Theor EcolDOI 10.1007/s12080-011-0122-4

ORIGINAL PAPER

Aquaculture-induced changes to dynamics of a migratoryhost and specialist parasite: a case study of pink salmonand sea lice

Jaime Ashander Martin Krkoek Mark A. Lewis

Received: 28 September 2010 / Accepted: 15 February 2011 The Author(s) 2011. This article is published with open access at Springerlink.com

Abstract Exchange of diseases between domesticatedand wild animals is a rising concern for conservation.In the ocean, many species display life histories thatseparate juveniles from adults. For pink salmon (On-corhynchus gorbuscha) and parasitic sea lice (Lepeoph-theirus salmonis), infection of juvenile salmon in earlymarine life occurs near salmon sea-cage aquaculturesites and is associated with declining abundance of wildsalmon. Here, we develop a theoretical model for thepink salmon/sea lice hostparasite system and use itto explore the effects of aquaculture hosts, acting as

J. Ashander M. A. LewisCentre for Mathematical Biology,Department of Math and Statistical Sciences,University of Alberta,Edmonton, AB, Canada

M. KrkoekSchool of Fisheries and Applied Sciences,University of Washington,Seattle, WA, USA

M. A. LewisDepartment of Biological Sciences,University of Alberta,Edmonton, AB, Canada

Present Address:J. Ashander (B)Center for Population Biology,Department of Environmental Science and Policy,University of California,Davis, CA, USAe-mail: jashander@ucdavis.edu

Present Address:M. KrkoekDepartment of Zoology, University of Otago, Dunedin,New Zealand

reservoirs, on dynamics. Because pink salmon have a2-year lifespan, even- and odd-year lineages breed inalternate years in a given river. These lineages can haveconsistently different relative abundances, a phenom-enon termed line dominance. These dominance rela-tionships between host lineages serve as a useful probefor the dynamical effects of introducing aquaculturehosts into this hostparasite system. We demonstratehow parasite spillover (farm-to-wild transfer) and spill-back (wild-to-farm transfer) with aquaculture hosts caneither increase or decrease the line dominance in anaffected wild population. The direction of the effectdepends on the response of farms to wild-origin infec-tion. If aquaculture parasites are managed to a constantabundance, independent of the intensity of infectionsfrom wild to farm, then line dominance increases. Onthe other hand, if wild-origin parasites on aquaculturehosts are proportionally controlled to their abundancethen line dominance decreases.

Keywords Hostparasite population dynamics Discrete-time models Spillover Spillback Aquaculture Salmon Sea lice

Introduction

In recent decades, exchange of diseases and parasitesbetween wild and domesticated animals has become aprominent concern for conservation and disease emer-gence (Daszak et al. 2000), as well as management ofpest species in food production (Bengis et al. 2002;Costello 2009). The occurrence of spillover, trans-fer of infection from wildlife to domestic hosts, andspillback, transfer from domestic hosts to wildlife,has been demonstrated in a wide variety of taxa

Theor Ecol

with migratory life histories, including terrestrial mam-mals (Cervus canadensis, Bison bison, Saiga tatarica;Cheville et al. 1998; Morgan et al. 2005), birds (Anser-iformes:Anatidae; Gilbert et al. 2006; Muzaffar et al.2006), and fishes (Oncorhynchus gorbuscha, O. keta;Krkoek et al. 2006, 2007b). Host migration plays animportant role in determining the effects of spilloverand spillback on conservation (Morgan et al. 2005;Krkoek et al. 2007b) and disease spread (Kilpatricket al. 2006). Foundational hostparasite and epidemio-logical models, however, do not include host movementor migration (Morgan et al. 2004). Thus, efforts tounderstand the role of host migration in parasite anddisease systems often employ complex tools, e.g. data-intensive statistical models (Kilpatrick et al. 2006) orparameter-heavy, spatially explicit simulation models(Morgan et al. 2007).

One application where simple models are successfulin understanding the effects of host migration and dis-ease exchange is in linking the declines of wild salmonpopulations in Europe and North America to theirassociation with sea-cage salmon aquaculture produc-tion (Ford and Myers 2008). One proposed explana-tion for the declines, at least in the Pacific, is thatspillover and spillback of parasitic sea lice (Lepeoph-theirus salmonis) between wild salmonids and farmsalmon leads to infections of juvenile wild salmon inearly marine life (Krkoek et al. 2006, 2007a; Costello2006, 2009). Without farms, infections do not occuruntil later in life because the migratory behaviour ofsalmon results in a spatial separation of juveniles inearly marine life from adults and their parasites, a char-acteristic termed migratory allopatry; host migrationcreates a barrier to adultjuvenile transmission duringearly marine entry (Krkoek et al. 2007b). Studies ofsea lice and salmon have described how aquaculturecan break down this migratory barrier when farms con-taining adult fish are placed on migration corridors forwild juvenile salmonids (recent reviews: Krkoek 2010;Costello 2009). Rather than develop complex, spatiallyexplicit models of hostparasite population dynamics,these studies have focused on the consequences ofinfections during early marine life (that are due to theinteraction of host migration and aquaculture). Usingthis approach, Krkoek et al. (2007a) demonstrateddeclines in pink salmon populations associated withepizootics in aquaculture regions. Another theoreti-cal paper reports a probabilistic analysis of infectionpressures generated by various stocking levels of farmsto explain these declines (Frazer 2009). To date theprimary focus, however has been effects on equilibrium

abundance of wild hosts. Salmon farming effectivelyaugments host diversity, and according to epidemiolog-ical theory, this should affect behaviour of populationdynamics as well (Dobson 2004).

Here, to understand the implications of parasitespillover and spillback with farm hosts for popula-tion dynamics, we introduce a model for the hostparasite system of wild pink salmon and parasitic sealice. In keeping with the approach of earlier work, weuse the classic Ricker (1954) model for pink salmonpopulation dynamics, which we couple with a simpletransmission model derived under the assumption ofmigratory allopatry. This permits a rather simple, spa-tially implicit, formulation of the model as a system ofdiscrete-time equations. Because pink salmon have a2-year lifespan, even- and odd-year lineages breed inalternate years in a given river. These lineages can haveconsistently different relative abundances of adults, aphenomenon termed line dominance in the salmonliterature (Groot and Margolis 1991). Mathematically,line dominance arises in our model through a period-doubling bifurcation, which links the degree of dom-inance with the strength of inter-lineage interactions.To understand the effects of introduced aquaculturehosts on wild host population dynamics, we focus onchanges in this dominance relationship, demonstratingthat a line dominance naturally maintained by negativedensity-dependent interactions between lineages canbe altered by the introduction of farm hosts.

Models

Sea lice are native copepod ectoparasites common onboth adult wild hosts and in sea-cage salmon aqua-culture (Costello 2009). These parasites have a directlife cycle, being transmitted primarily as planktoniclarvae that become infective copepodids after survivingthrough a naupliar period lasting between 1 to 9 days(Johnson and Albright 1991). Once the copepoditeattaches to a host fish, lice develop through a sequenceof attached chalimus stages before becoming motilepre-adults and adult lice. Adult lice reproduce sexuallyon a host fish, with the female extruding eggstringsfrom which nauplii hatch, completing the life cycle. Thefeeding activity of lice on host surface tissues can causemorbidity and mortality of host fish, as well as sublethaleffects that may influence survival via predation risk.

Pink salmon, like many fish, display migratoryallopatry in which juvenile fish are spatially sepa-rated from adult fish due to differences in habitat

Theor Ecol

requirements, food supply, natural enemies, and mi-gration. With a 2-year generation time, a pink salmonpopulation consists of two distinct lineages, or yearclasses, that use river (breeding) and ocean (maturing)habitat sequentially in time. Without farms (Fig. 1a),pink salmon of different lineages potentially interactthrough two means: (1) through effects on the environ-ments (river and ocean habitat) that are sequentiallyused by the juvenile and adult age classes in alternateyears, and (2) through transmission of a specialist par-asite from adult hosts to juvenile hosts, i.e. betweenlineages.

Interactions of the first type occur through changesin the biotic or abiotic environment of the river orocean habitat due to host density and include a varietyof mechanisms proposed by Ricker (1962) to explainline dominance in pink salmon, including direct sup-pressive interactions between lineages, fouling of therearing environment by large runs, and competitionfor food in ocean habitat (Groot and Margolis 1991).Interactions of the second type occur due to parasitetransmission, when the offspring of parasites associatedwith adults of one lineage infect juveniles of the otherlineage as the two age classes of host temporarily sharespace during migration. These infection windowsare shown as shaded regions of Fig. 1a between theriver and ocean habitat. We assume that the density-dependent interactions of both types act to increasemortality.

Introduction of farmed hosts to migration routesbetween river and ocean habitats (Fig. 1b) can leadt