Temporal and spatial refugia modify predation risk for non-native crabs in rocky intertidal habitats

Introduction

A prominent factor thought to facilitate the proliferation of non-native populations is enemy release (Heger & Jeschke, 2018), wherein non-native species benefit from a reduction in natural predators, competitors, and parasites in the naïve systems they invade (Colautti et al., 2004). While enemy release may confer an advantage to non-native species (Antonini et al., 2019; Roznik et al., 2020), invasion success is dependent on a myriad of ecological and environmental interactions within native communities (Weis, 2010; Prior et al., 2015).

Across taxa and environments, the availability of temporal and spatial refuge can act to constrain predation risk (Prugh & Golden, 2014; Palmer et al., 2022; Suraci et al., 2022). Habitat structure, the arrangement of biotic and abiotic substrate that supports plant and animal communities (Carvalho & Barros, 2017), can mediate interactions between predators and prey by providing spatial refuge in which prey can more easily avoid capture (Warfe & Barmuta, 2004; Lei, Lin & Zhang, 2014; Mendez, Schwindt & Bortolus, 2015; Pozzebon, Loeb & Duso, 2015). In addition to habitat-specific structural refuge, predation risk can change over time (Sperry et al., 2008). For example, risk and refuge for prey can exhibit diel variation (Clark, Ruiz & Hines, 2003). More to the point, the availability of refuge is a consequential determinant of predation risk (Smith et al., 2019), but rarely explored as a factor influencing enemy release (Soifer & Ackerman, 2019).

Evidence suggests a recent invader to the rocky shores of the North American Atlantic coast, the Asian shore crab, Hemigrapsus sanguineus, may have experienced enhanced success as a result of enemy release. The abundance of H. sanguineus in invaded habitats can far exceed densities found along the native Asian-Pacific region (Takahashi et al., 1985; Lohrer et al., 2000). Today, H. sanguineus is the most abundant intertidal crab species in southern New England and Long Island Sound (Kraemer et al., 2007; O’Connor, 2014) and can reach densities of >300 crabs m⁻² at some locations (O’Connor, 2018). The species’ success since its introduction to the northeast US in the 1980s is thought to be explained, in part, by reduced impact of natural enemies compared to the populations in native habitats (Pushchina & Panchenko, 2002; Brousseau et al., 2008).

The relationship between the non-native H. sanguineus and native predators, however, remains ambiguous in the absence of direct field experimentation and laboratory tests with adult crabs. In its native range, H. sanguineus is known to be consumed by two species of sculpins, Myoxocephalus stelleri and M. brandti (Pushchina & Panchenko, 2002). Species thought to prey on non-native H. sanguineus include those that utilize the rocky intertidal zone to forage, and species that are adapted to eating hard-shelled benthic invertebrate prey, including fish and likely bird species (Epifanio, 2013).

Of the potential predators of non-native H. sanguineus present throughout its range in northeast North America, only predation by fish on larvae and juveniles has been examined (Kim & O’Connor, 2007; Rasch & O’Connor, 2012). Little is known about predation on larger, sexually mature H. sanguineus. In one laboratory choice experiment, Tautog (Tautoga onitis) consumed juvenile H. sanguineus, but less often than native prey species (Savaria & O’Connor, 2013). Conversely, other laboratory experiments found T. onitis, cunner (Tautoga adspersus), and black sea bass (Centropristis striata) preferentially preyed upon H. sanguineus when given the choice with other local crab species as prey and that substrate influenced predator preference (Heinonen & Auster, 2012). Additionally, factors that modify predator–prey relationships, like spatial and temporal refuge, should be more fully explored to clarify the factors influencing enemy release in invaded systems.

The purpose of this study was to measure predation of sexually mature H. sanguineus in the field to examine temporal refuge from predation risk and use laboratory experiments to assess the influence of spatial refuge on predation. We hypothesized that relative predation of adult H. sanguineus is greatest during daytime high tides, when visual fish predators are active, and that predation risk by the fish predator T. onitis is mitigated by high levels of habitat refuge.

Materials & Methods

Laboratory experiment

Crab and fish collection

The tautog (T. onitis) is a temperate reef fish that plays an important role in the structure of nearshore marine communities as a specialized predator of hard-shelled benthic invertebrates, including crabs (Liem & Sanders, 1986; Clark et al., 2006). The 27 T. onitis (25.5–37 cm total length) used in the laboratory experiment were caught in New Bedford Harbor, MA using unbaited traps May 13–June 4, 2021. Fish were collected during annual trap surveys conducted by the Massachusetts Division of Marine Fisheries; traps were checked at least every three days. Fish were transported in a 50 L insulated cooler with fresh seawater and continuous aeration. Fish were held in groups of <10 for a two-week acclimation period ahead of experimentation. During acclimation, fish were fed crabs and cracked clams to satiation. Fish and crabs were housed, and experiments were conducted, in 1.8 m diameter tanks (tank floor area = 2.6 m²) continuously supplied with ambient seawater from Clarks Cove, New Bedford (water temp 20.5–24.5 °C, salinity 33–34, depth 1 m) and artificially lit to match natural light-dark cycles. Tanks were cleaned daily. Fish were provided pieces of large PVC pipe (10.2 cm diameter) for shelter during acclimation. Each fish was used for a single trial, then returned to the wild.

The 720 H. sanguineus (14–20 mm CW) used in these tests were collected from the rocky intertidal in Clark’s Cove 12–24 h prior to experimentation. Crabs were held in one tank in mesh-sided 0.5 L Tupperware containers in groups of <6 crabs. Crabs were not fed during this time. Only non-gravid crabs with all ten limbs were used in this study. Each crab was used for a single experiment.

Lab experiment treatment construction

Habitat structure provided during the lab experiment was constructed from concrete pavers (paving stones) (L = 40 cm, W = 20 cm, H = 5 cm). The experiment included a No Refuge control treatment without structure, as well as a control treatment without a fish predator. Low Refuge Habitat consisted of two pavers laid flat on the bottom of the tank (Fig. 1). High Refuge Habitat consisted of two pavers, modified to create sloping structure by the addition of twelve quartzite river stones (2–3 cm) glued to one long edge of the pavers using saltwater resistant Seachem cyanoacrylate Reef Glue™ . This created 1,984 cm³ of refuge space under each paver (Fig. 1). All materials were rinsed with fresh water and allowed >24 h to air dry before use. The glue was given >24 h to cure. Stones remained glued in place throughout the duration of the experiment.

Each tank was outfitted with a shelter for the fish made of three large PVC tubes (10.2 cm diameter), suspended eight cm above the tank floor in the center of the tank (Fig. 1). Fish utilized the inside and the outside of the PVC tubes as shelter. These PVC tubes were not accessible to the crabs. Fish were maintained and housed under University of Massachusetts Dartmouth Institutional Animal Care and Use Committee protocol # 21-02 approved July 19, 2021.

Lab experimental procedure

Four experimental tanks were randomly assigned habitat treatments for each trial. Nine trials were conducted for each habitat treatment (High Refuge, Low Refuge, No Refuge, and no fish control). One fish was used per trial and allowed 24 h to acclimate in the experimental tank prior to the experiment, during which time the fish was not fed. T. onitis require 8 h to process and evacuate food (Olla, Bejda & Martin, 1975). To begin the experiment, habitat treatments were lowered into the tanks (pavers placed >0.5 m away from each other and from tank walls) and 20 crabs were added to the tank. A dip net was used to quarantine the fish for 15 min to allow the crabs to acclimate to the experimental tank. Fish were then given 6 h to feed. Results of pilot work showed, when starved for 24 h, one T. onitis (35 cm total length) could consume between 20 and 30 H. sanguineus (10 mm–18 mm CW) in 6 h. All trials were conducted during daylight hours, approximately 0700–1300.

At the end of each trial, the surviving crabs were counted. Fish were returned to storage tanks and observed for an hour after experimentation, then transported in a cooler with sea water to Clark’s Cove. No fish showed signs of illness or injury, and all 27 fish used in experiments were released. Tanks were drained and cleaned following each trial.

Lab experimental analysis

The proportion of crabs eaten was calculated for all trials. To test whether habitat refuge treatments influenced the proportion of crabs eaten, all habitat treatments were compared using a one-way ANOVA. A Tukey post hoc test was then performed to determine any differences among levels of habitat complexity. Other independent variables including fish size, water temperature, and experimental tank (A, B, C, D) were tested using an expanded ANOVA model, none of which had significant influence on the proportion of crabs eaten, so were excluded from the final analysis. Data met the assumptions of normality and homogeneity of variance. Significance of factors was evaluated with type-III sums of squares using the R package car (Fox & Weisberg, 2019).

Results

Field tethering

One male and one female crab were found damaged (lost multiple limbs) after experiments, and although injuries were likely the result of predation attempts by a small predator, injured crabs were excluded from the analysis. Of the remaining 32 crabs, 17 were missing and presumed eaten, and 15 crabs were recovered unharmed. There was a significantly higher probability of predation with increased time of submersion (df = 30, X² = 30.997, p < 0.001), while the amount of daylight did not influence the probability of predation (df = 29, X² = 30.987, p = 0.92) (Fig. 2).

Laboratory experiment

In the control (no fish) treatment, all crabs survived without injury. Predation in laboratory feeding trials varied significantly with habitat complexity (df = 2, F = 35.99, p < 0.001) (Fig. 3). The proportion of crabs eaten was significantly lower in the presence of High Refuge Habitat compared to both Low Refuge Habitat (p < 0.001) and No Refuge (p < 0.001). There was no difference in predation between the Low Refuge Habitat and No Refuge treatments (p = 0.63) (Fig. 3). One fish consumed all 20 crabs available in the experiment (No Refuge, fish total length = 31 cm), and one fish consumed no crabs (High Refuge Habitat, fish total length = 35.5 cm).

Discussion

This study investigated temporal refuge by explicitly removing the spatial refuge, and spatial refuge by eliminating temporal refuge. Without spatial refuge predation of H. sanguineus was influenced by tide and not affected by time of day in field experiments. Crabs had a higher probability of being eaten when they were submerged in water. However, predation rates were similar during the day and night both at high tide and low tide. The tethering experiment in the field likely enhanced actual predation risk, because the tether apparatus restricted the crabs’ ability to flee predators, and the experiment was conducted in the absence of refuge for the crab. Nevertheless, results show predation of adult H. sanguineus can occur during high tide and low tide, and during daytime and nighttime, suggesting that diurnal as well as nocturnal marine and terrestrial species can prey on the non-native H. sanguineus.

The fish species T. onitis and T. adspersus were likely the primary fish species responsible for predation at high tide during this study. The native fishes are specialized for consuming hard-shelled invertebrates in the intertidal zone during high tide (Olla et al., 1974), and are known to feed on crabs, including H. sanguineus (Clark et al., 2006). T. onitis is a member of the Labridae family, named for their terminal protractile mouths and jaw morphology associated with hard-prey diets (Liem & Sanders, 1986), and feeds predominantly on mussels and other shellfish, including Atlantic rock crabs (Cancer irroratus), Jonah crabs (Cancer borealis) and small American lobsters (Homarus americanus) (Smith, 1907; Steimle & Ogren, 1982; Richards, 1992). At the study site, in >20 h of video footage of tethered crabs recorded during pilot work (July 18–August 23, 2020) there were three predation attempts observed during daytime high tides and all appeared to involve T. onitis and/or T. adspersus (R. Montanaro, pers. obs., 2020 ).

Higher probability of predation during high tide may, in part, be explained by what is known about the behavior of H. sanguineus. Both male and female H. sanguineus are highly mobile and can travel up to 16 m per day (Brousseau et al., 2002). Thought to be most active at high tide, laboratory experiments conducted during daytime found crabs were more likely to move under and out from under shelter when submerged in water (Towne, Judge & O’Connor, 2023). Activity at high tide may expose crabs to visual predators, like T. onitis and T. adspersus, which use a scan-and-pick foraging strategy to feed in the intertidal zone (Dew, 1976; Deacutis, 1982) where the crabs occur (Brousseau et al., 2002; Epifanio, 2013; Towne, Judge & O’Connor, 2023).

Given greater relative predation at high tide, periods of low tide appear to offer, to some extent, temporal refuge from predation. While at lower levels, predation did occur during field tethering experiments at low tide, demonstrating that species that are not strictly aquatic also pose a predation risk to adult H. sanguineus. Other invertebrate species could have been responsible for the predation observed in this study at high tide or low tide. European green crabs, Carcinus maenas, Atlantic rock crabs, C. irroratus, and blue crabs, Callinectes sapidus, co-occur with H. sanguineus on the US east coast (DeRivera et al., 2005).

Benthic species in rocky intertidal communities are commonly prey for birds (Edwards, Conover & Sutter, 1982; Wootton, 1992). Previous experiments that have excluded avian predators from rocky intertidal habitat found the absence of bird predation caused a significant increase in the density of intertidal crabs like Cancer borealis (Ellis et al., 2007). Predation of non-native H. sanguineus specifically by avian predators has not been well documented but merits additional investigation. Similarly, the relationship between H. sanguineus and coastal mammals has yet to be investigated. Further research should include field studies that specifically measure the impact of non-fish predators and the variation in predation risk and refuge throughout the tidal cycle, which could have an effect on the degree of enemy release experienced by non-native H. sanguineus.

In the field experiments conducted here, the probability of predation was similar during daytime and nighttime. This is counter to the hypothesis that predation risk would be constrained to the daytime period of the diel cycle, because of the propensity of high tide predators to restrict their activity to daytime. For example, T. onitis cease feeding at night (Dew, 1976; Deacutis, 1982).

Predation at nighttime may have been enhanced because of the proximity of the field experiment to a lighted dock. The dock could have encouraged aggregation of T. onitis and T. adspersus which are known to affiliate with structures like dock pilings (Olla, Bejda & Martin, 1975). Illumination from lights on the dock at night could improve the predators’ ability to see prey and forage throughout nighttime, particularly on a dark night, like under cloudy skies or during a new or crescent moon. Nevertheless, man-made structures like the dock described here are increasingly common in coastal habitats (Ruiz et al., 2000), and such human influence does affect predation efficiency and prey choice in nearshore ecosystems (Montalvo, 2020).

In the laboratory study, High Refuge Habitat significantly reduced the proportion of crabs eaten by T. onitis. In addition, there was no difference in the proportion of crabs eaten when provided Low Refuge Habitat and when provided No Refuge. H. sanguineus is most abundant in mid and lower intertidal zone where there is high structural complexity (Ledesma & O’Connor, 2001; Brousseau et al., 2002; Gilman & Grace, 2009; Epifanio, 2013). A limitation of this study is an under-representation of the true complexity of the rocky intertidal zone, which is more often composed of multiple layers of rocks and cobble that could provide more spatial refuge to further minimize the risk from predators (Lohrer et al., 2000). Nevertheless, evidence presented here demonstrates that T. onitis will readily consume adult (sexually mature) H. sanguineus, and that the availability of structural refuge modifies the risk of predation.

Given the high density of non-native H. sanguineus, relative to other intertidal crab species (O’Connor, 2014; O’Connor, 2018), H. sanguineus appears particularly adept at avoiding predation. The high abundance of H. sanguineus in many coastal communities throughout the invaded range suggests that predators do not strongly impact the population size of this species, perhaps, in part, because of the temporal and spatial refuge within rocky intertidal habitats.

This study shows that adult H. sanguineus can be eaten throughout the diurnal and tidal cycles in the rocky intertidal zone, and spatial refuge strongly modifies predation threat by a common fish predator, T. onitis. This study suggests non-native H. sanguineus predation risk can be modified by habitat refuge. Subsequent research should combine the factors examined here, spatial and temporal refuge, to directly test if predation pressure experienced by H. sanguineus populations is different among coastal habitats with varying structural complexity. Healthy populations of fish could lead to a decrease of H. sanguineus abundance.

Supplemental Information