Predators suppress prey populations and elicit defensive phenotypes in prey. The magnitude of predator effects depends upon several factors, including the density of predators, and their cue concentrations, in the environment. Predator density manipulations have often relied on laboratory studies that use unrealistic densities of predators and prey over unnatural temporal and spatial scales. Field studies can provide insights into predator-prey interactions under more realistic scenarios. However, field studies linking predator density and prey populations are limited by the challenge of manipulating predator densities or predicting predator densities in dynamic/stochastic environments. We exploited a somewhat predictable rise in predatory crab communities associated with ocean warming to evaluate the impacts of swimming crab density on ecologically important horn snails. Our approach combined long-term monitoring of crabs and snails with snail behavior surveys and snail tethering experiments repeated during and after a marine heat wave. Near the end of a marine heat wave in 2016, swimming crabs were found in 25% of marsh tidal creeks. No swimming crabs were found in these tidal creeks during cooler water conditions in 2012 and 2021. When swimming crabs were more abundant (e.g. 2016), horn snails experienced 7,533% more mortality and were 91% less abundant. The proportion of snails climbing vegetation was 454% higher when swimming crabs were more abundant. Thus, higher predator densities occurring during a marine heat wave were associated with changes in snail abundance, mortality, and behavior. Such changes could influence marsh food webs and nutrient cycling. Our findings highlight the value of exploiting climatic anomalies to understand ecological patterns linked to the top-down effects in ecosystems - a critical need in predator-prey ecology.

Study system and site

California horn snails (hereafter, horn snails) are ubiquitous in tidal marshes and mudflats across the Southern California Bight (Lorda & Lafferty 2012). Throughout this region, horn snails are prey for several predators (Armitage & Fong 2006; Lorda et al. 2016)and are intermediate hosts for parasitic trematodes (Hechinger & Lafferty 2005; Lafferty et al. 2006b). Horn snails can be predated upon by predatory swimming crabs during high tides, when tidal inundation facilitates crab movements into mud flat and marsh habitats (Cote et al. 2001; Belgrad & Smith 2014). In southern California, the abundance of these swimming crabs (e.g., the Xantus’ swimming crab, Portunus xantusii, and the arched swimming crab, Callinectes arcuatus; hereafter swimming crabs) is connected to sea water temperature. Warm water events (e.g., marine heat waves, El Niños) are associated with influxes of swimming crabs (Zedler et al. 1992; Williams et al. 2001).

Consistent with observations during warm water events in 1982-83 and 1997-98 (Zedler et al. 1992; Williams et al. 2001), the San Onofre Nuclear Generating Station (SONGS) Mitigation Monitoring Program found elevated abundances of swimming crabs in San Dieguito Lagoon (hereafter SDL) during the 2014-2016 Pacific marine heat wave. SDL is a restored 123 acre brackish tidal marsh located in Del Mar, California (32° 58’ 40.4’’N, 117° 14’ 32.8’’W) comprised of marsh, mud flats, and subtidal habitats (e.g., tidal creeks). Tidal marsh is the dominant habitat in SDL and is defined by robust populations of cordgrass (Spartina foliosa) and pickleweed (Sarcocornia pacifica) that were started via transplants in 2008 (Walker et al. 2021; Beheshti et al. 2022). Mud flats and subtidal habitats account for 18 and 31 acres of SDL, respectively (Beheshti et al. 2022). Horn snails and swimming crabs co-occur on mud flats and marsh habitats, particularly at high tide.

Lab Assay-Consumption:

To assess the relative predation-risk of swimming crabs and Lined shore crabs on horn snails, we conducted a short feeding trial. We offered horn snails to individual adult swimming crabs [n = 5; biomass:  127 ± 19 g (mean ± 1SE)] and Lined shore crabs (n = 5; biomass:  15 ± 1 g) in plastic containers (30 x 25 x 25 cm, l x w x h) with flow through seawater. To standardize the number of snails offered to these different sized crabs, we added 95 and 10 horn snails to swimming crab and Lined shore crab treatments, respectively. The horn snails offered to crabs were 5– 28 mm in length (measured as the distance between the shell aperture and apex). This range is representative of the natural range of sizes observed at SDL. We systematically added snails to each replicate to make sure large and small snails were evenly distributed across swimming crab and Lined shore crab treatments. We assessed crab predation rates on horn snails 24 hours later.

Lab Assay- Crab Abundance:

In August 2016, we used laboratory mesocosms to understand if waterborne chemical cues from swimming crabs elicit defensive behaviors in horn snails. We exposed snails to one of three cues: No Crab, Swimming Crab, and Lined shore Crab (n = 7). We included cues from Lined shore crabs (Pachygrapsus crassipes) to compare to cues from swimming crabs because Lined shore crabs consume horn snails, have been reported to elicit defensive phenotypes in horn snails, and are abundant at SDL (Armitage & Fong 2006; Lorda et al. 2016; Walker et al. 2021). All invertebrates used in our laboratory assays were collected from SDL. We collected horn snails and Lined shore crabs by hand, and we collected swimming crabs with baited, vinyl-coated crab traps [65 × 49 × 25 cm, l x w x h; covered with mesh (3.8 cm openings)] with 2 trap door entrances (16 × 8 cm, w x h).

All replicates consisted of an upstream (9 x 16.5 x 9 cm, l x w x h) and downstream container (15 x 14cm, h x diameter) that received flow through seawater from San Diego Bay (32° 43’ 48.1’’N, 117° 12’ 51.0’’W). Upstream containers (1 L) drained into downstream containers (2.36 L) at a rate of ~1 L minute-1. Four drainage holes (2.5 cm diameter each) were drilled into the downstream containers 9.5 cm above the container floor. Each downstream container held five horn snails (≥ 10 mm length). To prevent snail escape, holes and the top of the containers were covered with plastic mesh (6 mm opening). We randomly assigned upstream containers to one of three treatments (No Crab, Swimming Crab, Lined shore crab). No Crab treatments were left empty. In Swimming Crab and Lined shore crab treatments, we included a single adult swimming crab or Lined shore crab, respectively [Swimming crab biomass: 138 ± 14 g individual-1 (mean ± 1SE); Lined shore crab biomass: 15 ± 1 g individual-1].  Crabs were not fed during the assay but were provided horn snails ad libitum prior to the study.

We exposed horn snails in the downstream buckets to waterborne cues from upstream containers for 48 h. Then, we quantified the proportion of snails found above the waterline (i.e., the proportion of snails with their entire foot above the waterline). We monitored the proportion of snails climbing above the waterline because previous studies suggested that horn snails may seek refuge by crawling out of water containing predators and their associated cues (Belgard and Smith 2014).

Lab Assay- Crab Biomass:

Because adult swimming crabs are considerably larger than Lined shore crabs [138 ± 14 g individual-1 and 15 ± 1 g individual-1 (mean ± 1SE); respectively], we conducted a second behavioral assay in September 2016 to test if horn snails respond to Lined shore crab cues when crab biomass is increased. We included two treatments: No Crab (n = 7) and Lined Shore Crab (n = 6).

All replicates consisted of an upstream and downstream container and were set up as described in the Lab Assay-Abundance Methods section. However, in Lined Shore Crab treatments, instead of adding a single crab, we added 113 ± 5 g of Lined shore crabs, equal to 12 ± 1 individual Lined shore crabs. We targeted this biomass because we observed strong effects of individual swimming crabs in our previous assays and their biomass was similar (105-194 g, n = 9).

We exposed horn snails in the downstream containers to waterborne cues from upstream containers for 48 hours. After 48 hours, we quantified 1) if each snail was climbing and 2) if the snail had climbed above the waterline (i.e., climbed > 9.5 cm).

Temporal monitoring data:

            Horn snail and swimming crab relative abundance data at SDL were obtained from the SONGS Mitigation Monitoring Program. This monitoring effort began in 2012 and surveys sites on an annual basis in the late summer or early fall (July-November; Beheshti et al. 2022). Each year, six tidal creeks were monitored except in 2016 when eight tidal creeks were monitored. Within each tidal creek, snail and crab presence were monitored at five sampling stations spaced 10-20 m apart. At each sampling station within each tidal creek, horn snail presence was evaluated along two transects placed perpendicular to the water mark. Along each transect, three quadrats (0.06 m2) were placed between the lower limit of vegetation (~ 1.3 ft NGVD) and the thalweg of the tidal creek (~50-60 cm water depth). This resulted in a total of six quadrats per station, and 30 quadrats per tidal creek. Swimming crab presence (presence/absence) was assessed at all five sampling stations in each of the six tidal creeks using a circular enclosed trap (0.9 m tall x 0.74 m diameter; 0.43 m2 area) and a beach seine (2 m tall x 7.6 m wide, 3.2 mm mesh) covering a ~46 m 2 area (see (Steele et al. 2006b, a). The proportional abundance of horn snails and swimming crabs was then calculated by dividing the number of tidal creeks where each species was observed in any quadrat (snails) or any trap/seine (crabs) by the total number of tidal creeks sampled (n = 6 or n = 8 in 2016).

Mortality study:

In September 2016 and 2021, we evaluated horn snail predation-risk at three tidal elevations of a common site (32° 58’ 39.8’’N, 117° 14’ 40.9’’W) using a tethering experiment. Specifically, we deployed 30 - 0.5 m2 PVC frames containing 25 tethered horn snails each. Each PVC frame included five lengths of braided fishing line (rated to 50lbs) running parallel across the frame at 10 cm intervals. Five horn snails were tethered to each of the five fishing lines attached to every PVC frame. Horn snails were individually tethered using a 5 cm length of braided fishing line (rated to 50lbs) attached to the apex of each snails’ shell with cyanoacrylate.

We deployed tethered snails at three horizontal distances landward of MLLW (1.5, 3.0, and 4.5 m). These horizontal distances were associated with an increase in elevation. At each distance, plots were either open to allow predators access to tethered snails or were protected by a plastic mesh cage (1 cm opening) to act as mortality controls (open plots, n = 5 both years; caged plots, n = 5 and n = 2 for 2016 and 2021, respectively). We deployed fewer caged plots in 2021 because no snails were lost from these in 2016 when predation was high in open plots. We surveyed the number of live tethered horn snails remaining in each plot after 48 h. Additionally, we counted the number of tethered horn snails that had buried into the soil in caged plots because previous studies suggested that horn snails bury in response to predatory crabs (Lorda et al. 2016).

Mortality Study- Crab Traps:

To characterize the subtidal predator community in tidal creeks adjacent to our tethering plots, we deployed two crab traps within one week of our tethering studies in both 2016 and 2021. Vinyl-coated crab traps [65 × 49 × 25 cm, l x w x h; covered with mesh (3.8 cm openings)] with 2 trap door entrances (16 × 8 cm, w x h) were placed in the center of the tidal creek adjacent to the location of our tethering site. Traps were placed 10 m apart and were baited with squid (0.23 kg each). After 48 h, we counted the number of trapped swimming crabs (P. xantussii and C. arcuatus).

Field Survey & Field Survey- Climbing Logit:

To understand how changes in swimming crab abundance may be linked to the abundance, distribution, and climbing behavior of horn snails in the field, we conducted transect surveys spanning the intertidal zone at a single site (32° 58’ 39.8’’N, 117° 14’ 40.9’’W) at SDL in 2016 and 2021. Specifically, we placed five - 10 m transects perpendicular to the waterline starting at MLLW and extending landward. Individual transects were spaced 2 m apart. Along each transect, we recorded the 1) density of non-climbing horn snails, 2) density of climbing horn snails, and 3) percent cover of plants (mainly pickleweed) every meter using a 0.5 x 0.5m quadrat. Horn snails were considered climbing if their entire foot was attached to vegetation, rather than the bare sediment. Since horn snails are known to bury in the presence of crab predators (Armitage & Fong 2006; Lorda et al. 2016), we searched the top 2 cm of soil in quadrats for buried horn snails.  We also evaluated the propensity of horn snails to migrate landward from the MLLW in 2016 and 2021 using a mark and recapture study (see Horn snail migration study in Appendix S2).

Migration Study:

We evaluated the propensity of horn snails to migrate out of “high-risk” habitats close to the MLLW using a mark-and-capture study. In September 2016 and 2021, we collected 100 horn snails from San Dieguito Lagoon (SDL). Snails were transported to the Coastal and Marine Institute Laboratory where their shells were dabbed dry and painted with red nail polish. The nail polish dried for at least 12 hours before snails were returned to SDL and released inside a 0.5 x 2.0 m plot at Mean Lower Low Water (MLLW). The corners of the plot were marked with PVC stakes. We surveyed snail distribution and behavior 48 hours later. We systematically searched the area around the plot (see Fig. S2 for search area schematic), raking our fingers through the top 1-2cm of soil to recover any snails that may have buried themselves. For each recovered snail, we estimated the minimum horizontal distance traveled. The minimum horizontal distance was defined as the distance between the snail’s foot and the closest 2.0 m side of the deployment plot.