Dataset: Oyster survival differences in mesocosms

Oyster culturing
Oysters (Crassostrea virginica) were cultured as spat-on-shell at the Auburn University Shellfish Laboratory (AUSL) on Dauphin Island, AL starting in late May 2019 using standard techniques (Congrove et al. 2009). Oyster larvae were settled onto sun-bleached oyster shells to create spat-on-shell. After 3 days, when oyster spat were approximately 1.0 millimeters, they were exposed to either exudate from predatory blue crabs or empty cage controls in four flow-through holding tanks (length = 2.4 meters, width = 0.9 meters, water depth = 0.4 meters) supplied with unfiltered seawater pumped directly from the Gulf of Mexico. The number of spat per shell varied from approximately 5 – 40 and we elected to not alter the initial density to mimic natural settlement during the induction period. Oysters were suspended above the tank bottom in oyster aquaculture baskets (64 x 23 x 14 centimeters with 140 spat-covered per shell basket) to prevent sediment buildup from suffocating oysters. Seven oyster baskets were present in each tank (28 total).

Spat were exposed to blue crab predator cues by holding four live caged adult blue crabs (Callinectes sapidus) in two of the tanks (8 crabs total), whereas the remaining two tanks contained empty cages (control) to mimic conditions where oysters regularly experience predator cues or are limited in their exposure from cues. Water volumes and crab densities were informed from established procedures (Belgrad et al. 2021). Crabs in each tank were held in four separate cages (32 x 23 x 14 centimeters) to prevent crabs from consuming the experimental oysters or each other. Every crab was fed one adult oyster daily (approximately 5.0 centimeters in length) to maximize predator cue intensity as experimental oysters would be exposed to exudates from predators and damaged conspecifics. This ensured that oysters were exposed to the most natural set of cues indicative of a predation event, which produces a strong response in oysters (Scherer et al. 2016). Crabs were replaced during the experiment as needed due to mortality. Experimental oyster baskets were rotated around the crab cages daily to reduce differences in oyster growth due to proximity to predator cues, and no differences among cages were found. The induction period was 2 months.

Oyster survival experiment
After the 2-month spat grow-out period, we conducted a mesocosm experiment to determine how induced predator defenses (i.e., changes in shell strength) altered oyster survival under different ecological contexts of habitat shelter (within reef shelter vs outside reef) and predator regime (apex predator present vs absent). Spat-covered shells from the above oyster culturing were scraped so that each shell contained either four induced oysters grown with predator cues or four control oysters grown without predator cues to standardize individual predation risk. Eight shells from each of these two treatments were placed into six circular flow-through seawater tanks (diameter = 1.15 meters, water height = 40 centimeters; 16 shells per tank; 64 spat per tank).

An artificial reef was located in the center of each tank, and four shells containing spat of each type were placed in each of the “reefs” that provided a refuge for oyster consumers (mud crabs, Panopeus herbstii) and an additional four shells containing oysters of each type were placed roughly 15 centimeters (cm) from the tank wall. Artificial reef shelters were composed of a plastic basket (30 cm length x 20 cm width x 11 cm height) turned upside down and covered in sun-bleached oyster shells that were epoxied to envelop the outer edges of the basket. This roughly mimics the effect of a robust healthy reef with a 3-dimensional structure where mud crabs can occupy interstitial spaces difficult for blue crabs to access. Each tank also contained seven mature mud crabs (mean ± SD carapace width = 2.52 ± 0.43 cm) to serve as intermediate consumers with these densities being consistent with field measurements (Hill and Weissburg 2013). Every mud crab cohort contained at least two individuals of each sex to match natural fine-scale sex ratios.

Apex predators (blue crabs) were added to half of the tanks whereas the remaining three tanks lacked blue crabs. Tanks with the predator treatment contained a single adult blue crab (mean ± SD carapace width = 14.8 ± 1.4 centimeters), which at this size commonly feeds on mud crabs but rarely feeds upon oyster spat (Hines 2007, personal observations of all authors). As this experiment was focused on identifying how predator cues cause cascading effects through food webs, all blue crabs had their claws taped closed throughout the duration of the experiment so that blue crabs could provide chemical, visual, and mechanical predation risk cues without actually consuming the mud crabs. Blue crabs were fed a diet consisting of a single mud crab every day for a week prior to the start of the experiment to help ensure that blue crabs would produce urine containing metabolites mud crabs perceive as risk cues. Diet was standardized since cue perceptibility is affected by the amount of prey biomass consumed by a predator. Every blue crab was replaced with a new, recently fed blue crab each day to ensure that the apex predator would continue to release chemical cues. Preliminary experiments found that blue crabs did not consume spat under these conditions.

The experiment began by allowing the blue crabs and oysters to acclimate in the tank for 30 minutes. Mud crabs were released in the tank center near the artificial reef after the acclimation period whereupon they immediately began traveling in all directions. Oyster survival was recorded every 24 hours for three days. This experiment commenced on July 30th, 2019, and was repeated two additional times within that same week (9 replicate tanks distributed across 3 blocks). No individuals were used more than once (n = 1,152 spat, n = 126 mud crabs, n = 27 blue crabs total).

See the Supplemental file for crab survival data, which is a subset of the mesocosm data.