Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Orthohantavirus wikipedia , lookup
Middle East respiratory syndrome wikipedia , lookup
Hepatitis B wikipedia , lookup
Ebola virus disease wikipedia , lookup
West Nile fever wikipedia , lookup
Marburg virus disease wikipedia , lookup
Herpes simplex virus wikipedia , lookup
Henipavirus wikipedia , lookup
ICES Journal of Marine Science ICES Journal of Marine Science (2015), 72(Supplement 1), i139 –i146. doi:10.1093/icesjms/fsu209 Contribution to the Supplement: ‘Lobsters in a Changing Climate’ Original Article Modelling the spread and connectivity of waterborne marine pathogens: the case of PaV1 in the Caribbean Andrew S. Kough 1*, Claire B. Paris 1, Donald C. Behringer 2,3, and Mark J. Butler, IV4 1 Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, Miami, FL 33149, USA Fisheries and Aquatic Sciences Program, University of Florida, Gainesville, FL 32653, USA 3 Emerging Pathogens Institute, University of Florida, Gainesville, FL 32611, USA 4 Department of Biological Sciences, Old Dominion University, Norfolk, VA 23529, USA 2 *Corresponding author: tel: +1 240 447 0426; e-mail: [email protected] Kough, A. S., Paris, C. B., Behringer, D. C., and Butler, M. J.IV Modelling the spread and connectivity of waterborne marine pathogens: the case of PaV1 in the Caribbean. – ICES Journal of Marine Science, 72: i139 – i146. Received 25 August 2014; revised 6 October 2014; accepted 28 October 2014; advance access publication 21 November 2014. The PaV1 virus infects spiny lobsters (Panulirus argus) throughout most of the Caribbean, where its prevalence in adult lobsters can reach 17% and where it poses a significant risk of mortality for juveniles. Recent studies indicate that vertical transmission of the virus is unlikely and PaV1 has not been identified in the phyllosoma larval stages. Yet, the pathogen appears subclinically in post-larvae collected near the coast, suggesting that lobster post-larvae may harbour the virus and perhaps have aided in the dispersal of the pathogen. Laboratory and field experiments also confirm the waterborne transmission of the virus to post-larval and early benthic juvenile stages, but its viability in the water column may be limited to a few days. Here, we coupled Lagrangian modelling with a flexible matrix model of waterborne and post-larval-based pathogen dispersal in the Caribbean to investigate how a large area with complex hydrology influences the theoretical spread of disease. Our results indicate that if the virus is waterborne and only viable for a few days, then it is unlikely to impact both the Eastern and Northwestern Caribbean, which are separated by dispersal barriers. However, if PaV1 can be transported between locations by infected post-larvae, then the entire Caribbean becomes linked by pathogen dispersal with higher viral prevalence in the North. We identify possible regions from which pathogens are most likely to spread, and highlight Caribbean locations that function as dispersal “gateways” that could facilitate the rapid spread of pathogens into otherwise isolated areas. Keywords: disease, lobster, Panulirus argus, physical oceanography, virus. Introduction Within the dense and stable medium of seawater, many animals have evolved complex life cycles that include a distinctive planktonic larval phase that facilitates population dispersal. Marine larvae can harness powerful currents and potentially travel long distances which led to the once widely accepted paradigm that marine populations are “open” (Roughgarden et al., 1985). However, we now know that many larvae have specializations and behaviours that limit their dispersal (Paris et al., 2007; Butler et al., 2011; Staaterman et al., 2012). These adaptations enhance the local retention of larvae (Jones et al., 1999; Paris and Cowen, 2004; Almany et al., 2007), resulting in a continuum of dispersal potentials. Although this theory was promulgated on meroplankton, the dispersal of other important ecosystem components, such as marine pathogens, may be similar but remain understudied (Harvell et al., 2004). # International Pathogens play an important role in shaping marine communities (McCallum et al., 2001), where their dispersal and spread are quite different from that of the terrestrial environment. Marine pathogens tend to spread more rapidly than in terrestrial systems (McCallum et al., 2003, 2004) and can fundamentally alter ecosystems (Harvell et al., 1999). For example, an unknown pathogen that rapidly spread around the Caribbean in the 1980s killed longspined sea urchins (Diadema antillarium) en masse, removed an important herbivore from the system, and contributed to a shift towards macroalgae-dominated reefs (Lessios et al., 1984; Lessios, 1988). Emerging diseases that afflict habitat engineers, such as corals or seagrasses, may become more prevalent as the climate shifts (Harvell et al., 1999; Ward and Lafferty, 2004) and could alter their range in tandem with their hosts. Unfortunately, the dispersal, epidemiology, and population dynamics of most marine Council for the Exploration of the Sea 2014. All rights reserved. For Permissions, please email: [email protected] i140 diseases are unknown, and many diseases are never detected (Harvell et al., 2004). Many pathogens affect both aquacultured and wild species of interest to fisheries, so understanding the dynamics of disease in both is crucial (Daszak et al., 2000). One such pathogen, the virus Panulirus argus Virus 1 (PaV1), infects the Caribbean spiny lobster P. argus (Shields and Behringer, 2004) and is found throughout much of the wider Caribbean (Moss et al., 2013). PaV1 is a large, icosahedral, non-enveloped, DNA virus, and is primarily pathogenic to juvenile lobsters (Shields and Behringer, 2004; Butler et al., 2008; Behringer et al., 2012a, b). The high mortality that juveniles suffer adds another challenge to managing this fully exploited fishery (Chavez, 2009; Ehrhardt et al., 2010). Caribbean spiny lobsters have phyllosoma larvae which have a long pelagic larval duration of 6 months or more (Goldstein et al., 2008) and thus great dispersal potential. Research using genetics (Silberman et al., 1994; Naro-Maciel et al., 2011) and biophysical modelling (Briones-Fourzan et al., 2008; Butler et al., 2011; Kough et al., 2013) suggest that the international exchange of larvae is common so that populations around the Caribbean are linked through larval transport. Likewise, strains of PaV1 from disparate countries share phenotypes (Moss et al., 2013) and some post-larvae that arrive onshore in Florida are infected with PaV1 (Moss et al., 2012), making infected larvae a potential dispersal pathway for the pathogen. Experiments testing vertical transmission from female to larvae have proven negative (Behringer and Butler, unpublished data), so the mechanism or vector of dispersal for PaV1 among distant populations remains elusive. Thus far, no PaV1-infected phyllosoma larvae have been detected in field collections (Lozano-Alvarez et al., unpublished data). However, some post-larvae collected near the Florida coast (Moss et al., 2012) and others many kilometres off of the continental shelf of the Yucatan peninsula (Lozano-Alvarez et al., unpublished data) were found subclinically infected with PaV1. Post-larval lobsters are incapable of feeding (Wolfe and Felgenhauer, 1991) and live off concentrated lipids stored during the long larval phase (Lemmens, 1994; Lemmens and Knott, 1994), so ingestion of the PaV1 pathogen by post-larvae is not a viable transmission pathway. Whether late-stage phyllosoma larvae acquire the virus through ingestion of infected prey or marine aggregates and then retain the virus after metamorphosis into post-larvae, or if only post-larvae are susceptible to infection and the virus is acquired offshore is as yet unknown. Regardless, the presence of PaV1-infected post-larvae means that the virus can potentially spread over long distances within infected, planktonic hosts. Our goal was to compare the genetic relatedness of PaV1 strains in the Caribbean (Moss et al., 2013) with modelled patterns of pathogen spread. We used biophysical Lagrangian stochastic modelling coupled to a stepping-stone model of pathogen spread to describe the probable dispersal of a waterborne pathogen compared with behaviourally adroit P. argus post-larvae in the Caribbean Sea. We tested two hypotheses for how PaV1 may be spread, as: (i) a passive waterborne particle within the top 10 m of the ocean, or (ii) a lobster post-larvae infected with PaV1. Our approach evaluates how hydrology could affect the potential spread of a marine pathogen throughout the Caribbean. Methods We used an advanced multi-scale open-source Lagrangian biophysical model of dispersal, the Connectivity Modeling System (CMS; Paris et al., 2013), to probabilistically describe Caribbean-wide connections between reef sites for 5 years, from the start of 2004 through A. S. Kough et al. the end of 2008. We then used the resulting matrices to randomly and repeatedly seed an ensemble of simulations with pathogen emergence and subsequent spread. Habitat sites We chose habitat sites containing coral reefs from those identified in the Millennium Coral Reef Mapping Project (Andréfouët et al., 2006). We overlaid an 8 × 8 km grid across the Caribbean and created 3202 sites intersecting the reef extent, following Holstein et al. (2014). Real-world spatial data do not always translate into operational hydrodynamic models, and can rest on boundaries between land and water in the model, rendering them ineffective. We removed 806 sites that were impacted by the structure of the modelling framework. Further, we restricted our simulations to sites that both exchanged and received larvae (removing another 714 sites) so that we were working with a theoretical network with no disconnected sites. The resulting 1682 habitat sites stretched across the Caribbean, from the Barbados to the northern edge of the Florida Keys (USA) reef track (Figure 1). Lagrangian tracking We coupled and nested offline operational hydrodynamic models with the CMS. More specifically, we used the 1/128 Hybrid Coordinate Ocean Model (HyCOM) with NCODA assimilated and the 1/248 HyCOM Gulf of Mexico models (Bleck, 2002) that are publicly available and cover our study location. Both models yield daily output, resolve mesoscale eddy features, and have been widely applied to many different ocean transport scenarios. The habitat seascape and hydrodynamic models were integrated within the CMS for simulations of particle transport. We used highfrequency daily particle releases to model connectivity, corresponding to the daily windforcing of the HyCOM models (Bleck, 2002). A sensitivity analysis indicated that 100 particles released daily from each reef habitat site achieved probabilistic saturation at particle tracking times longer than those used in this study (Kough, 2014). Simulations We tested two different hypotheses on how PaV1 virions may be transported around the Caribbean: in seawater or in infected postlarvae. Our first hypothesis was that the pathogen is passively advected as a waterborne particle among habitat locations in the surface (,10 m) water layer (Waterborne simulation: WB). In the WB scenario, we used a pelagic duration of 5 d, which corresponds to laboratory estimates of how long PaV1 can be infectious in seawater (Butler and Behringer, unpublished data). Considering PaV1 virions lack a structural envelope (Shields and Behringer, 2004), they are apt to be more environmentally stable (see Satheesh Kumar et al., 2013 and references therein), which supports the possibility that PaV1 spreads in the water column itself. Unlike non-enveloped viruses such as PaV1, the lipoprotein layer that typically covers enveloped viruses and functions in host cell recognition and penetration, is fragile, and can limit their persistence. The 5 d duration in seawater that we modelled for PaV1 also fits well within the range of viabilities known for several other marine viruses. The viability of all viruses depends on environmental characteristics, particularly salinity and temperature, but the examples that follow all come from experiments that were conducted in seawater (36 psu) and under temperatures likely to be experienced by a virus in the tropical Caribbean Sea (20–338C). Some of these viruses have envelopes, such as white spot syndrome virus, a large DNA virus of decapod crustaceans that can remain viable in the Marine pathogen dispersal i141 Figure 1. Habitat map and mean pathogen spread times around the Caribbean. The habitat map (a) shows the centroids of the 8 × 8 km square coral reef habitat locations (n ¼ 1682) used in the simulations. The mean number of steps in the model (equivalent to time) for PaV1 to spread to each habitat location from random origins is shown for two scenarios: waterborne (WB; b) and post-larvae (PL; c). Colour is used to identify locations on the map and links the position of the habitat on the x-axis in the panels (b and c) of the mean pathogen spread time to a spatial location (a). The position of the habitat identified here is consistent throughout the paper. seawater for 12 d (Satheesh Kumar et al., 2013), or the viral haemorrhagic septicaemia virus, an enveloped RNA virus of fish that can remain viable in seawater for 4 d (Hawley and Garver, 2008). Other viruses such as PaV1 are non-enveloped or “naked”, such as the poliovirus, echovirus, and coxsackie virus; all are RNA mammal viruses that remain viable in seawater for 1–12 d (Fujioka et al., 1980; Labelle and Gerba, 1980; Block, 1983), 3 d (Labelle and Gerba, 1980), and 30 d (Nasser et al., 2003), respectively. Our second hypothesis was that after metamorphosis from the phyllosoma larval phase, presumably seaward of the continental shelf (McWilliam and Phillips, 2007; Phillips and McWilliam, 2009), P. argus post-larvae become infected with PaV1 and the pathogen spreads among locations during the post-larval competency period (Postlarvae simulation: PL). In the PL scenario, we used a variable pelagic duration of 5 –14 d based on estimates of the average duration (Goldstein et al., 2008) and metabolic i142 A. S. Kough et al. constraints (Wilkin and Jeffs, 2011) of the post-larval stage. For the PL, we included a daily vertical migration behaviour from the surface layer at night to depths of at least 100 m during the day as hypothesized by Phillips et al. (1978) for P. cygnus and Butler et al. (2011) for P. argus. Connectivity probability matrices M(i,j) were constructed for the WB and PL scenarios using the modelled particles that successfully arrived at a habitat site ( j) from a habitat site (i), after the pelagic duration. Particlesij Mij = . Particlesi measure of how important a habitat location is to the entire connectivity network. It was formulated by calculating how many times a given habitat location is involved in the fastest pathway of pathogen spread between any other two locations in the network. Thus, we used the time to infection as an edge weight, to assess how rapidly pathogens from one location spread throughout the system. In this context, locations in the Caribbean with the highest betweenness centrality played crucial roles in rapidly sending pathogens to other areas. Calculations were made in MATLABw using the BGL toolbox developed by Gleich (2006). Results Waterborne Stepping-stone pathogen spread model We used the probability connectivity matrices from the Lagrangian tracking simulations in a MATLABw model (Supplementary material) that seeded a randomly selected location in the Caribbean with disease and then spread the pathogen elsewhere based on either the WB or PL simulation scenario (Figure 2). The generality of the model is improved by incorporating stochasticity in the spatial origin and connectivity of the pathogen, which is pertinent for PaV1 as its origin is unknown and the probability of exchange between any two locations is likely to change from year to year based on factors such as temporal variability in the prevalence of PaV1 (Behringer et al., 2011) or changing atmospheric and oceanic conditions (e.g. ENSO, current positions and meanders, mesoscale eddy formation, etc.). Gateway habitats To evaluate how important any given habitat location is in the disease network, we considered its potential for facilitating further spread using betweenness centrality. Betweenness centrality is a In the WB scenario, the Eastern and Central Caribbean were not reached by pathogens originating in the North (Figure 3). Indeed, many regions do not exchange disease with other regions in this scenario and were isolated. This isolation delayed (i.e. more model steps and longer mean time for pathogen spread) or prevented the spread of the pathogen to other locations. Spread-time greatly increased near major oceanographic barriers: for example, near the Turks and Caicos Islands and the island of Hispanola (see Figures 1b and 3WB). The locations with the most betweenness centrality were clustered near the northwest coast of Cuba, and the northeast coast of the Yucatan (Figure 4). The connections at these sites represent a pathway for disease to spread from the Yucatan into the Northern Caribbean and would expand the network. Post-larvae The modelled pathogen spread faster throughout the Caribbean in the PL scenario than in the WB scenario, regardless of origin (see Figure 1c vs. b). The pathogen spread more widely between the Northern and the Central and Eastern Caribbean (see Figure 3PL) and spread more rapidly from the Central and Eastern Caribbean into the Northwestern Caribbean than vice versa. In both the PL and WB scenarios, the Lesser Antilles remained isolated, because it took more model steps for pathogen to spread to them or it never reached them. However, the Florida Keys took longer to receive the pathogen in the PL scenario than it did in the WB scenario (see Figure 3; the leftmost columns are darker in WB than in PL). The locations with the highest betweenness centrality in the PL scenario were next to oceanographic features such as the Windward, Mona, and Anegada passages, the Gulf of Colombia and Gulfo de los Mosquitos, the Nicaraguan Rise, and the Yucatan Channel (see Figure 4). Discussion Figure 2. A flowchart summarizing the steps in the pathogen stepping-stone simulations. A pathogen was seeded 10 000 times from a random starting location in the Caribbean. For each seeding event, the pathogen could spread between locations for 300 model steps. During each step, the probability of dispersal away from an infected location was taken from a matrix of the 5-year mean modelled particle transport specific to either a waterborne or post-larval scenario. This probability was augmented by a stochastic addition to better capture unpredictability between years. Our findings support the expected pattern of linkages among areas in the Caribbean based on well-studied oceanic circulation. The Northern, Central, and Eastern Caribbean were isolated from each other when the pathogen was waterborne and short-lived. Particles have to travel against the prevailing flow to spread from the Northern into the Central and Eastern Caribbean, so this was expected. It is unlikely that the two regions would rapidly share a disease outbreak that originated in the North, although some exchange is possible through pathways from the Greater into the Lesser Antilles and on into the gyres off of Colombia and Panama. Thus, even short-term waterborne transport could spread PaV1 through the Caribbean over long enough periods of time. However, when we modelled the pathogen travelling within an Marine pathogen dispersal i143 Figure 3. Time matrices for the Waterborne (WB) pathogen simulation (left) and Post-larval (PL) pathogen simulation (right) that show how fast the pathogen spread from one specific location (y-axis) to another specific location (x-axis). The position of habitat on both axes corresponds to previously used colour (see Figure 1a) and x-axis position (see Figure 1b and c). The shade of grey represents how many model steps it took for the pathogen to spread from one location to the next. There was no spread of disease among locations shown in white, slow spread between areas shown in lighter shades of gray, and rapid spread between areas shown in black. Figure 4. Map showing habitats that functioned as “gateways” from which the modelled pathogens spread most rapidly to other habitats. Locations denoted with white stars (waterborne scenario) and grey squares (post-larval scenario) were the areas we defined as “gateways” based on the betweenness centrality values of their pathogen spread time. These were the sites with the highest betweenness centrality (same order of magnitude) in the networks of each scenario. The prevailing surface currents in the Caribbean are shown with the dotted line, and major oceanographic features are labelled. infected post-larval host, there was considerably more exchange between the Northern and Southern Caribbean. Regardless of the pathogen transport scenario, it took more model time-steps for the pathogen to spread to the Lesser Antilles, South American Coast, and Central America. Thus, a highly virulent disease that is mainly limited by oceanographic dispersal should spread to i144 these locations last. We also identified certain locations that acted as “gateways” through which the pathogen spread rapidly among habitats. Gateways tended to be next to major currents and to other oceanographic divides. These gateways facilitated connections between localities that were otherwise isolated, and therefore are important locations to focus potential management for future epizootics. How do the two scenarios that we modelled (e.g. WB vs. PL) compare with the geographic structure of relatedness between strains of PaV1 around the Caribbean (Moss et al., 2013)? The WB scenario better captured the isolation and absence of PaV1 in the Lesser Antilles, and in the South and Central Caribbean. Yet, only the PL scenario was consistent with genetic linkages between the PaV1 virus in Panama and the Northern Caribbean (Moss et al., 2013), as it created increased exchange of the pathogen from Central and South America into the Central and Northern Caribbean. However, the number of infected lobsters found in Panama and the Southern Caribbean was low and the authors noted that they had the most diversity where they had the highest sample size (Moss et al., 2013). Still, the presence of shared variants between Central America and the Greater Antilles indicates that the dispersal method for PaV1 requires a waterborne or vectored pathogen that is viable in seawater for more than a few days. From the perspective of the pathogen, post-larvae represent an idealized method of dispersal. Post-larvae use chemical cues (Herrnkind and Butler, 1986; Goldstein and Butler, 2009) and tidal signals (Kough et al., 2014) to reach nursery habitat—an optimal location for a pathogen to reach susceptible juvenile hosts. Still, how do the non-feeding post-larvae acquire the pathogen, if they are not first infected as larvae? Early benthic juvenile lobsters are highly susceptible to PaV1 and can be readily infected through viral particles in the water column (Butler et al., 2008). Spiny lobster post-larvae come onshore during the dark phase of the lunar month with nocturnal flood tides (Acosta et al., 1997; Yeung et al., 2001), navigating from offshore using chemical, tidal, and probably other cues (Goldstein and Butler, 2009; Kough et al., 2014). Post-larvae orient towards coastal water masses, which may expose them to viral particles or other post-larvae already infected within the nursery habitat. Indeed, the model results suggest that the pathogen could spread across the Caribbean travelling within post-larvae if they remain subclinically infected and offshore for an extended time. In addition, stronger linkages between the North and South would exist and be maintained if the pathogen had origins in South or Central America. The transport of pathogens within marine aggregates is perhaps another way that coastal communities around the Caribbean may be linked. Our results in the WB scenario demonstrate that passive particle dispersal connects otherwise disparate areas, even over short time-scales. Many organisms, chemicals, and gradients are transported throughout the water column (Maddox, 2006) and drifting debris can harbour infectious pathogens (Lizarraga-Partida et al., 2011) and a diverse microbial community (Lyons et al., 2010). Marine microbial communities can transition into opportunistic pathogens that afflict plants and animals, especially in a changing ocean (Burge et al., 2013). The floating assemblages of biological materials such as mucilaginous aggregates and algae offshore of coastal communities can play an important role for larval recruitment as shelter and perhaps as a signal of nearby benthic habitat (Wells and Rooker, 2004; Goldstein and Butler, 2009). If such aggregates are used by meroplankton transitioning from the pelagic into benthic habitats, these could be locations where concentrations of A. S. Kough et al. susceptible animals come into contact with pathogens such as PaV1. Marine aggregates may also function as vessels for keeping the virus protected (e.g. from UV light) and viable (Bongiorni et al., 2007), and thus aid in pathogen transport for longer times and potentially larger distances. Our discovery of key gateways that facilitate pathogen connectivity among localities that would otherwise be isolated is meaningful for marine pathogen epidemiology and potential management of the spread of disease. The Greater Antilles have the highest PaV1 genetic diversity (Moss et al., 2013) and also contained the highest number of gateway habitats in the PL simulation scenario, supporting the contention that such locations acted as points connecting a large number of origins and destinations. In the WB scenario, the Caribbean network was split into two relatively separate areas: northwest and southeast of the prevailing current. The major gateways in the WB scenario straddled the primary pathway for linkage between these two areas and occurred at the northeastern tip of Mexico and on the west coast of Cuba. Very few gateway locations were shared between the two scenarios, although there were overlaps at the Cay Sal Bank (western Bahamas) and southeastern Cuba. The gateways in the Cay Sal Bank facilitated connections into the Bahamas and Florida, and were an important intermediary location next to the powerful Florida current. Active management of fishery activities that potentially increase pathogen prevalence or transmission (Behringer et al., 2012a, b) within gateway sites may help to forestall the spread of pathogens among regions of the Caribbean. As with any study, there are certain caveats to our results that bear mention. The time-scale of the Lagrangian modelling was 5 years, which does not encompass the full length of time since PaV1 was first detected. However, our aim was not to recreate the actual pattern of the spread of PaV1 disease, which is unknown, but to instead explore the potential way in which water- or vector-borne pathogens may transition between locations. By tying a stochastic stepping-stone model to the connectivity matrices generated by the Lagrangian particle tracking model, we added variation to simulate periodic, but important events. This construct could miss rare events that happen with a periodicity of ,5 years, but the mean connectivity measured over shorter periods is comparable to the connectivity over longer time-scales (Berglund et al., 2012). We also simulated the spread of a waterborne pathogen using only a 5 d duration, a time frame believed relevant to the PaV1 virus. The viability of many seawater-borne pathogens is poorly known, but their duration will clearly affect patterns of connectivity, as will other factors that we did not simulate (e.g. alternate hosts, dormant stages, host population dynamics, etc.). The CMS also used general ocean circulation models that best represented geostrophic flow (Bleck, 2002); the nearshore tidal and local wind-driven dynamics were not well represented in the general ocean circulation. However, we were interested in the transport through pelagic oceans that could advect a pathogen across greater distances than possible by nearshore processes. When the modelled pathogen originated in Florida in the PL scenario, it spread in a countercurrent direction similar to the well-documented invasion of the Caribbean by lionfish (Morris and Whitfield, 2009). Therefore, although the PL scenario used a shorter dispersal time than a lionfish larva, our model’s results were comparable to a known pattern of Caribbean-wide invasion. In addition, using only open-access tools, such as HyCOM and CMS, makes our work easily reproducible and accessible for other researchers to build on while investigating dispersal of other marine pathogens. Marine pathogen dispersal Although we were inspired to model pathogen dispersal by the discovery and subsequent Caribbean-wide interest in PaV1, our work is relevant to a myriad of different scenarios and organisms affected by marine pathogens. Distant pathogen dispersal could be tied into well-studied disease and ecosystem model systems (Cáceres et al., 2014; Dolan et al., 2014), therefore linking the oceanographic environment to population-based models and yielding a more comprehensive epidemiological model (McCallum et al., 2004). Indeed, considering distant dispersal in tandem with disease dynamics has better explained the outbreak of Aspergillosis in seafans throughout the Caribbean (Bruno et al., 2011) and the persistence of PaV1 in lobster populations in Florida (Dolan et al., 2014). Still, our understanding of the large-scale epidemiology of marine diseases is in its infancy and will likely require analyses that integrate physical dynamics, anthropogenic influence, and disease history so as to identify potential hotspots of marine pathogens in the world’s oceans, similar to work in terrestrial environments (Jones et al., 2008). Benthic marine ecosystems receive new recruits, and potentially pathogens, from ocean currents that unite habitat over broad spatial and temporal scales, making Lagrangian modelling an appropriate tool for examining the potential connectivity of marine diseases. Supplementary data Supplementary material is available at ICESJMS online version of the manuscript. Acknowledgements This project greatly benefited from advice and direction given by Michael Schmale, Tom Dolan, and two anonymous reviewers. Funding was provided through the National Science Foundation awards OCE-0928930 (CBP), OCE-0929086 (MJB), and OCE0928398 (DCB). References Acosta, C. A., Matthews, T. R., and Butler, M. J., IV. 1997. Temporal patterns and transport processes in recruitment of spiny lobster (Panulirus argus) postlarvae to South Florida. Marine Biology, 129: 79 –85. Almany, G. R., Berumen, M. L., Thorrold, S. R., Planes, S., and Jones, G. P. 2007. Local replenishment of coral reef fish populations in a marine reserve. Science, 316: 742– 744. Andréfouët, S., Muller-Karger, F. E., Robinson, J. A., Kranenburg, C. J., Torres-Pulliza, D., Spraggins, A., and Murch, B. 2006. In 10th ICRS. Global Assessment of Modern Coral Reef Extent and Diversity for Regional Science and Management Applications: a View from Space, pp. 1732– 1745. Japanese Coral Reef Society, Okinawa, Japan. Behringer, D. C., Butler, M. J., IV, Shields, J. D., and Moss, J. 2011. Review of Panulirus argus Virus 1 (PaV1)—a decade after its discovery. Diseases of Aquatic Organisms, 94: 153 – 160. Behringer, D. C., Butler, M. J., IV, and Stentiford, G. D. 2012a. Disease effects on lobster fisheries and ecology. Diseases of Aquatic Organisms, 100: 89 –93. Behringer, D. C., Moss, J., Shields, J. D., and Butler, M. J., IV. 2012b. PaV1 infection in the Florida spiny lobster fishery and its effects on trap function and disease transmission. Canadian Journal Fisheries and Aquatic Sciences, 69: 136– 144. Berglund, M., Jacobi, M. N., and Jonsson, P. R. 2012. Optimal selection of marine protected areas based on connectivity and habitat quality. Ecological Modeling, 240: 105 – 112. Bleck, R. 2002. An oceanic general circulation model framed in hybrid isopycnal-cartesian coordinates. Ocean Modeling, 4: 55 – 88. i145 Block, J. C. 1983. Viruses in environmental waters. In Viral Pollution of the Environment. Ed. by G. Berg. CRC Press, Boca Raton. Bongiorni, L., Armeni, M., Corinaldesi, C., Dell’Anno, A., Pusceddu, A., and Danovaro, R. 2007. Viruses, prokaryotes and biochemical composition of organic matter in different types of mucilage aggregates. Aquatic and Microbial Ecology, 49: 15– 23. Briones-Fourzan, P., Candela, J., and Lozano-Alvarez, E. 2008. Postlarval settlement of the spiny lobster Panulirus argus along the Caribbean coast of Mexico: patterns, influence of physical factors, and possible sources of origin. Limnology and Oceanography, 53: 970– 985. Bruno, J. F., Ellner, S. P., Vu, I., Kim, K., and Harvell, C. D. 2011. Impacts of aspergillosis on sea fan coral demography: modeling a moving target. Ecological Monographs, 81: 123– 139. Burge, C. A., Kim, C. J. S., Lyles, J. M., and Harvell, C. D. 2013. Special issue oceans and humans health: the ecology of marine opportunists. Microbial Ecology, 65: 869 – 879. Butler, M. J., IV, Behringer, D. C., and Shields, J. D. 2008. Transmission of Panulirus argus virus 1 (PaV1) and its effect on the survival of juvenile Caribbean spiny lobster. Diseases of Aquatic Organisms, 79: 173– 182. Butler, M. J., IV, Paris, C. B., Goldstein, J. S., Matsuda, H., and Cowen, R. K. 2011. Behavior constrains the dispersal of long-lived spiny lobster larvae. Marine Ecology Progress Series, 422: 223– 237. Cáceres, C. E., Tessier, A. J., Duffy, M. A., and Hall, S. R. 2014. Disease in freshwater zooplankton: what have we learned and where are we going? Journal of Plankton Research, 36: 1 – 8. Chavez, E. A. 2009. Potential production of the Caribbean spiny lobster (Decapoda, Palinura) fisheries. Crustaceana, 82: 1393– 1412. Daszak, P., Cunningham, A. A., and Hyatt, A. D. 2000. Wildlife ecology—emerging infectious diseases of wildlife—threats to biodiversity and human health. Science, 287: 443– 449. Dolan, T. W., Butler, M. J., IV, and Shields, J. D. 2014. Host behavior and habitat structure alter disease dynamics in a social marine species: the Caribbean spiny lobster. Ecology, 95: 2346– 2361. Ehrhardt, N. M., Puga, R., and Butler, M. J., IV. 2010. Large ecosystem dynamics and fishery management concepts: the Caribbean spiny lobster, Panulirus argus, fisheries. In Towards Marine EcosystemBased Management in the Wider Caribbean. Ed. by L. Fanning, R. Mahon, and P. McConney. Amsterdam University Press, Amsterdam. Fujioka, R. S., Loh, P. C., and Lau, L. S. 1980. Survival of human enteroviruses in the Hawaiian ocean environment: evidence for virus-inactivating microorganisms. Applied and Environmental Microbiology, 39: 1105– 1110. Gleich, D. 2006. Matlab BGL v4.0. http://dgleich.github.io/matlabbgl/ (last accessed 18 August 2014). Goldstein, J. S., and Butler, M. J., IV. 2009. Behavioral enhancement of onshore transport by postlarval Caribbean spiny lobster (Panulirus argus). Limnology and Oceanography, 54: 1669 – 1678. Goldstein, J. S., Matsuda, H., Takenouchi, T., and Butler, M. J. I. V. 2008. The complete development of larval Caribbean spiny lobster (Panulirus argus) in culture. Journal of Crustacean Biology, 28: 306– 327. Harvell, C. D., Kim, K., Burkholder, J. M., Colwell, R. R., Epstein, P. R., Grimes, D. J., Hofmann, E. E., et al. 1999. Emerging marine diseases—climate links and anthropogenic factors. Science, 285: 1505– 1510. Harvell, D., Aronson, R., Baron, N., Connell, J., Dobson, A., Ellner, S., Gerber, L., et al. 2004. The rising tide of ocean diseases: unsolved problems and research priorities. Frontiers in Ecology and the Environment, 2: 375 – 382. Hawley, L. M., and Garver, K. A. 2008. Stability of viral hemorrhagic septicemia virus (VHSV) in freshwater and seawater at various temperatures. Diseases of Aquatic Organisms, 82: 171– 178. Herrnkind, W. F., and Butler, M. J., IV. 1986. Factors regulating settlement and microhabitat use by juvenile spiny lobsters, Panulirus argus. Marine Ecology Progress Series, 34: 23 – 30. i146 Holstein, D. M., Paris, C. B., and Mumby, P. J. 2014. Consistency and inconsistency in multispecies population network dynamics of coral reef ecosystems. Marine Ecology Progress Series, 499: 1 – 18. Jones, G. P., Millicich, M. J., Emslie, M. J., and Lunow, C. 1999. Selfrecruitment in a coral reef fish population. Nature, 402: 802– 804. Jones, K. E., Patel, N. G., Levy, M. A., Storeygard, A., Balk, D., Gittleman, J. L., and Daszak, P. 2008. Global trends in emerging infectious diseases. Nature, 451: 990 – 994 Kough, A. S. 2014. Population connectivity and larval dispersal of Caribbean spiny lobster. Dissertation, University of Miami, Miami, FL, 177 pp. Kough, A. S., Paris, C. B., and Butler, M. J., IV. 2013. Larval connectivity and the international management of fisheries. PLoS ONE, 8: e64970. Kough, A. S., Paris, C. B., and Staaterman, E. 2014. In situ swimming and orientation behavior of spiny lobster (Panulirus argus) postlarvae. Marine Ecology Progress Series, 504: 207– 214. LaBelle, R. L., and Gerba, C. P. 1980. Influence of estuarine sediment in viral survival under field conditions. Applied Environmental Microbiology, 39: 749– 755. Lemmens, J. W. T. J. 1994. Biochemical evidence for absence of feeding in puerulus larvae of the Western rock lobster Panulirus cygnus (Decapoda: Palinuridae). Marine Biology, 118: 383 – 391. Lemmens, J. W. T. J., and Knott, B. 1994. Morphological changes in external and internal feeding structures during the transition phyllosoma-puerulus-juvenile in the Western Rock Lobster (Panulirus cygnus, Decapoda: Palinuridae). Journal of Morphology, 220: 271–280. Lessios, H. A. 1988. Mass mortality of Diadema antillarum in the Caribbean: what have we learned? Annual Reviews of Ecological Systems, 19: 371 –393. Lessios, H. A., Robertson, D. R., and Cubit, J. D. 1984. Spread of Diadema antillarum mass mortality through the Caribbean. Science, 226: 335– 337. Lizarraga-Partida, M. L., Mendez-Gomez, E., Rivas-Montano, A. M., Vargas-Hernandez, E., Portillo-Lopez, A., Gonzalez-Ramirez, A. R., Huq, A., et al. 2011. Association of Vibrio cholera with plankton in coastal areas of Mexico. Environmental Microbiology, 11: 201–208. Lyons, M. M., Ward, J. E., Gaff, H., Hicks, R., Drake, J., and Dobbs, F. C. 2010. Theory of island biogeography on a microscopic scale: are organic aggregates islands for aquatic pathogens? Aquatic Microbial Ecology, 60: 1 – 13. Maddox, E. D. 2006. Temporal patterns in the ingress of settlementstage gray snapper (Lutjanus griseus) and other fishes to Biscayne Bay, Florida, and the Caicos Bank, Turks and Caicos Islands. Dissertations from ProQuest. Paper 2468. McCallum, H. I., Barlow, N., and Hone, J. 2001. How should pathogen transmission be modelled? Trends in Ecology and Evolution, 16: 295– 300. McCallum, H. I., Harvell, C. D., and Dobson, A. P. 2003. Rates of spread of marine pathogens. Ecology Letters, 6: 1062– 1067. McCallum, H. I., Kuris, A. M., Harvell, C. D., Lafferty, K. D., Smith, G. W., and Porter, J. 2004. Does terrestrial epidemiology apply to marine systems? Trends in Ecology and Evolution, 19: 585 –591. McWilliam, P. S., and Phillips, B. F. 2007. Spiny lobster development: mechanisms inducing metamorphosis to the puerulus: a review. Reviews in Fish Biology and Fisheries, 17: 615 – 632. Morris, J. A., Jr, and Whitfield, P. E. 2009. Biology, Ecology, Control and Management of the Invasive Indo-Pacific Lionfish: an Updated Integrated Assessment. NOAA Technical Memorandum NOS NCCOS 99. 57 pp. Moss, J., Behringer, D. C., Shields, J. D., Baeza, J. A., Aguilar-Perera, A., Bush, P. G., and Dromer, C., IV.et al. 2013. Distribution, prevalence, and genetic analysis of Panulirus argus virus I from the Caribbean Sea. Diseases of Aquatic Organisms, 104: 129– 140. A. S. Kough et al. Moss, J., Butler, M. J., IV, Behringer, D. C., and Shields, J. D. 2012. Genetic diversity of the Caribbean spiny lobster virus, Panulirus argus virus 1, and the discovery of PaV1 in lobster post-larvae. Aquatic Biology, 14: 223 – 232. Naro-Maciel, E., Reid, B., Holmes, K. E., Brumbaugh, D. R., Martin, M., and DeSalle, R. 2011. Mitochondrial DNA sequence variation in spiny lobsters: population expansion, panmixia, and divergence. Marine Biology, 158: 2027– 2041. Nasser, A. M., Zaruk, N., Tenenbaum, L., and Netzan, Y. 2003. Comparative survival of Cryptosporidium, coxsackievirus A9 and Escherichia coli in stream, brackish and seawaters. Water Science and Technology, 47: 91 –96. Paris, C. B., Cherubin, L. M., and Cowen, R. K. 2007. Surfing, spinning, or diving from reef to reef: effects on population connectivity. Marine Ecology Progress Series, 347: 285– 300. Paris, C. B., and Cowen, R. K. 2004. Direct evidence of a biophysical retention mechanism for coral reef larvae. Limnology and Oceanography, 49: 1964– 1979. Paris, C. B., Helgers, J., Van Sebille, E., and Srinivasan, A. 2013. The connectivity modeling system: a probabilistic modeling tool for the multi-scale tracking of biotic and abiotic variability in the ocean. Environmental Modelling Software, doi:10.1016/j.envsoft.2012. 12.006 Phillips, B. F., and McWilliam, P. S. 2009. Spiny lobster development: where does successful metamorphosis to the puerulus occur?: a review. Reviews in Fish Biology and Fisheries, 19: 193 – 215. Phillips, B. F., Rimmer, D. W., and Reid, D. D. 1978. Ecological investigations of the late-stage phyllosoma and puerulus larvae of the Western rock lobster Panulirus longipes cygnus. Marine Biology, 45: 347 – 357. Roughgarden, J., Iwasa, Y., and Baxter, C. 1985. Demographic theory for an open marine population with space-limited recruitment. Ecology, 66: 54 – 67. Satheesh Kumar, S., Ananda Bharathi, R., Rajan, J. J. S., Alavandi, S. V., Poornima, M., Balasubramanian, C. P., and Ponniah, A. G. 2013. Viability of white spot syndrome virus (WSSV) in sediment during sun-drying (drainable pond) and under non-drainable pond conditions indicated by infectivity to shrimp. Aquaculture, 402: 119– 126. Shields, J. D., and Behringer, D. C. 2004. A new pathogenic virus in the Caribbean spiny lobster Panulirus argus from the Florida Keys. Diseases of Aquatic Organisms, 59: 109– 118. Silberman, J. D., Sarver, S. K., and Walsh, P. J. 1994. Mitochondrial DNA variation and population structure in the spiny lobster Panulirus argus. Marine Biology, 120: 601 – 608. Staaterman, E., Paris, C. B., and Helgers, J. 2012. Orientation behavior in fish larvae: a missing piece to Hjort’s critical period hypothesis. Journal of Theoretical Biology, 304: 188 – 196. Ward, J. R., and Lafferty, K. D. 2004. The elusive baseline of marine disease: are diseases in ocean ecosystems increasing? PLoS Biology, 2: e120. doi:10.1371/journal.pbio.0020120 Wells, R. J. D., and Rooker, J. R. 2004. Spatial and temporal patterns of habitat use by fishes associated with Sargassum mats in the northwestern Gulf of Mexico. Bulletin of Marine Science, 74: 81 – 99. Wilkin, J. L., and Jeffs, A. G. 2011. Energetics of swimming to shore in the puerulus stage of a spiny lobster: can a postlarval lobster afford the cost of crossing the continental shelf? Limnology and Oceanography: Fluids and Environments, 1: 163– 175. Wolfe, S. H., and Felgenhauer, B. E. 1991. Mouthpart and foregut ontogeny in larval, postlarval, and juvenile spiny lobster, Panulirus argus Latreille (Decapoda, Palinuridae). Zoologica Scripta, 20: 57 – 75. Yeung, C., Jones, D. L., Criales, M. M., Jackson, T. L., and Richards, W. J. 2001. Influence of coastal eddies and counter-currents on the influx of spiny lobster, Panulirus argus, postlarvae into Florida Bay. Marine and Freshwater Research, 52: 1217– 1232. Handling editor: C. Brock Woodson