Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Multiple abiotic changes and species interactions mediate responses to climate change on rocky shores Gooding, Rebecca Ann 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2014_spring_gooding_rebecca.pdf [ 4.28MB ]
JSON: 24-1.0165672.json
JSON-LD: 24-1.0165672-ld.json
RDF/XML (Pretty): 24-1.0165672-rdf.xml
RDF/JSON: 24-1.0165672-rdf.json
Turtle: 24-1.0165672-turtle.txt
N-Triples: 24-1.0165672-rdf-ntriples.txt
Original Record: 24-1.0165672-source.json
Full Text

Full Text

Multiple Abiotic Changes and SpeciesInteractions Mediate Responses toClimate Change on Rocky ShoresbyRebecca Ann GoodingA THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate and Postdoctoral Studies(Zoology)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)November 2013c Rebecca Ann Gooding 2013AbstractAnthropogenic climate change poses a serious threat to biodiversity. Accurate pre-dictions of the ecological consequences of future abiotic change will require a broadperspective that takes into account multiple climate variables, species-specific re-sponses, and intra- and interspecific dynamics. I addressed these issues in the con-text of a marine rocky intertidal community to determine how abiotic and bioticfactors can mediate the e?ects of climate change. I began with two studies on theorganismal-level e?ects of multiple abiotic variables. In the first study, I foundthat acute exposure to low salinity reduced the survival of littorine snails facingthermal stress, but that ocean acidification (OA) had no such e?ect. In a secondstudy, I showed that sustained exposure to increased temperature and OA had pos-itive and additive e?ects on the growth and feeding of the purple ochre sea star.These findings demonstrate that studies of multiple climate variables will be impor-tant not only to identify additive and non-additive e?ects, but also to determinewhich climate variables will be detrimental for a given species. Next, I measuredhow species-specific responses to climate change can alter species interactions. Byquantifying the e?ects of body size on the feeding behaviours of sea stars preying onmussels, I demonstrated that climate-driven changes in body size can have profoundimpacts on the strength of this interaction. Finally, I investigated how population-level responses to multiple abiotic variables can be a?ected by the presence of aninteracting species. I built a predator-prey model that simulates the ecologically im-iiportant interaction between the purple ochre sea star and its preferred prey, mussels.Using empirical estimates of sea star and mussel responses to increased temperatureand OA, I simulated their interaction under various climate scenarios. I found thatpredation exacerbated the e?ects of climate change on mussel populations, and thatclimate change increased the strength of the sea star-mussel interaction. My workdemonstrates that the e?ects of climate change will likely be mediated by a com-bination of biotic and abiotic factors, and that these factors should be consideredwhen making predictions about the ecological consequences of climate change.iiiPrefaceChapter 2 was conceived of and designed by myself. I collected and analyzed thedata, and wrote the manuscript. Andrea Au contributed some of the salinity andthermal tolerance experiments. Chris Harley edited the manuscript.Chapter 3 was conceived of by myself in collaboration with Chris Harley. I de-signed the experiments, collected and analyzed the data, and wrote the manuscript.Emily Tang assisted in conducting the experiments. Chris Harley contributed towriting the manuscript. A modified version of this chapter has been published:Gooding, R. A., Harley, C. D. G., & Tang, E. (2009). Elevated watertemperature and carbon dioxide concentration increase the growth of akeystone echinoderm. Proceedings of the National Academy of Sciences,106(23), 9316 - 9321.I conceived of and designed the study in Chapter 4, collected and analyzed thedata, and wrote the manuscript. Chris Harley edited the manuscript. A modifiedversion of this chapter has been accepted for publication in the journal BiologicalBulletin.Chapter 5 was conceived of and designed by myself in collaboration with TravisIngram. I created the conceptual model, provided the model parameters, ran thesimulations, analyzed the results, and wrote the manuscript. Travis Ingram helpeddevelop the model, generated the computer code for the simulations and assistedin writing the manuscript. Chris Harley contributed to interpreting the results andedited the manuscript.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Study system: rocky shores . . . . . . . . . . . . . . . . . . . . . . . 31.1.1 Climate change on rocky shores . . . . . . . . . . . . . . . . 61.1.2 Study species . . . . . . . . . . . . . . . . . . . . . . . . . . 91.2 Thesis objectives and overview . . . . . . . . . . . . . . . . . . . . . 112 Littorine Snail Responses to Multiple Abiotic Variables . . . . . 142.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18v2.3.1 Littorina spp. from Burrard Inlet . . . . . . . . . . . . . . . 182.3.2 Littorina spp. from Padilla Bay . . . . . . . . . . . . . . . . 242.3.3 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 252.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.4.1 Burrard Inlet snails . . . . . . . . . . . . . . . . . . . . . . . 262.4.2 Padilla Bay snails . . . . . . . . . . . . . . . . . . . . . . . . 282.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292.5.1 Thermal and salinity tolerances . . . . . . . . . . . . . . . . 292.5.2 Thermal tolerance with salinity and pH . . . . . . . . . . . . 312.5.3 Thermal coping behaviours with salinity and pH . . . . . . . 322.5.4 Broader implications and conclusions . . . . . . . . . . . . . 343 Sea Star Responses to Temperature and Ocean Acidification . . 443.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483.3.1 Study species and collection site . . . . . . . . . . . . . . . . 483.3.2 Sea star responses to temperature . . . . . . . . . . . . . . . 493.3.3 Sea star responses to temperature and OA . . . . . . . . . . 503.3.4 Statistical analyses . . . . . . . . . . . . . . . . . . . . . . . 513.4 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 524 Sea Star Feeding Behaviours with Predator and Prey Size . . . 654.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664.3 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . 694.3.1 Animal collection and morphometrics . . . . . . . . . . . . . 69vi4.3.2 Adult and large juvenile feeding behaviours . . . . . . . . . . 714.3.3 Juvenile feeding behaviours . . . . . . . . . . . . . . . . . . . 724.3.4 Statistical analyses . . . . . . . . . . . . . . . . . . . . . . . 744.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754.4.1 Adult feeding behaviours . . . . . . . . . . . . . . . . . . . . 754.4.2 Juvenile feeding behaviours . . . . . . . . . . . . . . . . . . . 764.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775 Modelling a Predator-Prey Interaction Under Climate Change . 855.1 Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865.3 Study system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895.3.1 Sea star and mussel responses to climate change . . . . . . . 905.4 Model presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 915.4.1 Mussel population dynamics . . . . . . . . . . . . . . . . . . 925.4.2 Sea star population dynamics . . . . . . . . . . . . . . . . . 945.4.3 Prey selection and consumption . . . . . . . . . . . . . . . . 955.4.4 Sea star growth . . . . . . . . . . . . . . . . . . . . . . . . . 975.4.5 Initial conditions . . . . . . . . . . . . . . . . . . . . . . . . . 995.4.6 E?ects of ocean warming and acidification . . . . . . . . . . 995.4.7 Model runs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1015.5 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 1025.5.1 Predator-prey dynamics with climate change . . . . . . . . . 1025.5.2 Mussel population dynamics with climate change . . . . . . 1055.5.3 Relative e?ects of single and combined climate change vari-ables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075.5.4 Conclusions and broader impacts . . . . . . . . . . . . . . . 108vii6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266.1 Direct e?ects of multiple climate variables on individual organisms . 1276.2 E?ects of climate change on species interactions . . . . . . . . . . . 1296.3 Direct and indirect e?ects of climate change via intra- and interspe-cific dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1316.4 Conclusions and recommendations . . . . . . . . . . . . . . . . . . . 134Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138viiiList of Tables2.1 Seawater chemistry parameters from Experiments 3, 4 and 6 . . . . 422.2 Upper lethal temperatures for Littorina spp. . . . . . . . . . . . . . 433.1 Review of growth responses to ocean acidification . . . . . . . . . . . 634.1 Juvenile sea star behaviours and energetics with prey size . . . . . . 845.1 Mussel responses to climate change . . . . . . . . . . . . . . . . . . . 1175.2 Sea star responses to climate change . . . . . . . . . . . . . . . . . . 1195.3 Model parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.4 Climate scenario scaling factors . . . . . . . . . . . . . . . . . . . . . 1245.5 Sea star body sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125ixList of Figures2.1 Tile temperatures during thermal tolerance experiment . . . . . . . . 352.2 Thermal tolerance of Littorina plena . . . . . . . . . . . . . . . . . . 362.3 Snail activity levels at di?erent salinities . . . . . . . . . . . . . . . . 372.4 E?ect of salinity and pH on snail activity levels . . . . . . . . . . . . 382.5 E?ects of salinity and pH on snail responses to thermal stress . . . . 392.6 E?ect of salinity on snail crawl-out activity . . . . . . . . . . . . . . 402.7 E?ects of salinity and pH on snail responses to thermal stress . . . . 413.1 Sea star growth and feeding rates temperature . . . . . . . . . . . . 593.2 Sea star growth with temperature and ocean acidification . . . . . . 603.3 Number of mussels consumed daily per sea star . . . . . . . . . . . . 613.4 Proportion of sea star wet mass consisting of calcified material . . . 624.1 Feeding behaviours with sea star and mussel size . . . . . . . . . . . 824.2 Juvenile sea star feeding and energetics with prey size . . . . . . . . 835.1 Schematic of the dynamics of the predator-prey simulation . . . . . . 1125.2 Mussel population density over time . . . . . . . . . . . . . . . . . . 1135.3 Mussel population size-frequency distributions . . . . . . . . . . . . . 1145.4 Interaction strength under di?erent climate scenarios . . . . . . . . . 1155.5 Mussel population responses to climate and predation . . . . . . . . 116xAcknowledgementsAlthough my name stands alone on the front page of this thesis, it was by no meansa solo endeavour.My work was partially funded by the National Science Foundation, the Univer-sity of British Columbia, the Padilla Bay National Estuarine Research Reserve, thePacific Northwest Shell Club, and the National Science and Research Council.Many people helped me run my experiments: Nicola Smith, Manon Picard,Melanie Overhill, Karen Lee, Sarah Neinhuis, Garth Covernton, Theraesa Coyle,Jean Rangyn Lim, Nicole Lee, Dan Budgell, Max Maliska, Gerald Singh, On LeeLau, Rebecca Guenther, Jocelyn Nelson, Arleigh Lambert, Mandy Wong, JanetMaclean, and almost certainly many more who have slipped my mind. They drilled1000?s of holes, hunted for sea stars at midnight in the freezing rain, hauled powertools into the intertidal, and braved sunburns and rogue waves, all in the name offriendship and science. I couldn?t have done it without their help.Doug Bulthuis and the rest of the sta? at the Padilla Bay National EstuarineResearch Reserve provided me with access and transportation to field and collectionsites.Huge thanks go to Travis Ingram, who provided intellectual conversation overbeer, saved me from the depths of R, listened to my many frustrated rants, helpedin the lab, collaborated with me, and has been a great friend all around.My current and former lab-mates in the Harley lab provided fruitful discussionxiand feedback, as well as lots of laughter during long days in the lab. Katie Lotter-hos and Andrew MacDonald provided statistical advice; Stillianos Louca and KyleDemes introduced me to the joys of LaTex; and Michael O?Donnell, Karen Chan,and many others provided intellectual help and feedback.Many thanks to my committee members - Mary O?Connor, Colin Brauner andIsabelle Co?te? - for feedback and guidance, particularly during the writing phase.My thought-processes and my work benefited greatly from each of their uniqueperspectives. An especially big thanks goes to my supervisor, Chris Harley, whowas instrumental in my development as an academic and a researcher. He providedguidance while respecting my stubborn independent streak, and rescued me manytimes from the depths of analysis-paralysis.Thanks to the Zoology sta?, who often go overlooked despite the critical rolethey play in the success and completion of any thesis. Andy LeBlanc and AlistairBlachford rescued my computer and my sanity several times, while Alice Liou en-sured I didn?t forget important paperwork and deadlines along the way. Bruce andVince in the shop helped me to MacGyver some pretty awesome science equipment.I also want to thank my undergraduate research supervisor, Mike Behrens. Al-though he did not directly contribute to my thesis, his encouragement, support andbelief in my abilities gave me the confidence to pursue a PhD.Thanks to my mom for proofreading my entire thesis in one night.And finally, thank you to Alex Graf, my amazing, wonderful, generous husband,who patiently tolerated my long work hours and dead-snail conversations at thedinner table, and who ensured I always had a steady supply of co?ee, chocolate andsupport to keep me going through the final stages of writing.xiiDedicationI dedicate this thesis to my brother, Joshua.His strength of spirit in the face of enormous challenges inspired and motivated me.I will never forget his gentleness, genius, and love of learning.xiiiChapter 1IntroductionAnthropogenic climate change poses one of the most serious threats to global biodi-versity (Halpern et al., 2008). Numerous biological changes, including species rangeshifts, ecosystem collapses, and local extinctions, have already been attributed to cli-mate change (Gosling et al., 2011; Parmesan, 2006; Walther, 2010). Future changesare predicted to be even greater, with an alarming number of species and ecosystemspredicted to disappear this century (Bellard et al., 2012; Thomas et al., 2004).The continuing accumulation of atmospheric carbon dioxide (CO2) (Solomonet al., 2007) is driving changes to multiple physical and chemical variables, oftenat rates far greater than ever seen before (Loarie et al., 2009). Global warming,ocean acidification, sea level rise, altered precipitation patterns, increased frequencyand severity of extreme weather events, and greater climate variability are amongthe many changes that are predicted or have already been seen (Hsieh et al., 2005;Solomon et al., 2007).Biological responses to simultaneous changes in multiple abiotic variables canbe complex and sometimes unexpected (Darling & Co?te?, 2008; Paine et al., 1998).In simple additive cases, simultaneous changes to two or more environmental vari-ables produce e?ects that can be well predicted from the e?ects of each variable inisolation. Responses can also be more complex, where the e?ect of one variable ismediated by changes to additional variables. These interactive e?ects can be eithersynergistic or antagonistic, where the combined e?ect of two or more variables is1greater than or less than, respectively, the sum of their individual e?ects (Crainet al., 2008; Darling & Co?te?, 2008). The potential for these combined e?ects todi?er drastically from those predicted by additive models of change highlights theimportance of empirically testing multiple climate variables together.The diversity of biological responses adds yet another layer of complexity to at-tempts to predict the consequences of climate change. Several recent reviews havehighlighted the presence of species-specific responses to various climate variables(Deutsch et al., 2008; Kroeker et al., 2010; Twomey et al., 2012). Intraspecific vari-ation can also occur, with responses to climate change sometimes di?ering betweenpopulations or even life history stages of a single species (Angilletta, 2001; Byrne,2011; Hamer, 2010). Attempts to make general predictions regarding responses toclimate change based on phylogeny, life history stage, or body size have been metwith mixed results, and the presence of numerous exceptions make clear predictionsdicult.The large variation in species? responses to climate change means that speciesinteractions will almost certainly be altered (Gilman et al., 2010; Lavergne et al.,2010; Van der Putten et al., 2010). If responses of two (or more) interacting speciesdi?er in magnitude and/or direction, the strength of their interaction is likely tochange, whereby the e?ect one species has on another is mediated by climate condi-tions (e.g. Sanford, 1999). On the other hand, a species? response to climate changemay be mediated by the presence of an interacting species, which can exacerbateor mitigate the direct e?ects of climate (Kordas et al., 2011; Louthan et al., 2013;O?Connor et al., 2011; Singer et al., 2012). Changes to species interactions oftenhave cascading e?ects on populations and communities, and could ultimately leadto ecosystem-wide consequences (Frelich et al., 2012; Singer et al., 2012).Accurate predictions of the e?ects of future climate change will require a broad2perspective that takes into consideration possible interactions between multipleclimate variables, species-specific responses, and intra- and interspecific dynam-ics (Doney et al., 2012; Van der Putten et al., 2010). However, manipulations ofmultiple climate variables usually require closely controlled mesocosms that limitthe degree to which population and community dynamics can be studied, whereasspace-for-time experiments or long-term observational data sets which allow intactcommunities to be studied are usually limited to a single abiotic variable. There is anurgent need for studies that combine the direct e?ects of multiple climate variableson individual organisms with indirect e?ects via species interactions and populationdynamics. In this dissertation, I address this issue using a combination of empiricaland modelling approaches, with a focus on several important invertebrate speciesfrom the rocky intertidal shores of British Columbia, Canada.1.1 Study system: marine intertidal rocky shoresRocky intertidal shores are one of the most physically stressful and dynamic envi-ronments on earth (Hawkins et al., 2008; Tomanek & Helmuth, 2002). Life in theintertidal is dominated by the constant ebb and flow of the tides, and the residentorganisms must cope with both marine and terrestrial conditions. Physical param-eters such as temperature, salinity, pH and pCO2 can vary widely within the spanof a few hours or days, reaching extremes that put many organisms at the edgeof their physiological tolerances (Tomanek & Helmuth, 2002; Truchot & Duhamel-Jouve, 1980). The ability of organisms to survive and even thrive under such severeconditions has been the topic of many decades of research (Hawkins et al., 2008;Thompson et al., 2002; Underwood, 2000).The ability of intertidal animals to withstand thermal and desiccation stress isparticularly notable. Radiative heating during low tide can drive the temperature of3rocky substrates as high as 45 - 50 C during hot summer days (Firth & Williams,2009; Harley & Helmuth, 2003). Many intertidal species have evolved physiologi-cal mechanisms, such as heat shock protein production and metabolic depression,to survive extreme temperatures (Sokolova & Portner, 2001; Tomanek, 2008). Or-ganisms also reduce their body temperature through behavioural changes such asmicrohabitat choice (Fly et al., 2012; Garrity, 1984), body orientation (Miller &Denny, 2011; Mun?oz et al., 2005), and changes to internal water volume (Pince-bourde et al., 2009). Nonetheless, severe aerial heat waves can cause mass mortalityevents (Harley, 2008), and thermal and desiccation stress often set the upper limitof species distributions on rocky shores (Benson, 2002).Most rocky shore organisms are also tolerant of short-term exposure to a widerange of salinity and pH conditions. Precipitation and terrestrial freshwater inputcan cause large salinity changes (Firth & Williams, 2009; Sutherland et al., 2011),with some areas experiencing short periods of almost freshwater conditions (Held& Harley, 2009; Sokolova et al., 2000a). Many intertidal crustaceans are able toregulate their internal volume and osmotic concentration (Findley et al., 1978);some crab and barnacle species can tolerate salinities as low as 5 psu for severaldays or even weeks (Bergen, 1968; Urbina et al., 2010), although the energeticdemands of living in these conditions for a prolonged period can be detrimentalto other physiological processes (Bergen, 1968). In contrast, gastropods and mostbivalves must rely primarily on behavioural mechanisms to isolate themselves fromacute salinity changes (Berger & Kharazova, 1997; Berger et al., 1978; Garrity,1984; Sokolova et al., 2000a), and often exhibit reduced activity and metabolic rates(Sokolova et al., 2000a; Todd, 1964). Nonetheless, some gastropods from estuarineenvironments can survive several weeks at salinities as low as 10 psu (Todd, 1964)or even freshwater conditions for several days (Sokolova et al., 2000a).4Diurnal changes in biological activity, as well as seasonal and storm-driven up-welling of CO2-rich water, can cause the pH to vary by as much as 4.0 units intidepools and 1.0 units in coastal waters (Feely et al., 2008; Marliave et al., 2011;Truchot & Duhamel-Jouve, 1980). Furthermore, organisms exposed for prolongedperiods during low tide can experience internal pH levels as low as 7.0 due to thebuild up of CO2 and anaerobic products in their tissues (Sokolova et al., 2000b;Truchot, 1990). Crustaceans are particularly good at regulating their internal acid-base balance (Spicer et al., 2006; Truchot, 1990), but most molluscs also have somedegree of regulatory capabilities (Michaelidis et al., 2005; Rumsey, 1973; Truchot,1990). However, acid-base regulation is energetically demanding and usually re-duces metabolic scope; during long-term exposure, growth and reproduction mayalso su?er (Michaelidis et al., 2005; Po?rtner et al., 2004; Reipschlager & Portner,1996).Abiotic conditions such as thermal stress have traditionally been thought to setthe upper vertical limits of rocky shore species, while their lower limits are oftendetermined by species interactions such as predation or competition (Benson, 2002;Connell, 1961). However, biotic factors are increasingly being shown to interactwith abiotic stressors; for example, mutualisms and commensalisms can extend up-per distributional limits through ecosystem engineering and amelioration of stressfulconditions (Bertness et al., 1999; Moore et al., 2007). Given their importance tothe functioning and survival of many organisms in the intertidal zone, species inter-actions are likely to play a key role in determining the outcomes of environmentalchange on rocky shores.51.1.1 Climate change on rocky shoresThe wealth of knowledge about the species, mechanisms, and patterns occurring onrocky shores make them ideal model systems for studying how ecological dynamicswill interact with changing environmental conditions in the future (Hawkins et al.,2008; Thompson et al., 2012). Furthermore, intertidal organisms may be particularlysusceptible to the e?ects of future climate change. Although intertidal species haveevolved to withstand harsh physical conditions, many are at or near the limits oftheir physiological tolerances (Helmuth et al., 2006b; Tomanek & Helmuth, 2002).Organisms from habitats with high ambient temperatures may have less room forfurther acclimation or adaptation (Somero, 2010; Stillman, 2003), and the additionof new stressors or exacerbation of existing ones due to climate change may pushorganisms beyond their tolerances (Thompson et al., 2002).Compared to many terrestrial habitats, rocky shores and other marine systemsare expected to experience especially large climate-driven abiotic changes (Harleyet al., 2006b; Wethey et al., 2011). The ocean acts as a sink that takes up largeamounts of CO2 and heat, resulting in profound physical and biological changes(Doney et al., 2012; Sabine, 2004). Although climate change will alter many abioticvariables in marine systems, rising water temperatures and ocean acidification arelikely to have some of the greatest biological impacts (Doney et al., 2012; Halpernet al., 2008).Increased sea surface temperature (SST) and its biological consequences are welldocumented. In the past 150 years, the global mean SST has risen approximately0.76 C and is predicted to rise an additional 1 - 4 C or more before stabilizing(Doney et al., 2012; Dufresne et al., 2013; Solomon et al., 2007). Because tem-perature influences nearly every biological process (Brown et al., 2004), increasedSST can have profound impacts on multiple levels of organization. Temperature-6dependent metabolic processes can alter organismal-level processes such as growth,feeding, and reproductive rates (Fearman & Moltschaniwskyj, 2010; Sanford, 2002;Vucic-Pestic et al., 2011). Suboptimal temperatures can lead to species range shifts(Southward et al., 1995), while extreme temperatures can cause catastrophic mortal-ity events (Baker et al., 2008; Harley, 2008). These individual and population-levelchanges often result in large-scale changes to community structure and functioning(Barry et al., 1995; Kordas et al., 2011).Less understood are the consequences of increased oceanic CO2 concentrationsand the resultant chemical changes, which are collectively referred to as ocean acid-ification or OA (Fabry et al., 2008; Feely et al., 2004). The global atmospheric CO2level has risen over 100 ppm since the Industrial Revolution, increasing from 280to nearly 400 ppm (Francey et al., 2013; Solomon et al., 2009), and is projectedto double or even triple by the end of this century (Dufresne et al., 2013; Hajimaet al., 2012). As a result, the mean seawater pH has declined more than 0.1 units,with a further decrease of 0.3 - 0.4 pH units expected by the year 2100 (Feely et al.,2009; Gosling et al., 2011). Ocean acidification is also reducing the saturation ofcarbonate, a component of calcium carbonate (CaCO3) required by many marinecalcifying organisms (Fabry et al., 2008; Feely et al., 2004). Prolonged exposure (e.g.weeks to months) to OA has been shown to reduce growth and calcification rates inheavily calcified organisms such as molluscs, echinoderms and corals (reviewed byKroeker et al., 2013); though there are some notable exceptions such as increasedgrowth rates of some sea star species (Dupont et al., 2010a; Gooding et al., 2009).Other groups, such as crustaceans and cephalopods, experience relatively few e?ectson growth (Kroeker et al., 2013). In many species, larval stages are more susceptiblethan juveniles or adults (Byrne, 2011; Dupont & Thorndyke, 2009). OA can alsoinfluence organismal behaviour (Bri?a et al., 2012); for example, littorine snails at7pH 6.62 showed stronger predator avoidance behaviours (Bibby et al., 2007), whilehermit crabs at pH 6.8 (? 1200 ppm CO2) were less able to detect food odors (de laHaye et al., 2012).It is unclear exactly what drives the observed e?ects of OA on growth andbehaviour, since OA involves multiple chemical changes including increased pCO2reduced carbonate availability, and reduced pH levels. However, many organismsexperience reduced or impaired metabolic functioning when exposed to high CO2conditions (Po?rtner et al., 2004); when combined with the increased demands ofregulating internal acid-base status, this could leave insucient energy for growthor other processes (Sokolova, 2013). Regardless of the precise driver of the mostlynegative responses to OA, they will likely cause profound population and commu-nity changes, as has been seen in coastal areas with naturally occurring high-CO2conditions (Hall-Spencer et al., 2008; Wootton et al., 2008).In addition to oceanic changes such as increased SST and ocean acidification,rocky shores will be influenced by terrestrial climate change (Helmuth et al., 2006b).Many of these terrestrial impacts involve greater variability in abiotic conditions; forexample, increasingly unpredictable patterns of freshwater runo? and precipitationmay alter salinity regimes (Morrison et al., 2002), while shifts in weather patternsmay increase the frequency and severity of aerial heat waves (Meehl, 2004). Thus,climate change in rocky shores will involve multiple large-scale changes to relativelystable physical variables as well as stochastic and increasingly extreme abiotic eventsthat may be more localized in time and space (Helmuth et al., 2010; Hsieh et al.,2005).The potential consequences of changes to multiple abiotic variables has only re-cently begun to receive attention. In marine systems, much of this research has fo-cused on interactions between increased SST and altered seawater chemistry such as8OA and salinity. Exposure to elevated pCO2 levels ranging from 550 to 1000 ppm (?0.06 - 0.25 unit reduction in pH) reduces the thermal tolerance of urchins (O?Donnellet al., 2008), crabs (Walther et al., 2009), and abalone (Zippay & Hofmann, 2010),while some species of corals and molluscs are more susceptible to the e?ects of OAwhen simultaneously exposed to increased temperatures (Melatunan et al., 2013;Parker et al., 2009; Rodolfo-Metalpa et al., 2011). Similarly, reduced salinity canexacerbate the e?ects of thermal stress, and vice-versa, in a variety of organismsincluding crustaceans, gastropods, and bivalves (Cheung & Lam, 1995; Heilmayeret al., 2008; McLeese, 1956; Todd & Dehnel, 1960; Yaroslavtseva & Sergeeva, 2006).Metabolic mechanisms are believed to contribute to these combined e?ects, sincetemperature, salinity, pH and pCO2independently a?ect metabolic scope and, whencombined, might leave organisms without sucient energy for other vital processesor behaviours (Po?rtner & Farrell, 2008; Sokolova, 2013).1.1.2 Study speciesOne of the diculties in predicting the ecological consequences of climate change isthe complexity of biological communities. It is not feasible to experimentally testthe e?ects of climate change on every species and interaction within a community.However, in some cases the structure of a community is largely controlled by onlya few keystone or dominant species. These can be ideal study subjects becausepredictions of broad community responses to climate change can be made withoutneeding to individually study every species in the community. In my work, I focuson four species that play key ecological roles on mid and high-intertidal rocky shoresin the Northeastern Pacific.The first two are a pair of littorine snails - Littorina plena and L. sitkana - thatare dominant grazers in the high intertidal zone. Along with barnacles and algae, lit-9torine snails are among the most abundant organisms in this section of the intertidal(Stephenson & Stephenson, 1949). Selective grazing by littorine snails can alter theabundance, diversity and composition of algal assemblages (Aquilino & Stachowicz,2012; Tarpley, 1992; Williams et al., 2013), and consumption of algal sporelingscan prevent the establishment of canopy forming species (Aquilino & Stachowicz,2012; McQuaid, 1996). Because algae are primary producers and ecosystem engi-neers, littorine snails can indirectly a?ect small invertebrates that rely on algae forfood and shelter, as well as barnacles and other species that may be outcompetedor smothered by algae Harley 2006; McQuaid 1996. Most of the e?ects of Litto-rina spp. are density-dependent (Aquilino & Stachowicz, 2012; McQuaid, 1996),so any climate-driven changes to snail abundance could have consequences for thecomposition and/or productivity of the surrounding community.The other two focal species of my work are Pisaster ochraceus and its preferredprey, the mussel Mytilus trossulus. As a keystone species, P. ochraceus controls thediversity and structure of mid-intertidal rocky shore communities through preda-tion on Mytilus spp. (Paine, 1969, 1974). In the absence of P. ochraceus, Mytilusoutcompetes other large species for space on the rocky substrate, but provides mi-crohabitat for a high diversity of small infaunal organisms (Harley, 2011; Paine,1974). When predation by P. ochraceus is high, in contrast, Mytilus abundance iskept in check and the diversity of larger species such as kelps, barnacles, and otherspace-holding organisms increases (Paine, 1969, 1974). Thus, the e?ects of climatechange on P. ochraceus, Mytilus spp., and their interaction will likely determine thee?ects of climate on the community as a whole.101.2 Thesis objectives and overviewAs I have highlighted here, the field of marine climate change faces complex andchallenging questions regarding the e?ects on species and communities. The nextstep is to expand our knowledge of individual, organism-level responses to broaderpopulation and interspecific changes. This will require consideration of how bothabiotic and biotic factors will influence the outcomes of climate change. I approachthese questions using several invertebrate species from the rocky shores of the SalishSea in British Columbia, Canada, and Washington State, USA.Broadly, I ask how abiotic and biotic factors mediate the e?ects of climatechange. This question can be broken down into three components:1. Direct e?ects of climate change via abiotic factors: How do single and multipleclimate variables directly a?ect individual organisms?2. Indirect e?ects of climate change via biotic factors: How might climate changeindirectly a?ect organisms via altered species interactions?3. Interplay between direct (abiotic) and indirect (biotic) climate factors: Howmight climate-driven changes to species interactions mediate the direct e?ectsof climate change on populations?I address the first component in Chapters 2 and 3, where I present two casestudies demonstrating the organismal-level e?ects of combined abiotic changes. Iaddress the second and third components in Chapters 4 and 5, where I build uponmy findings from Chapter 3 to present an in-depth study of how multiple stressorsimpact a predator-prey interaction. Below, I go into more depth regarding thecontent of each of my chapters.In Chapter 2, I investigate how multiple climate variables a?ect littorine snailsinhabiting the uppermost zone of the rocky shore. The high intertidal is subject to11severe abiotic conditions, especially during midday summer low tides when desicca-tion and thermal stress can be extreme (Mislan et al., 2009). The two snail species Istudied, Littorina plena and L. sitkana, are highly tolerant of these conditions andhave developed specific behaviours to help them cope with these stressors. However,climate change is predicted to increase both the frequency and intensity of manystochastic environmental conditions, increasing the likelihood that multiple stressfulabiotic events will temporally coincide in the future (Denny et al., 2009). If the pres-ence of additional stressors alters the ability of snails to withstand thermal stress,they may be unable to tolerate future climate conditions. I exposed snails for shortperiods (24 hours) to factorial combinations of low pH (due to increased pCO2) andlow salinity, and then measured how this a?ected their subsequent behaviour andsurvival under thermally stressful conditions.In Chapter 3, I ask how multiple stressors a?ect a keystone intertidal predator,the sea star Pisaster ochraceus. Specifically, I measured how factorial combinationsof increased seawater temperature and OA a?ect the growth and feeding rates ofthis species.In Chapter 4, I manipulate the body size of sea stars and their preferred prey,the mussel Mytilus trossulus, to gauge how climate-driven changes in body size ofeither species would impact their interaction. These data were then used to helpparameterize a predator-prey simulation in Chapter 5.In Chapter 5, I present a predator-prey model that simulates the trophic interac-tion between sea stars and mussels under various climate scenarios. Using my ownand previously published empirical estimates of the e?ects of climate change on seastars and mussels, I scale their vital rates (e.g. growth and recruitment) in orderto simulate the direct e?ects of climate on each species. I then incorporate thesescaled vital rates into simulated population and interspecific dynamics, enabling me12to predict how mussel populations may respond under several future scenarios ofincreased temperature and OA. Using these predictions, I test the hypothesis thatthe e?ects of climate change on mussel populations will be mediated by the presenceof an interacting species (sea stars).In Chapter 6, I discuss the broader implications of my work, summarize my mainconclusions, and highlight areas requiring future study. I also discuss potential weak-nesses in my methodologies and possible refinements that could be made. Finally, Iend with a few broad recommendations regarding priorities for future research.13Chapter 2Littorine Snail Responses toThermal Stress are Mediated byAdditional Stressors2.1 SynopsisThermal stress during low tide is one of the greatest stressors experienced by high-intertidal organisms, and is expected to become even more severe in the future.Because coincidental changes to other abiotic variables can reduce thermal tolerance,future increases in the temporal overlap of heat waves and other stressors could haveserious consequences for the ability of intertidal organisms to withstand climatechange. I tested whether acute exposure to factorial combinations of salinity (15,20, 28, and 30 psu) and CO2-driven pH changes (pH 7.4, 7.6, 7.9, and 8.0) a?ectedthe behaviour and survival of two littorine snail species (Littorina plena and L.sitkana) exposed to thermal stress. I found that snails kept in low salinity (15 or 20psu) for 24 hours had significantly lower survival rates during subsequent thermalstress, whereas exposure to low pH (7.4 or 7.6; ? 1500 - 1700 ppm CO2) had nosuch e?ect. However, neither salinity nor pH altered snail behavioural responses tothermal stress, suggesting the impact of salinity on snails? ability to survive thermalstress may be driven by physiological mechanisms and not behavioural changes. I14also found no consistent evidence of an interaction between salinity and pH on snailsurvival or behaviour. My findings suggest that the survival of littorine snails duringsevere heat waves could be mediated by low salinity events, leading to much greatermortality rates than would be predicted based on thermal stress alone. Coincidentalalignment of multiple acute abiotic stressors could have significant impacts on theability of organisms to withstand future climate change, and must be taken intoaccount when predicting the biological impacts of future change.2.2 IntroductionThermal stress is one of the greatest stressors faced by high-intertidal organisms. Onrocky shores, littoral organisms are subject to prolonged periods of desiccation andextreme temperatures during low tide (Chapperon & Seuront, 2010; Helmuth et al.,2006b). Many animals have developed physiological and behavioural adaptationsto survive these severe conditions (Garrity, 1984). For example, littorine snailscan reduce their body temperature by several degrees through the use of copingbehaviours such as aggregation, shell orientation, and microhabitat selection (Jones& Boulding, 1999; Miller & Denny, 2011; Rojas et al., 2013). Nonetheless, severeheat waves can drive mass mortality events (e.g. Harley, 2008), and are predicted toincrease in frequency and magnitude with future climate change (Denny et al., 2009;Meehl, 2004). Furthermore, recent work suggests animals in thermally-stressfulhabitats may have limited scope for future acclimation or adaptation to increasedtemperatures (Somero, 2010; Stillman, 2003).Littoral organisms may be even more vulnerable to thermal stress if their ex-posure coincides with stressful extremes in other abiotic variables (Denny et al.,2009; Harley & Paine, 2009). Rocky shores are highly variable environments, wherephysical conditions such as seawater pH or salinity can vary widely over the span of15only a few hours or days (Marliave et al., 2011; Truchot & Duhamel-Jouve, 1980).Climate change is expected to not only shift the baseline or average abiotic condi-tions, but also to increase the magnitude of variation around the ambient conditions(Hsieh et al., 2005; Solomon et al., 2007). These linear stochastic trends will likelyincrease the frequency with which thermal stress events coincide with other acuteenvironmental extremes (Denny et al., 2009; Harley & Paine, 2009), potentiallyexacerbating the e?ects of increased thermal stress on littoral species in the future.Concurrent exposure to otherwise minor stressors reduces the ability of manyorganisms to withstand thermal stress. Crabs exposed to high-CO2 conditions areless tolerant of subsequent thermal stress (Metzger et al., 2007), while intertidallimpets experience larger sublethal e?ects when high aerial temperatures coincidewith freshwater input from heavy rainfall (Williams et al., 2010). In many casesthese responses are believed be driven by compounded e?ects of multiple stressorson metabolic functioning (Sokolova, 2013), but impacts can also occur indirectlywhen one stressor alters an organism?s behaviour in a way that makes it moresusceptible to a second stressor (Bri?a et al., 2012).In the present study, I investigated whether acute changes to salinity and pHmediated the e?ects of thermal stress on littorine snails from rocky shores in theSalish Sea, a temperate sheltered coastal area in the Northeastern Pacific. Summerlow tides in this region occur mid-day when aerial temperatures and sun exposureare at their greatest (Helmuth et al., 2006a), leading to rock temperatures as highas 46.5 C (Harley & Helmuth, 2003). Several terrestrial freshwater sources drivelarge seasonal variations in salinity (Held & Harley, 2009; Todd & Dehnel, 1960),while localized upwelling of high-CO2 water and other oceanographic processes cangenerate sudden and large changes in seawater pH (Marliave et al., 2011; Wootton &Pfister, 2012). As climate change progresses, summers are predicted to become drier16and hotter (Mote et al., 2008; Solomon et al., 2007), precipitation and subsequentfreshwater input may become more variable (Doney et al., 2012; Johannessen &Macdonald, 2009), and oceanographic processes may drive ocean acidification atrates much greater than predicted for the ocean as a whole (e.g. Feely et al., 2008;Marliave et al., 2011; Wootton & Pfister, 2012). Thus, snails inhabiting this area arelikely to experience increasingly variable conditions in the future, with more frequentoverlap of transient and potentially stressful conditions such as low salinity, low pH,and thermal stress.Littorine snails can often tolerate short-term exposure to a wide range of pHand salinity conditions. High-intertidal species can experience internal pH levels aslow as 7.0 during periods of aerial exposure due to the build up of CO2 in theirtissues (Sokolova et al., 2000b; Truchot, 1990). Many species are highly tolerantof these conditions and some can regulate their intracellular pH for short periods(Sokolova et al., 2000b; Truchot, 1990), although chronic exposure to pH 7.6 orlower can reduce metabolic, growth and calcification rates (Bibby et al., 2007; Elliset al., 2009). Many species can also survive several days at salinities as low as 5 to15 psu, depending on the species (Hoyaux et al., 1976; Sokolova et al., 2000a; Todd,1964). However, littorine snails are poor osmoregulators and often exhibit strongsublethal responses such as reduced activity levels and metabolic depression at lowsalinities (Sokolova et al., 2000a; Todd, 1964). These possible sublethal e?ects ofacute reductions in salinity and/or pH could mediate snail responses to thermalstress, either through direct physiological impairment, or by reducing their activitylevels and subsequent use of thermal coping behaviours.I exposed two snail species, Littorina plena and L. sitkana, to factorial combi-nations of seawater salinity and pH conditions for 24 hours followed by thermallystressful low tide simulations, in order to determine the e?ects of these transient17stress events on their behaviour and survival. I predicted that (a) snails exposedto low salinity conditions (15 or 20 psu) prior to thermal stress will exhibit fewercoping behaviours and higher mortality rates, due to snails negative sublethal e?ectsof salinity on snail behaviour and metabolic functioning; (b) low pH (7.4 or 7.6) willhave no e?ect on snail coping behaviours or survival with thermal stress, since snailshave a greater tolerance for low pH conditions; (c) snails exposed to concurrent re-ductions in salinity and pH will experience synergistic negative e?ects on behaviourand survival due to the increased metabolic demands of simultaneously regulatingboth internal osmotic and ionic concentrations.2.3 MethodsI conducted a series of experiments on snail populations from two separate geo-graphic sites. Due to small di?erences in methodology, I will describe them sepa-rately for each site. The salinity and pH levels for these experiments were chosenbased on the local conditions experienced by each of the two snail populations and,therefore, the experiments used slightly di?erent treatment levels for each site.2.3.1 Littorina spp. from Burrard InletCollection site, snail handling and species identificationBurrard Inlet (BI) is located on the eastern edge of the Strait of Georgia, in southernBritish Columbia, Canada. The nearby Fraser River generates large fluctuations insalinity coinciding with annual cycles of the snowmelt (Held & Harley, 2009; Todd &Dehnel, 1960). The salinity of the upper 0.5 m of the water column ranges from 8.7psu in the summer to 29.5 psu in the winter, with an average annual salinity of 21psu (Held & Harley, 2009). Like many coastal areas, the mean pH in Burrard Inlet18is relatively low and highly variable compared to open oceans due to large changes inpCO2(Hofmann et al., 2011), with pH of the upper water column typically rangingfrom 7.3 - 7.9 (Marliave et al., 2011).Littorina spp. were collected from beneath small boulders on a rocky beach inthe high intertidal in West Vancouver, British Columbia, Canada (49.34, -123.22),in March 2012 for the temperature and salinity tolerance experiments, and December2012 for the factorial thermal stress and crawl-out experiments. The salinity wasaround 25 psu during both of these collection periods. Snails were kept withoutfood in a recirculating seawater system (12 C, salinity 28 psu) in the laboratoryfor 10 - 14 days prior to any experiments, depending on the collection date. Duringthis time, the snails were removed from the water every 1-3 days and left at 22.5 Cfor 3-4 hours at a time to simulate a low tide. This allowed them to acclimate towarmer air temperatures prior to the thermal stress trials.Based on multiple subsamples, the Littorina population on rocky shores in Bur-rard Inlet consists of ?85% L. plena and ?5% L. scutulata, with the remainder noteasily identifiable (R. Gooding, unpublished data). Species identification was basedon tentacle morphology (as described in Reid 1996). Because I could not confidentlyidentify all the individuals and this population is dominated by L. plena, I did notseparate them by species for the experiments and hereafter refer to the snails fromthis population as L. plena. Only medium-size individuals (approximately 4.5 to6.5 mm shell length) were used for this study.Experiment 1: Thermal toleranceI exposed Littorina plena to gradually increasing temperatures representative ofconditions during a summer low tide, in order to determine the temperature atwhich 50% of individuals died (hereafter referred to as LTemp50). Twenty snails19were patted dry and placed in each of 27 tightly sealed, dry 30 mL glass vials, witha cotton ball in the bottom to absorb any excess water. The vials were then placedin a circulating water bath that was gradually warmed from 15 C to 48 C overa period of 315 minutes (?0.1 C/minute). Three random vials were removed at 1C intervals beginning at 40 C, and the snails from these vials were immediatelyplaced in cold recirculating seawater to recover for 24 hours. Mortality was thenrecorded; snails were assumed dead if they did not open their operculum during a30 minute observation period in chilled seawater, or did not retract their foot whenpoked with a blunt dissecting needle.Experiment 2: Snail activity with acute low salinityIn order to determine sublethal e?ects of acute low salinity, particularly in regardto activity levels in the absence of thermal stress, I measured the crawl-out activityof Littorina plena. Although crawl-out behaviour does not necessarily have directfitness consequences, it is a commonly used metric of overall functioning and activitylevels because littorine snails have a strong behavioural tendency to climb to justabove the water line whenever they are able to (Jacobsen & Stabell, 1999; Todd,1964). I used six salinity treatments: 10, 12.5, 15, 17.5, 22.5, and 32.5 psu. Idid not include treatments with salinities between 22.5 and 32.5 psu because I wasspecifically interested in the e?ects of salinities near the lower end of what theyexperience, and they regularly experience salinity levels between 20 and 30 psu. Foreach treatment, I diluted natural seawater (30 psu) with dechlorinated water, withthe exception of the 32.5 psu treatment which was created by vigorously bubblingseawater with air for several days until sucient water had evaporated to reach thedesired salinity. Salinity was measured to ? 0.5 psu with a refractometer (FisherScientific). All treatments were gently bubbled with an airstone for at least one20hour prior to each experiment to ensure sucient oxygenation, but were not bubbledduring the experiment because this can dislodge snails from the side of the flask.For each replicate (n = 4 replicates per treatment), a 125 mL Erlenmeyer flaskwas filled with 100 mL (45 mm depth) of a randomly assigned salinity treatment.Sixteen snails were added to each flask, after being shaken for ? 15 seconds. Thisshaking step simulates wave action, encouraging snails to extend their foot andattach to the substrate (Jones & Boulding, 1999). The number of snails that hadcrawled above the water line in each flask was recorded after 30 minutes.Experiment 3: Snail crawl-out activity with salinity and pHI exposed snails to combinations of salinity and pH conditions for 24 hours to de-termine the e?ects on their activity levels. Snails were kept in one of four factorialseawater treatments (15 or 28 psu salinity, and 7.4 or 7.85 pH). A salinity of 28 psuand pH of 7.85 are representative of ambient conditions commonly found in BurrardInlet, while 15 psu and 7.4 pH are representative of the occasional drops in salin-ity and pH found in this region (Held & Harley, 2009; Marliave et al., 2011). Thetreatments were randomly assigned to plastic cups, which were filled with 300 mLof the appropriate seawater salinity and partially submerged in a 12 C seawatertable (n = 4 cups per treatment). The pH of the seawater was manipulated bythe addition of CO2 since acute pH reductions in rocky shore habitats are usuallydriven by the upwelling of CO2-rich water or by CO2 from biological processes (suchas respiration) (Feely et al., 2008; Truchot & Duhamel-Jouve, 1980). Each cup wasbubbled with either ambient air (7.85 pH treatment) or air supplemented with CO2regulated by a mass flow controller (7.4 pH treatment).Once the water had reached the appropriate pH level, twenty snails were placedin each cup. A mesh barrier prevented snails from crawling above the water line.21After acclimating to the treatments for 24 hours, the snails were removed, shakenfor ? 15 seconds, and placed in a 125 mL Erlenmeyer flask filled with seawater fromthe same salinity and pH treatment. Crawl-out rate was then measured as describedin Experiment 2.The pH of each cup was measured to ? 0.01 units at the beginning and end ofthe experiment, using a pH meter (Oakton 6 pH meter with automatic temperaturecorrection) with a double-junction electrode calibrated with NIST bu?ers. Immedi-ately prior to the crawl-out trial, a 30 mL sample of water was collected from eachtreatment cup and kept chilled in a tightly sealed glass vial with no head space. Thedissolved inorganic carbon (DIC) content in each sample was measured within twohours of being collected. For each water sample, the average of three consecutiveDIC measurements was used (Dissolved Inorganic Carbon Analyzer model AS-C3;Apollo SciTech Inc.). The pH and DIC were then used to estimate the total alkalin-ity and pCO2for each replicate cup, using CO2calc (Robbins et al., 2010) with theconstants of Mehrbach et al. (1973) as refitted by Dickson & Millero (1987).Experiment 4: Responses to thermal stress with salinity and pHTo test the simultaneous e?ects of salinity and pH on snail responses to thermalstress, snails were exposed to factorial salinity and pH treatments followed by asimulated summer low tide. I used the same four treatment levels and experimentalset up as described in Experiment 3, except the replicates were spread across fourrepeat trials on consecutive days, with the treatments spread evenly between days(n = 7 or 8 cups per treatment).After 24 hours in the seawater treatments, the snails were exposed to a simulatedlow tide during which the temperature of the substrate was gradually increased untilit reached their LTemp50 (determined from Experiment 1). Unglazed ceramic tiles22(10 ? 10 cm) were used as artificial substrate; each tile had 22 small depressions(approximately 7 mm wide ? 3 mm deep) drilled on the surface to simulate micro-habitats such as crevices or empty barnacle tests. The tiles were initially submergedin a 12 C recirculating seawater system for one week prior to any experiments; theywere then submerged for at least 16 hours prior to each trial in order to chill andfully saturate the porous surface with seawater. At the start of each tide simulation,the tiles were placed in a grid on a dry aluminum tray that was floating in a waterbath. Each tile was randomly assigned a cup from the first phase of the experiment.The snails from each cup were vigorously shaken for ? 15 seconds and then placedhaphazardly (all upright, no clumping) on the tile. Copper barriers around each tileprevented the snails from escaping.The tile temperatures were gradually increased using a combination of the waterbath underneath and heat lamps overhead. In this way, the surface of the tile reacheda much higher temperature than the surrounding air in a manner similar to howrocks in the high intertidal are heated by the sun during midday low tides. Thetile temperatures were carefully monitored every 5 - 10 minutes with an infraredthermometer (Sperry IRT200); a small temperature logger (Thermochron ibuttonDS1921-F6) was also placed on each tile to record the surface temperature at 1-minute intervals (Figure 2.1).At the end of each trial, the coping behaviours of the snails were recorded.These behaviours were chosen based on past studies of snail behaviours duringthermally-stressful low tides (Jones & Boulding, 1999; Miller & Denny, 2011; Rojaset al., 2013), and included propping on the side or lip of the shell, aggregating withother snails, and using depressions in the substrate. After each trial, the snailswere immediately transferred to recirculating seawater (12 C, 28 psu, ? 7.9 pH)for a 24-hour recovery period. Snail mortality was then recorded as described in23Experiment 1.Water samples were taken immediately prior to each thermal stress trial, andthe pH and DIC were measured as described in Experiment Littorina spp. from Padilla BayCollection site, snail handling and species identificationPadilla Bay (PB) is located on the southernmost edge of the Strait of Georgia, inWashington State, USA (48.54, -122.56). The Skagit River was gradually divertedover the past 200 years to just south of this bay (Collins, 1998), and Padilla Bay nowonly receives seasonal freshwater input from the Fraser, Skagit, and Nooksack Rivers(Bulthuis, 1996). Salinity in the top 0.5 meters ranges from 14.2 - 31.5 psu, with amean of 28.8; salinity seldom drops below 25 psu, usually during acute precipitationevents or peak river outflow in June and July. Similar to Burrard Inlet, the pHvaries diurnally and seasonally, ranging from 7.5 - 9.0 with a mean of 8.0 (NationalOceanic and Atmospheric Administration, Oce of Ocean and Coastal ResourceManagement, 2013).Littorina spp. were collected from a cobble beach in the high intertidal onSaddlebag Island in September and October 2010. They were acclimated for 10 -14 days in a recirculating seawater system (12 C and 30 psu salinity) in the samemanner as the snails from BI. L. sitkana, L. plena, and L. scutulata are commonlyfound in equal proportions in Padilla Bay. Unlike the populations in BI, L. plenaand L. scutulata from Padilla Bay can be reliably distinguished. For all of thefollowing experiments, I separated the three species according to shell morphology(L. sitkana) or tentacle morphology (L. plena and L. scutulata) (Reid, 1996), usingonly medium-sized individuals for the experiments.24Experiment 5: Snail activity with acute low salinityThe e?ect of salinity on snail crawl-out behaviour was tested separately for Litto-rina plena and L. sitkana, using the methods described for Experiment 2 with thefollowing changes. Four salinity treatments were used: 10, 15, 20, and 25 psu. Eachflask (n = 4 flasks per treatment, per species) contained 15 snails, and the numberof snails on the side of each flask was recorded after 60 minutes.Experiment 6: Responses to thermal stress with salinity and pHThe e?ects of salinity and pH on snail responses to thermal stress were tested usingthe methods from Experiment 4, with the following changes. The salinity levelswere 20 and 30 psu, while the pH levels were 7.6 and 7.95; these values were chosento represent the ambient and occasional low values found in Padilla Bay. The pHof each cup was measured, but DIC was not. Separate trials were conducted on L.plena and L. sitkana (n=3 or 4 replicate cups for each treatment and species). Thelow tide simulations were shorter than those used in Experiment 3a, lasting only 160minutes, and reached lower final temperatures (Figure 2.1). Three ibutton temper-ature loggers were placed on randomly selected tiles for each trial, and temperatureswere recorded at 5-minute intervals.2.3.3 Data analysisThe snail behaviour and mortality responses from all of the experiments generatedproportion data, so they were analyzed using generalized linear models (GLM) withbinomial error distributions (Warton & Hui, 2011).The LTemp50 of Littorina plena was predicted using a GLM fitted to the mor-tality data from Experiment 1. The e?ects of salinity on crawl-out behaviour wereanalyzed using separate GLMs for Experiments 2 and 5; for Experiment 5, the ef-25fect of snail species was also included in the model. When the main e?ects weresignificant, Tukey-HSD tests were used to determine where the di?erences lay.The e?ects of salinity and pH on the crawl-out behaviour (Experiment 3), ther-mal coping behaviours and survival (Experiments 4 and 6) were analyzed using sep-arate GLMs for each snail species, collection site (BI or PB), and response variable.Each model included the main e?ects of salinity and pH as well as the interac-tion between them. The analyses for Experiment 4 also included the maximum tiletemperature as a random factor, since the temperature varied between tiles; tem-perature data was not available for individual tiles from Experiment 6. The trial(date) was initially included as a blocking factor for Experiments 4 and 6, but wassubsequently removed because it was highly non-significant (p > 0.5) in all cases.In order to ensure that the CO2 addition resulted in significantly di?erent pH levelsin the control and low pH treatments, the mean pH levels were compared with a2-way ANOVA for each experiment.The statistical analyses were run in R Statistical Software, version 2.15.3 (RCore Team, 2013). Over-dispersion was present in all of the binomial GLMs butwas corrected for using version 1.1 of the disp.mod package (Scrucca, 2012) in RStatistical Software.2.4 Results2.4.1 Burrard Inlet snailsExperiment 1: Thermal toleranceThe LTemp50 ? SE of Littorina plena from BI was 46.26 ? 0.25 C (Figure 2.2;GLM; 21 = 67.89, p < 0.0001).26Experiment 2: Snail activity with low salinitySalinity had a strong e?ect on the activity level of L. plena from Burrard Inlet, withfewer snails crawling above the waterline at 10, 12.5 or 15 psu, compared to 17.5,22.5, and 32.5 psu (Figure 2.3; 25 = 70.51, p < 0.0001).Experiment 3: Snail activity with salinity and pHThe addition of CO2 led to a significantly lower pH; there was also a significantinteraction in that there was a greater di?erence between the pH levels in the lowsalinity treatments compared to the control salinity treatments (Table 2.1; 2-wayANOVA; salinity: F1,12 = 9.09, p = 0.011; pH-treatment: F1,12 = 2654.2, p < 0.001;salinity ? pH-treatment: F1,12 = 9.09, p = 0.011).Low salinity reduced snail activity rates, with 20% fewer snails crawling abovethe waterline in the low salinity treatment;, however pH had no e?ect, nor was thereany interaction between salinity and pH (Figure 2.4; GLM; salinity: 21 = 13.25,p = 0.04; pH-treatment: 21 = 13.21, p = 0.84; salinity ? pH-treatment: 21 =12.45, p = 0.38).Experiment 4: Responses to thermal stress with salinity and pHThe addition of CO2 led to a significantly lower pH but also interacted with salinity,where the di?erence between the pH treatments was greater at low salinity (Table2.1; 2-way ANOVA; salinity: F1,26 = 11.49, p = 0.002; pCO2 F1,26 = 169.63, p< 0.0001; salinity ? pCO2 F1,26 = 4.64, p = 0.041).When Littorina plena was exposed to thermal stress, low salinity reduced sur-vival by 17% while low pH had no e?ect on survival (Figure 2.5a; GLM; salinity: 21= 3.98, p = 0.0484; pH-treatment: 21 = 0.02, p = 0.913; salinity ? pH-treatment:21 = 0.26, p = 0.611). Neither low salinity nor low pH a?ected the frequency27of snail coping behaviours (Figure 2.5b; GLM; salinity: 21 = 0.95, p = 0.330;pH-treatment: 21 = 1.47, p = 0.2262; salinity ? pH-treatment: 21 = 0.05, p =0.83).2.4.2 Padilla Bay snailsExperiment 5: Snail activity with low salinityLow salinity reduced the activity levels of both Littorina plena and L. sitkana, withsignificantly fewer snails crawling above the waterline at 15 psu compared to 20, 25,or 30 psu (Figure 2.6; GLM; salinity: 23 = 76.72, p < 0.0001; species: 22 = 3.30,p = 0.1917).Experiment 6: Responses to thermal stress with salinity and pHIn the Littorina plena experiment, the addition of CO2 led to a significantly lowerpH while salinity had no e?ect on pH (Table 2.1; 2-way ANOVA; salinity: F1,10 =1.22, p = 0.30; pCO2 F1,10 = 146.27, p < 0.0001; salinity ? pCO2 F1,10 = 0.89,p = 0.37). There was a similar result in the L. sitkana experiment, though therewas also a marginally significant interaction where the addition of CO2 had a largere?ect on pH in the low salinity treatment (Table 2.1; 2-way ANOVA; salinity: F1,12= 0.52, p = 0.48; pCO2 F1,12 = 585.52, p < 0.0001; salinity ? pCO2 F1,12= 4.70,p = 0.05).When Littorina plena was exposed to thermal stress, low salinity reduced sur-vival by 5%, while low pH had no e?ect (Figure 2.7a; GLM; salinity: 21 = 7.26,p = 0.007; pH-treatment: 21 =2.72, p = 0.099; salinity ? pH-treatment: 21 =0.5457, p = 0.4601). L. plena?s coping behaviours were una?ected by salinity orpH alone, but there was a significant interaction where low salinity reduced copingbehaviours at the control pH level but increased them at the low pH level (Figure282.7b; GLM; salinity: 21 = 0.04, p = 0.8427; pH-treatment: 21 = 0.48, p = 0.4864;salinity ? pH-treatment: 21 = 6.58, p = 0.0103).Low salinity also led to a 7% reduction in the survival of Littorina sitkana afterthermal stress, while low pH had no e?ect (Figure 2.7a; GLM; salinity: 21 = 4.98,p = 0.0256; pH-treatment: 21 = 0.37, p = 0.5427; salinity ? pH-treatment: 21= 0.05, p = 0.82). Low salinity and low pH had no e?ect on L. sitkana?s copingbehaviours, nor was there any interaction between these variables (Figure 2.7b;GLM; salinity: 21 = 0.01, p = 0.911; pH-treatment: 21 = 0.03, p = 0.86; salinity? pH-treatment: 21 = 0.02, p = 0.886).2.5 Discussion2.5.1 Thermal and salinity tolerancesLittorine snails are generally quite tolerant of high temperatures (Clarke et al.,2000; Fraenkel, 1960; Stirling, 1982). Although the lethal upper temperature forLittorina plena or closely related species (e.g. L. scutulata or L. sitkana) has notbeen previously measured, I found that L. plena?s thermal tolerance of 46.2 C fellwithin the range of other Littorina spp., whose lethal thermal tolerances range from35.0 - 53.0 C (Davenport & Davenport, 2007; Evans, 1948; Fraenkel, 1960; Stirling,1982). However, caution should be exercised when comparing thermal tolerancesbecause a variety of methodologies have been used to test thermal tolerances. Forinstance, many studies either exposed snails to a series of constant temperaturesfor a set period of time or to a single constant temperature for varying periods oftime, whereas I exposed snails to a gradual increase in temperature similar to thethermal conditions they would experience in their natural habitat (see Table 2.2 fora comparison of some of these studies). These methodological di?erences can result29in very di?erent estimates of an organism?s lethal thermal tolerance, with methodsinvolving gradual aerial temperature increases yielding the most ecologically relevantestimates (Terblanche et al., 2011).When exposed to acute salinity reductions, I found that Littorina plena andL. sitkana both experienced a substantial drop in crawl-out activity between 20and 15 psu. Upon initial exposure to 15 psu or lower, most snails retracted theirfoot and kept their operculum tightly closed during the first few hours. Snailsexposed to 15 psu for 24 hours were alive and had open operculums, but many wereunable to attach well enough to climb the sides of the container. This pattern ofbehaviour has been shown in several other Littorina species exposed to this salinityrange (e.g. Todd, 1964). Since gastropods are poor osmoregulators, their primarycoping mechanism during acute salinity changes is to close their operculum or clamptheir shell to the substrate to avoid unfavourable conditions (Hoyaux et al., 1976).However, within 24 - 48 hours snails? internal tissues usually become isosmotic withthe surrounding seawater due to incomplete isolation from the environment (Avens& Sleigh, 1965; Hoyaux et al., 1976; Sokolova et al., 2000a), after which snails mayopen their operculum but still exhibit reduced activity levels (Todd, 1964).In contrast to their response to salinity, snails exposed to low pH for 24 hourshad similar activity levels as snails from control pH conditions. This could be dueto the fact that littorine snails in their natural habitat are regularly exposed tohighly variable external pH and pCO2 conditions (Truchot & Duhamel-Jouve, 1980;Wootton et al., 2008), and some species can regulate their acid-base status for shortperiods of time (i.e. 24 - 48 hours, Sokolova et al. 2000b). However, if metabolicrates or activity were reduced for a prolonged period due to chronic low-pH/high-CO2 conditions, this could lead to reductions in growth or other energetically-costlyactivities, as has been shown in L. littorea and L. obtusata exposed to ocean acid-30ification for 15 - 24 days (pH 6.45, Bibby et al. 2007; pH 7.6, Ellis et al. 2009).Additionally, the snails in these experiments, as well as the current study, were ex-posed to low pH driven by high pCO2 tensions; low pH conditions driven by othermechanisms could elicit a very di?erent response.2.5.2 Responses to thermal stress with salinity and pHThe negative e?ect of low salinity on the survival of Littorina spp. after thermalstress in my study agrees with previous findings on other species. Interactionsbetween salinity and temperature have been observed in a variety of invertebrates,including nemerteans (Zhao & Sun, 2006), crustaceans (McLeese, 1956; Todd &Dehnel, 1960), and gastropods (Clarke et al., 2000; Przeslawski et al., 2005). In someorganisms, such as crustaceans, salinity and temperature both have independente?ects on metabolic rates (Salvato et al., 2001; Todd & Dehnel, 1960), and severalstudies have shown that low salinity in combination with high temperature cantrigger metabolic depression (Salvato et al., 2001; Yagi et al., 1990). It is unclearwhether these mechanisms contributed to the e?ects of salinity on thermal tolerancein my study, but the possibility merits further study.Reductions in seawater pH driven by high pCO2 levels have also been shownto exacerbate thermal stress in organisms ranging from crustaceans (Metzger et al.2007; Walther et al. 2009) to echinoderms (O?Donnell et al., 2008), possibly due tocombined negative e?ects of hypercapnia (high pCO2) and temperature on aerobicscope (Po?rtner & Farrell, 2008). However, I found no e?ect of short-term exposureto low-pH on the thermal tolerance of Littorina plena or L. sitkana. It may bethat snails are well adapted to experiencing thermal stress and low pH conditionsconcurrently, since they frequently experience internal pH levels as low as 7.0 dueto the build up of CO2 or other metabolic products when they are retracted in their31shells during unfavorable conditions (Sokolova et al., 2000b; Truchot, 1990).Relatively little is known about the combined e?ects of low salinity and low pHon intertidal organisms. Nevertheless, there is some reason to expect interactionsbetween salinity and pH, since salinity a?ects acid-base regulation in organismssuch as crabs (Whiteley et al., 2001) and lugworms (Toulmond & Jouin, 1992). Itis unclear whether there is a similar e?ect of salinity on acid-base regulation and/orinternal pH in gastropods, who have weaker osmoregulatory abilities (Findley et al.,1978). Some snails reduce their internal pH in response to acute and extreme salinityreductions below ? 5 psu, possibly to depress their metabolic rate during periods ofsevere stress (Reipschlager & Portner, 1996; Sokolova et al., 2000a). However, thesalinity reductions in my study were transient and relatively moderate compared tothese past studies and unlikely to elicit such an extreme response, which may explainthe lack of an interaction between salinity and pH on L. plena crawl-out activity orsurvival. Given the negative e?ects of chronic exposure to reduced pH (? 6.5 - 7.6)due to ocean acidification on some snails (Bibby et al., 2007; Ellis et al., 2009), it ispossible that an interaction between salinity and pH would have appeared if snailsin my experiment had been exposed to the treatment conditions for a longer periodof time. Additional studies testing the e?ects of varying time-scales of exposure tocombined salinity and pH or OA would help address this question.2.5.3 Thermal coping behaviours with salinity and pHBased on the negative e?ects of low salinity on Littorina?s crawl-out activity andability to survive thermal stress, I expected to also see a decline in the number ofthermal coping behaviours used by these snails. Surprisingly, low salinity had noe?ect on the frequency of thermal coping behaviours, despite reducing snail crawl-out behaviours in the absence of thermal stress. It could be that in the absence of32thermal stress, snails voluntarily reduced their crawl-out activity under low salinityconditions in an e?ort to reduce their metabolic rate and reserve energy, but whenfaced with the more immediate threat of thermal stress the snails chose to resumeactivity levels in order to help cope with potentially lethal temperatures. Similartrade-o?s between sublethal conditions and immediate survival have been found incrabs and birds who chose to remain in suboptimal abiotic conditions in order togain protection from predation (McGaw, 2001; Tieleman et al., 2008).Equally surprising was the fact that snail mortality increased even though cop-ing behaviours remained the same. I did not directly test the impact of copingbehaviour on survival, but previous work has shown a strong correlation betweencoping behaviours, body temperature and survival (Chapperon & Seuront, 2010;Jones & Boulding, 1999; Miller & Denny, 2011). Although it is unclear why I didnot see a similar e?ect in my study, it does suggest that salinity likely impactedthermal tolerance directly through physiological mechanisms rather than indirectlyvia altered behaviour.The apparent antagonistic interaction between salinity and pH on the thermalcoping behaviours of Littorina plena from PB is puzzling, given the lack of suchan e?ect on any of the other snails in this study. For L. plena from PB, lowsalinity resulted in fewer coping behaviours under ambient pH but had no e?ectwhen combined with low-pH conditions. It is possible that this interactive e?ectonly occurs under certain combinations of salinity and pH, or the result could be anexperimental artifact or, perhaps, arose by chance. These findings suggest that thepossibility of an interaction should be explored, especially since salinity is known toa?ect acid-base status in some organisms (Toulmond & Jouin, 1992; Whiteley et al.,2001). However, at this point I cannot draw any definitive conclusions regardingthe nature or implications of this potential interaction.332.5.4 Broader implications and conclusionsMy findings support a growing body of literature showing the importance of multiplestressors to an organism?s response to climate change. Although short-term exposureto low salinity was not lethal by itself, it exacerbated the e?ect of thermal stress onsnail survival. In the future, stochastic events such as heat waves and low salinityconditions are expected to become more frequent, increasing the likelihood that theywill occur together (Denny et al., 2009; Harley & Paine, 2009). Thus, snails mayexperience greater mortality from thermal stress even if the severity of heat wavesand low-salinity events remains unchanged.The ability of organisms to tolerate future climate change may largely dependon the degree to which multiple climate stressors overlap in time and space. Giventhe stochastic nature of many climate-related variables, it will be important toconsider how various combinations of acute stressors will impact organisms? abilitiesto withstand future climate change. Predictions that do not take into accountcoincidental alignment of multiple stressors could lead to underestimates of thebiological impact of climate change.34Elapsed Time (minutes)Mean Tile Temperature (?C)10203040500 30 60 90 120 150 180 Max. temp.Experiment 4 (BI)Experiment 6 (PB)Figure 2.1: Mean tile temperature over time during thermal tolerance experiments.The points on the right-hand side of the figure represent the mean maximum tiletemperature, and the error bars represent ? 1 SD of the mean.3540 42 44 46 480. (?C)Proportion AliveLTemp50=46.26 CFigure 2.2: Thermal tolerance of Littorina plena from Burrard Inlet (N = 27). Thedashed line represents the temperature at which 50 % of snails did not survive.36Salinity (psu)Proportion Above Water10 12.5 15 17.5 22.5 A AB B BFigure 2.3: Mean proportion of Littorina plena from Burrard Inlet crawling out ofthe water after 30 minutes, at di?erent salinities (N = 24). Error bars are ? 1 SEof the mean. Treatments sharing a letter are not significantly di?erent.37Proportion Above Water15 psu 28 psu 15 psu 28 psu0.       Control pH                           Low pH        Figure 2.4: E?ect of salinity (15 or 28) and pH (control = pH 7.8; high = pH 7.4)on the mean proportion of Littorina plena Burrard Inlet crawling out of the waterafter 30 minutes (N = 16). Error bars are ? 1 SE of the mean.38Proportion Alive0. Coping15 psu 28 psu 15 psu 28 psu0.            Control pH                  Low pHFigure 2.5: E?ects of salinity (15 or 28) and pH (control = pH 7.8; high = pH 7.4)on the responses of Littorina plena from Burrard Inlet to thermal stress (N = 30).(a) Mean proportion surviving; (b) mean proportion using coping behaviours. Errorbars are ? 1 SE of the mean.39Salinity (psu)Proportion Above Water15 20 25 300. plenaL. sitkanaFigure 2.6: E?ect of salinity on mean proportion of Littorina plena and L. sitkanafrom Padilla Bay crawling out of the water after 60 minutes (N = 16). Note thatthe points for the two species are o?set slightly for clarity; treatments were 15, 20,25 and 30 psu salinity for both species. Error bars are ? 1 SE of the mean.40Proportion Alive0.800.850.900.951.00L. plenaL. sitkana(a)Proportion Coping20 psu 30 psu 20 psu 30 psu0.            Control pH                  Low pHFigure 2.7: E?ects of salinity (20 or 30 psu) and pH (control = pH 8.0; high = pH7.6) on the responses of Littorina plena (N = 14) and L. sitkana (N = 16) fromPadilla Bay to thermal stress. (a) Mean proportion surviving; (b) mean proportionusing coping behaviours. Error bars are ? 1 SE of the mean.41Table 2.1: Seawater parameters from Experiments 3, 4 and 6. Seawater salinity, pH, temperature, and dissolved inorganiccarbon (DIC) were measured directly. Total alkalinity (TA) and pCO2 were calculated using CO2Calc. Values are mean? SE, unless otherwise noted.Experiment Salinity(psu)CO2levelpHNIST DIC(?mol/kgSW) TA(?mol/kgSW) pCO2 (ppm)Crawl-out,no thermal stress(Burrard Inlet)15281528ControlControlHighHigh7.84 ? 0.017.90 ? 0.017.37 ? 0.027.37 ? 0.01917.0 ? 9.71368.1 ? 11.11065.5 ? 8.61467.8 ? 20.2921.0 ? 9.51420.2 ? 10.41003.7 ? 5.71412.2 ? 19.6449.6 ? 7.4502.5 ? 9.51534.3 ? 60.71829.4 ? 45.7Thermalchallenge(Burrard Inlet)15281528ControlControlHighHigh7.85 ? 0.027.94 ? 0.027.40 ? 0.027.42 ? 0.01879.2 ? 57.91334.2 ? 85.41070.4 ? 10.21469.4 ? 91.3883.5 ? 56.61393.1 ? 86.81013.9 ? 11.11425.0 ? 87.7435.4 ? 41.1454.8 ? 35.21437.1 ? 66.41635.7 ? 108.4Thermalchallenge,L. plena(Padilla Bay)20302030ControlControlHighHigh7.94 ? 0.027.98 ? 0.047.64 ? 0.017.65 ? 0.01No data No data No dataThermalchallenge,L. sitkana(Padilla Bay)20302030ControlControlHighHigh7.95 ? 0.027.99 ? 0.017.64 ? 0.017.62 ? 0.01No data No data No data42Table 2.2: Upper lethal temperatures (50% mortality) of Littorina spp. under di?erent experimental conditions.Species Duration Thermal Conditions Environment LethalLimit (C)ReferenceLittorina littorea 60 min. Constant temperature Seawater 41.0 Fraenkel 1960L. littorea 330 min. Constant temperature Seawater 39.0 Fraenkel 1960L. littorea 132 min. Gradual increase Seawater 46.0 Evans 1948L. littorea 60 min. Constant temperature Seawater 35.0 Davenport & Daven-port 2007L. littorea 60 min. Constant temperature Air 37.5 Davenport & Daven-port 2007L. littorea 360 min. Constant temperature Seawater 37.5 - 39.0 Newell et al. 1971L. neritoides 150 min. Gradual increase Seawater 46.3 Evans 1948L. kraussi 60 min. Constant temperature Seawater 50.0 Stirling 1982L. kraussi 270 min. Gradual increase Seawater 53.0 Stirling 198243Chapter 3Elevated Water Temperatureand Ocean AcidificationIncrease the Growth of aKeystone Echinoderm3.1 SynopsisAnthropogenic climate change poses a serious threat to biodiversity. In marineenvironments, multiple climate variables such as temperature and ocean acidifica-tion (OA) are changing simultaneously. Although temperature has well documentedecological e?ects, and many heavily calcified marine organisms experience reducedgrowth with ocean acidification, little is known about the combined e?ects of tem-perature and OA, particularly on species that are less dependent on calcified shellsor skeletons. I manipulated water temperature and OA to determine the e?ects onthe sea star Pisaster ochraceus, a keystone predator. I found that sea star growthand feeding rates increased with water temperature from 5 C to 21 C. A doublingof the current CO2 levels also increased growth rates with and without a concur-rent temperature increase from 12 C to 15 C. Increased CO2 had a positive butnonsignificant e?ect on sea star feeding rates, suggesting that CO2 may be acting44primarily at the physiological level to increase growth rates. As in past studies ofother marine invertebrates, increased CO2 reduced the relative calcified mass in seastars, although this e?ect was only observed at the lower experimental temperature.The positive relationship between growth and CO2 found here contrasts with themajority of previous studies, which have shown primarily negative e?ects of CO2 onmarine species; this may be due to sea stars having a less calcified body than theseother species. My findings demonstrate that increased CO2 will not have directnegative e?ects on all marine invertebrates, suggesting that predictions of bioticresponses to climate change should consider how di?erent types of organisms willrespond to changing climate variables.3.2 IntroductionAnthropogenic climate change is expected to generate environmental conditions thathave not been reached in many millions of years (Burrows et al., 2011; Feely et al.,2004). Although the ecological e?ects of these changes are already being felt (Harleyet al., 2006b; Parmesan, 2006; Wernberg et al., 2011), the full repercussions fornatural ecosystems and human societies remain poorly understood.In marine environments, increased water temperature and elevated carbon diox-ide concentrations (CO2) will be among the most important abiotic changes (Harleyet al., 2006b). Both positive and negative biological responses to increased temper-ature have been well documented and include vertical and latitudinal range shifts,altered feeding and growth rates, and acute responses such as coral bleaching (Ko-rdas et al., 2011; Parmesan, 2006; Sanford, 2002; Southward et al., 1995). Far lessis known about the biological e?ects of ocean acidification (OA), which collectivelyrefers to increased concentrations of oceanic CO2 and the resultant reductions inpH and carbonate availability (Fabry et al., 2008; Feely et al., 2004). The reduc-45tion in the availability of carbonate, which is a component of calcium carbonate(CaCO3) required by many marine calcifying organisms (Fabry et al., 2008; Feelyet al., 2004), is believed to be a major driver of decreased growth rates in mollusks,gastropods, coccolithophorids, and other heavily calcified species with experimentalincreases in CO2 (see review by Kroeker et al. 2013). However, the potential e?ectsof OA on species that are less dependent on calcified shells or skeletons have receivedcomparatively little attention.In addition to the important e?ects of individual climatic variables, simulta-neous changes in multiple climate variables have the potential to yield surprisingphysical and biological responses that could not be predicted by responses to sin-gle climate variables alone. For instance, CO2 solubility is temperature dependent,and therefore pH may show more or less extreme changes when increases in CO2and temperature are combined (Feely et al., 2004). Similarly, multiple stressorsmay have synergistic, antagonistic, or additive e?ects on marine species (Darling& Co?te?, 2008; Ho?man et al., 2003; Todgham & Stillman, 2013). Changes to oneabiotic variable may also increase an organism?s susceptibility to other stressors; forexample, OA reduces thermal tolerance in crabs (Metzger et al., 2007) and echinoidlarvae (O?Donnell et al., 2008), while warming has been shown to mediate the re-sponse of some corals to OA (Johnson & Carpenter, 2012; Reynaud et al., 2003).Additional studies on the presence of additive and/or interactive e?ects of multiplevariables are urgently needed.Pisaster ochraceus (Brandt, 1835), a primarily intertidal sea star, is an idealstudy organism to address the aforementioned questions. The importance of tem-perature on P. ochraceus biology has already been established; upwelling-associatedcooling of seawater has been shown to reduce P. ochraceus feeding rates and al-ter growth rates (Sanford, 2002). However, the extent to which these biological46rates will increase with incremental increases in water temperature and the limitsto this increase remain unknown. P. ochraceus also presents an interesting testof the e?ects of OA on less-calcified members of a phylum that has so far shownprimarily negative responses to OA. Rather than a continuous, heavily calcifiedendo- or exoskeleton, P. ochraceus instead has hundreds of tiny calcareous ossiclesembedded within and connected by soft tissue (Cavey & Markel, 1994; LeClair,1993); these ossicles make up a relatively small proportion of P. ochraceus? totalbody mass. Finally, P. ochraceus plays a key role in determining rocky intertidalcommunity structure along the western coast of North America (Navarrete et al.,2000; Paine, 1966, 1969; Sanford, 1999, 2002) and commonly feeds on heavily calci-fied species, such as mussels, that are predicted to experience reduced growth withOA. (Hiebenthal et al., 2013; Michaelidis et al., 2005; Parker et al., 2013). If P.ochraceus? response to climate change di?ers substantially from its prey?s response,this could have important implications for the strength of the predator-prey inter-action. Overall community responses to climate change will likely be contingentupon responses of key ecological species and how species-specific responses alter keyinterspecific interactions.In this study, I reared juvenile Pisaster ochraceus in the laboratory at watertemperatures ranging from 5 C to 21 C to determine whether growth and feedingrates will increase across the range of water temperatures this species is likely toexperience this century throughout much of its geographic range. I then rearedadditional juvenile P. ochraceus in factorial combinations of 2 temperature and 2CO2 treatments to determine the e?ects of simultaneous and climatically realisticincreases in water temperature and CO2 on the growth, feeding, and calcified massof P. ochraceus. I hypothesized the following: (1) P. ochraceus growth and feedingrates will increase with temperature toward an as-of-yet unmeasured maximum;47(2) elevated CO2 alone will have no e?ect on sea star growth or feeding rates, butthe proportion of sea stars total body mass consisting of calcareous material willdecline with increased CO2; and (3) increased temperature and CO2 will have nointeractive e?ect on P. ochraceus? overall growth but will have an antagonistic e?ecton the proportion of calcified material.3.3 Methods3.3.1 Study species and collection siteThe sea star Pisaster ochraceus is a marine intertidal predator found from Alaskato Baja California (Morris et al., 1980). P. ochraceus commonly feeds on mussels,barnacles, and gastropods, with its dominant prey source being mussels of the genusMytilus (Harley et al., 2006a). In protected areas such as the Strait of Georgia wheremy study was conducted, M. trossulus makes up the majority of P. ochraceus diet(Harley et al., 2006a). For this reason, and to facilitate comparisons with earlierwork (Southward et al., 1995), M. trossulus was used as prey in this study.Juvenile sea stars (3 - 7 g initial wet mass; 24 - 34 mm initial arm length) wereused for all experiments in this study. Juveniles were chosen because they exhibitgreater scope for growth. Furthermore, because they are not yet reproductivelymature (reproductive maturity generally is not achieved until at least 70 - 95 g wetmass; Menge & Menge 1974), excess energy is not put toward reproductive growth.All animals used in this study were collected from Jericho Beach in Vancouver,British Columbia, Canada (49.27 N, 123.2 W), in January 2008 (temperatureexperiment) and April 2008 (temp ? CO2 experiment). The water temperature inthis area ranges from a monthly mean of 6 C in February to 16 C in August and ispredicted to increase overall by ?1.5 C by 2040 (Mote et al., 2008). Once collected,48all sea stars were held in a recirculating seawater system maintained at 13 C for atleast 4 weeks before experimentation.3.3.2 Sea star growth and feeding rates with temperatureJuvenile Pisaster ochraceus (wet mass = 4.65 ? 0.19 g; all values reported aremean ? SE) were randomly assigned to one of 19 aquaria, which were set to dif-ferent temperatures ranging from 5 C to 21 C. Each 246-L tank was an indepen-dent unit with recirculating natural seawater (obtained from Burrard Inlet) bubbledconstantly with ambient air and equipped with a multistage filter system. Watertemperatures were maintained to ? 0.5 C of the set temperature using externalchillers and were measured at least 3 times a week with a mercury thermometer.The mean temperature for each tank was used for statistical analyses. Each sea starwas housed in its own 8 x 10 x 10 cm plastic container with mesh sides and top toallow water to flow through. Two containers were then randomly assigned to eachtank; there was no trend between the experimental temperature and initial sea starsize (simple linear regression: p = 0.1, df = 37, R2 = 0.02). At the beginning of theexperiment, initial wet mass was measured to the nearest 0.01 g. For all wet massmeasurements, each sea star was removed from the water, gently patted dry with apaper towel, immediately weighed on a scale, and then returned to the water. Thesea stars acclimated in their assigned tanks at 12.8 C for 4 days without food, thenthe tanks were changed to their experimental temperatures over an 8-h period, andfinally the sea stars acclimated for an additional 6 days without food.After the acclimation period, 20 small mussels (15 ? 2 mm shell length) wereplaced in each container. Empty shells were removed, recorded, and replaced withlive mussels every other day. No sea star ran out of mussels during the course ofthe experiment. Wet mass was measured weekly. After 21 days of feeding, sea stars49were wet weighed and all mussels were removed. To control for any e?ect of watertemperature on water retention in the sea stars, tank temperatures were broughtback to 12.8 C over an 8-h period and sea stars reacclimatized to this temperaturefor an additional 48 h without food before their final wet mass was measured.3.3.3 Sea star growth and feeding rates with temperature andCO2Juvenile sea stars (wet mass = 4.25 ? 0.10 g) were randomly assigned to one offour treatments: 12 C and 380 ppm CO2 (n = 5), 12 C and 780 ppm CO2 (n= 6), 15 C and 380 ppm CO2 (n = 6), and 15 C and 780 ppm CO2 (n = 5).These combinations were chosen to approximate current and predicted future levelsof change by the year 2100 (Intergovernmental Panel on Climate Change IS92aemissions scenario). The tanks and containers were the same as those described inthe previous experiment (section 3.3.2). Tanks were assigned to treatments using astratified random design. Temperature was maintained using external tank chillersas above, while CO2 concentrations were maintained using mass flow controllers toconstantly bubble the tanks with either ambient air (containing ?380 ppm CO2)run through an air compressor or the appropriate mixture of compressed CO2 (2%CO2 with balance air; Praxair) and ambient air from an air compressor. The tankswere covered with lids to help maintain the desired CO2 in the tank headspace andseawater.Two sea stars, each in their own container, were randomly assigned to eachtank. The mean initial sea star wet mass did not di?er between treatments (1-wayANOVA: F3,40 = 1.02, p = 0.395). Sea stars acclimated in their tanks without foodfor 9 days while the tanks equilibrated to experimental conditions. The initial wetweights were measured, and then sea stars were fed mussels (shell length = 17 ?502 mm) ad libitum for the remainder of the experiment. Tank temperatures weremeasured at least 3 times a week. Tank pH levels were also measured frequentlyto ? 0.01 pH units with a portable pH meter (YSI 556-MPS) calibrated at theappropriate experimental temperatures using NIST bu?ers. The mean standarddeviation of within-tank pH over time was ? 0.015, which was far smaller than theSD between tanks as well as between treatments. The mean pH of each tank wasused to determine treatment means. Sea stars were wet weighed weekly. After 10weeks, final wet mass was measured. Sea stars were then dried to constant massin an oven at 70 C and placed in 125 mL of a 10% bleach solution for 48 - 72hours to remove their soft tissue. The solution was then vacuum filtered on No. 1Whatman filter paper to collect the calcified material, which was then dried at 70C to constant mass and reweighed.3.3.4 Statistical analysesTo avoid pseudoreplication, all biological variables measured were averaged for the2 sea stars in each tank, and these tank means were used in all statistical analyses.The only exceptions to this were the analyses to determine if the mean initial seastar wet weights were the same across all treatments; individual sea star wet weightswere used to calculate these means in both experiments. In the temperature-onlyexperiment, separate simple linear regressions were used to determine the e?ect ofwater temperature on relative sea star growth rate [(final grams wet mass - initialgrams wet mass)/initial grams wet mass x 100] and per capita feeding rate (numberof mussels consumed daily per sea star). Although a second-order polynomial modelwas initially used to analyze the relationship between temperature and feeding rate,the polynomial term was nonsignificant. Therefore, I present a linear model in theresults.51In the 2 x 2 factorial experiment, a 2-way ANOVA was also used to determinethe e?ects of temperature and CO2 on relative sea star growth (percentage of gainrelative to initial sea star wet mass). Separate ANCOVAs were used to determinethe e?ects of temperature and CO2 on sea star feeding rates (number of musselsconsumed daily per sea star), percentage of wet mass consisting of calcareous mate-rial, and the ratio of dry soft tissue mass to water mass. Because absolute sea starsize can a?ect all of these variables, sea star wet mass was initially included as acovariate in all analyses. When size e?ects were nonsignificant (p > 0.1), they wereremoved from the analyses. Feeding rate and relative growth data for the factorialexperiment were log transformed to equalize variances, and all data were analyzedusing JMP 8 (SAS Institute).3.4 Results and discussionUnder manipulated water temperature alone, the relative growth (change in wetmass/initial wet mass) of juvenile Pisaster ochraceus increased linearly with tem-perature from 5 C to 21 C (Figure 3.1a; linear regression, p < 0.001, df = 18,R2 = 0.84), with no indication of a peak in growth over the measured temperaturerange. As no apparent growth maximum was reached in my experiment, my re-sults suggest that juvenile P. ochraceus growth rates will likely increase with futureoceanic warming throughout much of this species? range (Blanchette et al., 2007).Temperature also had a positive e?ect on feeding rates (Figure 3.1b; linear regres-sion, p < 0.001, df = 18, R2 = 0.69). It is somewhat less clear if an optimumfeeding rate exists within my manipulated range. Although a nonlinear model wasnot significantly better than a linear model, P. ochraceus feeding did appear to levelout around 16 C, and an optimum may lie somewhere between 15 C and 20 C.A similar trend was found in adult P. ochraceus feeding behaviours, although aerial52temperature also had a strong e?ect (Pincebourde et al., 2012). Further studies ofsea star feeding and growth responses to temperature at multiple life stages andwith greater temperature resolution above 15 C would help elucidate the nature ofthe temperature, growth and feeding relationship.In a factorial experiment, both temperature and CO2 had significant e?ects onseawater pH, with no interaction between them (2-way ANOVA; temperature, F1,19= 6.45, p = 0.021; CO2, F1,19 = 4.91, p = 0.039; temperature ? CO2 interaction,F1,19 = 2.09, p = 0.165). In the control (380 ppm) CO2 treatments, the seawaterpH was 7.85 ? 0.02 at 12 C and 7.88 ? 0.02 at 15 C; in the high (780 ppm) CO2treatments, the pH was 7.79 ? 0.01 at 12 C and 7.82 ? 0.04 at 15 C.Increased temperature and CO2 had positive and additive e?ects on sea stargrowth rates (Figure 3.2; 2-way ANOVA; temperature, F1,18 = 9.81, p = 0.006;CO2, F1,18 = 5.04, p = 0.038; temp ? CO2, F1,18 = 0.002, p = 0.967). Relativeto control treatments, high CO2 alone increased relative growth by ?67% over 10weeks, while a 3 C increase in temperature alone increased relative growth by?110%.The e?ects of temperature and CO2 on sea star feeding rates were more complexsince sea star feeding rates generally increase with sea star size. When the e?ectof sea star size was taken into account as a statistical covariate, elevated CO2 hadonly a marginally positive e?ect on feeding rates, while a 3 C temperature increasesignificantly raised sea star feeding rates by ?47% (Figure 3.3; ANCOVA; temper-ature, F1,17 = 12.3, p = 0.003; CO2, F1,17 = 3.18, p = 0.081; temp ? CO2, F1,17 =0.104, p = 0.707; mean wet mass, F1,17 = 96.8, p < 0.001).I also observed changes in the relative proportions of the three main sea starbody components, which were dry soft tissue, calcified tissue, and water. The rela-tive calcified mass (of total wet mass) showed no change with temperature alone, but53declined significantly overall with increased CO2, from a mean of 11.5% at controlCO2 to 10.9% at high CO2. However, there was also a significant interaction be-tween temperature and CO2, where the e?ect of increased CO2 on relative calcifiedmaterial was reduced in the high temperature treatments (Figure 3.4; ANCOVA;temperature, F1,17 = 2.21, p = 0.115; CO2, F1,17 = 4.84, p = 0.042; temp ? CO2,F1,17 = 4.84, p = 0.042; sea star wet mass, F1,17 = 8.48, p = 0.01). Althoughthe relative noncalcified wet mass ((dry soft tissue mass + water mass)/total wetmass) increased with CO2, the ratio of dry soft tissue mass to water mass remainedconstant at ?1:4 regardless of temperature or CO2 (2-way ANOVA; temperature,CO2, and temp ? CO2, all p > 0.3).My findings show similarities but also key di?erences from previous studies on thee?ects of climate change on marine organisms. Increased CO2 reduces calcificationrates in a variety of marine invertebrates, leading to reduced growth rates (Table 3.1;also see review by Kroeker et al. 2013). Although I found that the relative calcifiedmass of sea stars declined with increased CO2, Pisaster ochraceus ? overall growthrate did not su?er as a consequence. This seeming disagreement with the responsesof many marine invertebrates to elevated CO2 could be explained by di?erences inthe amount and location of their calcareous tissue. Unlike urchins and mollusks,P. ochraceus lacks a continuous calcified test, shell, or skeleton that encases a largeportion of its soft tissue, making it less likely that a reduction in the growth of P.ochraceus? calcareous material would physically limit soft tissue growth or function.Furthermore, P. ochraceus? calcified ossicles make up a relatively small proportionof its total body mass (R. Gooding, unpublished data). It may be that elevatedCO2 decreased the rate at which P. ochraceus added calcareous material as it doesin other species, but the lack of a continuous calcified shell or test in P. ochraceusallowed soft tissue growth to continue despite reduced calcification. The growth54rates of adults from two other asteroid species have also been to be robust to OA,though the e?ects on calcified material were not measured (Appelhans et al., 2012;Schram et al., 2011).Whatever the cause, I found no apparent negative e?ects ofreduced calcification on the growth, feeding, and survival of P. ochraceus during myexperiment. However, the long-term fitness consequences of reduced calcification insea stars are unknown.Despite the reduction in relative calcified mass with increased CO2, the overalle?ect of CO2 on growth was positive. The reasons for this are somewhat unclear.The ratio of dry soft tissue mass to water mass remained unchanged by temperatureor CO2, suggesting that the change in relative calcified mass must have been causedat least in part by an increase in the rate of wet soft tissue growth. Because I couldnot measure change in calcified mass over the course of the experiment, it is unclearwhether the rate of calcified tissue growth simply remained the same as that of seastars reared at control CO2 (thereby failing to keep pace with the increased softtissue growth) or declined compared to that of control CO2 sea stars. Experimentsspecifically testing sea star calcification rates under control and high CO2 conditionswill be necessary to answer this question.Although the unchanged ratio of dry soft tissue mass to water mass demonstratesthat the greater growth of sea stars reared at high CO2 was primarily because ofincreased wet soft tissue growth, it does not explain the mechanism behind thisincrease. The nonsignificant trend of increased feeding with increased CO2 suggeststhat although feeding rate may be partially responsible for the increase in growthrate, there are likely additional factors contributing to this change. It is possiblethat elevated CO2 increases resource use eciency; for example, the slightly lowerpH of high-CO2 seawater could aid in the digestion of prey tissue, making feedingless energetically costly. Alternatively, low level stressors such as low doses of toxins55can elicit positive responses such as increased growth in plants, invertebrates, andvertebrates, a phenomenon referred to as hormesis (Calabrese & Baldwin, 2001);the stress of reduced pH or carbonate availability may elicit a similar response insea stars. Identification of the precise mechanism driving the increase in wet softtissue growth with elevated CO2 will require further, more physiologically basedexperiments.Several studies have found predominantly negative and nonadditive e?ects ofmultiple climate variables on the growth and survival of marine organisms (Darling& Co?te?, 2008; Ho?man et al., 2003; Reynaud et al., 2003; Todgham & Stillman,2013). The lack of a similar negative or synergistic response in Pisaster ochraceuscould be explained by the relative thermal tolerances and environments of sea starsvs. many previously studied species. For example, tropical corals often live neartheir thermal tolerance for water temperature (Hughes et al., 2003) and are gener-ally experimentally manipulated at or near these levels (e.g., Reynaud et al. 2003).Under such conditions, the e?ect of any additional stressors may be magnified. P.ochraceus, in contrast, was well within its thermal range in my experiments and, asI have shown here, is unlikely to surpass its optimal water temperature with futureclimate change in much of its geographic range. My findings also suggest that thenature of an interaction between climate variables depends on the response variablebeing measured, even for the same species. In the case of P. ochraceus, I found thattemperature and CO2 had an antagonistic interaction in their e?ects on relativecalcified material, whereas they had a positive and additive e?ect on overall growthrates. These within-species di?erences in the interaction between and e?ects of com-bined temperature and CO2 add an additional level of complexity when attemptingto categorize the interactions between multiple climate variables.My findings suggest that caution should be exercised when predicting species?56responses to climate change on the basis of broad phylogenetic relationships alone.Negative responses to OA in ophiuroids and echinoids (reviewed by Dupont et al.2010b), both in the phylum Echinodermata, have led to overgeneralized predictionsthat echinoderms will respond negatively to OA. However, the lack of a negativee?ect of OA on sea star growth in my study demonstrates that this prediction cannotbe extended to all echinoderms. I also suggest that the di?erences in responses toOA in Pisaster ochraceus vs. previously studied echinoderms could be because of thelack of a continuous calcified endo- or exoskeleton in P. ochraceus. Further studiesshould be conducted on the responses of other less calcified members of taxa in whichother members have shown negative responses (see Kroeker et al. 2013 for examples).Although obvious examples exist, such as nudibranchs and many cephalopods withinthe predominantly shelled Mollusca, it could be that even subtle variation in thelocation or relative amount of calcareous tissue as is seen among species of bivalves,for example, is an important consideration when predicting biological responses toOA.The ecological implications of my findings should also be considered, becausePisaster ochraceus plays a keystone role in rocky intertidal communities (Navarreteet al., 2000; Paine, 1966, 1969; Sanford, 1999, 2002). An increase in P. ochraceusgrowth, even if only within the juvenile life stage, could lead to higher lifetime feedingrates because faster-growing sea stars would likely reach adult size classes sooner,thereby spending greater time in larger size classes that have higher per capitafeeding rates. This increase in predation rates will be even more pronounced if P.ochraceus prey, many of which are heavily calcified, respond negatively to climatechange, potentially resulting in a mismatch between predator and prey throughchanges in their relative sizes. Increased sea star growth rates could also havepopulation-level consequences. Faster-growing sea stars would spend less time in57vulnerable small size classes (Sewell & Watson, 1993), potentially increasing survivalrates and lifetime fecundity. I acknowledge, however, that these predicted responsescould be moderated by density-dependent e?ects and negative feedbacks on sea starsurvival and growth. Needless to say, changes in sea star population growth will becomplex and dicult to fully predict.As I have demonstrated here, responses to anthropogenic climate change, includ-ing ocean acidification, will not always be negative. This is an especially importantconsideration when attempting to make taxon-specific predictions about responsesto OA. Furthermore, species-specific responses could have serious ecological con-sequences when interacting species show di?erent or opposing responses to climatechange (Kroeker, 2013; Poore et al., 2013; Winder & Schindler, 2004; Wootton et al.,2008). In the rapidly expanding study of the biological consequences of ocean acid-ification, there is an understandable tendency to focus on calcified organisms thatare likely to show easily measured and generally negative responses to experimentalacidification. Some ecologically important species, however, may directly benefitfrom acidification, even within phyla that have traditionally been assumed to re-spond negatively to OA.58Relative growth (%)0306090120(a)   Feeding rate (mussels/day)5 10 15 (?C)Figure 3.1: Sea star growth and feeding rates increased linearly with water tem-perature. (a) Mean relative sea star growth (percentage of change from initial wetmass) with water temperature. (b) Mean number of mussels consumed daily per seastar with water temperature. Each data point is a within-tank mean of 2 sea stars(N = 19 tanks).590 10 20 30 40 50 60 700100200300400DayRelative growth (%)15 ?C, 780 ppm15 ?C, 380 ppm12 ?C, 780 ppm12 ?C, 380 ppmFigure 3.2: Mean sea star growth (percentage of change from initial wet mass) overtime with increased water temperature and CO2 (N = 22). Error bars represent ?1 standard error (SE) of the mean.60Per capita feeding rate (mussels per day)380 ppm 780 ppm 380 ppm 780 ppm0. ?C                                 15 ?CFigure 3.3: Mean number of mussels consumed daily per sea star under 4 factorialtemperature and CO2 combinations (N = 22). Error bars represent ? SE of themean.61Calcified mass (%)380 ppm 780 ppm 380 ppm 780 ppm9101112131412 ?C                                 15 ?CFigure 3.4: Mean proportion of sea star wet mass consisting of calcified materialunder 4 factorial temperature and CO2 combinations (N = 22). Error bars represent? 1 SE of the mean. To account for the confounding e?ect of sea star size on feedingrate, data were adjusted to the approximate median sea star wet mass (12 g).62Table 3.1: Growth responses of juvenile and adult marine invertebrates to ocean acidification with regard to taxon andskeletal type. Note this is not intended to be an exhaustive list of studies, but rather a representative sample of studiesusing climatically realistic increases in CO2.Phylum, class Species CaCO3skeleton1CO2(ppm)2E?ect(Calc.)3E?ect(Growth)ReferenceEchinodermataEchinoidea Lytechinus variegatus A 800 - - Albright et al. 2012Echinoidea Echinometra mathaei A 550 ND - Shirayama & Thornton 2005Ophiuroidea Amphiura filiformis A/B ? 1,000 + + / - Wood et al. 2008Asteroidea Pisaster ochraceus B 780 - + Present studyAsteroidea Asterias rubens B 1,250 ND = Appelhans et al. 2012MolluscaBivalvia Crassostrea virginica A 750 ND - Talmage & Gobler 2011Bivalvia Mytilus edulis A 740 - ND Gazeau et al. 2007Gastropoda Strombus luhuanus A 550 ND - Shirayama & Thornton 2005Gastropoda Littorina littorea A 1,000 - - Melatunan et al. 2013(continued on next page)63Table 3.1: Growth responses (continued)Phylum, class Species CaCO3skeleton1CO2(ppm)2E?ect(Calc.)3E?ect(Growth)ReferenceCephalopoda Sepia ocinalis B 4,000 = = Gutowska et al. 2008ArthropodaCrustacea Palaemon pacificus A 1,000 ND = Kurihara et al. 2008bCrustacea ParalithodescamtschaticusA 792 = - Long et al. 2013Crustacea Semibalanus balanoides A 1,000 - = Findlay et al. 2009CnidariaAnthozoa Montipora capitata A ?745 - - Jokiel et al. 2008Anthozoa Acropora cervicornis A 750 ND - Renegar & Riegl 20051 I distinguish between two skeletal types: A = most somatic and reproductive tissue is enclosed in skeleton (e.g.,shells and tests); B = nonencasing skeletal structures embedded within the soft tissues (e.g., spicules and ossicles).2 In studies where pH alone was reported, I estimated CO2 based on studies with similar pH changes.3 E?ect on calcified material.64Chapter 4Quantifying the E?ects ofPredator and Prey Body Sizeon Sea Star Feeding Behaviours4.1 SynopsisBody size plays a crucial role in determining the strength of species interactions,and the e?ects of body size on key interspecific relationships can be importantdeterminants of population and community structure. I measured how changes inbody size a?ect the trophic relationship between the sea star Pisaster ochraceusand its prey, the mussel Mytilus trossulus. I tested the e?ects of a wide range ofpredator and prey sizes on sea stars? prey size preference, feeding rate and preytissue consumption. I found that feeding rates and preferred prey size increasedwith sea star size. Juvenile sea stars preferred the most profitable prey sizes interms of handling time versus tissue consumed. While the number of mussels eatendaily declined with increasing mussel size, sea stars generally consumed the sameamount of mussel tissue overall regardless of the prey size o?ered. This suggests theremay be a limit to sea stars? overall rate of tissue intake, either due to physiologicallimitations such as digestion time, a behavioural strategy to limit the risk associatedwith increased feeding time, or a combination of the two.654.2 IntroductionBody size is one of the fundamental drivers of an organism?s biology, a?ecting ev-erything from metabolism to behaviour (Brown et al., 2004). It is also a majordeterminant of the strength of species interactions (Emmerson & Ra?aeli, 2004).Body size can a?ect predator feeding rates, prey susceptibility, and the outcomes ofcompetitive interactions (see review in Woodward et al. 2005), as well as food weband community structure (Heckmann et al., 2012; Jansson et al., 2007; Warren &Lawton, 1987). As such, any factor that alters body size will likely have cascadinge?ects on populations and communities (e.g. Rudolf & Rasmussen, 2013; Springeret al., 2003).It has long been known that body size is influenced by a wide range of environ-mental and biological factors. Growth rates and adult body size often scale withtemperature (Atkinson, 1995; Brown et al., 2004), while physical stressors such asacute thermal stress and hydrodynamic disturbance can cause size-dependent mor-tality (Denny et al., 1985; Peck et al., 2009). At the population level, predationsusceptibility, competitive interactions, and mate choice can all select for or againstcertain size classes (Green, 1992; McClintock & Robnett, 1986; Menge, 1972). An-thropogenic activities also have both direct and indirect impacts on body size; theserange from size-selective fisheries practices (Fenberg & Kaustuv, 2008), to climate-induced changes in growth rates (e.g. Drinkwater et al., 2010).The direction and magnitude of natural and anthropogenic e?ects on body sizeare often species-specific (Brose et al., 2012; Jochum et al., 2012). As a result,any factor that di?erentially a?ects the relative body size of one or more interact-ing species could shift the nature of their interaction, with potential population andcommunity consequences (Aljetlawi et al., 2004; Emmerson & Ra?aeli, 2004). Gain-ing a quantitative understanding of how body size a?ects a given interaction would66enable a more complete understanding of the role body size may play in structuringthe community as a whole. This is especially important for interactions involvingkeystone species or ecosystem engineers, where body size-driven changes to speciesinteractions could drastically alter community-level processes (Brose et al., 2012;Rudolf & Rasmussen, 2013; Woodward et al., 2005).One interaction with known size-dependencies involves the sea star Pisasterochraceus and its preferred prey, mussels. P. ochraceus is a keystone predator thatis common on rocky shores from Alaska to Baja California (Morris et al., 1980). Itexerts strong impacts on the structure and diversity of mid-intertidal communitiesthrough predation on mussels, which are the dominant space holders and ecosystemengineers (Harley, 2011; Paine, 1966, 1974). Where predation by P. ochraceus islow, mussels occupy much of the available substratum, outcompeting other macro-organisms for space while also providing habitat for microorganisms (Harley, 2011;Paine, 1974). In contrast, high P. ochraceus feeding rates keep mussel abundancein check, allowing other species to colonize the resulting open space (Paine, 1966,1974).The influence of prey size on the Pisaster-Mytilus interaction has received muchattention (Donahue et al., 2011; Paine, 1976; Robles et al., 2009, 2010; Sommeret al., 1999). Pisaster ochraceus exhibits a preference for mussels over other preyitems (Landenberger, 1968), as well as a strong preference for certain sizes of mus-sels (McClintock & Robnett, 1986; Robles et al., 2009). It is unclear whether P.ochraceus? prey size preference is due to physical limitations on the mussel sizesthat can be consumed, a cost-benefit decision that weighs handling time againsttissue consumed, or a combination of these two factors. On wave-exposed shores,P. ochraceus is often either unwilling or unable to physically consume the musselMytilus californianus beyond a shell length of 8 - 10 cm (Paine, 1976; Robles et al.,672009). However, P. ochraceus feeds primarily on mussels well below this apparentsize-refuge, and strongly prefers medium-sized individuals M. californianus to largerand more energy-rich mussels (McClintock & Robnett, 1986; Robles et al., 2009),possibly due to the energetic costs of dislodging larger individuals from the sub-stratum in order to consume them (McClintock & Robnett, 1986). Experimentalmanipulations of the size-frequency distributions of mussel beds have shown that seastars often remain in areas with high densities of preferred mussel sizes but emigrateelsewhere when only large, less-preferred mussels are available (Robles et al., 1995,2009).On sheltered rocky shores, Pisaster ochraceus feeds primarily on the musselMytilus trossulus. Unlike the closely related mussel Mytilus californianus, uponwhich most of the past research has focused, adult M. trossulus seldom grow largeenough to avoid predation by sea stars (Kozlo?, 1996), nor do their relatively weakbyssal thread attachments (Bell & Gosline, 1997) appear to pose a significant ob-stacle for P. ochraceus while feeding (personal observation). Thus, P. ochraceus ?prey size preference and feeding behaviours on M. trossulus are probably primarilydriven by the relative profitability (handling time versus tissue consumed; Krebs,1980) of di?erent mussel sizes. However, little quantitative data is available regard-ing the size-dependency of the interaction between P. ochraceus and M. trossulus,particularly in regard to the feeding behaviours of juvenile sea stars.Recent work suggests that future environmental change is likely to alter thegrowth of both Pisaster ochraceus and Mytilus spp., making it even more importantto understand how changing body sizes will a?ect their interaction. Although somevariables such as ocean warming may increase the growth rates of both sea starsand mussels (Gooding et al., 2009; Menge et al., 2008; Sanford, 2002), other factorssuch as ocean acidification may have opposite e?ects on these two species (Gooding68et al., 2009; Michaelidis et al., 2005). Although these predictions involve growthrates rather than body size, altered growth rates could change the size-frequencydistributions of some populations (Havens & DeCosta, 1985; Kirkpatrick, 1984).In this study, I quantified P. ochraceus? feeding behaviours across a range ofbody sizes of sea stars and their prey, the mussel Mytilus trossulus in order tobetter understand how their interaction might be a?ected by possible climate-drivenchanges in sea star or mussel size-frequency distributions (and therefore changes inthe availability of particular prey sizes). Specifically, I predicted that (1) sea starswill consume smaller mussels more quickly and therefore consume more of themdaily; (2) the balance between handling time and tissue gained will di?er basedon mussel size, and will result in some mussel sizes being more profitable thanothers; (3) sea stars will preferentially consume particular mussel sizes, and thattheir preferred prey size will increase with sea star size; (4) sea stars will preferthe prey size that provides the greatest tissue intake per unit of handling time, andwill therefore consume the greatest amount of mussel tissue when feeding on theirpreferred prey size.4.3 Materials and methods4.3.1 Animal collection and morphometricsSea stars (Pisaster ochraceus) and mussels (Mytilus trossulus) were collected atlow tide from Jericho Beach (49.27, -123.19), which is a sheltered rocky shorehabitat in Burrard Inlet near Vancouver, British Columbia, Canada. Sea stars weremaintained submerged in a laboratory recirculating seawater system at 13 C and28 psu salinity for 1 - 3 months prior to any experiments; during this time seastars were fed haphazardly sized mussels ad libitum approximately every two weeks.69Juvenile P. ochraceus (arm length = 25 - 30 mm) used for the experiments insection 4.3.3 (below) were collected in November 2007, while juvenile and adultP. ochraceus (arm length = 31 - 125 mm) used in section 4.3.2 were collected inDecember 2010. A range of sizes of M. trossulus were collected from Jericho Beachand maintained in a recirculating seawater system for 1 - 2 weeks prior to each set ofexperiments. New mussels and juvenile P. ochraceus were used for all experimentsin section 4.3.3. I was unable to collect sucient quantities of sea stars in 2010,so the same sea stars were used for both the feeding rate and prey size preferenceexperiments in section 4.3.2. However, sea stars were given two weeks to recoverbetween experiments and were randomly reassigned to treatments and containers.Additionally, prey size preference was measured first in order to prevent e?ects ofexposure to any one prey size during the feeding rate experiment. For all sizes ofsea stars, a flexible tape was used to measure arm length the nearest mm, from themouth to the tip of the ray nearest to the madreporite.The relationship between mussel size and tissue mass was quantified in orderto allow calculations of Pisaster ochraceus? prey tissue intake. Mytilus trossulusranging from 8 to 57 mm shell length (n = 33) were collected from Jericho Beachin December 2007 and kept in a recirculating seawater system for approximatelytwo weeks prior to dissection. Each mussel was pried open (by cutting the adductormuscle) and excess water was drained from the mantle cavity. The mussel was thenpatted dry and weighed, then the wet tissue was dissected out and the empty shellwas weighed again. The empty shell mass was subtracted from the total wet massto obtain wet tissue mass. Mussel shell length was measured to the nearest 0.1mm using vernier calipers. The relationship between mussel shell length and wettissue mass was determined using a 3rd-order polynomial regression. There was astrong relationship between mussel shell length and wet tissue mass (R2 = 0.99; p70< 0.0001), which was described by the following equation:Wet mass (g) = (0.042) ? (shell length (mm))2.43 (4.1)4.3.2 Part I: Sea star feeding behaviours with predator and preysizeTo determine how the preferred prey size scaled with sea star size, Pisaster ochraceusacross a range of sizes (n = 24 sea stars, arm length 31 to 125 mm) were placed inseparate plastic containers (28 ? 15 ? 11 cm) with large holes to allow adequatewater exchange with the seawater table (13 C and 28 psu salinity). After a 48 houracclimation period without food, each sea star was given two mussels from each ofsix approximate size classes: 10, 20, 30, 40, 50, and 60 mm shell length (? 2 mm; n= 12 mussels in each container). Size classes were selected in order to span the rangeof mussel sizes available to sea stars in the field, while still keeping the number oftreatments manageable. Sea stars were checked daily and empty mussel shells wereremoved and measured to the nearest 1 mm; they were then replaced by live musselsfrom the same size class (? 2 mm). This was repeated for nine days. The preferredmussel size for each sea star was calculated by taking the mean shell length of themussels chosen by that individual.To measure sea star feeding rate on di?erent mussel sizes, sea stars were groupedinto sets of three similarly sized individuals (8 sets total); each sea star in the setwas then housed separately and assigned one of three mussel size classes: 20 mm, 40mm, or 55 mm (all mussels were within ? 2 mm of assigned size class). All sea starswere fed ad libitum on haphazardly sized mussels for 5 days prior to the experimentto standardize their hunger levels. Their arm length was then measured and theywere placed in individual containers (28 ? 15 ? 11 cm) and given six mussels from71the appropriate prey size class. Empty mussel shells were removed and replacedwith similarly sized live mussels daily. One of the smallest sea stars, which wasassigned a 55 mm mussel, did not feed during the experimental period, and wastherefore removed from the subsequent analyses. After 13 days, feeding rates forthe remaining sea stars declined and the experiment was terminated.Due to diculties with getting mussels to remain attached to the substrate, es-pecially smaller mussels that tended to move around the experimental containers,sea stars were fed unattached mussels. Although byssal thread attachment maycontribute to handling costs of Pisaster ochraceus fed Mytilus californianus (Mc-Clintock & Robnett, 1986), sea stars in both the laboratory and the field often feedon M. trossulus that are still attached. This is especially common in juvenile seastars, who orient their body against the side of the mussel shell, insert their stom-ach into the small gap between the valves where the byssus and foot emerge, andconsume the entire mussel without detaching it from the substrate.4.3.3 Part II: Juvenile sea star feeding behaviours with prey sizeIn order to standardize hunger levels, juvenile sea stars (?25 - 30 mm arm length)were held without food for one week prior to experimentation, except for the feedingrate experiment for which they were fed mussels ad libitum for three days prior.During all experiments, sea stars were kept in the recirculating seawater table (13C and 28 psu) in small plastic containers (12 ? 12 ? 8.5 cm) with mesh sides toallow adequate water flow. Sea star arm lengths (mean of all 5 arms) were measuredat the beginning of each experiment. Mussels were sorted to within 1 mm of oneof four size classes: 10 mm, 15 mm, 20 mm, and 25 mm shell length. These sizeclasses were selected to be representative of the range of mussel sizes I had observedsea stars of this size consuming in the field.72The preferred prey size for juvenile sea stars was determined by giving sea stars(n = 13 sea stars) two mussels from each of the four prey size classes, for a total ofeight mussels per sea star. These mussels were placed haphazardly inside each ex-perimental chamber to minimize any e?ects of proximity on sea star prey selections.A sea star was assumed to have selected a mussel once it had initiated feeding (at-tachment of tube feet with mussel shell positioned against the sea star?s mouth), atwhich point the remaining mussels were removed. Once the sea star had consumedthe chosen mussel, the shell was removed and the sea star was provided with newmussels. This was repeated four times for each sea star over the course of threedays.To test sea stars? feeding rates on di?erent mussel sizes, two sea stars wereplaced in each of 8 small plastic containers (described above). Mean arm lengthwas measured to the nearest mm (mean ? SD = 25.2 ? 2.0 mm) at the beginningof the experiment. Each container was then randomly assigned to one of four preysize treatments, with two containers per treatment. Each container began with 20mussels of the appropriate size class; this number of mussels was far more than whatsea stars of this size can consume daily (personal observation), thus ensuring thatsea stars never ran out of mussels during the experiment. On the second day, emptymussel shells were removed and replaced with live mussels of the same size. Feedingrate was recorded for a total of 4 days, and mean daily per capita consumption rate(mussels consumed per day, per sea star) was calculated for each container. Thiswas repeated in three separate trials for a total of 6 replicate containers for eachprey size class.The handling time required for each prey size class was determined by randomlyassigning individual sea stars to be fed one of the four prey size classes. Handlingtime was defined as the time interval from the time the sea star touched the mussel73to the point when it released the empty mussel shell. It should be noted that mostsea stars touched and then left the mussel 2 - 4 times before commencing feeding.These initial contacts were ignored, and handling time was recorded as the single,continuous contact period that included feeding. This allowed an estimate of the truehandling time required for a given prey size, while eliminating time spent foragingor otherwise selecting prey. Sea stars sometimes hold onto prey shells even afterthey have finished eating, potentially leading me to overestimate handling time.However, small juvenile sea stars like those used in this experiment seem to be lessapt to do this; whenever I closely examined a juvenile sea star holding a mussel shell,its stomach was everted. Some sea stars did not feed during the 12 hour period of theexperiment and were subsequently removed from the analysis, resulting in unequalsample sizes (n = 4, 4, 3, and 2; for 10, 15, 20, and 25 mm mussels, respectively).4.3.4 Statistical analysesFor experiments from Part I, the relationship between prey size preference andsea star size was analyzed using a linear regression. The e?ects of prey size andarm length on feeding rate and daily prey tissue consumption were analyzed usingseparate ANCOVAs that included the interaction between prey size and arm length.The interaction term was highly non-significant (p > 0.5) for both analyses, so itwas removed. Normal-quantile and residual plots showed that feeding rate and dailytissue consumption data were non-normally distributed and had unequal variances;this was corrected for by log-transforming these variables.For experiments from Part II, I conducted a 2 test comparing juvenile sea stars?preference for 10-15 mm versus 20-25 mm mussels; I pooled the four original preysize classes into two groups in order to increase the sample size and power of thestatistical test since my original sample size was relatively small. The e?ects of prey74size on juvenile sea stars? handling time (minutes/mussel) and feeding rate (musselseaten/day) were analyzed with separate mixed-e?ects ANOVAs that included seastar size as a random e?ect, since I found that even small variations in arm lengtha?ected some feeding parameters. The prey profitability was calculated by dividingthe wet tissue mass of each mussel size class (derived from Equation 4.1) by thecorresponding handling time for that size class; this was then analyzed with a mixed-e?ects ANOVA that included arm length as a random e?ect. Daily prey tissueconsumption for juvenile sea stars was calculated by multiplying the wet tissue massof a mussel of a given size by the daily per capita feeding rate for that prey size; themeans for each prey size class were then compared using a mixed-e?ects ANOVAwith arm length as a random e?ect. The estimated daily feeding activity of juvenilesea stars was calculated for each prey size class by multiplying the handling timefor a given size class by the mean daily per capita feeding rate for that size class;again, this was analyzed with a mixed-e?ects ANOVA with arm length as a randome?ect. Handling time, feeding rate, and prey profitability were log transformed tocorrect for unequal variances.In analyses where a significant e?ect of prey size was found, a post-hoc TukeyHSD test was used to determine where the di?erences lay. All analyses were con-ducted using JMP 9.0.2 statistical software.4.4 Results4.4.1 Part I: Sea star feeding behaviours with predator and preysizePisaster ochraceus? preferred prey size increased with sea star arm length (Figure4.1a; linear regression; F1,22 = 30.1, p < 0.001, R2 = 0.578). Their daily feeding rate,75expressed as mussels consumed per day, also increased with sea star size; however,as prey size increased, sea stars consumed fewer mussels (Figure 4.1b; ANCOVA;arm length: F1,17 = 16.0, p ? 0.001; mussel size: F2,17 = 33.3, p < 0.001). Sea starsate significantly more small mussels than medium or large mussels (Tukey HSD; p< 0.05).The amount of tissue consumed daily also increased with sea star size, but preysize had only a small e?ect. Sea stars fed small mussels consumed less tissue thanthose fed medium or large mussels, though the e?ect of prey size was marginallynon-significant (Figure 4.1c; ANCOVA; arm length: F1,17 = 16.0, p < 0.001; musselsize: F1,17 = 2.10, p = 0.077).4.4.2 Part II: Juvenile sea star feeding behaviours with prey sizeJuvenile Pisaster ochraceus had a significantly greater preference for 20 - 25 mmmussels compared to 10 - 15 mm mussels (Figure 4.2a; 2test: 2= 6.53, df = 1, p= 0.011).When o?ered only a single prey size class, sea stars consumed significantly fewer20 and 25 mm mussels daily compared to the two smaller size classes (Table 4.1;mixed-e?ects ANOVA; F3,19 = 36.2, p < 0.001). Handling time increased signifi-cantly with prey size (4.1; mixed-e?ects ANOVA; F3,8 = 14.9, p = 0.001); sea starstook significantly longer to consume 20 and 25 mm mussels than they did 10 and15 mm ones.Profitability also varied with prey size (Table 4.1; Figure 4.2b; mixed-e?ectsANOVA; F3,8 = 12.2, p =0.002), with the largest size class of mussels providingmore tissue per unit handling time than the three smaller sizes, and the smallestsize providing far less compared to the three larger sizes. However, juvenile sea starsdid not consume significantly di?erent amounts of mussel tissue when fed di?erent76sized prey (Table 4.1; mixed-e?ects ANOVA; F3,19 = 1.88, p = 0.167).When daily feeding activity (mean handing time required to consume their meandaily number of mussels in a given size class) was estimated, sea stars feeding on10 mm mussels would require significantly more handling time each day than theywould if they fed on 20 - 25 mm mussels (Figure 4.2c; mixed-e?ects ANOVA; F3,16= 11.9, p < 0.001).4.5 DiscussionBody size plays a prominent role in determining the strength of most predator-preyinteractions (Brose et al., 2012). As such, any factor that alters the relative sizesof two interacting species could have far-reaching e?ects ranging from populationdynamics to community functioning (Emmerson & Ra?aeli, 2004; Jochum et al.,2012). This is especially true for interactions involving species that have domi-nant e?ects on community structure, such as the predator-prey interaction betweenPisaster ochraceus and its mussel prey (Paine, 1966).My findings support and expand upon past studies showing that Pisaster ochraceus?feeding behaviours scale with the size of both the sea stars and the mussels (McClin-tock & Robnett, 1986; Paine, 1976; Sommer et al., 1999). I found this pattern tobe true across a wide range of sea star sizes including small juveniles, about whichrelatively little is known. The higher activity rates of juvenile P. ochraceus alsoallowed me to quantify the e?ects of prey size on profitability and handling time.Juvenile sea stars preferred the largest prey sizes o?ered, which were also the mostprofitable based on the tissue gained per unit handling time for individual mussels.However, on a daily basis neither juvenile nor adult sea stars consumed significantlymore tissue when feeding on their preferred prey sizes; their daily tissue intake onlydeclined when feeding on the smallest prey sizes.77The fact that prey size had little e?ect on the daily tissue intake of both juvenileand adult Pisaster ochraceus was somewhat surprising. I had expected that seastars would prefer mussel sizes that allowed the greatest tissue intake in the shortestamount of time (i.e. the most profitable sizes; Hixon 1982), and that therefore seastars would consume more mussel tissue daily when feeding on the most profitableprey sizes. P. ochraceus experiences indeterminate growth, and their body size de-pends much on their energetic intake (Feder, 1956). Larger P. ochraceus are at lowerrisk of predation (Sewell & Watson, 1993), while reproductive output is positivelycorrelated with sea star size and energetic reserves (Sanford & menge, 2007). Basedon these traits, P. ochraceus would be predicted to be an energy-maximizer, seekingto consume as much prey tissue as possible (Schoener, 1971). Instead, I found thatP. ochraceus ? feeding behaviour is more characteristic of strategies involving limitson feeding time or consumption rates (see Hixon 1982 for a description of these).Although several species of sea stars have been shown to be energy-maximizers (e.g.Gaymer et al., 2004), the sea star Heliaster helianthus appears to limit its foodintake in a similar way to P. ochraceus, possibly as a means to reduce the risk ofbeing dislodged from the substrate by waves while feeding (Tokeshi, 1989).The reasons why prey size did not substantially a?ect Pisaster ochraceus? dailytissue intake are unclear, but may include physiological rate limits, environmentalfactors, or a combination of both of these. Digestion rate and stomach volumeconstrain the food intake of whelks (Burrows & Hughes, 1991) and fish (Hart &Gill, AB, 1992). However, P. ochraceus conducts much of its digestion externallywhile actively handling their prey (Feder, 1955), making it less likely that stomachvolume or digestion rate would be a significant rate-limiting factor. Alternatively,P. ochraceus ? assimilation eciency may decline sharply beyond a certain thresholdof tissue consumption. Under this scenario, feeding activity would become less and78less profitable as tissue was consumed beyond a certain threshold (Barnier et al.,1975; Vahl, 1984). Unfortunately little is known about this aspect of P. ochraceus ?physiology, so it is dicult to judge its likelihood as a driving factor.A second, though not mutually exclusive, possibility is that Pisaster ochraceusactively seeks to minimize its daily feeding time. It could be that once P. ochraceusreaches a threshold of minimum or optimum daily energy intake, the potential costsof continued feeding outweigh the benefits. Actively feeding sea stars often assume astationary, hunched position that leaves them only weakly attached to the substrate(Feder, 1955; Tokeshi, 1989). This may make them more vulnerable to dislodgementby waves, or predation by gulls and otters. By selecting prey items that minimizethe feeding time required to meet their daily energetic requirements, sea stars wouldalso have more time for non-feeding activities such as seeking refuge from desicca-tion or thermal stress during low tide. Similar trade-o?s in foraging time versusnon-feeding activities have been found in whelks avoiding wave action (Burrows &Hughes, 1991), birds protecting their territories (Krebs, 1980), and squirrels avoid-ing predation (Lima et al., 1985). I found that juvenile P. ochraceus preferred preysizes that would require the least daily handling time to reach their daily tissueintake, supporting the hypothesis that P. ochraceus ? prey size preference might bedriven at least in part by attempts to minimize feeding time while still meetingtheir daily energetic requirements. That said, I was only able to directly quantifythe profitability of di?erent prey sizes for juvenile P. ochraceus; thus, future studieson the profitability of preferred prey sizes for adult P. ochraceus will be necessarybefore broader conclusions can be made about this.It seems plausible that Pisaster ochraceus? prey size preference and daily tissueconsumption are influenced by a combination of physiological rate limits and time-minimizing behaviours. Physiological factors such as assimilation eciency might79set a maximum rate of tissue intake, but sea stars select the prey sizes that allowthem to reach this limit in the shortest amount of time. This would allow them tooptimize their daily energetic intake while limiting their feeding time and associatedrisks, and maximizing the time available for non-feeding activities. Furthermore, itis possible that the relative contribution of these two mechanisms varies over thecourse of a sea star?s lifetime, since the relative risks of feeding likely depend on seastar body size. Studies measuring the digestion and assimilation rates of juvenileand adult sea stars feeding on a variety of mussel sizes and daily tissue intakes wouldhelp test these hypotheses, as would detailed measurements of P. ochraceus ? dailyactivity budgets.Summary and implicationsUnderstanding which mechanisms drive Pisaster ochraceus? apparent limit on dailytissue intake could play an important role in predicting how increased risk or alteredprey size distributions might a?ect their behaviour and energetic intake. For ex-ample, if P. ochraceus? tissue consumption rate is primarily limited by behaviouralchoices, they may choose to spend less time feeding in the face of increased waveforce or high aerial temperatures, resulting in a lower net intake of prey tissue.However, if their tissue consumption is based primarily on a physiological limit, thisincreased risk may not be the first thing to cap their tissue intake. Similarly, if diges-tion or other physiological rates are the limiting factor to P. ochraceus? prey intake,changing environmental temperatures could influence this limit (Brown et al., 2004;Pincebourde et al., 2008), potentially altering the number of mussels and/or amountof tissue P. ochraceus could consume daily.My results corroborate previous findings that Pisaster ochraceus? feeding be-haviours scale with sea star and mussel size. However, my results also suggest a80previously undocumented dynamic - that despite their preference for the most prof-itable prey sizes, sea stars consume similar amounts of prey tissue regardless of theprey size eaten. This allows for potentially unexpected predictions, since a reductionin prey size or available feeding time will not necessarily equate to a change in P.ochraceus? overall energy consumption. Future studies will be required to determinewhich mechanisms drive this limit, as will in situ studies to help determine whetherenvironmental, tidal or seasonal changes influence P. ochraceus? feeding behaviours,especially their daily tissue consumption. For example, prior to spawning, sea starsmay switch to energy-maximizing behaviours to build their storage reserves (San-ford & menge, 2007). Finally, these conclusions may not apply to other prey items,including Mytilus californianus, since the relative costs and benefits of di?erent preysizes will likely depend on the specific traits of each prey species.The strong e?ects of body size on the Pisaster ochraceus-Mytilus trossulus in-teraction make it likely that changes to body size will be a major contributor tothe overall e?ect of environmental change on their community structure (see Broseet al. 2012, for examples of this dynamic in other systems). We may be able togenerate predictions regarding e?ects of environmental or anthropogenic change bycombining our knowledge of how these factors will directly a?ect body size withour understanding of how body size will a?ect the predator-prey interaction. On abroader scale, my findings reinforce the importance of body size in predator-preyinteractions, and suggest that these factors could play a large role in determiningorganism, population, and community responses to environmental or anthropogenicchange. Quantifying the interactions between species in appropriate levels of detailwill continue to be an important component of our ongoing quest to understand andpredict how spatial, temporal, and environmental gradients will a?ect species andcommunity functioning.81adult.pref$arm.length.mmPreferred prey size (mm)20304050 (a)adult.feed$arm.lengthMussels eaten day-102468(b)adult.feed$arm.lengthTissue eaten day-1(grams) 40 60 80 100 120 140246810(c)      Sea star arm length (mm)Figure 4.1: Feeding behaviours as a function of sea star and mussel size. (a) meanpreferred mussel shell length (N = 24 sea stars); (b) mussel consumption rate (N =23); (c) mussel tissue consumption rate (N = 23). For panels (b) and (c), musselsize classes were 20 mm (grey triangles, dotted line), 40 mm (black points, dashedline), and 55 mm (open circle, solid line).82Proportion selected0.  Profitability (g. tissue hr-1)   Feeding Time (hr. day-1 )10 15 20 254681012ABB B(c)    Mussel Size (mm)Figure 4.2: Juvenile sea star feeding and energetics as a function of prey size (N = 13sea stars); (a) preferred prey size; (b) mean profitability ; (c) mean estimated dailyhandling time required to meet daily tissue consumption rate. Error bars represent1 standard error of the mean. Treatments that share a letter are not significantlydi?erent from each other.83Table 4.1: Means and SE for juvenile sea star behaviours and energetics with prey size. Values are based on raw dataand not corrected for covariates such as small variations in sea star size.Mussel SizeExperiment N 10 mm 15 mm 20 mm 25 mmFeeding rate(mussels day1)24 4.43 ? 0.54 2.11 ? 0.48 0.86 ? 0.11 0.80 ? 0.11Handling time(minutes mussel1)13 132.2 ? 9.9 188.0 ? 28.6 354.7 ? 48.5 368.5 ? 11.5Tissue consumption(grams day1)24 0.48 ? 0.06 0.62 ? 0.14 0.50 ? 0.07 0.81 ? 0.11Profitability(g. tissue hour1)13 5.03e2 ? 4.27e3 9.84e2 ? 1.18e2 1.04e1 ? 1.04e2 1.65e1 ? 5.15e3Estimated dailyfeeding activity(minutes day1)13 567.34 ? 71.08 409.65 ? 93.30 307.25 ? 38.16 294.80 ? 41.7484Chapter 5Population Modelling PredictsPredator-Mediated E?ects ofClimate Change on anIntertidal Mussel Population5.1 SynopsisIt is becoming increasingly clear that the biological consequences of climate changewill not be driven by abiotic factors alone. The potential for climate change to alterspecies interactions means that organismal-level responses to abiotic change maybe poor predictors of population and community-level responses unless intra- andinterspecific dynamics are also considered. These complexities can be challenging toincorporate into empirical studies, particularly when the e?ects of multiple abioticvariables must be considered as well. Modelling approaches can help bridge thisgap, providing a framework in which to test and predict the relative contributionof biotic and abiotic factors to the biological outcomes of climate change. In thischapter, I present a predator-prey model that simulates the interaction between akeystone predator, the sea star Pisaster ochraceus, and its prey, the mussel Mytilustrossulus. Using previously published empirical estimates of the direct e?ects of in-85creased seawater temperatures and ocean acidification (OA) on each of these species,I modelled mussel populations under various climate change scenarios. I also ma-nipulated biotic factors such as the presence or absence of predation to determinewhat role, if any, the predator-prey interaction might play in determining popula-tion responses to climate change. My simulations predicted that in the absence ofpredators climate change had relatively little e?ect on the abundance, biomass orsize-frequency distributions of mussel populations. However, the e?ect of climatechange increased considerably when predators were present, with predominantlynegative consequences for mussel populations and especially strong e?ects of OA. Ialso found that climate change influenced the strength of the predator-prey interac-tion, with sea stars exerting a stronger e?ect on mussel populations in the presenceof OA. These model outcomes illustrate how species interactions can mediate the ef-fects of climate change, and how climate change can mediate the interaction strengthbetween species. More broadly, these results underscore the importance of includingboth biotic and abiotic factors, as well as potential interactions between them, whenmaking predictions about the biological consequences of climate change.5.2 IntroductionRecent research has made great strides toward understanding how climate changewill a?ect the vital rates and behaviour of a wide range of organisms. Biologi-cal responses are highly variable and often depend on the specific climate variable,species, and even life-history stage in question (Kroeker et al., 2013; Parmesan,2006; Twomey et al., 2012). Most studies on the e?ects of climate change involveeither laboratory-based experiments on individual organisms (e.g. Michaelidis et al.,2005; Pearce et al., 2005), or large-scale observational studies of communities andecosystems (e.g. Barry et al., 1995; Hall-Spencer et al., 2008). Fewer studies meet86in the middle, where direct climate e?ects on individual organisms are combinedwith indirect e?ects via species interactions and population dynamics. The handfulof such studies that have been conducted suggest that organismal-level responsesmay not be good predictors of population or community-level changes, and that thepresence of intra- and interspecific interactions can generate unexpected outcomes(Landes & Zimmer, 2012; Poloczanska et al., 2008). Mechanistic studies that in-clude e?ects of climate change via direct abiotic changes as well as indirect bioticchanges are necessary to more accurately understand and predict the populationand community-level outcomes of climate change (Brose et al., 2012; Doney et al.,2012).Concurrent changes to multiple climate variables add further complexity to at-tempts to predict the consequences of climate change. In marine systems, climatechange is expected to alter seawater temperature, ocean circulation, carbonate sat-uration and pH levels (hereafter referred to as ocean acidification, or OA), salinityregimes, and many other environmental factors (Caldeira & Wickett, 2003; Harleyet al., 2006b; Solomon et al., 2007). The e?ects of multiple climate variables onorganisms can be additive, where they are well predicted by the e?ects of the indi-vidual variables, or interactive e?ects can occur that would not have been predictedby additive models of change (Darling & Co?te?, 2008). This issue has been high-lighted by several authors (Crain et al., 2008; Darling & Co?te?, 2008; Paine et al.,1998), and empirical studies are increasingly manipulating multiple climate variablestogether.The combination of species-specific responses, intra- and interspecific dynamics,and multiple stressors makes it challenging to predict community-level responses toclimate change. Empirical studies that involve this level of complexity are costlyand logistically challenging, and it can be dicult to determine which aspects (e.g.87levels of organization, specific climate variables, species of interest) merit inclu-sion. Mathematical modelling approaches can help bridge the gap between empiri-cal organismal-level studies and population and community-level responses (Fulton,2011; Grith et al., 2012). By incorporating empirical findings on organismal-levelresponses into a biologically motivated representation of multispecies communities,models can serve as an important step toward a more complete understanding of cli-mate responses. They can help researchers to prioritize species or climate variablesfor further experimental study, to generate predictions to be tested by empiricalstudies, and to test hypotheses when real-world studies are not possible.A variety of modelling techniques have been used to predict responses to envi-ronmental change. Demographic models tailored to particular species and processesgenerate specific quantitative predictions (e.g. Jenouvrier et al., 2009; Simas et al.,2001), while multi-species models provide predictions of community-level responses(Cockrell & Sorte, 2013; Keith et al., 2008). At the broadest level, ?end-to-end?models can be used to predict ecosystem-level responses (Fulton, 2011; Grith et al.,2012). However, there is currently a lack of studies or models that predict popu-lation and community responses to multiple climate variables while also allowingspecific abiotic or biotic variables to be isolated in order to gain a more detailed andmechanistic understanding of the drivers behind the predicted responses.I present an individual-based predator-prey model that simulates the e?ects ofmultiple climate variables on an ecologically key pair of marine species: the seastar Pisaster ochraceus and its preferred prey, the mussel Mytilus trossulus. Mymodel uses experimental data to parameterize population dynamics and predation,and simulates climate change scenarios by scaling key parameters (e.g. growth andrecruitment) for each species based on empirical estimates of the direct e?ects oftemperature and ocean acidification. Using this model, I simulated the interaction88between sea star and mussel populations under varying climate scenarios to esti-mate the e?ects on mussel populations. The aim of this model was to generatequalitative predictions about the relative contributions of species interactions andabiotic change to the outcomes of climate change, as well as the influence of climatechange on the strength of species interactions. Specifically, I addressed the followingquestions:1. Do ocean acidification (OA) and/or increased seawater temperature a?ect thestrength of the interaction between sea stars and mussels?2. Does the presence of sea stars alter the e?ects of climate change on musselpopulation abundance, biomass and size-frequency distributions?3. What are the relative impacts of increased seawater temperatures, OA, andcombined temperature + OA scenarios on mussel populations?5.3 Study systemThe interaction between the sea star Pisaster ochraceus and its preferred prey,mussels in the genus Mytilus, is one of the best-studied species interactions. As akeystone species, P. ochraceus? predation on mussels strongly controls communitystructure and diversity of the mid-intertidal zone on rocky shores, where musselsare often the dominant space holders and ecosystem engineers (Paine, 1966, 1974).Predation by P. ochraceus can keep the abundance of mussels in check, allowingother large species to colonize the resulting open space (Paine, 1966, 1974).Pisaster ochraceus exerts most of its influence on mussel beds via strong percapita forces rather than a high abundance of individuals. One of the greatest driversof per capita predation by sea stars is body size (McClintock & Robnett, 1986; Paine,1976; Sommer et al., 1999). In Chapter 4, I demonstrated the importance of predator89and prey body size to sea star prey size selection, tissue consumption, and feedingrates on mussels. If climate change alters the body size distribution of either P.ochraceus or Mytilus trossulus, the resulting change in their interaction strength islikely to have profound community consequences.5.3.1 E?ects of increased seawater temperatures and OA onmussels and sea starsIncreased seawater temperature and OA are predicted to alter the growth and be-haviour of many marine species (Kordas et al., 2011; Kroeker et al., 2013). I providea brief overview of how sea stars and mussels are likely to respond to these climatechange variables in Table 5.1 and 5.2. I only included studies that used treatmentlevels that might occur by the year 2300 (? 2000 ppm CO2, +10 C, RCP8.5 sce-nario; Dufresne et al. 2013; Riahi et al. 2011). Because my model is primarilyconcerned with the interaction between sea stars and mussels, I limited my reviewto responses that are likely to a?ect body size and/or predator-prey interactions(e.g. growth, feeding, survival). Comprehensive reviews of the e?ects of temper-ature and OA on mussels can be found in Zippay & Helmuth (2012) and Parkeret al. (2013), respectively. There are no published reviews of sea star responses toincreased seawater temperatures or OA.Small increases in seawater temperature are generally beneficial to the growthand calcification rates of juvenile and adult mussels (Table 5.1, Zippay & Helmuth2012). However, the e?ects on larval mussels are more variable, with no clear con-sensus as to whether warming waters will have positive or negative consequences onthis sensitive life stage. Predictions regarding mussel responses to OA are far moreconsistent. With a few exceptions, the e?ects of OA on mussel growth, calcification,and survival are overwhelmingly negative (Table 5.1, Parker et al. 2013). These90e?ects have been documented across all life stages, though larval stages seem to beespecially sensitive.The combined e?ects of OA and increased seawater temperatures on musselshave received little attention. However, studies on oysters and clams have foundboth additive and non-additive e?ects (Matoo et al., 2013; Parker et al., 2009),depending on the species and response studied. Thus, it currently remains unclearwhether to expect interactive e?ects of temperature and OA on mussels.Similar to mussels, sea star responses to increased water temperature vary bylife stage. Larvae are more prone to detrimental e?ects, while juveniles and adultsoften experience positive e?ects such as increased growth and feeding rates (Table5.2). Relatively little is known about sea star responses to OA, especially for post-larval life stages. Studies on larval sea stars have found highly variable responses toOA, while the few available studies on juvenile and adult sea stars mostly show thatOA has either no e?ect or, in some cases, beneficial e?ects. Similar to mussels, fewstudies have investigated the combined e?ects of increased seawater temperatureand OA on sea stars. The handful of studies that have been done on this topic haveshown additive e?ects (Table 5.2, Chapter 3).5.4 Model presentationI used an individual-based predator-prey simulation that tracked the population sizeand size-structure of both predator (the sea star Pisaster ochraceus) and prey (themussel Mytilus trossulus) populations. The model is based on a daily cycle thattracks the recruitment, growth and mortality of individuals within each population,as well as the selection and consumption of mussels by sea stars (Figure 5.1). Themodel is spatially implicit: I did not track locations of individual mussels and seastars, but I use population densities, carrying capacities and recruitment rates based91on a focal area of approximately 30 m2 - hypothetically, a relatively isolated patchof low-intertidal rocky shore. Parameters are based on empirical estimates whereverpossible, and are subscripted M (mussels) or P (sea stars) where appropriate todistinguish them (Table 5.3).5.4.1 Mussel population dynamicsThe mussel population was modelled as an open population with density-dependentrecruitment, since available space for settlement is often a major factor controllingrecruitment of sessile organisms such as mussels (Gosselin & Qian, 1997; Petraitis,1995; Roughgarden et al., 1985). I did not track very small mussels (< 5mm shelllength), as only very small juvenile sea stars feed on them (personal observation)and they are unlikely to have much e?ect on mussel population dynamics such asbiomass and reproduction. Thus, I treated recruitment as including both settlementof larvae out of the planktonic larval pool and growth to the minimum size (5 mm)at which I tracked individuals in the population. Due to inadequate empirical dataspecifically tracking settlement rates of mussels of this size, I combined empiricalestimates of recruitment rates of very small recruits with estimates of early post-settlement mortality. The mortality rate of early settlers is generally thought tobe quite high but can also be extremely variable (Gosselin & Qian, 1997; Hunt &Scheibling, 1997; von der Med et al., 2012); therefore, I arbitrarily chose an earlysettlement mortality rate of 50%. The maximum possible recruitment rate was notinfluenced by the number of adults in the model population. The number of 5 mmrecruits joining the population at each daily time step was drawn from a Poissondistribution as follows: RM ? Poisson(rm(1 NMKM )) (5.1)92where rM is the maximum daily recruitment, and KM is the carrying capacityfor the mussel population; both parameters were scaled to the spatial extent of themodel (30 m2). I estimated KM from sea star-exclusion studies, where mussels oftencovered nearly 100% the available substrate (Table 5.3). My use of densities (NM andKM ) to represent the space occupied by the mussel population implies that musselssettle in a monolayer, and that mussel size can be ignored in calculating the amountof space occupied. Although Mytilus trossulus can and does form multilayer bedsin some areas, monolayer beds are also common. Additionally, as mussel densitiesnever reached carrying capacity in the model runs presented here, the simplifyingassumption that numbers of mussels equate to space occupied relative to carryingcapacity should have a limited impact on the model dynamics.Mussel growth, tracked using shell length in mm, followed the von Bertalan?ymodel using parameter estimates from Millstein & O?Clair (2001; Equation B1); thismodel was based on both age-length and tag-growth increment data from Mytilustrossulus in Prince William Sound, AK. I also accounted for variation in growthrates around the expected growth interval using the residual mean square from thefitted von Bertalan?y model of 4.33 mm2 over one year?s growth. This was convertedto a daily variance in growth of 0.0119 mm2, and a random normal deviate with thisvariance was added to the daily growth increment of each mussel. This growth modelresults in mussel growth that slows over time, and generally reaches a maximum sizeof around 60 mm after roughly 7 - 10 years.In addition to mortality due to sea star predation (see section 5.4.3, below),I simulated a low level of daily mussel mortality from other causes. The dailymortality rate for any given day was sampled from a binomial distribution, with theresulting number of mussels then randomly removed from the population. Due toinsucient empirical data on how non-predation mortality rates vary with mussel93size, I assumed that the mortality rate was independent of mussel size.5.4.2 Sea star population dynamicsI tracked sea star size using arm length rather than wet weight, as arm length is botha more stable measure and a better predictor of feeding behaviour. Adult sea starwet weight can fluctuate widely throughout the year due to energetic stores, repro-ductive status and water uptake (Mauzey, 1966; Pincebourde et al., 2008), whereasthere is little fluctuation in arm length aside from somatic growth (personal obser-vation). As a result, the correlation between wet weight and arm length becomesweaker as body size increases (Feder, 1956). Additionally, feeding behaviours suchas prey size preference and feeding rate are well predicted by arm length in adultsea stars (Chapter 4).I assumed for convenience that the population size of sea stars within the focalarea was a constant value NP over time. A relatively stable population of primarilyadult Pisaster ochraceus is supported by empirical studies (e.g. Paine, 1976; Rogers& Elliott, 2012). Furthermore, the strong per capita e?ect of P. ochraceus on mussels(Paine, 1974) means that even small changes in sea star density could trigger largechanges in mussel abundance, potentially overshadowing any climate e?ects.With a low daily probability dP , one sea star was randomly removed from thepopulation, due either to mortality or to emigration out of the focal area. Thisindividual was immediately replaced with a juvenile of arm length 20 mm. Juve-niles below this size were not included in the model because they consume primarilybarnacles and newly recruited (< 5 mm) mussels (Menge & Menge, 1974; Sewell& Watson, 1993), so they are unlikely to have much influence on mussel popula-tion dynamics. Juvenile sea star recruitment dynamics were not explicitly modelledbecause recruitment of P. ochraceus is a poor predictor of incorporation of juve-94niles into the population, and survival of early recruits is extremely low and highlyvariable (Sewell & Watson, 1993).5.4.3 Prey selection and consumptionI modelled predator-prey encounters on a daily cycle, with sea stars active for ? =20 hours out of every 24-hour period. The four hour period of inactivity represents alow tide during which sea stars cease feeding activity. At the beginning of the dailycycle, each sea star randomly encountered a mussel from the population, which thesea star either accepted and began to feed on, or rejected and randomly encounteredanother mussel. The probability of acceptance of a mussel depended on both sea starand mussel sizes, based on a preference function inferred from experimental datafrom Chapter 4. In the experiment, sea stars were allowed to select from a rangeof mussel sizes, following which the mean and standard deviation of mussel sizesselected were each regressed against sea star size. The linear regression coecientswere used to predict the center ?L and width L of a Gaussian prey preferencefunction for a given sea star of size L, where both ?L and L increase as a functionof sea star size.When a sea star encountered a mussel of size LM , the probability that it willaccept it is P [feeding] = Pmax exp?(LM  ?L)222L  (5.2)where the probability declines as the mussel size deviates from the sea star?s preferredprey size. I assumed Pmax, the probability with which a sea star will feed on a musselexactly matching its preferred prey size, to be 1. If the sea star rejected the mussel,it immediately encountered a new mussel at random, and the process was iterateduntil one was accepted (up to 20 iterations, at which point it was forced to accept the95most recently encountered mussel). When a mussel was accepted, it was removedfrom the mussel population and was thus unavailable to be selected by any othersea star.Once the sea star accepted a mussel, I calculated the amount of time until thesea star was ready to select its next prey item. This ?interfeeding time? includes thehandling time required to position and open the mussel, as well as the time requiredfor extra-oral digestion of the mussel tissue. It also includes time for unspecifiednon-feeding activities that occur in between bouts of feeding. As these individualdurations can be highly variable and challenging to estimate, I instead estimateinterfeeding time using feeding rates from experiments in Chapter 4.Using the data from Chapter 4, feeding rate (number of mussels per day) wasmodelled using a GLM (generalized linear model) with a quasipoisson error distri-bution and a log-link function. Sea star arm length, mussel shell length (15, 35 or55 ? 1 mm, treated as a continuous variable) and their interaction were used aspredictor variables. I then used the model coecients to calculate the expected log-transformed feeding rate log(f) for any combination of sea star size and mussel sizeE[log(f)]P,M . I incorporated error in this estimation by calculating SE[log(f)]P,M ,the standard error around the estimate E[log(f)]P,M . I sampled a random deviatelog(f) from a Gaussian distribution with mean E[log(f)]P,M and standard deviationSE[log(f)]P,M , then converted the sampled value of log(f) to the interfeeding timein hours. Thus, the interfeeding interval TF islog TF = I(f) + ?M(f)LM?+ ?P(f)LP?+ ?PM(f)LMLP? (5.3)The expected interfeeding time for a sea star eating a mussel can vary from afew hours (e.g. 4.8 hours for a 160mm sea star to feeding on a 5mm mussel) to far96longer than it would plausibly continue to feed (e.g. 1855 hours for a 20 mm sea starto consume a 65mm mussel). If the sea star did begin to consume the mussel, it wasonly allowed to finish consuming it and to obtain the resulting energetic intake ifthe interfeeding time TF was less than the remaining number of hours in the day. Ifthis was not the case, the sea star was forced to abandon the mussel, and its feedingactivity as terminated until the following day. Although this underestimated theenergetic intake of sea stars in some cases, sea stars strongly prefer mussels thatrequire an interfeeding interval of approximately 8 - 12 hours (based on experimentsfrom Chapter 4).Once all sea stars had accepted mussels, each of their interfeeding intervals wassampled as described above. The time of day t advanced until the first sea starfinished consuming its mussel; this sea star then selected another mussel from thepopulation, unless the time had passed the end of the daily cycle (? = 20). Thisprocess was repeated until t  ? , at which point all sea stars that had not finishedconsuming their current mussel were forced to abandon it. The daily cycle was thenreinitiated.5.4.4 Sea star growthWhenever a sea star finished consuming a mussel of length LM , it took in a quantityof mussel biomass BM calculated from the length-weight regression from Chapter 4(Equation 4.1).To calculate sea star growth in terms of arm length, BM had to be convertedinto a growth increment in mm arm length. Pisaster ochraceus? growth patternis characterized by an increase in arm length until it approaches an asymptoticsize beyond which energy is presumably allocated entirely to somatic maintenanceand reproduction (Feder, 1956). It was therefore necessary to derive a relationship97whereby the ?conversion factor? (mm arm length growth per gram of mussel tissueconsumed) changes as a function of sea star size. Unfortunately, because feedingtrials are rarely conducted on adult sea stars over sucient time frames for significantarm growth to be detected, it is not straightforward to fit a function directly to data.Instead, I inferred the function using a combination of estimates of arm growth overtime and measures of feeding rates for sea stars of di?erent sizes.I began with the best long-term record of Pisaster ochraceus arm length growth,by Feder (1956). I converted time to days rather than years, then fit various func-tions to model how arm length changes with time and sea star body size. I founda good fit of a breakpoint linear relationship in which arm length increases linearlywith slope b over time until it reaches a terminal length of 180 mm, then ceases tochange. This linear relationship better accounts for laboratory growth rates of juve-nile sea stars (R. Gooding, unpublished data) than logistic or asymptotic functions.In Feder?s (1956) data set, the daily growth increment for a sea star of arm lengthLP was therefore inferred to be b if LP < 180 mm and 0 if LP  180 mm.Next, I estimated the daily mussel biomass intake of sea stars in Feder?s (1956)study. These sea stars were fed ad libitum, and I made the additional simplifyingassumption that sea stars consumed only mussels matching their preferred prey size,as defined by Equation 5.2. By calculating expected feeding time on preferred musselsizes as a function of sea star arm length, I was able to estimate the daily rate ofmussel consumption, which I converted to mussel tissue biomass using Equation 4.1.By dividing the daily growth increment (b or 0, in mm/day) by the daily musselbiomass intake (in g/day), I obtained a conversion factor in units of mm/g thatchanges with sea star arm length. The growth in arm length per gram of musselmass consumed therefore relates to sea star length LP following:98rLP (1 LPLmax ) ?a(?L)b ? exp?E[log(f)]L,?L? (5.4)Every mussel consumed by a sea star thus contributed some small increment toits arm length growth, until growth stops at an arm length of LP = 180 and all preyintake is presumably directed toward body maintenance and reproduction. Thisallows for possible feedbacks between mussel abundance and sea star growth rates.5.4.5 Initial conditionsI used preliminary runs of the model to identify appropriate starting size distri-butions and densities for sea stars and mussels. Using a sea star population sizeof 24, I ran the model 5 times for a duration of 30 years each, sucient to reacha stable state. I repeated this with predators absent. I used the average musseldensity across runs as the initial density in subsequent runs (predators absent: NM= 5238 m2; predators present: NM = 4280 m2). The final mussel size frequencydistribution could be approximated by a truncated normal distribution (mean =20.4 mm, sd = 9.67, lower bound = 5). As most sea stars in the population at anygiven time are adults, I used a combination of a proportion of adult sea stars ofarm length 180 (proportion = 79) with a uniform distribution of sizes (20 - 180 mmarm length) for the non-adults. For each subsequent model run, I sampled initialsize-frequency distributions from these two distributions to ensure that initial con-ditions were somewhat near a steady state prior to the imposition of climate changescenarios.5.4.6 E?ects of ocean warming and acidificationThe climate scenarios in my model consist of two marine climate change variables:increased seawater temperature (+3 C) and ocean acidification (+400 ppm CO2).99These levels of change reflect moderate climate predictions for the year 2100 (RCP6.0scenario; Dufresne et al. 2013; Moss et al. 2010). I included four climate scenariosin my model: present conditions (i.e. no climate change), increased temperature,ocean acidification (OA), and increased temperature + OA. I simulated these cli-mate change scenarios by scaling four individual components of the model system:mussel recruitment, mussel growth, sea star interfeeding time, and sea star con-version rate. I set scaling factors using published empirical estimates (Table 5.4).For mussels, I preferentially used studies conducted on Mytilus trossulus, or if thesewere unavailable I used estimates from closely related species. For studies in whichtreatment levels were more extreme than the hypothetical conditions in my model(+3 C, +400 ppm CO2), I interpolated from their results to estimate e?ects at thetreatment levels in my model. Little is known about interactive e?ects of temper-ature and OA on mussels; however, based on the available data (see Table 5.1 andsection 5.3.1) I assumed that the e?ects were simply additive.I set the scaling factors for Pisaster ochraceus using data from Chapter 3 (alsoin Gooding et al. 2009), as this is the only work looking at the e?ects of temperatureand OA on P. ochraceus or closely related species. The results from Chapter 3 werepresented as growth and feeding rates; however, these changes are driven primarilyby prey consumption, feeding time, and conversion rate. Thus, I converted feedingand growth rate to interfeeding time (mussels eaten daily/24 hours) and conversionrate (sea star growth/g mussel tissue eaten). Because sea stars grew rapidly duringthe experiment, I also had to account for the e?ects of body size. To do this, I useda mixed-e?ects linear model that included the e?ect of sea star size (measured ona weekly basis) and time (to account for repeated measures of individual sea starsover time) on either feeding rate or conversion rate. The residuals from this modelrepresented the e?ects of temperature and OA on sea stars, as well as associated100error, with the confounding e?ect of size removed. I used these residuals to calculatethe mean responses (interfeeding time or conversion rate) under the four factorialtemperature ? OA treatments. Finally, I calculated the scaling factors for eachclimate scenario by scaling the response to each treatment relative to the controltreatment.5.4.7 Model runsI simulated a total of 14 di?erent scenarios. These consisted of the four climatescenarios, which were combined factorially with three biotic scenarios (where climatescaling factors were applied to only sea stars, only mussels, or both species). Somescenarios were also run in both the presence and absence of predators.For each scenario, I conducted ten replicate simulation runs, each lasting twentyyears. New mussel and sea star size distributions were randomly sampled for eachsimulation run using the baseline distributions described above. For each run, Itracked mussel abundance and size-structure every twenty days. I estimated totalbiomass of each mussel population using the mussel length-to-weight relationshipfrom Equation 4.1 (Chapter 4). I also tracked sea star size frequencies but did notfind any clear trends between climate scenarios (Table 5.5) so I do not discuss thoseresults.Because I was primarily interested in the relative e?ects of various climate sce-narios and predation on mussel populations, I calculated relative di?erences betweenvarious scenarios rather than presenting raw values. For all calculations, I used thevalues from the final day of each scenario run. I present two major comparisons:interaction strength and relative change. For interaction strength, I calculated asimple measure of the relative e?ect of that the sea star population had on the mus-sel population (Berlow et al., 1999). This measure of interaction strength provided101an estimate of the relative e?ect of predation on mussels, which could be comparedto the relative e?ects of climate change scenarios on mussels. I calculated meaninteraction strength, S, as S = ? 1n? nXi=1 Ni DiDi (5.5)where Ni is the mussel abundance in the ith run of a particular climate scenario withpredators present, Di is the mussel abundance in the ith run of the same climatescenario but with predators absent, and n = 10 replicate runs of each scenario. I usedthe ten calculated interaction strengths to also calculate the standard deviation (SD)of the mean for each scenario. I calculated the mean relative change, C, between twoscenarios in a similar manner, subtracting the response under the future scenariofrom the response under the present scenario, then dividing by the present scenarioresponse. Presenting both the mean and standard deviation provides an estimatethe average magnitude of the e?ects of a given scenario as well as the variation inthis magnitude across the ten replicate runs.5.5 Results and discussion5.5.1 Predator-prey dynamics under di?erent climate scenariosThe presence or absence of predators in my simulations resulted in drastically dif-ferent mussel populations in terms of both abundance and size-structure. Underpresent-day (ambient) climate conditions in the absence of predators, the musselpopulations were near their carrying capacity (95% of carrying capacity based onaverage of all 10 replicate runs, Figure 5.2), with relatively low recruitment ratesand size-frequency distributions that were heavily skewed toward larger individuals(Figure 5.3). Their population size was quite stable over time and across multiple102simulation runs (Figure 5.2). In contrast, when predators were present, simulatedmussel populations consisted primarily of small and medium-sized individuals aslarger mussels were preferentially eaten by predators (Figure 5.3). The additionof predators also led to large reductions in mussel biomass and abundance (Figure5.4), with mussel abundance reduced to 78% of carrying capacity, on average. Thisstrong e?ect of sea stars on mussel population dynamics was seen in Paine?s (1966,1974) classic field experiments on Mytilus californianus, where the mussel beds inthe absence of P. ochraceus consisted of many large individuals and covered nearlyall the available substratum, while those with predators present were greatly reducedin terms of both abundance and mean body sizes.When I compared simulations with and without predators to estimate the in-teraction strength, I found that climate change increased the e?ect of sea stars onmussel abundance and biomass. Compared to the present (ambient) climate sce-nario, interaction strength was slightly greater in the temperature-only and OA-onlyscenarios, and was strongest in the temperature+OA scenario (Figure 5.4). How-ever, it is worth noting that predators had an extremely large e?ect on musselsunder all scenarios, and the addition of climate change only increased interactionstrength by an average of 5-10%, depending on the scenario.Thus, under conditions simulated in my model, future climate conditions causedsea stars to exert a greater e?ect on mussel populations, but the presence of seastars is still the dominant driver of mussel dynamics regardless of the climate sce-nario. Although this dominant e?ect of sea star predation on mussel dynamics isnot surprising, it means that even small climate-driven changes in sea star densitiescould have profound impacts on mussels.Although several past studies have found similar increases in the strength oftrophic interactions with warming and OA (e.g. Asnaghi et al., 2013; O?Connor,1032009; Vucic-Pestic et al., 2011), the e?ects of climate change on species interactionscan be highly variable and largely dependent on the context and relative tolerancesof the interacting species (Kordas et al., 2011; Tylianakis et al., 2008; Van der Puttenet al., 2010). Studies on predator-prey and herbivore-plant interactions have shownthat experimental warming can lead to both positive (O?Connor, 2009; Vucic-Pesticet al., 2011) and negative (Laws & Joern, 2012; Tylianakis et al., 2008; Wagneret al., 2012) changes in interaction strength. In some cases, warming equally a?ectsboth interacting species, leading to no net change interaction strength (Landes &Zimmer, 2012; Tylianakis et al., 2008). Similar variability has been found in thee?ects of OA, where the strength of trophic interaction can increase (Asnaghi et al.,2013; Burnell et al., 2013; Ferrari et al., 2011), decrease (Busch et al., 2013; Russellet al., 2013), or remain unchanged (Allan et al., 2013; Appelhans et al., 2012; Landes& Zimmer, 2012).Few studies directly test which mechanisms might be driving climate e?ects onspecies interactions. However, the temperature-dependence of metabolic rates andhandling time likely contributes to the e?ects of warming on trophic interactions(Brown et al., 2004; O?Connor, 2009; Sanford, 2002; Vucic-Pestic et al., 2011), whilebehavioural and morphological changes driven by the increased demands of acid-base regulation and/or calcification under acidified conditions are often implicatedin OA-driven trophic changes (Bri?a et al., 2012; Ferrari et al., 2011; Manr??quezet al., 2013). However, the vast majority of the studies on the e?ects of warmingand OA on species interactions were conducted in mesocosms over relatively shorttime scales with a focus on organismal-level changes such as behaviour and phys-iology. Although these undoubtedly contribute to altered species interactions, myresults suggest that population-level responses such as recruitment, abundance, andsize-frequency distributions could also strongly influence the outcomes for species104interactions.5.5.2 Combined e?ects of predation and climate scenario onmussel populationsThe presence or absence of predators mediated the e?ects of climate change onsimulated mussel populations. In the absence of predators, mussels were relativelyuna?ected by climate change, regardless of the specific climate scenario (Figure5.5). This was somewhat surprising given the large direct e?ects climate change hasbeen shown to have on mussel growth and recruitment (Table 5.1). It may be thatthe individual-level e?ects of climate change are bu?ered or masked by populationdynamics; for example, when mussel abundances are near carrying capacity, density-dependence and competition for space may limit recruitment to such a degree thatthe e?ects of OA on recruitment will have little additional impact on the population(e.g. Paine, 2002).The e?ects of climate change on mussel populations became much more pro-nounced when predators were present (Figure 5.5). For example, OA reduced musselbiomass by only 4%, on average, in the absence of predators, but by nearly 25% whenpredators were present. This interaction between climate change and predation oc-curred regardless of whether predators are also directly a?ected by climate change.However, when sea stars were directly a?ected by climate change but mussels werenot, the resulting changes to mussel populations were comparatively small. Thissuggests the interaction between predation and climate change is driven primarilyby the combination of predation and direct climate e?ects on mussels, rather thanthe direct e?ects of climate on sea star interfeeding time. Simulations in which cli-mate change had direct e?ects on both species led to the greatest changes in musselpopulations (5 - 30% average reduction in biomass, depending on climate scenario),105though in most cases this e?ect was only slightly greater than scenarios where onlymussels were directly a?ected (Figure 5.5). Thus, it appears that, at least under theclimate scenarios and parameter values simulated in my model, the largest responsesto climate change are generated by the combination of direct e?ects of climate onmussels and indirect e?ects via the presence of predators.The striking di?erence in the e?ects of climate on mussel populations with andwithout predators may be driven largely by indirect e?ects of sea stars on the ratesof mussel growth and recruitment. In the absence of predators, low mussel mortal-ity and a lack of size-selective predation result in a population dominated by largemussels with little remaining scope for growth. Mussel abundance is near carryingcapacity, leading to density-dependent suppression of recruitment. Under these con-ditions, the two mussel parameters most a?ected by temperature and OA - growthand recruitment, respectively - occur at extremely low rates and play relatively mi-nor roles in the population dynamics. In contrast, the presence of predators leadsto mussel populations consisting of predominantly small and medium-sized individ-uals, since adult sea stars selectively consume larger mussels. Predation also drivesdown mussel density, facilitating higher recruitment rates. Under these conditions,climate-driven change to mussel growth and recruitment rates are likely to have amuch larger impact on population dynamics.In climate scenarios involving OA, the presence of predators may further exacer-bate the e?ects of climate change because both predation and OA influence musselsize (e.g. Table 5.1; Chapter 4). In mussel populations exposed to predation, adultmussels are selectively consumed by sea stars and replaced by recruits, leading toa population skewed toward smaller individuals, regardless of the climate scenario(Figure 5.3). The addition of OA reduces the growth rates of these mussels, caus-ing them to spend more of their lifetime in smaller size classes and further skewing106the size-frequency distribution toward smaller individuals. Sea stars consume thesesmaller mussels more quickly, further reducing the abundance of mussels. This cy-cle could conceivably continue to shift the population toward smaller and smallermussels, as lower mussel abundances lead to an influx of small recruits into thepopulation and sea stars then consume greater numbers of these small individuals.5.5.3 Relative e?ects of single and combined climate changevariablesAlthough I explicitly modelled the direct e?ects of combined temperature and OA asadditive, non-additive e?ects could conceivably arise at the population-level. How-ever, I found no evidence of non-additive e?ects based on qualitative comparisonsbetween the single and combined climate scenarios.The relative impacts of the temperature and OA scenarios on mussel populationsdepended largely on the population response measured. Mussel abundance wasa?ected almost solely by OA, whereas temperature and OA had nearly equal butopposite e?ects on mussel biomass.The relatively small e?ect of climate change on mussel abundance was unex-pected. I had anticipated seeing a large reduction in mussel abundance under futureclimate scenarios, but the changes turned out to be relatively small compared toe?ects on biomass (Figure 5.5). This may have been due partly to the fact that asmussel abundance declines, recruitment increases, so that as long as predation onmussels does not outpace maximum recruitment rates, the population may be fairlywell bu?ered against large changes in abundance (Petraitis, 1995). Thus, if musselabundance were the only response considered, one might conclude that climate?se?ects will be relatively minor and mostly driven by OA. In contrast, total biomassof the mussel population was strongly a?ected by both temperature and OA, some-107times changing by as much as five times more than abundance. This di?erence makessense if density-dependent recruitment is indeed the primary mechanism bu?eringchanges in abundance. In such cases, there could be nearly the same number ofindividuals in two populations, but vastly di?erent biomasses due to di?erences inthe size-frequency distributions. Population-level dynamics such as this illustratethe need to consider multiple response variables in order to ensure that importantclimate e?ects are not overlooked.5.5.4 Conclusions and broader impactsMy simulations indicate that the direct e?ects of climate on mussels may interactwith the indirect e?ects of sea star predation. In other words, climate change me-diates species interactions and species interactions mediate climate change e?ects.A likely driver of this dynamic is the combined e?ect of both climate change andpredation on the size distributions and abundance of mussel populations, and thesubsequent feedback cycle that developed between mussel size, sea star prey sizepreference, and sea star feeding rates. Given the important role that body sizeplayed in this outcome, it might be productive to test for a similar dynamic in othersize-structured population and interactions (e.g. Fong & Glynn, 1998; Maury et al.,2007)To my knowledge, no studies have measured or simulated the long-term in situe?ects of climate change on interacting populations of Pisaster and Mytilus. Thatsaid, a handful of studies have tracked intertidal and subtidal communities undernatural or experimentally manipulated climate conditions, particularly in regardto OA. Populations of bivalves, such as Mytilus spp., almost universally declinein abundance and average size under acidified conditions (Hale et al., 2011; Hall-Spencer et al., 2008; Wootton et al., 2008). Although these declines are similar to108those observed in my simulations, these studies did not quantify the abundance ore?ects of predators, so I cannot say whether or how much predation contributed tothe responses seen in these studies. Predator addition and removal experiments inthese acidified field locations would be extremely informative on the relative e?ectsof predators under di?erent climate scenarios.As is the case with any model, I made certain simplifying assumptions thatshould be kept in mind when interpreting the model outcomes. These include alack of size-specific natural mortality rates (Denny et al., 1985), sea star encounterrates that were independent of mussel density (Emlen, 1966), and open populationdynamics where the maximum potential mussel recruitment was unrelated to adultabundance (Fraschetti et al., 2002). Although I believe these assumptions are un-likely to have a large qualitative influence on the model outcomes, they may reducethe realism of my results.I also excluded certain spatial and temporal dynamics due to the level of com-plexity that would be required to incorporate them into my model. Mussel andsea star populations vary across gradients of intertidal height and wave exposure(e.g. Robles & Desharnais, 2002), while seasonal cycles of recruitment can influencepredator and prey population dynamics (Mauzey, 1966; Robles et al., 1995). Finally,infrequent but catastrophic events such as severe heat waves cause substantial popu-lation losses due to mortality (Tsuchiya, 1983). These dynamics represent potentialareas for future refinement in my model.The climate change scenarios in my model included two abiotic variables - in-creased water temperature and ocean acidification. In reality, several other climatevariables are likely to change as well (Harley et al., 2006b; Solomon et al., 2007).Perhaps most relevant to my model system is an increase in aerial temperatures,which has been shown to reduce mussel growth rates (Petes et al., 2007) and de-109press sea star growth and feeding rates (Petes et al., 2008; Pincebourde et al., 2008).Increased aerial temperatures will also likely increase the frequency of mass mortal-ity events due an increased incidence of extreme temperatures (Harley, 2008; Peteset al., 2008). Future versions of my model could integrate some of these e?ects, andmight be useful in qualitatively judging the relative importance of increased watertemperatures versus aerial temperatures for mussel and sea star dynamics.Based on the outcomes of my model, I can make some basic predictions andrecommendations for future research directions. The interaction between sea starsand mussels was clearly one of the fundamental drivers of responses to climatechange in my simulations. Furthermore, mussel populations as a whole may showvery di?erent responses to climate change than individual mussels, and predictionsshould not be based solely on the organismal-level responses to climate change. Insome cases, forces such as density-dependent recruitment and competition for spacemay overwhelm or limit climate?s e?ects on populations. However, the addition ofan interacting species could alter these population dynamics, allowing climate tohave a much larger impact.The e?ects of climate change on sea stars and mussels are likely to cause asignificant decline in mussel biomass. The consequences of such a change would befar-reaching. Paine?s (1966, 1974) sea star removal and reintroduction experimentsclearly demonstrate the impact that reduced mussel cover can have on communitystructure and diversity. Changes to mussel biomass would also alter its role as anecosystem engineer, and would likely be detrimental to species that rely on thebiogenic habitat to survive stressful abiotic conditions (Borthagaray & Carranza,2007). At the ecosystem level, lower mussel biomass would have consequences fornutrient cycles and community productivity as well as ecosystem services (Branchet al., 2012; Hutchings et al., 2012; Singh, 2010).110On a broader scale, my results add to the growing evidence that biotic andabiotic factors will likely interact to determine the population and community-leveloutcomes of climate change. Similar to past studies, I found that interaction strengthwas altered by warming and OA (e.g. Asnaghi et al., 2013; Wagner et al., 2012),and that the presence of interacting species can alter the e?ects of climate change(McCluney et al., 2011; Suttle et al., 2007). However, my findings also indicatethat population-level dynamics, such as density-dependent recruitment and size-frequency distributions, may be equally as important as organismal-level changesin behaviour and physiological rates. Little is known about the e?ects of climatechange on populations with size-structured species interactions, so further studyis needed to determine whether this pattern is applicable to other size-structuredsystems.Attempts to predict species, population, and community level responses to cli-mate change based solely on abiotic e?ects on single species are often unreliable. Thenext step toward more accurately understanding and predicting biological responsesto climate change is to incorporate direct abiotic e?ects with indirect changes tospecies interactions and population dynamics. As I have demonstrated here, onesuitable approach is to combine empirically-based predictions of climate e?ects onorganisms and interactions with theoretical techniques to extend these predictionsto population and community-level dynamics.111Figure 5.1: Simplified schematic illustrating the dynamics of the predator-prey sim-ulation.112Time (days)Density (mussels / m2  )0 1000 2000 3000 4000 5000 6000 700035004000450050005500CurrentT + OACurrentT + OAPredators absentPredators presentFigure 5.2: Mussel density over time for present-day climate scenario (black lines)and temperature+OA climate scenario (red lines). Each solid line represents a singlemodel run, with 10 replicate runs for each scenario. The dotted line indicates thepopulation carrying capacity. Note that the y-axis does not begin at zero.11310 20 30 40 50 60 10 20 30 40 50 60 10 20 30 40 50 6011  Predators  absent10 20 30 40 50 60 10 20 30 40 50 60 10 20 30 40 50 6011  Only mussels  affected10 20 30 40 50 60 10 20 30 40 50 60 10 20 30 40 50 6011  Only sea stars  affected10 20 30 40 50 60 10 20 30 40 50 60 10 20 30 40 50 6011 Both species  affectedMussel Size (mm)Relative FrequencyTemperature                                    OA                                  Temp+OA                                              Figure 5.3: Representative size-frequency distributions for mussel populations un-der di?erent climate and predation scenarios. The red overlay represents the size-distribution for the present-day scenario in either the absence of predators (top row)or the presence of predators (remaining rows); thus, the di?erence between the redand grey distributions illustrates the e?ect of the climate scenario on the population.The dashed line represents the median mussel size of the plotted population. Notethat relative frequencies are plotted, so absolute di?erences in abundance will notbe visible.114-0.28-0.24-0.20-0.16AbundancePresent Temperature OA Temp + OA-0.88-0.84-0.80-0.76BiomassInteraction StrengthFigure 5.4: Interaction strength between sea stars and mussels under present-dayand future climate scenarios. Each point represents the mean relative e?ect ofpredation on mussel populations compared to the corresponding predator-absentclimate scenario. Error bars represent ? 1 SD of the mean, and indicate the variationbetween replicate runs of the given scenario. Note that the y-axis does not begin atzero.115-0.15-0.10- + OAPredators  absentAffects only  musselsAffects only  sea starsAffects both  species-0.4- BiomassRelative change from controlFigure 5.5: E?ects of various climate change and predation scenarios on mussel pop-ulation abundance and biomass. Each point represents the mean relative di?erencebetween the present-day scenario and the corresponding climate change scenario.In the second and third columns, the direct e?ects of climate change were appliedto only the prey or the predator, respectively. Error bars represent ? 1 SD of themean, and indicate the variation between replicate runs of the given scenario.116Table 5.1: Mussel responses to climatically realistic increases in water temperature and ocean acidification.Climate variable Species Life stage Response E?ect ReferenceIncreased watertemperatureMytilusgalloprovincialisEmbryo GametogenesisFecundityEgg quality-==Fearman & Moltschani-wskyj 2010Increased watertemperatureM. trossulus Larvae Growth + Yaroslavtseva & Sergeeva2006Increased watertemperatureM. galloprovincialis Larvae GrowthSurvival+ / == / -His et al. 1989Increased watertemperatureM. galloprovincialis Larvae GrowthSurvival+ / =+ / =Lazo & Pita 2011Increased watertemperatureM. californianus Juvenile Growth + Almada-Villela et al. 1982Increased watertemperatureM. edulis Juvenile GrowthShell strength= / --Hiebenthal et al. 2013Ocean acidification M. galloprovincialis Embryo,larvaeDevelopmentSize--Kurihara et al. 2008a(continued on next page)117Table 5.1: Mussel responses to climate (continued)Climate variable Species Life stage Response E?ect ReferenceOcean acidification M. californianus Larvae Shell strengthShell areaTissue mass---Gaylord et al. 2011Ocean acidification M. edulis Juvenile GrowthShell strength==Hiebenthal et al. 2013Ocean acidification M. edulis Juvenile,adultCalcification rate - Gazeau et al. 2007Ocean acidification M. galloprovincialis Adult Calcification rate - Rodolfo-Metalpa et al.2011Ocean acidification M. edulis Adult Growth - Melzner:2011klTemperatureand OAM. edulis Juvenile GrowthShell strengthPossiblesynergyHiebenthal et al. 2013118Table 5.2: Sea star responses to climatically realistic increases in water temperature and ocean acidification.Climate variable Species Life stage Response E?ect ReferenceIncreased watertemperaturePatiriella regularis Embryo,larvaeFertilizationSurvivalSize=--Byrne et al. 2013Increased watertemperatureMeridiastra calcar Larvae DevelopmentSurvival--Nguyen et al. 2012Increased watertemperaturePisaster ochraceus Juvenile GrowthFeeding++Chapter 3, Gooding et al.2009Increased watertemperaturePisaster ochraceus Adult GrowthFeeding=+Sanford 2002Ocean acidification Odontaster validus Embryo,larvaeFertilizationSurvivalSize=--Gonzalez-Bernat et al.2012Ocean acidification Patiriella regularis Embryo,larvaeFertilizationSurvivalSize=--Byrne et al. 2013Ocean acidification Meridiastra calcar Larvae DevelopmentSurvival==Nguyen et al. 2012(continued on next page)119Table 5.2: Sea star responses to climate (continued)Climate variable Species Life stage Response E?ect ReferenceOcean acidification Crossaster papposus Larvae,juvenileGrowthSurvival+=Dupont et al. 2010aOcean acidification Pisaster ochraceus Juvenile GrowthFeeding++Chapter 3, Gooding et al.2009Ocean acidification Luidia clathrata Adult Righting responseArm regeneration==Schram et al. 2011Ocean acidification Asterias rubens Adult GrowthFeeding==Appelhans et al. 2012Temperatureand OAPatiriella regularis Embryo,larvaeFertilizationSurvivalAdditive Byrne et al. 2013Temperatureand OAMeridiastra calcar Larvae DevelopmentSurvivalAdditive Nguyen et al. 2012Temperatureand OAPisaster ochraceus Juvenile GrowthFeedingAdditive Chapter 3,Gooding et al.2009120Table 5.3: Model parametersParameter Description Units Value Reference / SourceMussel parametersLM(0) Mussel length at recruit-mentmm 5 See text (section 5.4.1)rM Max. mussel recruitment Individuals/day/m215 Hunt & Scheibling 1998;Menge et al. 2009KM Mussel carrying capacity Individuals/ m2 5500 Cusson & Bourget 2005; Pe-traitis 1995GrowthM Daily mussel shell growth mm/day See section5.4.1Millstein & O?Clair 2001dM Natural mussel mortalityrateProb. mortality/day/mussel1.37e4 Ceccherelli & Rossi 1984;Mallet & Carver 1995;Schneider et al. 2010Sea star parametersNP Sea star density Individuals/ m2 0.8 Feder 1956; Harley 1998;Paine 1976(continued on next page)121Table 5.3: Model parameters (continued)Parameter Description Units Value Reference / SourcedP Sea star mortality or emi-gration rateProb. mortality/day/sea star2.74e4 Connolly & Roughgarden1999; Ebert 1996; Menge1975LP (0) Sea star arm length at re-cruitmentmm 20 See section 5.4.2LP (max) Asymptotic arm length mm 180 Personal observationL Max. daily growth for seastars < 180mm mm/day 0.152 Feder 1956 (fitted breakpointlinear)Prey size preference coecients0(?L) Equation 5.2 mm M length 3.77 This chapter; Chapter 41(?L) Equation 5.2 M length /P arm length (mm)0.2310(L) Equation 5.2 mm M length 8.951(L) Equation 5.2 M length /P arm length (mm)0.051(continued on next page)122Table 5.3: Model parameters (continued)Parameter Description Units Value Reference / Sourcefmax Prob accepting preferredmussel size- 1Interfeeding time coecientsI(f) Equation 5.3 - 2.7474 This chapter; Chapter 4P (f) Equation 5.3 - -0.0066M (f) Equation 5.3 - -0.1196PM (f) Equation 5.3 - 0.00062123Table 5.4: Scaling factors applied to model parameters under di?erent climate sce-narios. Scaling factor of 1 = no change from control. Refer to the footnotes forsources.Parameter Temperature(+ 3 C)CO2 (OA)(+400 ppm)Temp.+(OA)Mussel recruitment 1 0.8a 0.8bMussel growth 1.15c 0.92d 1.07bSea star conversion rate 0.95e 1.11e 1.02eSea star feeding time 0.95e 0.98e 0.91eaGaylord et al. 2011; Talmage & Gobler 2011; Watson et al. 2009bAssumed to be additivecAlmada-Villela et al. 1982; Hiebenthal et al. 2013dGazeau et al. 2007; Hiebenthal et al. 2013; Rodolfo-Metalpa et al. 2011eChapter 3124Table 5.5: Mean sea star arm length (? SD) for each scenario. Values were calculatedusing the mean sea star size of the population (n = 24 sea stars), averaged acrossruns for each scenario (n = 10 runs for each scenario).Climate scenario Species a?ected Mean size (mm)Present-day NA 171 ? 3Temperature Mussels 173 ? 3OA Mussels 172 ? 6Temperature + OA Mussels 173 ? 6Temperature Sea stars 171 ? 3OA Sea stars 172 ? 5Temperature + OA Sea stars 173 ? 3Temperature Both 174 ? 3OA Both 172 ? 4Temperature + OA Both 174 ? 3125Chapter 6ConclusionClimate change is progressing at an alarming pace and has already caused bioticchanges ranging from shifts in species distributions to the collapse of entire biologicalcommunities (Bellard et al., 2012; Parmesan, 2006). The combination of multiplestressor e?ects, species-specific responses, and complex population and interspecificdynamics can make it challenging to predict how species and communities will re-spond to climate change. Nonetheless, understanding the biological consequences ofclimate change is a critically important endeavour.The complexity of predicting biological responses to climate change means thereis a need for studies addressing the e?ects of individual and combined climate vari-ables on multiple levels of biological organization, from organismal responses tointra- and interspecies dynamics. In my work, I addressed these issues in a ma-rine rocky shore ecosystem, with the overarching question of how abiotic and bioticfactors mediate the e?ects of climate change on rocky shores. I divided this ques-tion into three components: (1) direct organismal-level e?ects of multiple abioticchanges; (2) indirect e?ects of abiotic change on species interactions; and (3) e?ectsof combined direct (abiotic) and indirect (biotic) factors on populations. Using em-pirical and modelling techniques, I found that both abiotic and biotic factors canhave strong, and sometimes surprising, e?ects on species and population responsesto climate change.In this chapter, I will briefly summarize my approach and conclusions regard-126ing each of the three components listed above. I will then discuss the potentiallimitations of my methodologies and how they may have influenced my results andconclusions, suggest possible improvements or extensions for future studies, andconsider my results in the broader context of ecological responses to climate change.Finally, I will make a few broad recommendations regarding priorities for futureresearch in this field.6.1 Direct e?ects of multiple climate variables onindividual organismsIn Chapters 2 and 3, I presented two case studies of the organismal-level e?ects ofmultiple abiotic variables on rocky intertidal invertebrates. In both chapters I useda factorial experimental design to manipulate climate variables individually and intandem. This design allowed me to explicitly test for the presence of non-additivee?ects when multiple abiotic variables were combined. In Chapter 2, I measured howacute exposure to low salinity and a CO2-driven pH reduction a?ected the ability oftwo littorine snails (Littorina plena and L. sitkana) to tolerate acute thermal stress.I found that snails exposed 15 or 20 psu salinity for 24 hours prior to thermal stresshad lower survival rates compared to snails at 28 - 30 psu. However, a ? 0.4 unitreduction in pH (corresponding to a ? 1000 ppm increase in pCO2) had little orno e?ect on their response to thermal stress. In Chapter 3, I measured growthand feeding behaviours of juvenile sea stars (Pisaster ochraceus) during a 10-weekexposure to increased seawater temperature (+ 3 C) and ocean acidification (OA; -0.06 pH units). I found that temperature and OA had positive and additive e?ectson sea star growth and feeding rates.The lack of detrimental e?ects of OA on sea stars and snails in my studies127contrasts with the majority of past research on the e?ects of OA on marine inver-tebrates. Numerous studies have shown negative responses in a range of organisms(reviewed by Kroeker et al., 2013), making it tempting to assume that the e?ects ofOA will always be detrimental. My work demonstrates that this is not always thecase, and that some organisms may even benefit from OA. Several recent studieshave found similar positive responses to OA in other species (Byrne et al., 2009;Dupont et al., 2010a; Nguyen et al., 2012). The implications of these positive e?ectsextend far beyond the individual organisms, and have the potential to alter thestrength of species interactions, particularly between species that respond to OA inopposite directions.One aspect that should be considered when interpreting my results from Chap-ters 2 and 3 is the scale of my experimental manipulations, both in terms of theduration of the experimental period as well as the magnitude of abiotic change. InChapter 2, I exposed snails to relatively large but short-term (24 hours) changes insalinity and OA. These experiments were designed to detect the potential impactsof transient changes in pH and salinity that are experienced on many rocky shores(Helmuth et al., 2010; Iacarella & Helmuth, 2012), and the results may not be rep-resentative of chronically stressful conditions for these taxa. For example, previouswork has shown that chronic exposure to OA (0.4 - 1.3 unit reduction in pH) canalter the shell growth and morphology of gastropods (Bibby et al., 2007; Ellis et al.,2009).In contrast, my experiments in Chapter 3 simulated moderate but sustained in-creases in temperature and OA. By maintaining sea stars in experimental conditionsfor 3 or 10 weeks, depending on the experiment, I was able to measure responses suchas growth and calcification rates that would have been dicult to detect in shorterexperiments. That said, the sea stars remained in the juvenile stage throughout my128experiments, so I only measured the e?ects of climate on a single life-history stage.Exposure as larvae or adults could produce di?erent responses such as reduced sur-vival or fecundity (see Table 5.2 for examples). Some variables, such as OA, canalso have carry-over e?ects from one life-history stage to the next (Dupont et al.,2012; Parker et al., 2012). Experiments spanning multiple life-history stages arecurrently rare in the field of OA ecology, but will be necessary to gain a more com-plete understanding of how climate change will a?ect organisms throughout theirlifespans.An additional consideration in the interpretation of my findings from Chapter 3is that the sea stars were continually submerged during the experiments. Althoughsome Pisaster ochraceus live subtidally and therefore remain submerged throughouttheir lives, most individuals are emersed at least occasionally during low tide. Recentwork has shown that aerial temperatures can influence sea star growth and feedingrates, and that these e?ects can be mediated by the degree of temporal overlap withelevated water temperatures (Pincebourde et al., 2012, 2008). This suggests that myfindings regarding the e?ects of water temperature on sea stars should be consideredin the context of coincident changes in other abiotic conditions aside from OA.6.2 E?ects of climate change on species interactionsThe results from Chapter 3 and from similar studies on mussels (Table 5.1) suggestthat changes in growth rates due to future climate change could potentially alter thebody size-frequency distributions of populations of Pisaster ochraceus and Mytilusspp. In Chapter 4, I quantified the potential indirect e?ects of these changes on thepredator-prey interaction between the sea star Pisaster ochraceus and its preferredprey in the Strait of Georgia, the mussel Mytilus trossulus. I measured sea star feed-ing behaviours across a range of sea star and mussel body sizes in order to simulate129how climate-driven changes in body sizes might alter this species interaction.As I had anticipated, the body size of both predator and prey played a sig-nificant role in determining sea star feeding behaviours. Many past studies haveshow a similar link between sea star feeding behaviours and prey size (McClintock& Robnett, 1986; Paine, 1976; Sommer et al., 1999); however, to my knowledge,no studies have systematically manipulated both sea star and mussel body size asI did in order to quantify feeding behaviours across the size range of both species.Furthermore, relatively few studies on the Pisaster-Mytilus interaction have usedthe mussel Mytilus trossulus, despite the important role it plays on many shelteredrocky shore communities in the Northeastern Pacific. My findings provide a baselinefor understanding the size-dependencies of this species interaction. Studies of thistype are especially relevant given the significant role that climate-driven changes inbody size are likely to play in determining the ecological outcomes of climate change(Brose et al., 2012).Certain aspects of my experimental design in Chapter 4 may pose limitationsto the interpretation of my results. In particular, I did not allow mussels to attachto a substrate prior to being fed to sea stars. This decision was mainly due to thediculty I had in coaxing smaller mussels to attach to a suitable substrate in therecirculating seawater table. McClintock & Robnett (1986) suggested that Pisasterochraceus? preference for medium-sized individuals of the mussel Mytilus californi-anus is driven primarily by the disproportionate diculty in detaching larger musselsfrom the substrate. It is possible that the lack of byssal thread attachments in mystudy led me to underestimate handling time and overestimate feeding rates. How-ever, the mussel species used in my study, M. trossulus, has relatively weak byssalthread attachments (Bell & Gosline, 1997), and I have seen both small and large seastars feed on M. trossulus without detaching them from the substrate. This leads130me to believe that dislodgement of mussels is unlikely to be a dominant factor inthe handling and feeding costs for sea stars preying on M. trossulus. Nonetheless,it would be beneficial to conduct feeding trials using multiple mussel species andvarious degrees of attachment to the substrate to determine the costs to predators.This is of particular interest in light of recent work showing that OA may impactthe strength of byssal thread attachments (O?Donnell, 2013).6.3 Direct and indirect e?ects of climate change viaintra- and interspecific dynamicsIn Chapter 5, I simulated how the combination of multiple climate variables, organ-ismal level responses, and species interactions a?ects mussel population dynamics.By incorporating empirical estimates of individual sea star and mussel responsesto seawater temperature (+ 3 C) and OA (+ 400 ppm CO2) into a predator-preymodel, I was able to predict mussel population attributes under various biotic andabiotic scenarios. My model outcomes demonstrate that the direct e?ects of climatechange on mussel populations may be exacerbated by the presence of a predator,and that the e?ects of sea star predation on mussels may be magnified by the directe?ects of climate change on both species. I suggest two possible mechanisms for thisdynamic. The first possibility is that predation-driven changes in mussel populationsresult in higher rates of mussel growth and recruitment compared to populations inthe absence of predators. Since temperature and OA have direct e?ects on growthand recruitment rates in my model, climate change will have a greater impact onpopulations where these rates are relatively high. A second mechanism by whichpredation may mediate the e?ects of climate on mussels is via the combined e?ects ofclimate change and predation on mussel size. If climate-induced changes to mussel131growth rates alter the size-frequency distribution of the population, this will a?ectsea star feeding rates, which in turn will impact mussel size and abundance. Thesetwo potential mechanisms are not mutually exclusive, and I believe they probablyboth play a role.As happens with any model, I made certain assumptions when parameterizingmy model. Two assumptions in particular - the strength of density-dependent re-cruitment and density-independent larval supply - merit future sensitivity analysesto determine the degree to which their values influence the model outcomes. Thelack of a direct link between adult mussel abundance and larval supply in my modelcould lead to underestimates of the e?ects of climate and predation on mussel pop-ulations, since a large enough decline in the abundance of reproductively maturemussels could conceivably reduce larval supply to the point of limiting populationsize. However, there are several reasons why I believe this is unlikely to be of greatconcern. The relative contribution of pre- and post-settlement factors to adultabundance is under debate, and studies comparing the two mechanisms in musselpopulations present mixed results (see review by Fraschetti et al., 2002). In somecases, small to moderate decreases in adult abundance may actually benefit newlyrecruited mussels by reducing competition for space (Petraitis, 1995). It seems likelythat the issue of open versus closed population dynamics would primarily impactpopulations that are not subject to strong competition for space or other density-dependent factors that can limit recruitment (Fraschetti et al., 2002). Additionally,the magnitude of future changes to seawater temperature and OA will vary region-ally (Helmuth et al., 2006a; Solomon et al., 2007; Wootton & Pfister, 2012), makingit unlikely that all mussel populations will be equally impacted by climate change.Populations that are less a?ected could serve as larval sources for populations whereadult abundance has been more heavily impacted by climate, provided the distance132between the populations is not too great (Dias, 1996).Several other potentially important facets of real-life populations of mussels andsea stars were not included from my model. Although no model can fully representin situ dynamics, these factors should be taken into account when interpreting theoutcomes of my model. I have discussed some of these in section 5.5, but oneaspect in particular deserves further attention. I did not include temporal variationin the biotic or abiotic parameters in my model. This decision was primarily dueto the high degree of natural variation in the timing of these dynamics, whichmade it extremely dicult to select a single temporal scenario from the literature.However, I recognize that, in reality, both biotic and abiotic parameters can varywidely throughout the year. Growth, recruitment, and feeding activity often varyseasonally (Hunt & Scheibling, 1998; Mauzey, 1966; Nagarajan et al., 2006), andthe degree of temporal overlap between sea star and mussel cycles could have majore?ects on population dynamics.I also omitted seasonal and stochastic changes in abiotic conditions, limiting theclimate scenarios in my model to stable increases in seawater temperature and OA.Numerous studies have demonstrated the strong role that episodic and acute abioticconditions can play in population dynamics (Denny et al., 2009; Paine et al., 1998;Tomanek & Helmuth, 2002). In particular, periodic spikes in aerial temperatures arelikely to influence sea star feeding behaviours (Tsuchiya, 1983). Stochastic environ-mental conditions can also cause catastrophic mortality events of sessile organismssuch as mussels (Carrington et al., 2009; Denny et al., 2009; Harley, 2008). Mymodel includes the capacity to manipulate both the magnitude and frequency ofnatural mussel mortality events, but I elected not to include this variability in mycurrent model presentation because I felt it did not contribute to my main ques-tion. The mathematical complexity that would be required to incorporate other133temporal aspects into my model is beyond the scope of this thesis, but they serveas interesting areas for possible refinement to my model in the future.A variety of models have been developed to predict the biological consequencesof climate change. Demographic models often simulate climate e?ects on vital ratesto make specific predictions about populations of interest (e.g. Jenouvrier et al.,2009), while bioclimate envelope models use predicted temporal and spatial changesin habitat and climate conditions to forecast species ranges and phenology (e.g. Che-ung et al., 2009). However, both of these approaches tend to focus on single species,overlooking possible impacts of species interactions. End-to-end models integratenumerous abiotic and biotic dynamics to provide whole-ecosystem predictions (e.g.Fulton, 2011), but these broad-scale models provide little by way of a mechanis-tic understanding of climate responses. In contrast to these approaches, my modeluses organismal-level predictions in the context of intra- and interspecific dynam-ics, with the goal of understanding the relative importance of abiotic versus bioticfactors in driving climate responses. A handful of similar models have been usedin attempts to answer these questions. Sorte & White (2013) developed a multi-species demographic model to predict how climate alters community compositionand species interaction strengths. Poloczanska et al. (2008) used a size-structuredmatrix model to simulate competition between two barnacle species under variousscenarios of climate warming. In both of these studies, the authors concluded thatspecies interactions are likely to play a key role in determining the outcomes ofclimate change.6.4 Conclusions and recommendationsMy thesis demonstrates that multiple abiotic variables and the presence of interact-ing species can both play significant roles in determining species responses to climate134change. By combining empirical estimates of the e?ects of climate change on indi-vidual organisms and interspecific interactions with simulated population dynamicsin a predator-prey simulation model, I was able to predict population-level responsesto climate change. Few studies have taken such an approach to understanding andpredicting the ecological e?ects of climate change, and my work provides an exam-ple of how empirical and theoretical techniques can be combined to estimate andcompare the relative importance of direct (abiotic) and indirect (biotic) factors indetermining the biological consequences of climate change.Although I only studied the e?ects of climate change on a handful of species, thekey ecological roles that they play in their communities allow me to make predictionsregarding community-level changes. Littorine snails are often the dominant herbi-vores on high intertidal shores (McQuaid, 1996). Their grazing activities can alterthe structure of algal assemblages (Aquilino & Stachowicz, 2012; Tarpley, 1992;Williams et al., 2013) and indirectly a?ect species that rely on algae for food orshelter (Harley, 2006; McQuaid, 1996). Increased mortality of snails due to morefrequent temporal alignment of acute low salinity and thermal stress events couldalter the density of littorine snails, potentially driving changes to the structure ofthe overall community due to reduced grazing pressure. Similarly, climate-drivenchanges to the predator-prey interaction between Pisaster ochraceus and Mytilustrossulus would likely alter the structure of mid-intertidal communities by reducingthe amount of rocky substrate occupied by mussels. The increased availability ofopen space would facilitate a higher diversity of large organisms such as kelps, whilethe diversity and abundance of small infaunal organisms that rely on mussel bedsfor habitat would decline (Harley, 2011; Paine, 1974). Because of the keystone andfoundation roles that my study species play in their communities, their responsesto climate change could overwhelm the e?ects of less ecologically important species135and largely determine the overall community responses to climate change.The field of climate change ecology has made great progress toward understand-ing how organisms and communities will be impacted by climate change, but manychallenges still lie ahead. In the current chapter, I have mentioned several av-enues for future work. For example, it will be important to identify vulnerable lifehistory stages that may serve as bottlenecks and dominate a species? response toclimate change (Byrne, 2011, 2012). Longer-term experiments are needed to iden-tify whether the e?ects of climate change will carry over across multiple life historystages or even multiple generations (Dupont et al., 2012; Parker et al., 2012). Thereis also an urgent need to better understand the role that phenotypic and geneticvariation will play in organisms? ability to adapt to future abiotic conditions (Hueyet al., 2012; Schulte et al., 2011; Somero, 2010; Sunday et al., 2011).The single strongest recommendation I can make is for future studies to incorpo-rate both abiotic and biotic changes. The results of my work in Chapter 5 providean example of how the presence of an interacting species could mediate the e?ects ofclimate change, and other authors have shown similar results that support the needfor a multi-species approach in climate studies (Davis et al., 1998; Singer et al., 2012;Van der Putten et al., 2010). Understanding and predicting how climate change willa?ect biological systems can seem a formidable task at times. As I demonstrate inmy thesis, a multi-pronged approach that combines empirical and modelling tech-niques can be useful for understanding and predicting the consequences of climatechange at multiple levels of biological organization.On a broader scale, we should expand our e?orts to include the impacts ofnon-climatic stressors, since organisms and ecosystems are facing a myriad of otheranthropogenically-driven changes (Halpern et al., 2008). Overfishing (Dulvy et al.,1362004; Hutchings et al., 2012) and invasive species (Mainka & Howard, 2010) aredirectly driving biotic changes, while eutrophication and habitat degradation arealtering the abiotic conditions experienced by organisms (Carstensen et al., 2011;Halpern et al., 2008). Furthermore, these factors can interact with one another(Burnell et al., 2013; Crain et al., 2008; Mainka & Howard, 2010; Russell et al.,2009). For example, climate warming can help facilitate species invasions and diseasetransmission (Cockrell & Sorte, 2013; Lester et al., 2007; Wolkovich et al., 2013), andoverfishing can increase the susceptibility of corals to warming-induced bleachingevents (Carilli et al., 2009; Darling et al., 2013). It will be important to considerhow these non-climatic changes might mediate responses to climate change, andvice-versa.We are already committed to a minimum trajectory of climate change throughthe end of this century (Solomon et al., 2009), and it seems unlikely that politicaland social changes will occur in the near future on the global scale required to slow orhalt climate change. However, we may be able to manage other anthropogenically-driven changes (e.g. fishing, eutrophication) at smaller spatial and temporal scalesin order to reduce the overall impact on ecosystems that are particularly threatenedby climate change (see discussion by Brown et al. 2013 and Bille? et al. 2013). Thus, itmay be most fruitful to put our energies toward identifying species and habitats thatare particularly vulnerable to future climate change impacts, and then protectingthese from other anthropogenic stressors that might exacerbate these climate e?ects.Using techniques such as those presented in this thesis could help us determine howthese combined anthropogenically-driven abiotic (climatic as well as non-climaticstressors) and biotic changes will impact biological communities, as well as howthey might interact, allowing us to best prioritize our research and managemente?orts.137BibliographyAlbright, R., Bland, C., Gillette, P., Serafy, J. E., Langdon, C., & Capo, T. R.(2012). Juvenile growth of the tropical sea urchin Lytechinus variegatus exposed tonear-future ocean acidification scenarios. Journal of Experimental Marine Biologyand Ecology, 426-427, 12?17.Aljetlawi, A. A., Sparrevik, E., & Leonardsson, K. (2004). Prey-predator size-dependent response: derivation and rescaling to the real world. Journal of AnimalEcology, 73(2), 239?252.Allan, B. J. M., Domenici, P., McCormick, M. I., Watson, S.-A., & Munday, P. L.(2013). Elevated CO2 a?ects predator-prey interactions through altered perfor-mance. PLoS ONE, 8(3), e58520.Almada-Villela, P. C., Davenport, J., & Gru?ydd, L. D. (1982). The e?ects of tem-perature on the shell growth of young Mytilus edulis L. Journal of ExperimentalMarine Biology and Ecology, 59(2), 275?288.Angilletta, Jr, M. J. (2001). Thermal and physiological constraints on energy assim-ilation in a widespread lizard (Sceloporus undulatus). Ecology, 82(11), 3044?3056.Appelhans, Y. S., Thomsen, J., Pansch, C., Melzner, F., & Wahl, M. (2012). Sourtimes: seawater acidification e?ects on growth, feeding behaviour and acid?basestatus of Asterias rubens and Carcinus maenas . Marine Ecology Progress Series,459, 85?98.Aquilino, K. M. & Stachowicz, J. J. (2012). Seaweed richness and herbivory increaserate of community recovery from disturbance. Ecology, 93(4), 879?890.Asnaghi, V., Chiantore, M., Mangialajo, L., Gazeau, F., Francour, P., Alliouane,S., & Gattuso, J.-P. (2013). Cascading e?ects of ocean acidification in a rockysubtidal community. PLoS ONE, 8(4), e61978.Atkinson, D. (1995). E?ects of temperature on the size of aquatic ectotherms:exceptions to the general rule. Journal of Thermal Biology, 20(1/2), 61?74.Avens, A. C. & Sleigh, M. A. (1965). Osmotic balance in gastropod molluscs?Isome marine and littoral gastropods. Comparative Biochemistry and Physiology,16(1), 121?141.138Baker, A. C., Glynn, P. W., & Riegl, B. (2008). Climate change and coral reefbleaching: An ecological assessment of long-term impacts, recovery trends andfuture outlook. Estuarine, Coastal and Shelf Science, 80(4), 435?471.Barnier, G. W., Sheehan, M. V., & Williams, D. C. (1975). The production andsecretion of digestive enzymes in the Purple Seastar Pisaster ochraceus . MarineBiology, 29, 261?266.Barry, J. P., Baxter, C. H., Sagarin, R. D., & Gilman, S. E. (1995). Climate-related,long-term faunal changes in a california rocky intertidal community. Science,267(5198), 672?675.Bell, E. C. & Gosline, J. M. (1997). Strategies for life in flow: tenacity, morphometry,and probability of dislodgment of two Mytilus species. Marine Ecology ProgressSeries, 159, 197?208.Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., & Courchamp, F. (2012).Impacts of climate change on the future of biodiversity. Ecology Letters, 15(4),365?377.Benson, K. R. (2002). The study of vertical zonation on rocky intertidal shores?ahistorical perspective. American Zoologist, 42(4), 776?779.Bergen, M. (1968). The salinity tolerance limits of the adults and early-stage em-bryos of Balanus glandula Darwin, 1854 (Cirripedia, Thoracica). Crustaceana,15(3), 229?234.Berger, V. J. & Kharazova, A. D. (1997). Mechanisms of salinity adaptations inmarine molluscs. Hydrobiologia, 355(1-3), 115?126.Berger, V. Y., Khlebovich, V. V., Kovaleva, N. M., & Natochin, Y. V. (1978).The changes of ionic composition and cell volume during adaptation of molluscs(Littorina) to lowered salinity. Comparative Biochemistry and Physiology PartA: Physiology, 60(4), 447?452.Berlow, E. L., Navarrete, S. A., Briggs, C. J., Power, M. E., & Menge, B. A. (1999).Quantifying variation in the strengths of species interactions. Ecology, 80(7),2206?2224.Bertness, M. D., Leonard, G. H., Levine, J. M., Schmidt, P. R., & Ingraham, A. O.(1999). Testing the relative contribution of positive and negative interactions inrocky intertidal communities. Ecology, 80(8), 2711?2726.Bibby, R., Cleall-Harding, P., Rundle, S., Widdicombe, S., & Spicer, J. (2007).Ocean acidification disrupts induced defences in the intertidal gastropod Littorinalittorea. Biology Letters, 3(6), 699?701.139Bille?, R., Kelly, R., Biastoch, A., Harrould-Kolieb, E., Herr, D., Joos, F., Kroeker,K., La?oley, D., Oschlies, A., & Gattuso, J.-P. (2013). Taking Action AgainstOcean Acidification: A Review of Management and Policy Options. Environmen-tal Management, 52(4), 761?779.Blanchette, C. A., Helmuth, B., & Gaines, S. D. (2007). Spatial patterns of growth inthe mussel, Mytilus californianus, across a major oceanographic and biogeographicboundary at Point Conception, California, USA. Journal of Experimental MarineBiology and Ecology, 340(2), 126?148.Borthagaray, A. I. & Carranza, A. (2007). Mussels as ecosystem engineers: Theircontribution to species richness in a rocky littoral community. Acta Oecologica,31(3), 243?250.Branch, T. A., DeJoseph, B. M., Ray, L. J., & Wagner, C. A. (2012). Impactsof ocean acidification on marine seafood. Trends in Ecology & Evolution, 28(3),178?186.Bri?a, M., de la Haye, K., & Munday, P. L. (2012). High CO2 and marine animalbehaviour: Potential mechanisms and ecological consequences. Marine PollutionBulletin, 64(8), 1519?1528.Brose, U., Dunne, J. A., Montoya, J. M., Petchey, O. L., Schneider, F. D., &Jacob, U. (2012). Climate change in size-structured ecosystems. PhilosophicalTransactions of the Royal Society B: Biological Sciences, 367(1605), 2903?2912.Brown, C. J., Saunders, M. I., Possingham, H. P., & Richardson, A. J. (2013).Managing for Interactions between Local and Global Stressors of Ecosystems.PLoS ONE, 8(6), e65765.Brown, J. H., Gillooly, J. F., Allen, A. P., Savage, V. M., & West, G. B. (2004).Toward a metablic theory of ecology. Ecology, 85(7), 1771?1789.Bulthuis, D. A. (1996). Coastal habitats in Padilla Bay, Washington: A review.Washington State Dept. of Ecology, Mt. Vernon, WA.Burnell, O. W., Russell, B. D., Irving, A. D., & Connell, S. D. (2013). Eutrophicationo?sets increased sea urchin grazing on seagrass caused by ocean warming andacidification. Marine Ecology Progress Series, 485, 37?46.Burrows, M. & Hughes, R. N. (1991). Optimal foraging decisions by dogwhelks,Nucella lapillus (L.): influences of mortality risk and rate-constrained digestion.Functional Ecology, 5, 461?475.Burrows, M. T., Schoeman, D. S., Buckley, L. B., Moore, P., Poloczanska, E. S.,Brander, K. M., Brown, C., Bruno, J. F., Duarte, C. M., Halpern, B. S., Holding,140J., Kappel, C. V., Kiessling, W., O?Connor, M. I., Pandolfi, J. M., Parmesan, C.,Schwing, F. B., Sydeman, W. J., & Richardson, A. J. (2011). The Pace of ShiftingClimate in Marine and Terrestrial Ecosystems. Science, 334(6056), 652?655.Busch, D. S., Harvey, C. J., & McElhany, P. (2013). Potential impacts of oceanacidification on the Puget Sound food web. ICES Journal of Marine Science,70(4), 823?833.Byrne, M. (2011). Impact of ocean warming and ocean acidification on marineinvertebrate life history stages: vulnerabilities and potential for persistence in achanging ocean. Oceanography and Marine Biology: An Annual Review, 49, 1?42.Byrne, M. (2012). Global change ecotoxicology: Identification of early life historybottlenecks in marine invertebrates, variable species responses and variable ex-perimental approaches. Marine Environmental Research, 76, 3?15.Byrne, M., Gonzalez-Bernat, M., Doo, S., Foo, S., Soars, N., & Lamare, M. (2013).E?ects of ocean warming and acidification on embryos and non-calcifying larvaeof the invasive sea star Patiriella regularis. Marine Ecology Progress Series, 473,235?246.Byrne, M., Ho, M., Selvakumaraswamy, P., Nguyen, H. D., Dworjanyn, S. A., &Davis, A. R. (2009). Temperature, but not pH, compromises sea urchin fertiliza-tion and early development under near-future climate change scenarios. Proceed-ings of the Royal Society B: Biological Sciences, 276(1663), 1883?1888.Calabrese, E. J. & Baldwin, L. A. (2001). Hormesis: A generalizable and unifyinghypothesis. Critical Reviews in Toxicology, 31(4-5), 353?424.Caldeira, K. & Wickett, M. E. (2003). Oceanography: anthropogenic carbon andocean pH. Nature, 425(6956), 365.Carilli, J. E., Norris, R. D., Black, B. A., Walsh, S. M., & McField, M. (2009). Localstressors reduce coral resilience to bleaching. PLoS ONE, 4(7), e6324.Carrington, E., Moeser, G. M., Dimond, J., Mello, J. J., & Boller, M. L. (2009).Seasonal disturbance to mussel beds: Field test of a mechanistic model predictingwave dislodgment. Limnology and Oceanography, 54(3), 978?986.Carstensen, J., Sa?nchez-Camacho, M., Duarte, C. M., Krause-Jensen, D., & Marba`,N. (2011). Connecting the Dots: Responses of Coastal Ecosystems to ChangingNutrient Concentrations. Environmental Science & Technology, 45(21), 9122?9132.Cavey, M. J. & Markel, K. (1994). Microscopic Anatomy of Invertebrates: Echino-dermata, volume 14. Wiley-Liss, New York.141Ceccherelli, V. U. & Rossi, R. (1984). Settlement, growth and production of themussel Mytilus galloprovincialis. Marine Ecology Progress Series, 16(1), 173?184.Chapperon, C. & Seuront, L. (2010). Behavioral thermoregulation in a tropicalgastropod: links to climate change scenarios. Global Change Biology, 17(4), 1740?1749.Cheung, S. G. & Lam, S. W. (1995). E?ect of salinity, temperature and acclimationon oxygen consumption of Nassarius festivus (Powys, 1835)(Gastropoda: Nas-sariidae). Comparative Biochemistry and Physiology Part A: Physiology, 111(4),625?631.Cheung, W. W. L., Lam, V. W. Y., Sarmiento, J. L., Kearney, K., Watson, R., &Pauly, D. (2009). Projecting global marine biodiversity impacts under climatechange scenarios. Fish and Fisheries, 10(3), 235?251.Clarke, A. P., Mill, P. J., & Grahame, J. (2000). Biodiversity in Littorina species(Mollusca: Gastropoda): a physiological approach using heat-coma. Marine Bi-ology, 137, 559?565.Cockrell, M. L. & Sorte, C. J. B. (2013). Predicting climate-induced changes in pop-ulation dynamics of invasive species in a marine epibenthic community. Journalof Experimental Marine Biology and Ecology, 440, 42?48.Collins, B. (1998). Preliminary assessment of historic conditions of the Skagit Riverin the Fir Island area: Implications for salmonid habitat restoration. Reportprepared for the Skagit System Cooperative, LaConner, Washington.Connell, J. H. (1961). The influence of interspecific competition and other factorson the distribution of the barnacle Chthamalus stellatus . Ecology, 42(4), 710?723.Connolly, S. R. & Roughgarden, J. (1999). Theory of marine communities: com-petition, predation, and recruitment-dependent interaction strength. EcologicalMonographs, 69(3), 277?296.Crain, C. M., Kroeker, K., & Halpern, B. S. (2008). Interactive and cumulativee?ects of multiple human stressors in marine systems. Ecology Letters, 11(12),1304?1315.Cusson, M. & Bourget, E. (2005). Small-scale variations in mussel (Mytilus spp.)dynamics and local production. Journal of Sea Research, 53(4), 255?268.Darling, E. S. & Co?te?, I. M. (2008). Quantifying the evidence for ecological synergies.Ecology Letters, 11(12), 1278?1286.Darling, E. S., McClanahan, T. R., & Co?te?, I. M. (2013). Life histories predict coralcommunity disassembly under multiple stressors. Global Change Biology, 19(6),1930?1940.142Davenport, J. & Davenport, J. L. (2007). Interaction of thermal tolerance andoxygen availability in the eurythermal gastropods Littorina littorea and Nucellalapillus . Marine Ecology Progress Series, 332, 167?170.Davis, A. J., Jenkinson, L. S., Lawton, J. H., Shorrocks, B., & Wood, S. (1998).Making mistakes when predicting shifts in species range in response to globalwarming. Nature, 391, 783? la Haye, K. L., Spicer, J. I., Widdicombe, S., & Bri?a, M. (2012). Reduced pH seawater disrupts chemo-responsive behaviour in an intertidal crustacean. Journalof Experimental Marine Biology and Ecology, 412, 134?140.Denny, M. W., Daniel, T. L., & Koehl, M. (1985). Mechanical limits to size inwave-swept organisms. Ecological Monographs, 55(1), 69?102.Denny, M. W., Hunt, L. J., Miller, L. P., & Harley, C. D. (2009). On the predictionof extreme ecological events. Ecological Monographs, 79(3), 397?421.Deutsch, C. A., Tewksbury, J. J., Huey, R. B., Sheldon, K. S., Ghalambor, C. K.,Haak, D. C., & Martin, P. R. (2008). Impacts of climate warming on terrestrialectotherms across latitude. Proceedings of the National Academy of Sciences ofthe United States of America, 105(18), 6668?6672.Dias, P. C. (1996). Sources and sinks in population biology. Trends in Ecology &Evolution, 11(8), 326?330.Donahue, M. J., Desharnais, R. A., Robles, C. D., & Arriola, P. (2011). Mussel BedBoundaries as Dynamic Equilibria: Thresholds, Phase Shifts, and AlternativeStates. The American Naturalist, 178(5), 612?625.Doney, S. C., Ruckelshaus, M., Emmett Du?y, J., Barry, J. P., Chan, F., English,C. A., Galindo, H. M., Grebmeier, J. M., Hollowed, A. B., Knowlton, N., Polovina,J., Rabalais, N. N., Sydeman, W. J., & Talley, L. D. (2012). Climate changeimpacts on marine ecosystems. Annual Review of Marine Science, 4(1), 11?37.Drinkwater, K. F., Beaugrand, G., Kaeriyama, M., Kim, S., Ottersen, G., Perry,R. I., Po?rtner, H.-O., Polovina, J. J., & Takasuka, A. (2010). On the processeslinking climate to ecosystem changes. Journal of Marine Systems, 79(3-4), 374?388.Dufresne, J. L., Foujols, M. A., Denvil, S., Caubel, A., Marti, O., Aumont, O.,Balkanski, Y., Bekki, S., Bellenger, H., Benshila, R., Bony, S., Bopp, L., Bracon-not, P., Brockmann, P., Cadule, P., Cheruy, F., Codron, F., Cozic, A., Cugnet,D., Noblet, N., Duvel, J. P., Ethe?, C., Fairhead, L., Fichefet, T., Flavoni, S.,Friedlingstein, P., Grandpeix, J. Y., Guez, L., Guilyardi, E., Hauglustaine, D.,Hourdin, F., Idelkadi, A., Ghattas, J., Joussaume, S., Kageyama, M., Krinner,143G., Labetoulle, S., Lahellec, A., Lefebvre, M. P., Lefevre, F., Levy, C., Li, Z. X.,Lloyd, J., Lott, F., Madec, G., Mancip, M., Marchand, M., Masson, S., Meur-desoif, Y., Mignot, J., Musat, I., Parouty, S., Polcher, J., Rio, C., Schulz, M.,Swingedouw, D., Szopa, S., Talandier, C., Terray, P., Viovy, N., & Vuichard, N.(2013). Climate change projections using the IPSL-CM5 Earth System Model:from CMIP3 to CMIP5. Climate Dynamics, 40(9-10), 2123?2165.Dulvy, N. K., Freckleton, R. P., & Polunin, N. V. C. (2004). Coral reef cascadesand the indirect e?ects of predator removal by exploitation. Ecology Letters, 7(5),410?416.Dupont, S., Dorey, N., Stumpp, M., Melzner, F., & Thorndyke, M. (2012). Long-term and trans-life-cycle e?ects of exposure to ocean acidification in the green seaurchin Strongylocentrotus droebachiensis. Marine Biology, 160(8), 1835?1843.Dupont, S., Lundve, B., & Thorndyke, M. (2010a). Near future ocean acidificationincreases growth rate of the lecithotrophic larvae and juveniles of the sea starCrossaster papposus. Journal of Experimental Zoology Part B: Molecular andDevelopmental Evolution, 314(5), 382?389.Dupont, S., Ortega-Mart??nez, O., & Thorndyke, M. (2010b). Impact of near-futureocean acidification on echinoderms. Ecotoxicology, 19(3), 449?462.Dupont, S. & Thorndyke, M. C. (2009). Impact of CO2-driven ocean acidificationon invertebrates early life-history ? What we know, what we need to know andwhat we can do. Biogeosciences Discussions, 6(2), 3109?3131.Ebert, T. A. (1996). The consequences of broadcasting, brooding, and asexualreproduction in echinoderm metapopulations. Oceanologica Acta, 19(3-4), 217?226.Ellis, R., Bersey, J., Rundle, S., Hall-Spencer, J., & Spicer, J. (2009). Subtle butsignificant e?ects of CO2 acidified seawater on embryos of the intertidal snail.Aquatic Biology, 5, 41?48.Emlen, J. M. (1966). The role of time and energy in food preference. The AmericanNaturalist, 100(916), 611?617.Emmerson, M. C. & Ra?aeli, D. (2004). Predator-prey body size, interactionstrength and the stability of a real food web. Journal of Animal Ecology, 73(3),399?409.Evans, R. G. (1948). The lethal temperatures of some common British littoralmolluscs. The Journal of Animal Ecology, 17(2), 165?173.144Fabry, V. J., Seibel, B. A., Feely, R. A., & Orr, J. C. (2008). Impacts of oceanacidification on marine fauna and ecosystem processes. ICES Journal of MarineScience: Journal du Conseil, 65(3), 414?432.Fearman, J. & Moltschaniwskyj, N. A. (2010). Warmer temperatures reduce ratesof gametogenesis in temperate mussels, Mytilus galloprovincialis . Aquaculture,305(1), 20?25.Feder, H. M. (1955). On the methods used by the starfish Pisaster ochraceus inopening three types of bivalve molluscs. Ecology, 36(4), 764?767.Feder, H. M. (1956). Natural history studies on the starfish Pisaster ochraceus(Brandt, 1835) in the Monterey Bay area. PhD thesis, Stanford University.Feely, R. A., Doney, S. C., & Cooley, S. R. (2009). Ocean acidification: presentconditions and future changes in a high-CO2 world. Oceanography, 22(4), 36?47.Feely, R. A., Sabine, C. L., Hernandez-Ayon, J. M., Ianson, D., & Hales, B. (2008).Evidence for upwelling of corrosive ?acidified? water onto the continental shelf.Science, 320(5882), 1490?1492.Feely, R. A., Sabine, C. L., Lee, K., Berelson, W., Kleypas, J., Fabry, V. J., &Millero, F. J. (2004). Impact of anthropogenic CO2 on the CaCO3 system in theoceans. Science, 305(5682), 362?366.Fenberg, P. B. & Kaustuv, R. (2008). Ecological and evolutionary consequencesof size-selective harvesting: how much do we know? Molecular Ecology, 17(1),209?220.Ferrari, M. C. O., McCormick, M. I., Munday, P. L., Meekan, M. G., Dixson, D. L.,Lonnstedt, O?., & Chivers, D. P. (2011). Putting prey and predator into the CO2equation - qualitative and quantitative e?ects of ocean acidification on predator-prey interactions. Ecology Letters, 14(11), 1143?1148.Findlay, H. S., Kendall, M. A., Spicer, J. I., & Widdicombe, S. (2009). Post-larvaldevelopment of two intertidal barnacles at elevated CO2 and temperature. MarineBiology, 157(4), 725?735.Findley, A. M., Belisle, B. W., & Stickle, W. B. (1978). E?ects of salinity fluctuationson the respiration rate of the southern oyster drill Thais haemastoma and the bluecrab Callinectes sapidus. Marine Biology, 49(1), 59?67.Firth, L. B. & Williams, G. A. (2009). The influence of multiple environmentalstressors on the limpet Cellana toreuma during the summer monsoon season inHong Kong. Journal of Experimental Marine Biology and Ecology, 375(1), 70?75.145Fly, E. K., Monaco, C. J., Pincebourde, S., & Tullis, A. (2012). The influence ofintertidal location and temperature on the metabolic cost of emersion in Pisasterochraceus. Journal of Experimental Marine Biology and Ecology, 422, 20?28.Fong, P. & Glynn, P. W. (1998). A dynamic size-structured population model:does disturbance control size structure of a population of the massive coral Gar-dineroseris planulata in the Eastern Pacific? Marine Biology, 130(4), 663?674.Fraenkel, G. (1960). Lethal high temperatures for three marine invertebrates: Limu-lus polyphemus, Littorina littorea and Pagurus longicarpus. Oikos, 11(2), 171?182.Francey, R. J., Trudinger, C. M., van der Schoot, M., Law, R. M., Krummel, P. B.,Langenfelds, R. L., Steele, L. P., Allison, C. E., Stavert, A. R., Andres, R. J., &Roedenbeck, C. (2013). Atmospheric verification of anthropogenic CO2 emissiontrends. Nature Climate Change, 3(5), 520?524.Fraschetti, S., Giangrande, A., Terlizzi, A., & Boero, F. (2002). Pre-and post-settlement events in benthic community dynamics. Oceanologica Acta, 25(6),285?295.Frelich, L. E., Peterson, R. O., Dovciak, M., Reich, P. B., Vucetich, J. A., & Eisen-hauer, N. (2012). Trophic cascades, invasive species and body-size hierarchies in-teractively modulate climate change responses of ecotonal temperate-boreal forest.Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1605),2955?2961.Fulton, E. A. (2011). Interesting times: winners, losers, and system shifts underclimate change around Australia. ICES Journal of Marine Science, 68(6), 1329?1342.Garrity, S. D. (1984). Some adaptations of gastropods to physical stress on a tropicalrocky shore. Ecology, 65(2), 559?574.Gaylord, B., Hill, T. M., Sanford, E., Lenz, E. A., Jacobs, L. A., Sato, K. N., Russell,A. D., & Hettinger, A. (2011). Functional impacts of ocean acidification in anecologically critical foundation species. Journal of Experimental Marine Biologyand Ecology, 214(15), 2586?2594.Gaymer, C. F., Dutil, C., & Himmelman, J. H. (2004). Prey selection and predatoryimpact of four major sea stars on a soft bottom subtidal community. Journal ofExperimental Marine Biology and Ecology, 313(2), 353?374.Gazeau, F., Quiblier, C., Jansen, J. M., Gattuso, J.-P., Middelburg, J. J., & Heip,C. H. R. (2007). Impact of elevated CO2 on shellfish calcification. GeophysicalResearch Letters, 34(7).146Gilman, S. E., Urban, M. C., Tewksbury, J., Gilchrist, G. W., & Holt, R. D. (2010).A framework for community interactions under climate change. Trends in Ecology& Evolution, 25(6), 325?331.Gonzalez-Bernat, M. J., Lamare, M., & Barker, M. (2012). E?ects of reduced sea-water pH on fertilisation, embryogenesis and larval development in the Antarcticseastar Odontaster validus. Polar Biology, 36(2), 235?247.Gooding, R. A., Harley, C. D. G., & Tang, E. (2009). Elevated water temperatureand carbon dioxide concentration increase the growth of a keystone echinoderm.Proceedings of the National Academy of Sciences, 106(23), 9316?9321.Gosling, S. N., Warren, R., Arnell, N. W., Good, P., Caesar, J., Bernie, D., Lowe,J. A., van der Linden, P., O?Hanley, J. R., & Smith, S. M. (2011). A review ofrecent developments in climate change science. Part II: The global-scale impactsof climate change. Progress in Physical Geography, 35(4), 443?464.Gosselin, L. A. & Qian, P.-Y. (1997). Juvenile mortality in benthic marine inverte-brates. Marine Ecology Progress Series, 146, 265?282.Green, A. J. (1992). Positive allometry is likely with mate choice, competitivedisplay and other functions. Animal Behaviour, 43, 170?172.Grith, G. P., Fulton, E. A., Gorton, R., & Richardson, A. J. (2012). Predictinginteractions among fishing, ocean warming, and ocean acidification in a marinesystem with whole-ecosystem models. Conservation Biology, 26(6), 1145?1152.Gutowska, M., Po?rtner, H., & Melzner, F. (2008). Growth and calcification inthe cephalopod Sepia ocinalis under elevated seawater pCO2. Marine EcologyProgress Series, 373, 303?309.Hajima, T., Ise, T., Tachiiri, K., Kato, E., Watanabe, S., & Kawamiya, M. (2012).Climate change, allowable emission, and Earth system response to representativeconcentration pathway scenarios. Journal of the Meteorological Society of Japan.Ser. II, 90(3), 417?434.Hale, R., Calosi, P., McNeill, L., Mieszkowska, N., & Widdicombe, S. (2011). Pre-dicted levels of future ocean acidification and temperature rise could alter com-munity structure and biodiversity in marine benthic communities. Oikos, 120(5),661?674.Hall-Spencer, J. M., Rodolfo-Metalpa, R., Martin, S., Ransome, E., Fine, M.,Turner, S. M., Rowley, S. J., Tedesco, D., & Buia, M.-C. (2008). Volcanic carbondioxide vents show ecosystem e?ects of ocean acidification. Nature, 454(7200),96?99.147Halpern, B. S., Walbridge, S., Selkoe, K. A., Kappel, C. V., Micheli, F., D?Agrosa,C., Bruno, J. F., Casey, K. S., Ebert, C., Fox, H. E., Fujita, R., Heinemann, D.,Lenihan, H. S., Madin, E. M. P., Perry, M. T., Selig, E. R., Spalding, M., Steneck,R., & Watson, R. (2008). A global map of human impact on marine ecosystems.Science, 319(5865), 948?952.Hamer, K. C. (2010). The search for winners and losers in a sea of climate change.Ibis, 152(1), 3?5.Harley, C. (2008). Tidal dynamics, topographic orientation, and temperature-mediated mass mortalities on rocky shores. Marine Ecology Progress Series, 371,37?46.Harley, C. D. & Helmuth, B. S. (2003). Local-and regional-scale e?ects of waveexposure, thermal stress, and absolute versus e?ective shore level on patterns ofintertidal zonation. Limnology and Oceanography, 48(4), 1498?1508.Harley, C. D. & Paine, R. T. (2009). Contingencies and compounded rare per-turbations dictate sudden distributional shifts during periods of gradual climatechange. Proceedings of the National Academy of Sciences, 106(27), 11172?11176.Harley, C. D. G. (1998). Species-specific responses to environmental gradients de-termine regional scale pattern in an intertidal community. Padilla Bay NationalEstuarine Research Reserve.Harley, C. D. G. (2006). E?ects of physical ecosystem engineering and herbivory onintertidal community structure. Marine Ecology Progress Series, 317, 29?39.Harley, C. D. G. (2011). Climate change, keystone predation, and biodiversity loss.Science, 334(6059), 1124?1127.Harley, C. D. G., Pankey, M. S., Wares, J. P., Grosberg, R. K., & Wonham,M. J. (2006a). Color polymorphism and genetic structure in the sea star Pisasterochraceus. Biological Bulletin, 211, 248?262.Harley, C. D. G., Randall Hughes, A., Hultgren, K. M., Miner, B. G., Sorte, C.J. B., Thornber, C. S., Rodriguez, L. F., Tomanek, L., & Williams, S. L. (2006b).The impacts of climate change in coastal marine systems. Ecology Letters, 9(2),228?241.Hart, P. & Gill, AB (1992). Constraints on prey size selection by the threespinedstickleback: energy requirements and the capacity and fullness of the gut. Journalof Fish Biology, 40(2), 205?218.Havens, K. & DeCosta, J. (1985). The e?ect of acidification in enclosures on thebiomass and population size structure of Bosmina longirostirs. Hydrobiologia,122, 153?158.148Hawkins, S., Moore, P., Burrows, M., Poloczanska, E., Mieszkowska, N., Herbert,R., Jenkins, S., Thompson, R., Genner, M., & Southward, A. (2008). Complexinteractions in a rapidly changing world: responses of rocky shore communitiesto recent climate change. Climate Research, 37, 123?133.Heckmann, L., Drossel, B., Brose, U., & Guill, C. (2012). Interactive e?ects ofbody-size structure and adaptive foraging on food-web stability. Ecology Letters,15(3), 243?250.Heilmayer, O., Digialleonardo, J., Qian, L., & Roesijadi, G. (2008). Stress toler-ance of a subtropical Crassostrea virginica population to the combined e?ects oftemperature and salinity. Estuarine, Coastal and Shelf Science, 79(1), 179?185.Held, M. B. E. & Harley, C. D. G. (2009). Responses to low salinity by the sea starPisaster ochraceus from high- and low-salinity populations. Invertebrate Biology,128(4), 381?390.Helmuth, B., Broitman, B. R., Blanchette, C. A., Gilman, S., Halpin, P., Harley,C. D., O?Donnell, M. J., Hofmann, G. E., Menge, B., & Strickland, D. (2006a).Mosaic patterns of thermal stress in the rocky intertidal zone: implications forclimate change. Ecological Monographs, 76(4), 461?479.Helmuth, B., Broitman, B. R., Yamane, L., Gilman, S. E., Mach, K., Mislan, K.A. S., & Denny, M. W. (2010). Organismal climatology: analyzing environmen-tal variability at scales relevant to physiological stress. Journal of ExperimentalMarine Biology and Ecology, 213(6), 995?1003.Helmuth, B., Mieszkowska, N., Moore, P., & Hawkins, S. J. (2006b). Living on theedge of two changing worlds: forecasting the responses of rocky intertidal ecosys-tems to climate change. Annual Review of Ecology, Evolution, and Systematics,37, 373?404.Hiebenthal, C., Philipp, E. E., Eisenhauer, A., & Wahl, M. (2013). E?ects ofseawater pCO2 and temperature on shell growth, shell stability, condition andcellular stress of Western Baltic Sea Mytilus edulis (L.) and Arctica islandica(L.). Marine Biology, 160(8), 2073?2087.His, E., Robert, R., & Dinet, A. (1989). Combined e?ects of temperature and salinityon fed and starved larvae of the Mediterranean mussel Mytilus galloprovincialisand the Japanese oyster Crassostrea gigas . Marine Biology, 100(4), 455?463.Hixon, M. A. (1982). Energy maximizers and time minimizers: Theory and reality.The American Naturalist, 119(4), 596?599.Ho?man, J. R., Hansen, L. J., & Klinger, T. (2003). Interactions between UVradiation and temperature limit inferences from single-factor experiments. Journalof Phycology, 39(2), 268?272.149Hofmann, G. E., Smith, J. E., Johnson, K. S., Send, U., Levin, L. A., Micheli,F., Paytan, A., Price, N. N., Peterson, B., Takeshita, Y., Matson, P. G., Crook,E. D., Kroeker, K. J., Gambi, M. C., Rivest, E. B., Frieder, C. A., Yu, P. C., &Martz, T. R. (2011). High-frequency dynamics of ocean pH: A multi-ecosystemcomparison. PLoS ONE, 6(12), e28983.Hoyaux, J., Gilles, R., & Jeuniaux, C. (1976). Osmoregulation in molluscs of theintertidal zone. Comparative Biochemistry and Physiology Part A: Physiology,53(4), 361?365.Hsieh, C.-h., Glaser, S. M., Lucas, A. J., & Sugihara, G. (2005). Distinguishingrandom environmental fluctuations from ecological catastrophes for the NorthPacific Ocean. Nature, 435(7040), 336?340.Huey, R. B., Kearney, M. R., Krockenberger, A., Holtum, J. A. M., Jess, M., &Williams, S. E. (2012). Predicting organismal vulnerability to climate warming:roles of behaviour, physiology and adaptation. Philosophical Transactions of theRoyal Society B: Biological Sciences, 367(1596), 1665?1679.Hughes, T. P., Baird, A. H., Bellwood, D. R., Card, M., Connolly, S. R., Folke,C., Grosberg, R., Hoegh-Guldberg, O., Jackson, J. B. C., Kleypas, J., Lough,J. M., Marshall, P., Nystro?m, M., Palumbi, S. R., Pandolfi, J. M., Rosen, B., &Roughgarden, J. (2003). Climate change, human impacts, and the resilience ofcoral reefs. Science, 301(5635), 929?933.Hunt, H. L. & Scheibling, R. E. (1997). Role of early post-settlement mortality inrecruitment of benthic marine invertebrates. Marine Ecology Progress Series, 155,269?301.Hunt, H. L. & Scheibling, R. E. (1998). Spatial and temporal variability of patternsof colonization by mussels (Mytilus trossulus, M. edulis) on a wave-exposed rockyshore. Marine Ecology Progress Series, 167, 155?169.Hutchings, J. A., Co?te?, I. M., Dodson, J. J., Fleming, I. A., Jennings, S., Man-tua, N. J., Peterman, R. M., Riddell, B. E., & Weaver, A. J. (2012). Climatechange, fisheries, and aquaculture: trends and consequences for Canadian marinebiodiversity. Environmental Reviews, 20(4), 220?311.Iacarella, J. C. & Helmuth, B. (2012). Body temperature and desiccation constrainthe activity of Littoraria irrorata within the Spartina alterniflora canopy. Journalof Thermal Biology, 37(1), 15?22.Jacobsen, H. P. & Stabell, O. B. (1999). Predator-induced alarm responses in thecommon periwinkle, Littorina littorea: dependence on season, light conditions,and chemical labelling of predators. Marine Biology, 134, 551?557.150Jansson, M., Persson, L., De Roos, A. M., Jones, R. I., & Tranvik, L. J. (2007).Terrestrial carbon and intraspecific size-variation shape lake ecosystems. Trendsin Ecology & Evolution, 22(6), 316?322.Jenouvrier, S., Caswell, H., Barbraud, C., Holland, M., Str ve, J., & Weimerskirch,H. (2009). From the Cover: Demographic models and IPCC climate projectionspredict the decline of an emperor penguin population. Proceedings of the NationalAcademy of Sciences of the United States of America, 106(6), 1844?1847.Jochum, M., Schneider, F. D., Crowe, T. P., Brose, U., & O?Gorman, E. J. (2012).Climate-induced changes in bottom-up and top-down processes independentlyalter a marine ecosystem. Philosophical Transactions of the Royal Society B:Biological Sciences, 367(1605), 2962?2970.Johannessen, S. C. & Macdonald, R. W. (2009). E?ects of local and global changeon an inland sea: the Strait of Georgia, British Columbia, Canada. ClimateResearch, 40, 1?21.Johnson, M. D. & Carpenter, R. C. (2012). Ocean acidification and warming de-crease calcification in the crustose coralline alga Hydrolithon onkodes and increasesusceptibility to grazing. Journal of Experimental Marine Biology and Ecology,434, 94?101.Jokiel, P. L., Rodgers, K. S., Ku?ner, I. B., Andersson, A. J., Cox, E. F., & Macken-zie, F. T. (2008). Ocean acidification and calcifying reef organisms: a mesocosminvestigation. Coral Reefs, 27(3), 473?483.Jones, K. M. & Boulding, E. G. (1999). State-dependent habitat selection by anintertidal snail: the costs of selecting a physically stressful microhabitat. Journalof Experimental Marine Biology and Ecology, 242, 149?177.Keith, D. A., Akcakaya, H. R., Thuiller, W., Midgley, G. F., Pearson, R. G., Phillips,S. J., Regan, H. M., Araujo, M. B., & Rebelo, T. G. (2008). Predicting extinctionrisks under climate change: coupling stochastic population models with dynamicbioclimatic habitat models. Biology Letters, 4(5), 560?563.Kirkpatrick, M. (1984). Demographic models based on size, not age, for organismswith indeterminate growth. Ecology, 65(6), 1874?1884.Kordas, R. L., Harley, C. D. G., & O?Connor, M. I. (2011). Community ecologyin a warming world: The influence of temperature on interspecific interactions inmarine systems. Journal of Experimental Marine Biology and Ecology, 400(1-2),218?226.Kozlo?, E. N. (1996). Marine Invertebrates of the Pacific Northwest. University ofWashington Press.151Krebs, J. R. (1980). Optimal foraging, predation risk and territory defence. Ardea,68, 83?90.Kroeker, K. J. (2013). Ocean acidification causes ecosystem shifts via altered com-petitive interactions. Nature Climate Change, 3(2), 156?159.Kroeker, K. J., Kordas, R. L., Crim, R., Hendriks, I. E., Ramajo, L., Singh, G. S.,Duarte, C. M., & Gattuso, J.-P. (2013). Impacts of ocean acidification on marineorganisms: quantifying sensitivities and interaction with warming. Global ChangeBiology, 19(6), 1884?1896.Kroeker, K. J., Kordas, R. L., Crim, R. N., & Singh, G. G. (2010). Meta-analysisreveals negative yet variable e?ects of ocean acidification on marine organisms.Ecology Letters, 13(11), 1419?1434.Kurihara, H., Asai, T., Kato, S., & Ishimatsu, A. (2008a). E?ects of elevated pCO2on early development in the mussel Mytilus galloprovincialis . Aquatic Biology, 4,225?233.Kurihara, H., Matsui, M., Furukawa, H., Hayashi, M., & Ishimatsu, A. (2008b).Long-term e?ects of predicted future seawater CO2 conditions on the survivaland growth of the marine shrimp Palaemon pacificus. Journal of ExperimentalMarine Biology and Ecology, 367(1), 41?46.Landenberger, D. E. (1968). Studies on selective feeding in the Pacific StarfishPisaster in Southern California . Ecology, 49(6), 1062?1075.Landes, A. & Zimmer, M. (2012). Acidification and warming a?ect both a calcifyingpredator and prey, but not their interaction. Marine Ecology Progress Series, 450,1?10.Lavergne, S., Mouquet, N., Thuiller, W., & Ronce, O. (2010). Biodiversity andclimate change: Integrating evolutionary and ecological responses of species andcommunities. Annual Review of Ecology, Evolution, and Systematics, 41(1), 321?350.Laws, A. N. & Joern, A. (2012). Predator-prey interactions in a grassland foodchain vary with temperature and food quality. Oikos, 122(7), 977?986.Lazo, C. S. & Pita, I. M. (2011). E?ect of temperature on survival, growth and de-velopment of Mytilus galloprovincialis larvae. Aquaculture Research, 43(8), 1127?1133.LeClair, E. E. (1993). E?ects of anatomy and environment on the relative preserv-ability of asteroids; a biomechanical comparison. Palaios, 8(3), 233?243.152Lester, S. E., Tobin, E. D., & Behrens, M. D. (2007). Disease dynamics and thepotential role of thermal stress in the sea urchin, Strongylocentrotus purpuratus.Canadian Journal of Fisheries and Aquatic Sciences, 64(2), 314?323.Lima, S. L., Valone, T. J., & Caraco, T. (1985). Foraging-eciency-predation-risktrade-o? in the grey squirrel. Animal Behaviour, 33, 155?165.Loarie, S. R., Du?y, P. B., Hamilton, H., Asner, G. P., Field, C. B., & Ackerly,D. D. (2009). The velocity of climate change. Nature, 462(7276), 1052?1055.Long, W. C., Swiney, K. M., Harris, C., Page, H. N., & Foy, R. J. (2013). E?ectsof Ocean Acidification on Juvenile Red King Crab (Paralithodes camtschaticus)and Tanner Crab (Chionoecetes bairdi) Growth, Condition, Calcification, andSurvival. PLoS ONE, 8(4), e60959.Louthan, A. M., Doak, D. F., Goheen, J. R., Palmer, T. M., & Pringle, R. M. (2013).Climatic stress mediates the impacts of herbivory on plant population structureand components of individual fitness. Journal of Ecology, 101(4), 1074?1083.Mainka, S. A. & Howard, G. W. (2010). Climate change and invasive species: doublejeopardy. Integrative Zoology, 5(2), 102?111.Mallet, A. L. & Carver, C. E. (1995). Comparative growth and survival patternsof Mytilus trossulus and Mytilus edulis in Atlantic Canada. Canadian Journal ofFisheries and Aquatic Sciences, 52(9), 1873?1880.Manr??quez, P. H., Jara, M. E., Mardones, M. L., Navarro, J. M., Torres, R., Lardies,M. A., Vargas, C. A., Duarte, C., Widdicombe, S., Salisbury, J., & Lagos, N. A.(2013). Ocean acidification disrupts prey responses to predator cues but not netprey shell growth in Concholepas concholepas (loco). PLoS ONE, 8(7), e68643.Marliave, J. B., Gibbs, C. J., Gibbs, D. M., Lamb, A. O., & Young, S. J. (2011).Biodiversity stability of shallow marine benthos in Strait of Georgia, BritishColumbia, Canada through climate regimes, overfishing and ocean acidification.Biodiversity Loss in a Changing Planet?, ISBN, (pp. 979?953).Matoo, O. B., Ivanina, A. V., Ullstad, C., Beniash, E., & Sokolova, I. M. (2013).Interactive e?ects of elevated temperature and CO2 levels on metabolism andoxidative stress in two common marine bivalves (Crassostrea virginica and Mer-cenaria mercenaria). Comparative Biochemistry and Physiology, Part A, 164(4),545?553.Maury, O., Shin, Y.-J., Faugeras, B., Ben Ari, T., & Marsac, F. (2007). Modelingenvironmental e?ects on the size-structured energy flow through marine ecosys-tems. Part 2: Simulations. Progress in Oceanography, 74(4), 500?514.153Mauzey, K. P. (1966). Feeding behavior and reproductive cycles in Pisasterochraceus. Biological Bulletin, 131, 127?144.McClintock, J. B. & Robnett, T. J. (1986). Size selective predation by the asteroidPisaster ochraceus on the bivalve Mytilus californianus: A cost-benefit analysis.P.S.Z.N.I.: Marine Ecology, 7(4), 321?332.McCluney, K. E., Belnap, J., Collins, S. L., Gonza?lez, A. L., Hagen, E. M.,Nathaniel Holland, J., Kotler, B. P., Maestre, F. T., Smith, S. D., & Wolf, B. O.(2011). Shifting species interactions in terrestrial dryland ecosystems under al-tered water availability and climate change. Biological Reviews, 87(3), 563?582.McGaw, I. J. (2001). Impacts of habitat complexity on physiology: Purple shorecrabs tolerate osmotic stress for shelter. Estuarine, Coastal and Shelf Science,53(6), 865?876.McLeese, D. W. (1956). E?ects of temperature, salinity and oxygen on the survivalof the American lobster. Journal of the Fisheries Board of Canada, 13(2), 247?272.McQuaid, C. D. (1996). Biology of the gastropod family Littorinidae II. Role in theecology of intertidal and shallow marine ecosystems. Oceanography and MarineBiology: An Annual Review, 34, 263?302.Meehl, G. A. (2004). More intense, more frequent, and longer lasting heat waves inthe 21st century. Science, 305(5686), 994?997.Melatunan, S., Calosi, P., Rundle, S. D., Widdicombe, S., & Moody, A. J. (2013).E?ects of ocean acidification and elevated temperature on shell plasticity and itsenergetic basis in an intertidal gastropod. Marine Ecology Progress Series, 472,155?168.Menge, B. A. (1972). Competition for food between two intertidal starfish speciesand its e?ect on body size and feeding. Ecology, 53(4), 635?644.Menge, B. A. (1975). Brood or broadcast? The adaptive significance of di?erentreproductive strategies in the two intertidal sea stars Leptasterias hexactis andPisaster ochraceus . Marine Biology, 31(1), 87?100.Menge, B. A., Chan, F., & Lubchenco, J. (2008). Response of a rocky intertidalecosystem engineer and community dominant to climate change. Ecology Letters,11, 151?162.Menge, B. A., Chan, F., Nielsen, K. J., Lorenzo, E. D., & Lubchenco, J. (2009).Climatic variation alters supply-side ecology: impact of climate patterns on phy-toplankton and mussel recruitment. Ecological Monographs, 79(3), 379?395.154Menge, J. L. & Menge, B. A. (1974). Role of resource allocation, aggression andspatial heterogeneity in coexistence of two competing intertidal starfish. EcologicalMonographs, 44(2), 189?209.Metzger, R., Sartoris, F. J., Langenbuch, M., & Po?rtner, H. O. (2007). Influenceof elevated CO2 concentrations on thermal tolerance of the edible crab Cancerpagurus. Journal of Thermal Biology, 32(3), 144?151.Michaelidis, B., Ouzounis, C., Paleras, A., & Portner, H. O. (2005). E?ects oflong-term moderate hypercapnia on acid?base balance and growth rate in marinemussels Mytilus galloprovincialis . Marine Ecology Progress Series, 293, 109?118.Miller, L. P. & Denny, M. W. (2011). Importance of behavior and morphologicaltraits for controlling body temperature in littorinid snails. The Biological Bulletin,220(3), 209?223.Millstein, J. & O?Clair, C. E. (2001). Comparison of age-length and growth-increment general growth models of the Schnute type in the Pacific Blue Mussel,Mytilus trossulus Gould. Journal of Experimental Marine Biology and Ecology,262(2), 155?176.Mislan, K. A. S., Wethey, D. S., & Helmuth, B. (2009). When to worry about theweather: role of tidal cycle in determining patterns of risk in intertidal ecosystems.Global Change Biology, 15(12), 3056?3065.Moore, P., Hawkins, S., & Thompson, R. (2007). Role of biological habitat amelio-ration in altering the relative responses of congeneric species to climate change.Marine Ecology Progress Series, 334, 11?19.Morris, R. H., Abbot, D. P., & Haderlie, E. C. (1980). Intertidal Invertebrates ofCalifornia. Stanford University Press.Morrison, J., Quick, M. C., & Foreman, M. G. (2002). Climate change in the FraserRiver watershed: flow and temperature projections. Journal of Hydrology, 263(1),230?244.Moss, R. H., Edmonds, J. A., Hibbard, K. A., Manning, M. R., Rose, S. K., vanVuuren, D. P., Carter, T. R., Emori, S., Kainuma, M., Kram, T., Meehl, G. A.,Mitchell, J. F. B., Nakicenovic, N., Riahi, K., Smith, S. J., Stou?er, R. J., Thom-son, A. M., Weyant, J. P., & Wilbanks, T. J. (2010). The next generation ofscenarios for climate change research and assessment. Nature, 463(7282), 747?756.Mote, P., Salathe?, E., & Peacock, C. (2008). Scenarios of future climate for thePacific Northwest. Report prepared for King County, WA by the Climate ImpactsGroup, University of Washington, Seattle, WA.155Mun?oz, J. L. P., Randall Finke, G., Camus, P. A., & Bozinovic, F. (2005). Ther-moregulatory behavior, heat gain and thermal tolerance in the periwinkle Echi-nolittorina peruviana in central Chile. Comparative Biochemistry and Physiology- Part A: Molecular & Integrative Physiology, 142(1), 92?98.Nagarajan, R., Lea, S. E. G., & Goss-Custard, J. D. (2006). Seasonal variations inmussel, Mytilus edulis L. shell thickness and strength and their ecological impli-cations. Journal of Experimental Marine Biology and Ecology, 339(2), 241?250.National Oceanic and Atmospheric Administration, Oce of Ocean and CoastalResource Management (2013). National Estuarine Research Reserve System-WideMonitoring Program. Centralized Data Management Oce, Baruch Marine FieldLab, University of South Carolina.Navarrete, S. A., menge, B. A., & Daley, B. A. (2000). Species interactions inintertidal food webs: prey or predation regulation of intermediate predators?Ecology, 81(8), 2264?2277.Newell, R. C., Pye, V. I., & Ahsanullah, M. (1971). The e?ect of thermal acclimationon the heat tolerance of the intertidal Prosobranchs Littorina littorea (L.) andMonodonta lineata (da Costa). The Journal of experimental biology, 54(2), 525?533.Nguyen, H. D., Doo, S. S., Soars, N. A., & Byrne, M. (2012). Noncalcifying larvaein a changing ocean: warming, not acidification/hypercapnia, is the dominantstressor on development of the sea star Meridiastra calcar . Global Change Biology,18(8), 2466?2476.O?Connor, M. I. (2009). Warming strengthens an herbivore-plant interaction. Ecol-ogy, 90(2), 388?398.O?Connor, M. I., Selig, E. R., Pinsky, M. L., & Altermatt, F. (2011). Toward a con-ceptual synthesis for climate change responses. Global Ecology and Biogeography,21(7), 693?703.O?Donnell, M. J. (2013). Mussel byssus attachment weakened by ocean acidification.Nature Climate Change, 3(6), 587?590.O?Donnell, M. J., Hammond, L. M., & Hofmann, G. E. (2008). Predicted impactof ocean acidification on a marine invertebrate: elevated CO2 alters response tothermal stress in sea urchin larvae. Marine Biology, 156(3), 439?446.Paine, R. T. (1966). Food web complexity and species diversity. The AmericanNaturalist, 100(910), 65?75.Paine, R. T. (1969). A note on trophic complexity and community stability. TheAmerican Naturalist, 103(929), 91?93.156Paine, R. T. (1974). Intertidal Community Structure: experimental studies on therelationship between a dominant competitor and its principal prey. Oecologia, 15,93?120.Paine, R. T. (1976). Size-limited predation: an observational and experimentalapproach with the Mytilus-Pisaster interaction. Ecology, 57(5), 858?873.Paine, R. T. (2002). Trophic control of production in a rocky intertidal community.Science, 296(5568), 736?739.Paine, R. T., Tegner, M. J., & Johnson, E. A. (1998). Compounded perturbationsyield ecological surprises. Ecosystems, 1, 535?545.Parker, L., Ross, P., O?Connor, W., Po?rtner, H., Scanes, E., & Wright, J. (2013).Predicting the response of molluscs to the impact of ocean acidification. Biology,2(2), 651?692.Parker, L. M., Ross, P., O?Connor, W., Borysko, L., Raftos, D. A., & Po?rtner, H.-O. (2012). Adult exposure influences o?spring response to ocean acidification inoysters. Global Change Biology, 18(1), 82?92.Parker, L. M., Ross, P. M., & O?connor, W. A. (2009). The e?ect of ocean acidi-fication and temperature on the fertilization and embryonic development of theSydney rock oyster Saccostrea glomerata (Gould 1850). Global Change Biology,15(9), 2123?2136.Parmesan, C. (2006). Ecological and evolutionary responses to recent climatechange. Annual Review of Ecology, Evolution, and Systematics, 37(1), 637?669.Pearce, C. M., Williams, S. W., Yuan, F., Castell, J. D., & Robinson, S. M. C.(2005). E?ect of temperature on somatic growth and survivorship of early post-settled green sea urchins, Strongylocentrotus droebachiensis (Mu?ller). AquacultureResearch, 36(6), 600?609.Peck, L. S., Clark, M. S., Morley, S. A., Massey, A., & Rossetti, H. (2009). Animaltemperature limits and ecological relevance: e?ects of size, activity and rates ofchange. Functional Ecology, 23(2), 248?256.Petes, L. E., Menge, B. A., & Murphy, G. D. (2007). Environmental stress de-creases survival, growth, and reproduction in New Zealand mussels. Journal ofExperimental Marine Biology and Ecology, 351(1-2), 83?91.Petes, L. E., Mouchka, M. E., Milston-Clements, R. H., Momoda, T. S., & Menge,B. A. (2008). E?ects of environmental stress on intertidal mussels and their seastar predators. Oecologia, 156(3), 671?680.157Petraitis, P. S. (1995). The role of growth in maintaining spatial dominance bymussels (Mytilus edulis). Ecology, 76(4), 1337?1346.Pincebourde, S., Sanford, E., Casas, J., & Helmuth, B. (2012). Temporal coincidenceof environmental stress events modulates predation rates. Ecology Letters, 15(7),680?688.Pincebourde, S., Sanford, E., & Helmuth, B. (2008). Body temperature during lowtide alters the feeding performance of a top intertidal predator. Limnology andOceanography, 53(4), 1562?1573.Pincebourde, S., Sanford, E., & Helmuth, B. (2009). An intertidal sea star adjuststhermal inertia to avoid extreme body temperatures. The American Naturalist,174(6), 890?897.Poloczanska, E. S., Hawkins, S. J., Southward, A. J., & Burrows, M. T. (2008).Modeling the response of populations of competing species to climate change.Ecology, 89(11), 3138?3149.Poore, A. G. B., Graba-Landry, A., Favret, M., Sheppard Brennand, H., Byrne, M.,& Dworjanyn, S. A. (2013). Direct and indirect e?ects of ocean acidification andwarming on a marine plant?herbivore interaction. Oecologia, 173(3), 1113?1124.Po?rtner, H. O. & Farrell, A. P. (2008). Physiology and climate change. Science,322(5902), 690?692.Po?rtner, H. O., Langenbuch, M., & Reipschla?ger, A. (2004). Biological impact ofelevated ocean CO2 concentrations: Lessons from animal physiology and earthhistory. Journal of Oceanography, 60(4), 705?718.Przeslawski, R., Davis, A. R., & Benkendor?, K. (2005). Synergistic e?ects asso-ciated with climate change and the development of rocky shore molluscs. GlobalChange Biology, 11(3), 515?522.Reid, D. G. (1996). Systematics and Evolution of Littorina. Dorchestter, Dorset:Dorset Press.Reipschlager, A. & Portner, H. O. (1996). Metabolic depression during environmen-tal stress: the role of extracellular versus intracellular pH in Sipunculus nudus .Journal of Experimental Marine Biology and Ecology, 199, 1801?1807.Renegar, D. A. & Riegl, B. M. (2005). E?ect of nutrient enrichment and elevatedCO2 partial pressure on growth rate of Atlantic scleractinian coral Acropora cer-vicornis. Marine Ecology Progress Series, 293, 69?76.158Reynaud, S., Leclercq, N., Romaine Lioud, S., Ferrier-Page`s, C., Jaubert, J., &Gattuso, J.-P. (2003). Interacting e?ects of CO2 partial pressure and temperatureon photosynthesis and calcification in a scleractinian coral. Global Change Biology,9(11), 1660?1668.Riahi, K., Rao, S., Krey, V., Cho, C., Chirkov, V., Fischer, G., Kindermann, G.,Nakicenovic, N., & Rafaj, P. (2011). RCP 8.5?A scenario of comparatively highgreenhouse gas emissions. Climatic Change, 109(1-2), 33?57.Robbins, L. L., Hansen, M. E., Kleypas, J., & Meylan, S. C. (2010). CO2calc: Auser-friendly seawater carbon calculator for Windows, Mac OS X, and iOS. U.S.Geological Survey Open-File Report.Robles, C. & Desharnais, R. (2002). History and current development of a paradigmof predation in rocky intertidal communities. Ecology, 83(6), 1?16.Robles, C., Sherwood-Stephens, R., & Alvarado, M. (1995). Responses of a keyintertidal predator to varying recruitment of its prey. Ecology, 76(2), 565?579.Robles, C. D., Desharnais, R. A., Garza, C., Donahue, M. J., & Martinez, C. A.(2009). Complex equilibria in the maintenance of boundaries: experiments withmussel beds. Ecology, 90(4), 985?995.Robles, C. D., Garza, C., Desharnais, R. A., & Donahue, M. J. (2010). Landscapepatterns in boundary intensity: a case study of mussel beds. Landscape Ecology,25(5), 745?759.Rodolfo-Metalpa, R., Houlbre`que, F., Tambutte?, E?., Boisson, F., Baggini, C., Patti,F. P., Je?ree, R., Fine, M., Foggo, A., Gattuso, J.-P., & Hall-Spencer, J. M.(2011). Coral and mollusc resistance to ocean acidification adversely a?ected bywarming. Nature Climate Change, 1(6), 308?312.Rogers, T. L. & Elliott, J. K. (2012). Di?erences in relative abundance and sizestructure of the sea stars Pisaster ochraceus and Evasterias troschelii amonghabitat types in Puget Sound, Washington, USA. Marine Biology, 160(4), 853?865.Rojas, J. M., Castillo, S. B., Escobar, J. B., Shinen, J. L., & Bozinovic, F. (2013).Huddling up in a dry environment: the physiological benefits of aggregation in anintertidal gastropod. Marine Biology, 160(5), 1119?1126.Roughgarden, J., Iwasa, Y., & Baxter, C. (1985). Demographic theory for an openmarine population with space-limited recruitment. Ecology, 66(1), 54?67.Rudolf, V. H. W. & Rasmussen, N. L. (2013). Ontogenetic functional diversity:size structure of a keystone predator drives functioning of a complex ecosystem.Ecology, 94(5), 1046?1056.159Rumsey, T. J. (1973). Some aspects of osmotic and ionic regulation in Littorinalittorea (L.)(Gastropoda, Prosobranchia). Comparative Biochemistry and Physi-ology Part A: Physiology, 45(2), 327?344.Russell, A. D., Thompson, J.-A. I., Falkenberg, L. J., & Connell, S. D. (2009).Synergistic e?ects of climate change and local stressors: CO2 and nutrient-drivenchange in subtidal rocky habitats. Global Change Biology, 15(9), 2153?2162.Russell, B. D., Connell, S. D., Findlay, H. S., Tait, K., Widdicombe, S., &Mieszkowska, N. (2013). Ocean acidification and rising temperatures may in-crease biofilm primary productivity but decrease grazer consumption. Philosoph-ical Transactions of the Royal Society B: Biological Sciences, 368(1627).Sabine, C. L. (2004). The oceanic sink for anthropogenic CO2. Science, 305(5682),367?371.Salvato, B., Cuomo, V., Di Muro, P., & Beltramini, M. (2001). E?ects of environ-mental parameters on the oxygen consumption of four marine invertebrates: acomparative factorial study. Marine Biology, 138(4), 659?668.Sanford, E. (1999). Regulation of keystone predation by small changes. Science,283(5410), 2095?2097.Sanford, E. (2002). The feeding, growth, and energetics of two rocky intertidalpredators (Pisaster ochraceus and Nucella canaliculata) under water tempera-tures simulating episodic upwelling. Journal of Experimental Marine Biology andEcology, 273, 199?218.Sanford, E. & menge, B. A. (2007). Reproductive output and consistency of sourcepopulations in the sea star Pisaster ochraceus. Marine Ecology Progress Series,349, 1?12.Schneider, K. R., Van Thiel, L. E., & Helmuth, B. (2010). Interactive e?ects of foodavailability and aerial body temperature on the survival of two intertidal Mytilusspecies. Journal of Thermal Biology, 35(4), 161?166.Schoener, T. W. (1971). Theory of feeding strategies. Annual review of ecology andsystematics, 2, 369?404.Schram, J. B., McClintock, J. B., Angus, R. A., & Lawrence, J. M. (2011). Regener-ative capacity and biochemical composition of the sea star Luidia clathrata (Say)(Echinodermata: Asteroidea) under conditions of near-future ocean acidification.Journal of Experimental Marine Biology and Ecology, 407(2), 266?274.Schulte, P. M., Healy, T. M., & Fangue, N. A. (2011). Thermal performance curves,phenotypic plasticity, and the time scales of temperature exposure. Integrativeand Comparative Biology, 51(5), 691?702.160Sewell, M. A. & Watson, J. C. (1993). A ?source? for asteroid larvae?: recruitmentof Pisaster ochraceus, Pycnopodia helianthoides and Dermasterias imbricata inNootka Sound, British Columbia. Marine Biology, 117, 387?398.Shirayama, Y. & Thornton, H. (2005). E?ect of increased atmospheric CO2 onshallow water marine benthos. Journal of Geophysical Research: Oceans, 110(C9).Simas, T., Nunes, J. P., & Ferreira, J. G. (2001). E?ects of global climate changeon coastal salt marshes. Ecological Modelling, 139(1), 1?15.Singer, A., Travis, J. M., & Johst, K. (2012). Interspecific interactions a?ect speciesand community responses to climate shifts. Oikos, 122(3), 358?366.Singh, G. G. (2010). E?ects of sea otters on nearshore ecosystem functions withimplications for ecosystem services. PhD thesis, University of British Columbia.Sokolova, I. M. (2013). Energy-limited tolerance to stress as a conceptual frameworkto integrate the e?ects of multiple stressors. Integrative and Comparative Biology,53(4), 597?608.Sokolova, I. M., Bock, C., & Portner, H. O. (2000a). Resistance to freshwaterexposure in White Sea Littorina spp. I: Anaerobic metabolism and energetics.Journal of Comparative Physiology B: Biochemical, Systemic, and EnvironmentalPhysiology, 170(2), 91?103.Sokolova, I. M., Bock, C., & Portner, H. O. (2000b). Resistance to freshwater expo-sure in White Sea Littorina spp. II: Acid-base regulation. Journal of ComparativePhysiology B: Biochemical, Systemic, and Environmental Physiology, 170(2), 105?115.Sokolova, I. M. & Portner, H. O. (2001). Physiological adaptations to high intertidallife involve improved water conservation abilities and metabolic rate depressionin Littorina saxatilis . Marine Ecology Progress Series, 224, 171?186.Solomon, S., Plattner, G.-K., Knutti, R., & Friedlingstein, P. (2009). Irreversibleclimate change due to carbon dioxide emissions. Proceedings of the NationalAcademy of Sciences of the United States of America, 106(6), 1704?1709.Solomon, S., Qin, S., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor,M., & Miller, H. L., Eds. (2007). Climate change 2007: The physical sciencebasis., volume 4 of Working group I contribution to the fourth assessment reportof the IPCC. Cambridge University Press, Cambridge, United Kingdom and NewYork, NY, USA.Somero, G. N. (2010). The physiology of climate change: how potentials for acclima-tization and genetic adaptation will determine ?winners? and ?losers?. Journal ofExperimental Biology, 213(6), 912?920.161Sommer, U., Meusel, B., & Stielau, C. (1999). An experimental analysis of theimportance of body-size in the seastar-mussel predator-prey relationship. ActaOecologica, 20(2), 81?86.Sorte, C. J. B. & White, J. W. (2013). Competitive and demographic leverage pointsof community shifts under climate warming. Proceedings of the Royal Society B:Biological Sciences, 280(1762).Southward, A., Hawkins, S., & Burrows, M. (1995). Seventy years? observations ofchanges in distribution and abundance of zooplankton and intertidal organismsin the western English Channel in relation to rising sea temperature. Journal ofThermal Biology, 20(1), 127?155.Spicer, J. I., Ra?o, A., & Widdicombe, S. (2006). Influence of CO2-related seawa-ter acidification on extracellular acid?base balance in the velvet swimming crabNecora puber . Marine Biology, 151(3), 1117?1125.Springer, A. M., Estes, J. A., van Vliet, G. B., Williams, T. M., Doak, D. F., Danner,E. M., Forney, K. A., & Pfister, B. (2003). Sequential megafaunal collapse in theNorth Pacific Ocean: an ongoing legacy of industrial whaling? Proceedings of theNational Academy of Sciences, 100(21), 12223?12228.Stephenson, T. A. & Stephenson, A. (1949). The universal features of zonationbetween tide-marks on rocky coasts. The Journal of Ecology, 37(2), 289?305.Stillman, J. H. (2003). Acclimation capacity underlies susceptibility to climatechange. Science, 301(5629), 65.Stirling, H. P. (1982). The upper temperature tolerance of prosobranch gastropodsof rocky shores at Hong Kong and Dar Es Salaam, Tanzania. Journal of Experi-mental Marine Biology and Ecology, 63(2), 133?144.Sunday, J. M., Crim, R. N., Harley, C. D. G., & Hart, M. W. (2011). QuantifyingRates of Evolutionary Adaptation in Response to Ocean Acidification. PLoSONE, 6(8), e22881.Sutherland, D. A., MacCready, P., Banas, N. S., & Smedstad, L. F. (2011). A modelstudy of the Salish Sea Estuarine Circulation . Journal of Physical Oceanography,41(6), 1125?1143.Suttle, K. B., Thomsen, M. A., & Power, M. E. (2007). Species interactions reversegrassland responses to changing climate. Science, 315(5812), 640?642.Talmage, S. C. & Gobler, C. J. (2011). E?ects of elevated temperature and carbondioxide on the growth and survival of larvae and juveniles of three species ofNorthwest Atlantic bivalves. PLoS ONE, 6(10), e26941.162Tarpley, J. A. (1992). The e?ects of herbivorous gastropod species and densitieson early algal succession in the high intertidal zone. PhD thesis, San Jose StateUniversity.Terblanche, J. S., Ho?mann, A. A., Mitchell, K. A., Rako, L., le Roux, P. C., &Chown, S. L. (2011). Ecologically relevant measures of tolerance to potentiallylethal temperatures. Journal of Experimental Biology, 214(22), 3713?3725.Thomas, C. D., Cameron, A., Green, R. E., Bakkenes, M., Beaumont, L. J., Colling-ham, Y. C., Erasmus, B. F. N., De Siqueira, M. F., Grainger, A., Hannah, L.,Hughes, L., Huntley, B., Van Jaarsveld, A. S., Midgley, G. F., Miles, L., Ortega-Huerta, M. A., Peterson, A. T., Phillips, O. L., & Williams, S. E. (2004). Extinc-tion risk from climate change. Nature, 427(6970), 145?148.Thompson, R. C., Crowe, T. P., & Hawkins, S. J. (2002). Rocky intertidal commu-nities: past environmental changes, present status and predictions for the next 25years. Environmental Conservation, 29(02), 168?191.Thompson, S. A., Sydeman, W. J., Santora, J. A., Black, B. A., Suryan, R. M.,Calambokidis, J., Peterson, W. T., & Bograd, S. J. (2012). Linking predatorsto seasonality of upwelling: Using food web indicators and path analysis to infertrophic connections. Progress in Oceanography, 101(1), 106?120.Tieleman, B. I., Van Noordwijk, H. J., & Williams, J. B. (2008). Nest site selectionin a hot desert: Trade-o? between microclimate and predation risk? The Condor,110(1), 116?124.Todd, M. E. (1964). Osmotic balance in Littorina littorea, L. littoralis, and L.saxatilis (Littorinidae). Physiological Zoology, 37(1), 33?44.Todd, M. E. & Dehnel, P. A. (1960). E?ect of temperature and salinity on heat tol-erance in two grapsoid crabs, Hemigrapsus nudus and Hemigrapsus oregonensis.Biological Bulletin, 118(1), 150?172.Todgham, A. E. & Stillman, J. H. (2013). Physiological Responses to Shifts inMultiple Environmental Stressors: Relevance in a Changing World. Integrativeand Comparative Biology, 53(4), 539?544.Tokeshi, M. (1989). Development of a foraging model for a field population of theSouth American sun-star Heliaster helianthus. Journal of Animal Ecology, 58(1),189?206.Tomanek, L. (2008). The importance of physiological limits in determining bio-geographical range shifts due to global climate change: The heat-shock response.Physiological and Biochemical Zoology, 81(6), 709?717.163Tomanek, L. & Helmuth, B. (2002). Physiological ecology of rocky intertidal organ-isms: a synergy of concepts. American Zoologist, 42(4), 771?775.Toulmond, A. & Jouin, C. (1992). Seawater salinity and blood acid-base balance inthe lugworm, Arenicola marina (L.). Respiration physiology, 87(3), 429?446.Truchot, J.-P. (1990). Respiratory and ionic regulation in invertebrates exposed toboth water and air. Annual review of physiology, 52, 61?76.Truchot, J.-P. & Duhamel-Jouve, A. (1980). Oxygen and carbon dioxide in themarine intertidal environment: diurnal and tidal changes in rockpools. Respirationphysiology, 39(3), 241?254.Tsuchiya, M. (1983). Mass mortality in a population of the mussel Mytilus edulisL. caused by high temperature on rocky shores. Journal of Experimental MarineBiology and Ecology, 66(2), 101?111.Twomey, M., Brodte, E., Jacob, U., Brose, U., Crowe, T. P., & Emmerson, M. C.(2012). Idiosyncratic species e?ects confound size-based predictions of responsesto climate change. Philosophical Transactions of the Royal Society B: BiologicalSciences, 367(1605), 2971?2978.Tylianakis, J. M., Didham, R. K., Bascompte, J., & Wardle, D. A. (2008). Globalchange and species interactions in terrestrial ecosystems. Ecology Letters, 11(12),1351?1363.Underwood, A. (2000). Experimental ecology of rocky intertidal habitats: whatare we learning? Journal of Experimental Marine Biology and Ecology, 250(1-2),51?76.Urbina, M., Paschke, K., Gebauer, P., & Chaparro, O. R. (2010). Physiologicalenergetics of the estuarine crab Hemigrapsus crenulatus (Crustacea: Decapoda:Varunidae): responses to di?erent salinity levels. Journal of the Marine BiologicalAssociation of the UK, 90(2), 267?273.Vahl, O. (1984). The relationship between specific dynamic action (SDA) and growthin the common starfish, Asterias rubens L. Oecologia, 61(1), 122?125.Van der Putten, W. H., Macel, M., & Visser, M. E. (2010). Predicting speciesdistribution and abundance responses to climate change: why it is essential toinclude biotic interactions across trophic levels. Philosophical Transactions of theRoyal Society B: Biological Sciences, 365(1549), 2025?2034.von der Med, M., Porri, F., & McQuaid, C. D. (2012). New estimates of early post-settlement mortality for intertidal mussels show no relationship with meso-scalecoastline topographic features. Marine Ecology Progress Series, 463, 193?204.164Vucic-Pestic, O., Ehnes, R. B., Rall, B. C., & Brose, U. (2011). Warming up thesystem: higher predator feeding rates but lower energetic eciencies. GlobalChange Biology, 17(3), 1301?1310.Wagner, A., Hu?lsmann, S., Horn, W., Schiller, T., Schulze, T., Volkmann, S., &Benndorf, J. (2012). Food-web-mediated e?ects of climate warming: consequencesfor the seasonal Daphnia dynamics. Freshwater Biology, 58(3), 573?587.Walther, G. R. (2010). Community and ecosystem responses to recent climatechange. Philosophical Transactions of the Royal Society B: Biological Sciences,365(1549), 2019?2024.Walther, K., Sartoris, F. J., Bock, C., & Portner, H. O. (2009). Impact of anthro-pogenic ocean acidification on thermal tolerance of the spider crab Hyas araneus.Biogeosciences, 6(10), 2207?2215.Warren, P. H. & Lawton, J. H. (1987). Invertebrate predator-prey body size re-lationships: an explanation for upper triangular food webs and patterns in foodweb structure? Oecologia, 74(2), 231?235.Warton, D. I. & Hui, F. K. C. (2011). The arcsine is asinine: the analysis ofproportions in ecology. Ecology, 92(1), 3?10.Watson, S.-A., Southgate, P. C., Tyler, P. A., & Peck, L. S. (2009). Early larvaldevelopment of the Sydney rock oyster Saccostrea glomerata under near-futurepredictions of CO2-driven ocean acidification. Journal of Shellfish Research, 28(3),431?437.Wernberg, T., Russell, B. D., Moore, P. J., Ling, S. D., Smale, D. A., Campbell,A., Coleman, M. A., Steinberg, P. D., Kendrick, G. A., & Connell, S. D. (2011).Impacts of climate change in a global hotspot for temperate marine biodiversityand ocean warming. Journal of Experimental Marine Biology and Ecology, 400(1),7?16.Wethey, D. S., Woodin, S. A., Hilbish, T. J., Jones, S. J., Lima, F. P., & Brannock,P. M. (2011). Response of intertidal populations to climate: E?ects of extremeevents versus long term change. Journal of Experimental Marine Biology andEcology, 400(1-2), 132?144.Whiteley, N. M., Scott, J. L., Breeze, S. J., & McCann, L. (2001). E?ects of watersalinity on acid-base balance in decapod crustaceans. Journal of ExperimentalBiology, 204, 1003?1011.Williams, G. A., De Pirro, M., Cartwright, S., Khangura, K., Ng, W.-C., Leung, P.T. Y., & Morritt, D. (2010). Come rain or shine: the combined e?ects of physicalstresses on physiological and protein-level responses of an intertidal limpet in themonsoonal tropics. Functional Ecology, 25(1), 101?110.165Williams, S. L., Bracken, M. E. S., & Jones, E. (2013). Additive e?ects of physicalstress and herbivores on intertidal seaweed biodiversity. Ecology, 94(5), 1089?1101.Winder, M. & Schindler, D. E. (2004). Climate change uncouples trophic interac-tions in an aquatic ecosystem. Ecology, 85(8), 2100?2106.Wolkovich, E. M., Davies, T. J., Schaefer, H., Cleland, E. E., Cook, B. I., Travers,S. E., Willis, C. G., & Davis, C. C. (2013). Temperature-dependent shifts in phe-nology contribute to the success of exotic species with climate change. AmericanJournal of Botany, 100(7), 1407?1421.Wood, H. L., Spicer, J. I., & Widdicombe, S. (2008). Ocean acidification mayincrease calcification rates, but at a cost. Proceedings of the Royal Society B:Biological Sciences, 275(1644), 1767?1773.Woodward, G., Ebenman, B., Emmerson, M. C., Montoya, J. M., Olesen, J., Valido,A., & Warren, P. (2005). Body size in ecological networks. Trends in Ecology &Evolution, 20(7), 402?409.Wootton, J. T. & Pfister, C. A. (2012). Carbon system measurements and potentialclimatic drivers at a site of rapidly declining ocean pH. PLoS ONE, 7(12), e53396.Wootton, J. T., Pfister, C. A., & Forester, J. D. (2008). Dynamic patterns andecological impacts of declining ocean pH in a high-resolution multi-year dataset.Proceedings of the National Academy of Sciences of the United States of America,105(48), 18848?18853.Yagi, H., Ceccaldi, H. J., & Gaudy, R. (1990). Combined influence of temperatureand salinity on oxygen consumption of the larvae of the pink shrimp, Palaemonserratus (Pennant)(Crustacea, Decapoda, Palaemonidae). Aquaculture, 86(1), 77?92.Yaroslavtseva, L. M. & Sergeeva, E. P. (2006). Adaptivity of the bivalve Mytilustrossulus larvae to short-and long-term changes in water temperature and salinity.Russian Journal of Marine Biology, 32(2), 82?87.Zhao, Y. & Sun, S. (2006). E?ects of salinity, temperature and pH on the survivalof the nemertean Procephalothrix simulus Iwata, 1952. Journal of ExperimentalMarine Biology and Ecology, 328(2), 168?176.Zippay, M. L. & Helmuth, B. (2012). E?ects of temperature change on mussel,Mytilus. Integrative Zoology, 7(3), 312?327.Zippay, M. L. & Hofmann, G. E. (2010). E?ect of pH on Gene Expression and Ther-mal Tolerance of Early Life History Stages of Red Abalone (Haliotis rufescens).Journal of Shellfish Research, 29(2), 429?439.166


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items