Vol. 607: 1–18, 2018 MARINE ECOLOGY PROGRESS SERIES Published December 6 https://doi.org/10.3354/meps12793 Mar Ecol Prog Ser OPENPEN ACCESSCCESS FEATURE ARTICLE Individual and population level effects of ocean acidification on a predator−prey system with inducible defenses: bryozoan−nudibranch interactions in the Salish Sea Sasha K. Seroy1,2,*, Daniel Grünbaum1,2 1School of Oceanography, University of Washington, Seattle, Washington 98105, USA 2Friday Harbor Laboratories, University of Washington, Friday Harbor, Washington 98250, USA ABSTRACT: Ocean acidification (OA) from increased oceanic CO2 concentrations imposes significant phys- iological stresses on many calcifying organisms. OA effects on individual organisms may be synergistically amplified or reduced by inter- and intraspecies inter- actions as they propagate up to population and com- munity levels, altering predictions by studies of cal - cifier responses in isolation. The calcifying colonial bryo zoan Membranipora membranacea and the pre- datory nudibranch Corambe steinbergae comprise a trophic system strongly regulated by predator- induced defensive responses and space limitation, presenting a unique system to investigate OA effects on these regulatory mechanisms at individual and population levels. We experimentally quantified OA effects across a range of pH from 7.0 to 7.9 on growth, Predatory nudibranchs Corambe steinbergae (gelatinous, at the bottom) preying on zooids of the colonial bryo zoan Mem- calcification, senescence and predator-induced spine branipora membranacea. Zooids emptied by C. stein bergae, formation in Membranipora, with or without water- and C. steinbergae egg clutches, are visible in the upper part borne predator cue, and on zooid consumption rates of the photo. in Corambe at Friday Harbor Laboratories, San Photo: Sasha K. Seroy Juan Is land, WA. Membranipora exhibited maximum growth and calcification at moderately low pH (7.6), and continued spine formation in all pH treatments. due to competition. Our coupled experimental and Spines reduced Corambe zooid consumption rates, modeling results demonstrate the need to consider with lower pH weakening this effect. Using a spatially population-level processes when assessing ecological explicit model of colony growth, where colony area responses to stresses from changing environments. serves as a proxy for colony fitness, we assessed the population-level impacts of these experimentally KEY WORDS: Predator–prey interactions · Ocean determined individual-level effects in the context of acidification · Inducible defenses · Space competition · space limitation. The area-based fitness costs associ- Modeling · Membranipora membranacea · Corambe ated with defense measured at the individual level led steinbergae to amplified effects predicted for the population level © The authors 2018. Open Access under Creative Commons by *Corresponding author: [email protected] Attribution Licence. Use, distribution and reproduction are un - restricted. Authors and original publication must be credited. Publisher: Inter-Research · www.int-res.com 2 Mar Ecol Prog Ser 607: 1–18, 2018 INTRODUCTION these stresses occur within interconnected species networks that ultimately structure populations and Ocean acidification (OA) resulting from increased communities. Skeletal mineralogy alone is an incom- oceanic uptake of atmospheric carbon dioxide pres- plete predictor of OA sensitivity, especially in the con- ents a variety of environmental stresses to marine text of trophic interactions where direct and indirect orga nisms and ecosystems (Hofmann et al. 2010, effects associated with species interactions may be Kroeker et al. 2013). OA causes a decrease in both significant (Busch & McElhany 2017). Physiological pH and carbonate ion availability (Feely et al. 2009), and behavioral processes in both predators and prey often with negative physiological implications for may be impacted differently by OA. These alterations many organisms (Kroeker et al. 2013). Global pH lev- potentially amplify or reduce organism-level effects of els have been declining, with continued projections OA on population and community levels, highlighting of 0.3−0.4 units within the next century (Feely et al. the importance of considering predator− prey inter - 2004, 2009). Coastal areas, however, often experi- actions when conducting OA studies (Kroeker et al. ence natural pH fluctuations that currently exceed 2014). For example, OA may impair predator detection these predicted changes (Doney 2010, Hofmann et mechanisms, resulting in reduced predator avoidance al. 2010). In the Salish Sea and San Juan Archipel- behavior in fish (Dixson et al. 2010) and marine snails ago, WA, USA, naturally low seawater pH levels (Manríquez et al. 2013), but can reduce predation by result from regular upwelling and river input (Mur- crabs on oysters (Dodd et al. 2015). In these examples, ray et al. 2015). While open ocean pH averages well the differential effects of OA on distinct trophic levels above 8.0 (Doney 2010), the San Juan Archipelago indicate a potential for environmental stress to impact experiences pH around 7.8 and can reach as low as populations and communities to different extents 7.67 in San Juan Channel during seasonal up welling and in different directions than suggested by studies (Sullivan 2012, Murray et al. 2015). Consequently, of species in isolation. The potential of OA to alter the Salish Sea is a natural laboratory to investigate predator−prey interactions, and how these effects OA effects where organisms already experience pH may propagate to populations and communities, how- values that open ocean organisms will likely not ever, remains poorly understood. experience until the end of the century. Inducible defenses are an important subset of Organisms that make shells or skeletons from predator−prey interactions that confer protection to CaCO3 are common in the Salish Sea, with a third of organisms upon the detection of environmental cues Puget Sound species identified as calcifiers (Busch & of impending predation (Harvell 1990). When de- McElhany 2017). Calcifying organisms have been fenses incur significant costs, inducibility allows demonstrated to be generally vulnerable to acidifica- organisms to maximize the benefit of being defended tion (Kroeker et al. 2013). Decreasing availability of when under attack while avoiding the cost of defense 2− CO3 can impede calcification (Feely et al. 2009). In when predation risks are low (Tollrian & Harvell conjunction, the dissolution of calcified structures is 1999, Ferrari et al. 2010). Predator-induced defenses favored by the low-saturation states of CaCO3 poly- are common in marine calcifiers, e.g. shell thickening morphs, calcite and aragonite, that often accompany in oysters (Lord & Whitlatch 2012, Scherer et al. OA (Orr et al. 2005). Calcifier responses to OA have 2018), snails (Trussell & Nicklin 2002, Bourdeau been generally negative, with many organisms ex - 2010), and mussels (Leonard et al. 1999). OA has hibiting decreased growth, survival and calcification been found to affect inducible defenses in some rates (Kroeker et al. 2013). Some taxa, however, have species. Both in duced shell thickening in the snail displayed no effect or even positive effects, de - Littorina obtusata (Bibby et al. 2007) and predator- monstrating the importance of quantifying species- induced neck teeth in freshwater Daphnia spp. were specific responses in diverse taxa (Ries et al. 2009). reduced by high partial pressure of carbon dioxide Variable OA responses have been attributed to a (pCO2) (Weiss et al. 2018). While these studies have number of potential causes, including diversity of investigated the effects of acidification on inducible calcification mechanisms, presence and thickness of defenses, few have done so in the context of multiple protective organic membranes and solubility of dif- levels of biological organization. ferent CaCO3 polymorphs in addition to differences Inducible defenses can have effects beyond the in capacity for plasticity and prior exposure to low pH organism level, with potential propagating effects (Ries et al. 2009). through food webs and population dynamics of both While most studies document OA effects on calci- predators and prey (Miner et al. 2005), which sug- fiers in isolation, the true ecological consequences of gests that OA effects on these interactions must Seroy & Grünbaum: Ocean acidification effects on bryozoan−nudibranch interactions 3 then be assessed at population and community lev- tecting waterborne chemical cues from Corambe. els. Scaling organism-level interactions to estimate Spines are produced only on newly formed zooids at population-level responses can provide quantitative the growing margin of the colony, requiring the for- assessments of propagating effects of OA. In this mation of new calcified zooids (Harvell 1984). Spina- study, we used a well-known inducible defense sys- tion begins with formation of corner spines on zooid tem to develop a strategy for assessing OA effects vertices, followed by ad ditional spines along the in which we empirically measured individual-level walls of the zooids (Har vell 1984). Spines have been responses and used a numerical model to infer demonstrated to be effective in reducing predation population-level effects. (Harvell 1986). However, spines also have associated The encrusting bryozoan Membranipora membra - costs, including re duced colony growth as resources n acea (hereafter Membranipora) is an abundant colo-