentage of biomass loss resulting from uprooting vs. cutting/feeding by juvenile crabs). ANOVA assumptions of residual homoscedasticity and normality were checked with Bartlett's and Shapiro–Wilk tests, respectively. The results of the additional laboratory experiment prepared for video-recording were not tested statistically.

Results

In the field-inclusion experiment, the biomass of fresh floating eelgrass shoots in cages containing green crabs was at least an order of magnitude greater than in control cages lacking crabs (Figure 1; ANOVA, p = 0.034; Table 1). The biomass of eelgrass collected from the sediment within both cages (above ground biomass) was lower in the crab-included cages than in the control cages, but this difference was not significant (Figure 1; ANOVA, p = 0.296; Table 1).

Table 1.

Results of two- and one-way ANOVA summarizing comparisons conducted in field and laboratory experiments, with each comparison making reference to the results presented in Figures 1–3, and dependent variables and the main ANOVA components detailed.

Setting	Dependent variable	Source	Sum of squares	F-value	p-value
Field (Figure 1)	Floating eelgrass	Treatment	0.184	5.370	0.034
	Biomass	Location	0.049	1.890	0.188
		Location × treatment	0.013	0.77	0.393
		Error	0.266
	Bottom eelgrass	Treatment	5.131	1.170	0.296
	Biomass	Location	7.870	0.100	0.751
		Location × treatment	0.224	0.120	0.732
		Error	29.566
Laboratory (Figure 2)	Number of shoots uprooted by adult crabs	Treatment	224.450	134.22	<0.001
		Error	30.100
	Number of shoots uprooted by juvenile crabs	Treatment	1.800	3.86	0.065
		Error	8.400
Laboratory (Figure 3)	% of biomass uprooted by juvenile crabs	Treatment	0.155	0.230	0.153
		Error	0.126
	% of biomass cut by juvenile crabs	Treatment	0.157	9.600	0.006
		Error	0.285

Figure 1.

Results of the field inclusion experiment. (Top) Average (+1 s.e.) biomass of live uprooted eelgrass collected from inside the cages. (Bottom) Biomass of eelgrass shoots collected from the bottom. Filled and open bars correspond to cages with and without green crabs, respectively (one adult male green crab per cage).

Figure 2.

Results of laboratory inclusion experiments. (Top) Average (+1 s.e.) number of eelgrass shoots uprooted by an adult green crab. (Bottom) Average number of eelgrass shoots uprooted by a juvenile green crab. Filled and open bars correspond to cages with and without green crabs, respectively (one green crab per tank).

In the laboratory, adult green crabs uprooted ten times more eelgrass shoots than juvenile male green crabs (Figure 2). In contrast, shoots in control tanks remained unaltered. The number of uprooted shoots by juvenile crabs did not differ significantly from that estimated in the control tanks, but the probability value was marginal (Figure 2; ANOVA, p = 0.065; Table 1).

Eelgrass biomass loss as a result of juvenile burrowing/uprooting only reached ∼7%, whereas biomass loss through feeding/cutting totalled ∼19% (Figure 3). Feeding/cutting occurred evenly over the duration of the trials (6 d) and was observed in seven of the ten experimental tanks. Both losses contrast with the ∼1% of biomass lost in the control tanks as a result of natural tissue necrosis, an average significantly lower than the biomass lost by feeding/cutting (Figure 3; ANOVA, p = 0.006; Table 1).

Figure 3.

Results of laboratory inclusion experiments. (Top) Average (+1 s.e.) percentage of biomass uprooted by a juvenile green crab (in relation to intact eelgrass biomass). (Bottom) Average percentage of biomass cut by a juvenile green crab. Filled and open bars correspond to cages with and without green crabs, respectively (one green crab per tank).

In the additional laboratory experiment run for video-recording (results not plotted), juvenile green crabs grazed an average of 1.64/4 eelgrass shoots per tank (i.e. ∼41% of the planted shoots). Eelgrass shoots were considered grazed only if clear shear marks were visible on the basal meristem of the eelgrass or if shoots were completely cut in half. For all laboratory experiments, very few leaf epiphytes were detected, but grazing was always on the basal meristematic tissue, not on the leaves. Green crabs were recorded tearing the basal meristem and ingesting bits of plant tissue (Supplementary Video S1).

Discussion

As expected from direct observations and published evidence (e.g. Davis et al., 1998), green crabs were detrimental to eelgrass beds. The field and laboratory experiments also confirmed that such a deleterious role is mediated by at least two mechanisms, depending on the size/age of the crabs: uprooting by adults and grazing by juveniles.

Crustaceans have all the necessary enzymes to digest plants such as seagrass (Linton and Greenaway, 2007), so it is not surprising that eelgrass is indeed in the diet of large green crabs (Cohen et al., 1995). However, feeding on eelgrass is likely only incidental, because there is plenty of evidence suggesting that green crabs (Cohen et al., 1995; Davis et al., 1998) and other crab species (Freire and Gonzalezgurriaran, 1995) prefer soft-shell clams and a variety of other bivalve species over eelgrass shoots. These preferences are likely the consequence of the nutritional value of prey, which is substantially greater in animal tissue (Heck and Valentine, 2006). Large green crabs dig into the sediment in search of food and either uproot, expose, or weaken eelgrass roots and rhizomes, which may then be uprooted by the action of waves, currents, or other bioturbators. Shrimp, cownose rays, sand prawns, and several crab species also have the ability to damage eelgrass through sediment bioturbation while burrowing in search of food (Orth, 1975; Valentine et al., 1994; Dumbauld and Echeverria, 2003; Siebert and Branch, 2006). However, of the crab species found along the shores of the southern Gulf of St Lawrence, the green crab is the most prominent bioturbator and the only species that creates characteristic “feeding pits” that last from one to several days (PAQ, pers. obs.; Locke and Hanson, 2004).

In the laboratory trials examining uprooting, large green crabs could uproot a considerable quantity (nearly 84%) of the shoots available. These results need to be taken cautiously, however, because such an impact is likely not as severe in the field. In nature, green crabs are not confined to a small tank (or a field cage) and can easily dig for clams or alternative prey (Cohen et al., 1995) in sediments where eelgrass is absent. When green crabs do dig in eelgrass beds, waves and currents can amplify the effects observed both in the laboratory and in the field, facilitating the dislodgement of eelgrass shoots from the sediment. The extrapolation of uprooting rates estimated in the laboratory (and in the field experiment) is not necessarily straightforward. Our study calls for further studies that explicitly focus on two main venues: the use of larger-scale manipulations attempting to capture the landscape structure of eelgrass beds (cf. Hovel and Lipcius, 2001), and the manipulation of the number of green crabs to account for potential density-dependence.

Grazing by green crabs as a destruction mechanism for eelgrass was demonstrated for the first time in this study. Davis et al. (1998) reported seared sheaths in experimental eelgrass transplants, but did not attribute them to the feeding activity of green crabs. Video-recording conducted during our additional laboratory experiment suggests otherwise. Juvenile green crabs grazed at the basal meristem of the eelgrass shoots, where the tissues are younger and softer (Olesen and Sand-Jensen, 1994). In the Atlantic, the lady crab (Ovalipes ocellatus), a species of mud crab (Panopeus herbstii), and a species of hermit crab (Pagurus longicarpus) all graze on eelgrass, but they target either seeds or seedlings (Wigand and Churchill, 1988). In contrast, some species of isopod graze on eelgrass leaves and cause from 2 to 30% loss of tissue biomass (Bostrom and Mattila, 2005). Similarly, the gastropod Smaragdia viridis ingests primarily young leaf tissue near the junction of the leaves with the sheath (Rueda et al., 2009), an area in proximity to the basal meristems targeted by the juvenile green crabs in our experiments.

Although juvenile green crabs did uproot eelgrass at a lower rate than adults, they caused a combined biomass loss of 26% by grazing primarily on the eelgrass shoots. This may be attributed to the limited ability of small crabs to uproot eelgrass shoots, because they face more difficulty than adults digging through an eelgrass rhizome network. In the absence of alternative prey of greater nutritional value (cf. Chaves et al., 2010), juvenile green crabs are likely prompted to feed directly on the soft meristems of the eelgrass shoots. Similar to our estimates of uprooting by adult crabs, the rates of juvenile grazing need to be taken with caution because these results come from small-scale manipulations; several studies have indeed demonstrated that confinement of predators such as green crabs may alter their behaviour considerably (cf. Hulberg and Oliver, 1980; Hall et al., 1990).

Although the effect of native species seems to be sporadic or kept in check by natural predators, invaders such as green crabs may operate differently. The build-up of large populations of green crabs may imply a drastic change for eelgrass bed structure from pre-invasion situations. In areas where native bioturbators comparable with green crabs are not common, as in the shoreline system studied here, the effects of green crabs are most likely to be severe. In areas where there are native bioturbators, changes are still likely to occur because such species may interact and potentially be displaced by the growing populations of green crabs.

Supplementary material

Supplementary material is available at the ICESJMS online version of the paper in the form of a video of green crabs consuming eelgrass shoots.

Acknowledgements

We thank J. Greenan, J. Willis, V. Lutz-Collins, M. Wadowski, C. Pater, and D. McNeill for their help in the laboratory and/or the field. We also thank two anonymous reviewers whose comments improved the quality of an early version of the manuscript. This study was funded by a NSERC Discovery grant and ORD-UPEI grants to PAQ, and a NSERC USRA to AM.