J. J.J.CETACEAN CETACEAN RES. MANAGE. 11(1):1–8, 2009 CETACEAN RES.MANAGE. MANAGE.11(1):1–8, 11(1):1–8, 2009 CETACEANRES. MANAGE. 11(1):1–­­ 8,2009 2010 111 Modelling Modellingship shipstrike strikerisk riskto tofin, fin,humpback humpbackand andkiller killerwhales whalesin in British BritishColumbia, Columbia,Canada Canada *, +++ #,^ *,*, #,^ #,^ R OB ILLIAMS AND ARA RR OB W ILLIAMS AND PATRICK O’H ARA OBW W ILLIAMS ANDP PATRICK ATRICKO’H O’H ARA Contact Contact e-mail: r.williams@fisheries.ubc.ca Contacte-mail: e-mail:r.williams@fisheries.ubc.ca r.williams@fisheries.ubc.ca ABSTRACT ABSTRACT ABSTRACT Many Many cetacean species are susceptible totomortality mortality ororserious serious injury from vessel collisions. Spatially explicit assessments ofofrisk risk per whale Manycetacean cetaceanspecies speciesare aresusceptible susceptibleto mortalityor seriousinjury injuryfrom fromvessel vesselcollisions. collisions.Spatially Spatiallyexplicit explicitassessments assessmentsof riskper perwhale whale can can help identify potential problem areas totoguide guide appropriate mitigation measures. Canada’s Pacific waters host high cetacean densities canhelp helpidentify identifypotential potentialproblem problemareas areasto guideappropriate appropriatemitigation mitigationmeasures. measures.Canada’s Canada’sPacific Pacificwaters watershost hosthigh highcetacean cetaceandensities densities and and intense maritime traffic, and the issue ofofvessel vessel collisions has taken on high priority ininBritish British Columbia (BC) recently due totoseveral several andintense intensemaritime maritimetraffic, traffic,and andthe theissue issueof vesselcollisions collisionshas hastaken takenon onaa ahigh highpriority priorityin BritishColumbia Columbia(BC) (BC)recently recentlydue dueto several major major industrial development applications. Spatially-explicit statistical modelling and Geographic Information System (GIS) visualisation majorindustrial industrialdevelopment developmentapplications. applications.Spatially-explicit Spatially-explicitstatistical statisticalmodelling modellingand andGeographic GeographicInformation InformationSystem System(GIS) (GIS)visualisation visualisation techniques techniques identified areas ofofoverlap overlap between shipping activity and waters used by humpback, fin and killer whales. Areas ofofhighest highest risk techniquesidentified identifiedareas areasof overlapbetween betweenshipping shippingactivity activityand andwaters watersused usedby byhumpback, humpback,fin finand andkiller killerwhales. whales.Areas Areasof highestrisk risk were were far removed from areas with highest concentrations ofofpeople, people, suggesting that many beach-cast carcasses could go undetected. With werefar farremoved removedfrom fromareas areaswith withhighest highestconcentrations concentrationsof people,suggesting suggestingthat thatmany manybeach-cast beach-castcarcasses carcassescould couldgo goundetected. undetected.With With few few exceptions, high-risk areas were found iningeographic geographic bottlenecks, such asasnarrow narrow straits and passageways. Port expansion and fewexceptions, exceptions,high-risk high-riskareas areaswere werefound foundin geographicbottlenecks, bottlenecks,such suchas narrowstraits straitsand andpassageways. passageways.Port Portexpansion expansionand andaa a proposed proposed pipeline for carrying oil from Alberta totoBC’s BC’s north coast (with associated oil tanker traffic) would increase ship strike risk for all proposedpipeline pipelinefor forcarrying carryingoil oilfrom fromAlberta Albertato BC’snorth northcoast coast(with (withassociated associatedoil oiltanker tankertraffic) traffic)would wouldincrease increaseship shipstrike strikerisk riskfor forall all three three species. The risk assessments illustrate where ship strikes are most likely totooccur, occur, but cannot estimate how many strikes occur. threespecies. species.The Therisk riskassessments assessmentsillustrate illustratewhere whereship shipstrikes strikesare aremost mostlikely likelyto occur,but butcannot cannotestimate estimatehow howmany manystrikes strikesoccur. occur. Propeller killer whales are common in and have been reported in for all three species. Propeller wounds on live killer whales are common ininthe the region, and fatal collisions have been reported ininBC BC for all three species. Propeller woundson onlive livekiller killer whales are common theregion, andfatal fatal collisions havehave been reported for all three species. Propellerwounds wounds on live whales are relatively common inregion, the region, andcollisions fatal collisions been reported inBC BC for all three species. Procedures Procedures were used totoestimate estimate potential mortality limits ininaccordance accordance with wide range ofofquantitative quantitative management objectives from Procedureswere wereused usedto estimatepotential potentialmortality mortalitylimits limitsin accordancewith withaa awide widerange rangeof quantitativemanagement managementobjectives objectivesfrom from jurisdictions jurisdictions around the world. While the extent ofofunder-reporting under-reporting ofofship ship strikes has not been evaluated, the few known cases ofofcollisions collisions jurisdictionsaround aroundthe theworld. world.While Whilethe theextent extentof under-reportingof shipstrikes strikeshas hasnot notbeen beenevaluated, evaluated,the thefew fewknown knowncases casesof collisions involving involving fin whales suggest that mortality due totoship ship strike for this species may already bebeapproaching approaching ororeven even exceeding mortality limits involvingfin finwhales whalessuggest suggestthat thatmortality mortalitydue dueto shipstrike strikefor forthis thisspecies speciesmay mayalready alreadybe approachingor evenexceeding exceedingmortality mortalitylimits limits under objectives. hoped that risk maps may inform environmental impact assessments under the most risk-averse management objectives. isis hoped that risk maps may inform environmental impact assessments ofof marine underthe the most risk-averse management ItItItIt isis hoped that these maps may inform environmental impact assessmentsof ofmarine marine under themost mostrisk-averse risk-aversemanagement managementobjectives. objectives. hoped that riskrisk maps may inform environmental impact assessments marine traffic traffic because will bebeeasier easier totoplan plan new shipping lanes sosothat that they avoid high-density areas for whales than will bebeto totomove move the lanes trafficbecause becauseitititwill willbe easierto plannew newshipping shippinglanes lanesso thatthey theyavoid avoidhigh-density high-densityareas areasfor forwhales whalesthan thanitititwill willbe movethe thelanes lanes after after they become entrenched. afterthey theybecome becomeentrenched. entrenched. KEYWORDS: KEYWORDS: SHIP STRIKE; CONSERVATION; SPATIAL MODELLING; REGULATIONS; ABUNDANCE ESTIMATE; KEYWORDS: SHIP SHIP STRIKE; STRIKE; CONSERVATION; CONSERVATION; SPATIAL SPATIAL MODELLING; MODELLING; REGULATIONS; REGULATIONS; ABUNDANCE ABUNDANCE ESTIMATE; ESTIMATE; MANAGEMENT MANAGEMENT OBJECTIVES; RISK ASSESSMENT; NORTHERN HEMISPHERE; FIN WHALE; HUMPBACK WHALE: KILLER MANAGEMENTOBJECTIVES; OBJECTIVES;RISK RISKASSESSMENT; ASSESSMENT;NORTHERN NORTHERNHEMISPHERE; HEMISPHERE;FIN FINWHALE; WHALE;HUMPBACK HUMPBACKWHALE: WHALE:KILLER KILLER WHALE WHALE WHALE INTRODUCTION INTRODUCTION INTRODUCTION Collisions Collisions with vessels cause serious injury and mortality inin Collisionswith withvessels vesselscause causeserious seriousinjury injuryand andmortality mortalityin many many cetacean species. Quantifying the population-level manycetacean cetaceanspecies. species.Quantifying Quantifyingthe thepopulation-level population-level extent extent ofofship ship strike mortality, however, isisnotoriously notoriously extentof shipstrike strikemortality, mortality,however, however,is notoriously difficult; difficult; collisions are frequently unnoticed, and difficult; collisions collisions are are frequently frequently unnoticed, unnoticed, and and consequently consequently go unreported (Laist etetal., al., 2001; Panigada etet consequentlygo gounreported unreported(Laist (Laistet al.,2001; 2001;Panigada Panigadaet al., al., 2007; Vanderlaan and Taggart, 2007). Ship strikes can al.,2007; 2007;Vanderlaan Vanderlaanand andTaggart, Taggart,2007). 2007).Ship Shipstrikes strikescan can jeopardise jeopardise the viability ofofsmall small populations (Fujiwara and jeopardisethe theviability viabilityof smallpopulations populations(Fujiwara (Fujiwaraand and Caswell, Caswell, 2001), and the importance ofofthe the topic isisreflected reflected Caswell,2001), 2001),and andthe theimportance importanceof thetopic topicis reflected in ininits its appearance ininthe the terms ofofreference reference ofofboth both the itsappearance appearancein theterms termsof referenceof boththe the Scientific Scientific and Conservation Committees ofofthe the International Scientificand andConservation ConservationCommittees Committeesof theInternational International Whaling Whaling Commission (IWC). WhalingCommission Commission(IWC). (IWC). Important Important areas for research include developing methods Importantareas areasfor forresearch researchinclude includedeveloping developingmethods methods for for quantifying ship strike mortality, assessment ofofthe the forquantifying quantifyingship shipstrike strikemortality, mortality,assessment assessmentof the effects effects ofofsuch such mortality atatthe the population level and the effectsof suchmortality mortalityat thepopulation populationlevel leveland andthe the development development ofofappropriate appropriate mitigation measures. AAvaluable valuable developmentof appropriatemitigation mitigationmeasures. measures.A valuable exercise exercise totoinform inform the estimation ofofthe the potential size ofofthe the exerciseto informthe theestimation estimationof thepotential potentialsize sizeof the problem problem and the identification ofof mitigation mitigation measures, problem and and the the identification identification of mitigation measures, measures, involves involves spatially explicit risk assessment. Underlying this involvesspatially spatiallyexplicit explicitrisk riskassessment. assessment.Underlying Underlyingthis this premise premise isisaaacommon-sense common-sense view that minimising spatial premiseis common-senseview viewthat thatminimising minimisingspatial spatial overlap overlap between ships and whales isis the the best way toto overlap between between ships ships and and whales whales is the best best way way to minimise minimise ship strike risk. Although spatial overlap between minimiseship shipstrike strikerisk. risk.Although Althoughspatial spatialoverlap overlapbetween between ships ships and whales isisnot not equivalent totocollision collision risk, spatial shipsand andwhales whalesis notequivalent equivalentto collisionrisk, risk,spatial spatial overlap overlap isisobviously obviously prerequisite for ship strikes. overlapis obviouslyaaaprerequisite prerequisitefor forship shipstrikes. strikes. Canada’s Canada’s Pacific waters host high densities ofofcetaceans cetaceans Canada’sPacific Pacificwaters watershost hosthigh highdensities densitiesof cetaceans (Williams (Williams and Thomas, 2007) asaswell well asasintense intense maritime (Williamsand andThomas, Thomas,2007) 2007)as wellas intensemaritime maritime traffic traffic (O’Hara and Morgan, 2006), but there has been little traffic(O’Hara (O’Haraand andMorgan, Morgan,2006), 2006),but butthere therehas hasbeen beenlittle little effort effort towards estimating cetacean mortality due totoship ship efforttowards towardsestimating estimatingcetacean cetaceanmortality mortalitydue dueto ship strikes. strikes. There isisreason reason totobelieve believe that ininBritish British Columbia strikes.There Thereis reasonto believethat thatin BritishColumbia Columbia (BC) (BC) this issue has taken on greater urgency ininrecent recent years (BC)this thisissue issuehas hastaken takenon ongreater greaterurgency urgencyin recentyears years as asas considerable considerable industrial development isis occurring occurring considerable industrial industrial development development is occurring throughout throughout coastal BC, including inter alia: major port throughoutcoastal coastalBC, BC,including includinginter interalia: alia:major majorport port expansions expansions for Prince Rupert and Delta superport; planned expansionsfor forPrince PrinceRupert Rupertand andDelta Deltasuperport; superport;aaaplanned planned pipeline pipeline terminal for accepting condensate and dispensing pipelineterminal terminalfor foraccepting acceptingcondensate condensateand anddispensing dispensing crude crude oil totoand and from the Alberta tarsands oil fields, with crudeoil oilto andfrom fromthe theAlberta Albertatarsands tarsandsoil oilfields, fields,with with associated associated petroleum tanker traffic; and potential offshore oil associatedpetroleum petroleumtanker tankertraffic; traffic;and andpotential potentialoffshore offshoreoil oil and and gas exploration and production ininHecate Hecate Strait and andgas gasexploration explorationand andproduction productionin HecateStrait Straitand and Queen Queen Charlotte Sound (Fig. 1). All ofofthese these developments QueenCharlotte CharlotteSound Sound(Fig. (Fig.1). 1).All Allof thesedevelopments developments would would result ininaaaconsiderable considerable increase ininshipping shipping traffic and wouldresult resultin considerableincrease increasein shippingtraffic trafficand and consequently consequently an increase ininthe the risk ofofwhales whales being struck. consequentlyan anincrease increasein therisk riskof whalesbeing beingstruck. struck. Within Within Canada, there isisaaagrowing growing recognition ofofthe the need toto WithinCanada, Canada,there thereis growingrecognition recognitionof theneed needto assess assess the extent ofof cetacean cetacean mortality associated with assess the the extent extent of cetacean mortality mortality associated associated with with human human activities and totomitigate mitigate impacts where feasible. humanactivities activitiesand andto mitigateimpacts impactswhere wherefeasible. feasible. Canada Canada has not specified uniform set ofofquantitative quantitative Canadahas hasnot notspecified specifiedaaauniform uniformset setof quantitative management management objectives totoprotect protect marine mammal stocks managementobjectives objectivesto protectmarine marinemammal mammalstocks stocks from from anthropogenic mortality, but methods that take into fromanthropogenic anthropogenicmortality, mortality,but butmethods methodsthat thattake takeinto into account account uncertainty ininpopulation population vital rates and abundance accountuncertainty uncertaintyin populationvital vitalrates ratesand andabundance abundance estimates estimates have been proposed totoestimate estimate potential limits toto estimateshave havebeen beenproposed proposedto estimatepotential potentiallimits limitsto anthropogenic anthropogenic mortality ofofCanadian Canadian marine mammal stocks anthropogenicmortality mortalityof Canadianmarine marinemammal mammalstocks stocks (Johnston (Johnston etetal., al., 2000; Williams etetal., al., 2008). isisunclear unclear (Johnstonet al.,2000; 2000;Williams Williamset al.,2008). 2008).ItItItis unclear what what level ofof ship ship strike mortality would constitute what level level of ship strike strike mortality mortality would would constitute constitute aaa sufficiently sufficiently large fraction ofofaaacetacean cetacean population totowarrant warrant sufficientlylarge largefraction fractionof cetaceanpopulation populationto warrant legislative legislative management action ininCanada. Canada. legislativemanagement managementaction actionin Canada. Risk Risk assessment isis needed needed especially for humpback Risk assessment assessment is needed especially especially for for humpback humpback (Megaptera (Megaptera novaeangliae) and fin whales (Balaenoptera (Megapteranovaeangliae) novaeangliae)and andfin finwhales whales(Balaenoptera (Balaenoptera physalus), physalus), and for the small killer whale (Orcinus orca) physalus),and andfor forthe thesmall smallkiller killerwhale whale(Orcinus (Orcinusorca) orca) populations populations found ininthe the region. Commercial whaling inin populationsfound foundin theregion. region.Commercial Commercialwhaling whalingin British British Columbia (BC) brought baleen whale populations BritishColumbia Columbia(BC) (BC)brought broughtbaleen baleenwhale whalepopulations populations ** Marine *Marine Mammal Research Unit, University ofofBritish British Columbia, 2202 Main Mall, Room 247, AERL, Vancouver, BC, V6T 1Z4 Canada. MarineMammal MammalResearch ResearchUnit, Unit,University Universityof BritishColumbia, Columbia,2202 2202Main MainMall, Mall,Room Room247, 247,AERL, AERL,Vancouver, Vancouver,BC, BC,V6T V6T1Z4 1Z4Canada. Canada. + ++ Sea Sea Mammal Research Unit, Gatty Marine Laboratory, University ofofSt StStAndrews, Andrews, StStAndrews, Andrews, Fife, KY16 8LB, Scotland. SeaMammal MammalResearch ResearchUnit, Unit,Gatty GattyMarine MarineLaboratory, Laboratory,University Universityof Andrews,St Andrews,Fife, Fife,KY16 KY168LB, 8LB,Scotland. Scotland. ## Environment #Environment Canada Wildlife Service, 5421 Robertson Road, Delta, BC, V6T 1Z4, Canada. EnvironmentCanada Canada-- Canadian -Canadian CanadianWildlife WildlifeService, Service,5421 5421Robertson RobertsonRoad, Road,Delta, Delta,BC, BC,V6T V6T1Z4, 1Z4,Canada. Canada. ^^ Department ^Department ofofBiology, Biology, University ofofVictoria, Victoria, PO Box 3020, Station CSC, Victoria, BC, V8W 3N5, Canada. Departmentof Biology,University Universityof Victoria,PO POBox Box3020, 3020,Station StationCSC, CSC,Victoria, Victoria,BC, BC,V8W V8W3N5, 3N5,Canada. Canada. 2 WILLIAMS & O’HARA: MODELLING SHIP STRIKE RISK IN BRITISH COLUMBIA Fig. 1. Marine traffic vessel intensity (number of movements per grid cell) along the coast of BC for June, July and August, 2003. well below below historic historic levels levels of of abundance abundance (Government (Government of of well 2006). In In Canadian Canadian regulatory regulatoryframeworks, frameworks,ships ship Canada, 2006). strikes have have been been identified identified as as important important factors factors in in the the strikes humpback, blue blue (B. (B. musculus), musculus), fin fin and and sei sei (B. (B. borealis) borealis) humpback, whale whale recovery recovery plans plans (Government (Government of of Canada, Canada, 2006), 2006), and and for for both both resident resident and and transient transient ecotypes ecotypes of of killer killer whales whales (Fisheries (Fisheries and and Oceans Oceans Canada, Canada, 2007; 2007; 2008). 2008). Humpback, Humpback, gray gray (Eschrichtius (Eschrichtius robustus) robustus) and and fin fin whales whales have have occasionally occasionally been been reported reported to to be be struck struck by by ships ships transiting transiting the the ‘Inside ‘Inside Passage’ Passage’ (Douglas (Douglas et al., 2008), 2008), which which refers refers to to the the series series of of inland inland and and protected protected waterways waterways used used by by ships ships transiting transiting between between Johnstone Johnstone Strait Strait (northern (northern Vancouver Vancouver Island) Rupert (Fig. (Fig. 1). 1). Island) and and Prince Prince Rupert In In this this paper, paper, results results from from aa spatially-explicit spatially-explicit risk risk assessment that identifies areas of overlap assessment that identifies areas of overlap between between whales whales and BCBC coastal waters are reported. This and shipping shippingactivity activityin in coastal waters are reported. assessment is based on dataonfrom recenta systematic survey This assessment is based dataa from recent systematic of Inside Passage waters of BC that yielded survey of Inside Passage waters of BC that yieldedestimates estimatesof of distribution and abundance for six cetacean distribution and abundance for six cetacean species species (Williams (Williams and and Thomas, Thomas, 2007). 2007). The The assessment assessment also also includes includes shipping activity based on compiled and shipping activity based on compiled and summarised summarised information made available by the Canadian Coast Guard information made available by the Canadian Coast Guard (CCG) that tracks ship movements through their Exclusive (CCG) that tracks ship movements through their Exclusive Economic Zone (O’Hara and Morgan, 2006). A secondary Economic Zone (O’Hara and Morgan, 2006). A secondary goal was to use existing abundance estimates to assess goal wasmortality to use existing estimates assess potential limits forabundance three cetacean species.toFinally, potential mortality limits for three cetacean species. Finally, the frequency of vessel collisions and propeller strikes are the frequency and propeller reported (basedofonvessel scars collisions seen in photographs of strikes living are reported on reported scars seen living animals) that (based have been in in thephotographs region in theof primary animals) that have been reported in the region in the primary and grey literature. This minimum estimate of known vessel and grey literature. minimum estimate known vessel collisions is used to This assess, qualitatively, the of plausibility that collisions used to theexceeds plausibility that ship strikesiscould beassess, causingqualitatively, mortality that potential ship strikeslimits could be causing mortality exceeds potential mortality calculated for BC that waters according to mortality calculated forin BC according to procedureslimits that have been used otherwaters regions. procedures that have been used in other regions. METHODS Whale abundance and density surface fitting Whale abundance and density surface fitting Whale Whale distribution distribution data data Whale Whale data data used used in in the the analyses analyses were were collected collected from from aa systematic line transect survey designed (Thomas systematic line transect survey designed (Thomas et et al., al., 2007) and conducted (Williams and Thomas, 2007) 2007) and conducted (Williams and Thomas, 2007) in in BC BC coastal waters inin thethesummers summers (June-August) of 2004 coastal waters (June-August) of 2004 and and 2005. survey conductedusing using20m 20m boats boats and 2005. TheThe survey waswas conducted and covered covered coastal coastal waters waters (out (out to to approximately approximately 80 80 n.miles) n.miles) between between the the BC-Washington BC-Washington and and BC-Alaska BC-Alaska borders. borders. Methodology Methodology and and conventional conventional distance distance sampling sampling abundance abundance estimates estimates have have been been reported reported previously previously for for several several cetacean cetacean species species from from these these surveys surveys (Williams (Williams and and Thomas, Thomas, 2007; 2007; 2009). 2009). Additional Additional data data were were also also collected collected in in August August 2006, 2006, but but funding funding limitations limitations only only allowed allowed about about half half of of the the planned planned tracklines tracklines to to be be surveyed. surveyed. As As aa result, result, potential potential mortality mortality limits limits were were estimated estimated using using the the analytic analytic abundance and and variance variance estimates estimates previously previously reported reported from from abundance the design-unbiased design-unbiased survey, survey, i.e. i.e. data data collected collected during during the the the 2004 and and 2005 2005 field field seasons seasons (Williams (Williams and and Thomas, Thomas, 2007). 2007). 2004 However, data data collected collected in in 2006 2006 were were included included in in modelmodelHowever, based, density density surface surface fitting fitting models models (see (see below) below) for for based, distribution maps maps to to inform inform the the risk risk analyses, analyses, in in order order to to distribution benefit from from increased increased numbers numbers of of sightings sightings for for fitting the benefit fitting the J. CETACEAN MANAGE. 11(1):1–­­ 11(1):1–8, 2009 CETACEAN RES. MANAGE. 8, 2010 density surface model. For the case of northern resident killer whales, in which every individual is known from annual censuses conducted by Fisheries and Oceans Canada (Ford et al., 2000), known abundance with zero variance was used to estimate mortality limits rather than using abundance estimates from the survey data; although the conventional distance sampling abundance estimates agree well with the known population size (Williams and Thomas, 2009). For humpback and fin whales, abundance refers to the average number of whales in the study area at the time of the surveys, rather than biological population size, because the fraction of the stock(s) using BC waters in summer is unknown. Whale density surface fitting Animal density was modelled using the density surface modelling engine in Distance 6.0 Beta 5 (Thomas et al., 2006) following the four-stage approach outlined by Thomas et al. (2006): (1) fitting a detection function; (2) estimating whale abundance in each segment as a function of spatial covariates; (3) using the descriptive model to predict whale density throughout the study region; and (4) producing variance estimates. Candidate forms for the detection function were the hazard-rate and half-normal models (Buckland et al., 2001). Model selection was guided by Akaike’s Information Criterion (AIC) and goodness of fit statistics. Trackline detection probability was assumed to be certain (i.e. g(0) was assumed to be 1). The log of school size, ln(s), was regressed on the estimated detection probability at the perpendicular distance for each school. The predicted value of ln(s) at zero distance (where detection probability was assumed to be 1) was then backtransformed to provide the required estimate. Effort and sightings data were modelled using the ‘count’ method (Hedley et al., 1999; Williams et al., 2006), which has been packaged into the Density Surface Modelling (DSM) engine in Distance (Thomas et al., 2006). Tracklines were divided into segments approximately 1 n.mile in length. Depth of the midpoint of the segment was estimated by overlaying the tracklines on a bathymetry grid in ArcView 3.2. The saturated DSM model was of the general form: The DSM engine in Distance models abundance of whales in each segment using generalised additive models, using thin-plate regression splines (s) by calling the mgcv package in programme R (Wood, 2006). This saturated model was used unless a term was not significant at p<0.05, or if AIC favoured replacing the bivariate locational spline (slongitude, latitude) with two one-dimensional smooths. A gridded dataset was created, containing a value in every grid cell for each explanatory variable in the model. A square grid size of 2 n.miles on a side (i.e. 4 n.miles2) was chosen for prediction. Values for the explanatory variables (latitude, longitude and depth) were calculated using the value at the midpoint of each grid square. The prediction grid data were passed to the selected model for each species in Distance, which called the predict.gam function in mgcv. The output of the model was an estimate of the predicted number of whale schools in each grid cell, based on each cell’s latitude, longitude, depth and area. Animal abundance was calculated by multiplying the predicted density in each cell by expected school size from the size-bias regression in the detection function modelling step (Buckland et al., 2001) and by the area of each cell, and taking the sum of all values in the grid. The prediction grid was defined by the 3 same shapefile as that used for designing the original survey (Thomas et al., 2007), so the model prediction only interpolated density between tracklines and did not involve extrapolation beyond the survey region. Shipping movement data The Canadian and US Coast Guards monitor ship traffic using radio communication, radar detection and an Automatic Identification System (AIS). The only AIS data used for this study were collected by the US Coast Guard in the transboundary waters of the Strait of Juan de Fuca. The Canadian Coast Guard (CCG) documents ship position approximately every 4 minutes with ship-identification (registered name and Lloyd’s registry number), flag-state (country of registry), ship-type and size. Included in this database are ships over 20m in length, and ships engaged in towing or pushing any vessel or object more than 20m (other than fishing gear) that had a combined length of more than 45m. The database does not include vessels towing or pushing inside a log-booming ground, pleasure yachts <30m, or fishing vessels <24m and 150 tonnes gross, which are not required to report to the CCG. Shipping movement analyses were based on shipping information for the calendar year 2003 as provided by the CCG (Pacific Region). The first complete year of data archived by the CCG (Pacific Region) was 2003, and these data were assumed to be representative of ship movement patterns off the BC coast for all years considered in this study (2004-06). To minimise computer processing time for the analyses, observations were reduced to one uniquely identifiable ship observation per hour per cell in a grid of 5×5km cells using data manipulation procedures in SAS (Cary, North Carolina: SAS v9.3). Ship identification was based on vessel name, call-sign and Lloyd’s registry number. Shipping data were removed when ship movement between cells was not indicated (i.e. ensuring that data were from moving ships only). Finally, for each grid cell, data were summarised by calculating total number of uniquely identifiable ship observations (‘Proc Tabulate’: SAS v9.3), and these totals were used as an index of ship intensity throughout our study area (Fig. 1). This index of ship intensity is a minimum estimate of actual ship movements because a number of ships were not clearly identified in the dataset (i.e. ships tracked by radar were not always identified), and in regions where radar was not available some ships passing through Canadian waters were not tracked because they were not always required to call in (i.e. they were not destined for a Canadian port). Mapping relative ship strike risk A ship strike risk layer was created by multiplying the predicted whale density estimates at each grid point with the nearest value of shipping intensity. The resulting surface layers were explored for potential hotspots of elevated risk of ship strike for all three species of whales studied. These surfaces were created to quantify risk spatially, in relative terms within species, and no attempt was made to compare vulnerability to ship strike across species. Shipping movement patterns, predicted whale density estimates, and relative ship-strike grids were mapped using Inverse Distance Weighting (IDW) ArcGIS v9.3 (ESRI 2002), which is an interpolation technique that estimates focal cell or point values by averaging values for neighbouring cells or points. Average values were calculated using a fixed minimum number of neighbour-values and variable radius. The effect of distance of neighbour cell on the estimated average value of the focal cell is affected by 4 WILLIAMS & O’HARA: MODELLING SHIP 4 WILLIAMS & O’HARA: MODELLING SHIP distance of the neighbouring cell from the focal cell. distance of the neighbouring cell from the focal cell. Categories of gray-shading in the mapping were defined Categories of gray-shading in the mapping were defined using ‘Natural Breaks’ or ‘Jenks’ method in ArcGIS 9.3. using ‘Natural Breaks’ or ‘Jenks’ method in ArcGIS 9.3. Potential mortality limits Potential mortality limits Canada does not use a generic set of quantitative objectives Canada does not use a generic set of quantitative objectives to calculate allowable annual anthropogenic mortality to to calculate allowable annual anthropogenic mortality to marine mammal stocks. Consequently, a range of marine mammal stocks. Consequently, a range of conservation objectives were considered that have been conservation objectives were considered that have been specified in various national and international frameworks specified in various national and international frameworks (Wade et al., 2008; Williams et al., 2008). As an illustrative (Wade et al., 2008; Williams et al., 2008). As an illustrative example, the Potential Biological Removal (PBR) example, the Potential Biological Removal (PBR) calculations under the US Marine Mammal Protection Act calculations under the US Marine Mammal Protection Act were conducted using the default guidelines for assessing were conducted using the default guidelines for assessing marine mammal stocks in US waters (Wade and Angliss, marine mammal stocks in US waters (Wade and Angliss, 1997), and are described as follows: 1997), and are described as follows: Where Rmax is defined as the maximum theoretical or Where R is defined as the maximum theoretical or estimated maxnet productivity rate (default value for estimated net productivity rate (default value for cetaceans=0.04), F as the recovery factor, set to 0.5 for these cetaceans=0.04), F as the recovery factor, set to 0.5 for these stocks as recommended for depleted stocks and Nmin as the stocks as recommended for depleted stocks and Nmin as the 20th percentile of a log-normal distribution surrounding an 20th percentile of of aa log-normal log-normal distribution surrounding an 20th percentile abundance estimate: abundance estimate: where, N is the abundance estimate and CV(N) is the where, N is the abundance estimate and CV(N) is the coefficient of variation of the abundance estimate. coefficient of variation of the abundance estimate. This first step toward estimating potential mortality limits This first step toward estimating potential mortality limits is tentative because information is lacking for fin and is tentative because information is lacking for fin and humpback whales on stock definition and stock boundaries, humpback whales on stock definition and stock boundaries, and because this study lacks information for all three species and because this study lacks information for all three species on the proportion of the stock found in the study area in on the proportion of the stock found in the study area in summer months. Applying mortality limits such as those summer months. Applying mortality limits such as those estimated by PBR to a small area (more specifically to the estimated by PBR to a small area (more specifically to the average number of animals within an area), rather than to a average number of animals within an area), rather than to a biological population is a conservative approach. Lack of biological population is a conservative approach. Lack of information on seasonal patterns in distribution and information on seasonal patterns in distribution and abundance is a weakness that will affect the estimates of risk abundance is a weakness that will affect the estimates of risk (i.e. exposure to ships), but this is a precautionary first step (i.e. exposure to ships), but this is a precautionary first step and is the best that can be done with the existing information. and is the best that can be done with the existing information. A review of US and Canadian status reports and grey A review of US and Canadian status reports and grey literature was conducted to produce minimum estimates of literature was conducted to produce minimum estimates of STRIKE RISK IN BRITISH COLUMBIA STRIKE RISK IN BRITISH COLUMBIA known cases of ship strike and propeller wounds. Note that known cases of ship strike and propeller wounds. Note that current mortality data are presented from scattered records current mortality data are presented from scattered records reported throughout the year, but abundance, mortality limits, reported throughout the year, but abundance, mortality limits, distribution and risk analyses are restricted to a summer, distribution and risk analyses are restricted to a summer, three-month period. Despite this temporal mismatch, there is three-month period. Despite this temporal mismatch, there is no information available on seasonal variability in abundance no information available on seasonal variability in abundance of these species in the region. Consequently, the methods use of these species in the region. Consequently, the methods use all available information and, by including information on all available information and, by including information on known mortality events from outside the summer season, err known mortality events from outside the summer season, err on the side of being precautionary. on the side of being precautionary. RESULTS RESULTS Whale abundance and density surface fitting Whale abundance and density surface fitting Whale distribution, abundance and potential mortality Whale distribution, abundance and potential mortality limits limits Previously reported abundance estimates for fin and killer Previously reported abundance estimates for fin and killer whales suffered due to a lack of sightings (Williams and whales suffered due to a lack of sightings (Williams and Thomas, 2007; 2009). Including the effort and sightings data Thomas, 2007; 2009). Including the effort and sightings data from 2006 improved the fit of the detection function for both from 2006 improved the fit of the detection function for both fin and killer whales. Although model-based abundance fin and killer whales. Although model-based abundance estimates that incorporated the additional data collected in estimates that incorporated the additional data collected in 2006 had little effect on the point estimates of abundance for 2006 had little effect on the point estimates of abundance for any of the three species, analytic abundance and variance any of the three species, analytic abundance and variance estimates were used for estimating potential mortality limits estimates were used for estimating potential mortality limits due to known problems with reliability of variance estimates due to known problems with reliability of variance estimates from model-based abundance estimators (Hedley et al., from model-based abundance estimators (Hedley et al., 1999; Williams et al., 2006). Abundance estimates and 1999; Williams et al., 2006). Abundance estimates and associated CVs used in the analyses are shown in Table 1. A associated CVs used in the analyses are shown in Table 1. A comparison of six assessments of potential limits to annual comparison of six assessments of potential limits to annual anthropogenic mortality of fin, humpback and northern anthropogenic mortality of fin, humpback and northern resident killer whales are presented in Table 1. resident killer whales are presented in Table 1. The highest density regions predicted for fin whales (Fig. The highest density regions predicted for fin whales (Fig. 2) were found in Dixon Entrance and off the southern end of 2) were found in Dixon Entrance and off the southern end of Queen Charlotte Islands. Fin whale density in mainland Queen Charlotte Islands. Fin whale density in mainland inlets was generally low, with one exception on the central inlets was generally low, with one exception on the central coast. The highest-density regions for humpback whales coast. The highest-density regions for humpback whales were qualitatively similar to those of fin whales (Dixon were qualitatively similar to those of fin whales (Dixon Entrance and off the southern end of Queen Charlotte Entrance and off the southern end of Queen Charlotte Islands), but humpback whale density in mainland inlets Islands), but humpback whale density in mainland inlets was much higher than it was for fin whales. For northern was much higher than it was for fin whales. For northern resident killer whales, the highest density region was resident killer whales, the highest density region was Johnstone Strait, however, additional high-density areas Johnstone Strait, however, additional high-density areas were found in central coast waters. were found in central coast waters. Table 1 Table 1 A comparison of mortality limits estimated for three cetacean species using six conservation objectives A comparison of mortality limits estimated for three cetacean species using conservation objectives used in international conservation and management frameworks (aftersixWade et al. 2008). used in international conservation and management frameworks (after Wade et al. 2008). Fin whale Humpback whale Killer whale Fin whale Humpback whale Killer whale Abundance 496 1,313 235 Abundance 496 1,313 235 % CV 45.8 27.5 0 % CV 45.8 27.5 0 th Nmin (20th percentile) 332.3 1,024.9 235 Nmin (20 percentile) 332.3 1,024.9 235 Conservation approach: Formula: Conservation Formula: IWC Scientificapproach: Committee11 2% of Nbest 9.9 26.3 4.7 IWC Scientific‘unacceptable’ Committee 2% ofof Nbest 9.9 26.3 4.7 ASCOBANS 1.7% Nbest 8.4 22.3 4 ASCOBANS ‘unacceptable’ 1.7% of N 8.4 22.3 4 best ASCOBANS ‘precautionary’ 1% of Nbest 5 13.1 2.4 ASCOBANS ‘precautionary’ 1% R of Nbest 5 13.1 2.4 PBR ‘no bias or uncertainty’ 1/2 * N * 1.0 6.6 20.5 4.7 max min PBR ‘robust’ ‘no bias or uncertainty’ 1/2 1/2 R Rmax ** N Nmin ** 0.5 1.0 6.6 20.5 4.7 PBR 3.3 10.2 2.4 max min PBR ‘robust’ 1/2 R 3.3 10.2 2.4 max * Nmin * 0.5 New Zealand MALFIRM 1/2 Rmax * Nmin * 0.15 1 3.1 0.7 New Zealand MALFIRM 1/2 Rmax * Nmin * 0.15 1 3.1 0.7 1 1The IWC Scientific Committee cautions that bycatch levels >2% of the best abundance estimate are The IWC Scientific of theused bestbyabundance estimate are unacceptable, and thatCommittee takes of 1%cautions of Nbest that (i.e. bycatch the samelevels criteria>2% as those ASCOBANS) warrant unacceptable, that1996, takesp.89). of 1% of Nbest (i.e. the same criteria as those used by ASCOBANS) warrant close attentionand (IWC, close attention (IWC, 1996, p.89). J. CETACEAN MANAGE. 11(1):1–­­ 11(1):1–8, 2009 CETACEAN RES. MANAGE. 8, 2010 Fig. 2. Density surfaces for fin, humpback and northern resident killer whales (left-hand column), and intensity surfaces for whale-ship interactions (right-hand column) for fin, humpback and killer whales on the right (whale density x marine traffic vessel intensity: see Methods). Whale densities (numbers per grid cell) range from 0-5 for fin whales, 0-4 for humpback whales, and 0-1 for killer whales. Whale-ship interactions scale from 0-224 (fin whales), 0-841 (humpback whales), and 01,279 (killer (killer whales). 1279 whales). 5 666 WILLIAMS WILLIAMS & & O’HARA: O’HARA: MODELLING MODELLING SHIP SHIP STRIKE STRIKE RISK RISK IN IN BRITISH BRITISH COLUMBIA COLUMBIA WILLIAMS & O’HARA: MODELLING SHIP STRIKE RISK IN BRITISH COLUMBIA Table Table 22 Table 2 Known Known ship ship strikes strikes affecting affecting BC BC cetacean cetacean species, species, with with summaries summaries of of events events compiled compiled from from killer killer whale whale recovery recovery strategies strategies Known ship strikes affecting BC cetacean species, with summaries of events compiled from killer whale recovery strategies (Canada (Canada and and the the US), US), humpback humpback and and fin fin whale whale status status reports reports (COSEWIC) (COSEWIC) (Douglas (Douglas et et al., al., 2008; 2008; Jensen Jensen and and Silber, Silber, 2004), 2004), the the (Canada and the US), humpback and fin whale status reports (COSEWIC) (Douglas et al., 2008; Jensen and Silber, 2004), the newsletter newsletter of of Vancouver Vancouver Aquarium’s Aquarium’s BC BC Cetacean Cetacean Sightings Sightings Network Network (http://www.vanaqua.org), (http://www.vanaqua.org), and and personal personal observations observations newsletter of Vancouver Aquarium’s BC Cetacean Sightings Network (http://www.vanaqua.org), and personal observations from from Jackie Jackie Hildering Hildering (whalewatch (whalewatch naturalist naturalist working working in in Johnstone Johnstone and and Queen Queen Charlotte Charlotte Straits). Straits). from Jackie Hildering (whalewatch naturalist working in Johnstone and Queen Charlotte Straits). Year Species ID Location Fate Year Species ID Location Fate of of animal animal Year Species ID Location Fate of animal 1999 UNK British Fatal 1999 Fin Fin whale whale UNK British Columbia Columbia (BC) (BC) Fatal 1999 Fin whale UNK British Columbia (BC) Fatal 2002 UNK Puget Fatal 2002 Fin Fin whale whale UNK Puget Sound, Sound, Washington Washington (WA) (WA) Fatal 2002 Fin whale UNK Puget Sound, Washington (WA) Fatal 2002 UNK Puget Fatal 2002 Fin Fin whale whale UNK Puget Sound, Sound, WA WA Fatal 2002 Fin whale UNK Puget Sound, WA Fatal 2002 UNK Puget Fatal 2002 Fin Fin whale whale UNK Puget Sound, Sound, WA WA Fatal 2002 Fin whale UNK Puget Sound, WA Fatal 2002 UNK Puget Fatal 2002 Fin Fin whale whale UNK Puget Sound, Sound, WA WA Fatal 2002 Fin whale UNK Puget Sound, WA Fatal 2004 UNK West 2004 Fin Fin whale whale UNK West coast coast Vancouver Vancouver Island Island (VI) (VI) BC BC Fatal Fatal 2004 Fin whale UNK West coast Vancouver Island (VI) BC Fatal 2006 UNK Northwest Fatal 2006 Fin Fin whale whale UNK Northwest inland inland waters waters WA WA Fatal 2006 Fin whale UNK Northwest inland waters WA Fatal 2006 UNK Puget Fatal 2006 Fin Fin whale whale UNK Puget Sound, Sound, WA WA Fatal 2006 Fin whale UNK Puget Sound, WA Fatal 2004 UNK West Fatal 2004 Humpback Humpback whale whale UNK West coast coast WA WA Fatal 2004 Humpback whale UNK West coast WA Fatal 2006 UNK Knight Uncertain 2006 Humpback Humpback whale whale UNK Knight Inlet, Inlet, BC BC Uncertain 2006 Humpback whale UNK Knight Inlet, BC Uncertain 2006 UNK Swiftsure 2006 Humpback Humpback whale whale UNK Swiftsure Bank Bank (west (west coast coast VI, VI, BC) BC) Uncertain Uncertain 2006 Humpback whale UNK Swiftsure Bank (west coast VI, BC) Uncertain 2006 BCX0022 Seen 2006 Humpback Humpback whale whale BCX0022 calf calf Johnstone Johnstone Strait, Strait, BC BC Seen injured injured and and disappeared disappeared 2006 Humpback whale BCX0022 calf Johnstone Strait, BC Seen injured and disappeared 2006 BCY0177 Johnstone Serious 2006 Humpback Humpback whale whale BCY0177 Johnstone Strait, Strait, BC BC Serious injury injury 2006 Humpback whale BCY0177 Johnstone Strait, BC Serious injury 1995 UNK British Non-fatal 1995 Killer Killer whale whale (NR) (NR) UNK British Columbia Columbia Non-fatal injury injury 1995 Killer whale (NR) UNK British Columbia Non-fatal injury 2005 A60 Johnstone Non-fatal 2005 Killer Killer whale whale (NR) (NR) A60 Johnstone Strait, Strait, BC BC Non-fatal strike strike 2005 Killer whale (NR) A60 Johnstone Strait, BC Non-fatal strike 2006 A82 Campbell Injured 2006 Killer Killer whale whale (NR) (NR) A82 Campbell River, River, BC BC Injured and and died died following following year year 2006 Killer whale (NR) A82 Campbell River, BC Injured and died following year 2006 C21 Prince Fatal 2006 Killer Killer whale whale (NR) (NR) C21 Prince Rupert, Rupert, BC BC Fatal 2006 Killer whale (NR) C21 Prince Rupert, BC Fatal 2006 A59 Campbell Non-fatal 2006 Killer Killer whale whale (NR) (NR) A59 Campbell River, River, BC BC Non-fatal strike strike (calf (calf A82 A82 injured) injured) 2006 Killer whale (NR) A59 Campbell River, BC Non-fatal strike (calf A82 injured) 2006 G39 Johnstone Serious 2006 Killer Killer whale whale (NR) (NR) G39 Johnstone Strait, Strait, BC BC Serious injury injury 2006 Killer whale (NR) G39 Johnstone Strait, BC Serious injury 2007 Johnstone Serious 2007 Killer Killer whale whale (offshore) (offshore) UNK UNK Johnstone Strait, Strait, BC BC Serious injury injury (dorsal (dorsal cut cut off) off) 2007 Killer whale (offshore) UNK Johnstone Strait, BC Serious injury (dorsal cut off) 1998 UNK Haro Non-fatal 1998 Killer Killer whale whale (SR) (SR) UNK Haro Strait, Strait, BC BC Non-fatal strike strike 1998 Killer whale (SR) UNK Haro Strait, BC Non-fatal strike 2006 L98 Nootka Fatal 2006 Killer Killer whale whale (SR) (SR) L98 Nootka Sound Sound (west (west coast coast VI, VI, BC) BC) Fatal 2006 Killer whale (SR) L98 Nootka Sound (west coast VI, BC) Fatal 2005 UNK Haro Non-fatal 2005 Killer Killer whale whale (SR) (SR) UNK Haro Strait, Strait, BC BC Non-fatal strike strike 2005 Killer whale (SR) UNK Haro Strait, BC Non-fatal strike Ship strike risk Shipstrike strikerisk risk Ship Ship risk Areas of relatively high risk of ship strikes for fin whales Areasstrike of relatively relatively high risk risk of of ship ship strikes strikes for for fin fin whales whales Areas of high Areas of relatively high risk of ship strikesQueen for fin whales were found in Dixon Entrance (off northern Charlotte were found in Dixon Entrance (off northern Queen Charlotte were found in Dixon Entrance (off northern Queen Charlotte were found in two Dixon Entrance (off northern Queen Charlotte Islands), and areas coincidental with elevated shipping Islands), and two areas coincidental with elevated elevated shipping Islands), and two areas coincidental with shipping Islands), and two areas coincidental with elevated shipping movement patterns in Hecate Strait, and at the entrance to movement patterns in Hecate Strait, and at the entrance to movement patterns in Hecate Strait, and at the entrance to movement patternson Hecate Strait, at 2). the entrance one inlet system the central coast (Fig. Areas of one inlet inlet system system oninthe the central coastand (Fig. 2). Areas Areas of one on central coast (Fig. 2). of to one inlet system thestrikes central (Fig. 2). Areas of relatively high risk of ship for humpback whales were relatively high riskof ofon ship strikes forcoast humpback whales were relatively high risk ship strikes for humpback whales were relatively high risk of ship strikes for humpback whales were roughly similar to those for fin whales, but also occurred in roughly similar to those for fin whales, but also occurred in roughly similar to those for fin whales, but also occurred in roughly similar to those for fin whales, but also occurred Queen Charlotte Strait, Hecate Strait, and several inlet Queen Charlotte Strait, Hecate Strait, and several inlet Queen Charlotte Strait, Hecate Strait, and several inlet in Queen Charlotte Strait, Hecate Strait, and several inlet systems along the central coast. For killer whales, the region systems along thecentral central coast. Forkiller killerwhales, whales, theregion region systems along the coast. For the systems along the central coast. For killer whales, the region of highest ship strike risk was constrained to Johnstone of highest ship strike risk was constrained to Johnstone of highest ship strike risk was constrained to Johnstone of highest ship risk was constrained Strait, where risk was estimated to be about an order of Strait, where riskstrike was estimated estimated to be be about abouttoan anJohnstone order of of Strait, where risk was to order Strait, where risk was estimated to be about an order of magnitude higher than anywhere else along the coast. magnitude higher than anywhere else along the coast. magnitude higher than anywhere else along the coast. magnitude higher than anywhere else along the coast. Minimum estimates of mortality and serious injury due to Minimumestimates estimatesof ofmortality mortalityand andserious seriousinjury injurydue dueto to Minimum ship strikes Minimum estimates of mortality and serious injury due to ship strikes ship strikes Evidence of injuries and mortalities due to vessel collisions ship strikes Evidence of injuries and mortalities due to vessel collisions Evidence of injuries and mortalities due to vessel collisions is presented in Table 2. The number of cases reported for is presented presented in Table Tableand 2. The The number number ofto cases reported for Evidence of injuries mortalities due vessel collisions is in 2. of cases reported for each species probably does not represent relative frequency each species probably does not represent relative frequency is presented in Table 2. The number of cases reported for each species probably does not represent relative frequency of collisions, because killer whales are better studied in the ofcollisions, collisions, becausekiller killerwhales whales arebetter better studied inthe the each species probably does not represent relative frequency of because are studied in region than the other two species. Similarly, much of the region than the other two species. Similarly, much of the of collisions, because killer whales are better studied in region than the other two species. Similarly, much of the available information on collisions comes from Washington available information on collisions comes frommuch Washington region than the otheron two species.comes Similarly, of the available information collisions from Washington State, while the abundance estimates for assessing mortality State, while the abundance estimates for assessing mortality available information on collisions comes from Washington State, while the abundance estimates for assessing mortality limits apply only to BC waters. limitsapply applyonly only toBC BCwaters. waters. State, while the abundance estimates for assessing mortality limits to limits apply only to BC waters. DISCUSSION DISCUSSION DISCUSSION DISCUSSION This study presents an objective and quantitative framework This study presents an and quantitative Thisstudy studypresents presentsan an objective objectiveand andquantitative quantitative framework framework This for identifying areas ofobjective elevated risk of ship strikeframework for whales for identifying areas of elevated risk of ship strike for whales for identifying areas of elevated risk of ship strike forwhales whales for identifying areasdata of elevated riskdistribution of ship strike based on existing on whale andforshipping based on existing data on whale distribution and shipping based on existing data on whale distribution and shipping based existing on emerges whale distribution and shipping traffic on intensity. A data pattern that is consistent among traffic intensity. A emerges that is consistent among traffic intensity. intensity.A A pattern pattern emerges emerges that that is is consistent consistent among among traffic the three species pattern of whales (humpback, fin and killer), the three species of whales (humpback, fin and killer), the three species of whales (humpback, fin and killer), the threeareas species ofthewhales (humpback, fin and whereby with highest relative risk (i.e. risk killer), of ship whereby areas with the highest relative risk (i.e. risk of ship whereby areas with the highest relative risk (i.e. risk of ship whereby areas species) with the are highest relative risk (i.e. risk regions of ship strike within found in ‘bottlenecks’; strike within species) are found in ‘bottlenecks’; regions strike within species) are found in ‘bottlenecks’; regions strike within and species) are found in ‘bottlenecks’; regions where whale boat densities are both concentrated (Fig. 2). where whale and boat densities are both concentrated (Fig. 2). wherewhale whaleand andboat boatdensities densitiesare areboth bothconcentrated concentrated(Fig. (Fig.2). 2). where Ship strike risk to killer whales is highest in Johnstone Strait, Ship strike risk to killer whales is highest in Johnstone Strait, Ship strike risk to killer whales is highest in Johnstone Strait, Ship strike risk to killer whales is highest in Johnstone Strait, and for humpback whales, the Queen Charlotte and Johnstone and for humpback whales, the Queen Charlotte and Johnstone andfor forhumpback humpbackwhales, whales,the theQueen QueenCharlotte Charlotteand andJohnstone Johnstone and Straits (northeast of Vancouver Island) and the narrow Straits (northeast of Vancouver Island) and the narrow Straits (northeast (northeast of of Vancouver Vancouver Island) Island) and and the the narrow narrow Straits passages of the central coast are relatively high-risk areas for passages of the central coast are relatively high-risk areas for passages of the central coast are relatively high-risk areas for passages of the central coast relatively for both species. Although thearewaters offhigh-risk southernareas Queen both species. Although the waters off southern Queen both species. species. Although Although the the waters waters off off southern southern Queen Queen both Charlotte Islands host the highest densities fin whales, risk Charlotte of CharlotteIslands Islandshost hostthe thehighest highestdensities densitiesof offin finwhales, whales,risk risk Charlotte of ship strike isis relatively low because of the low levels of of ship of ship strike relatively low because of the low levels of of ship strike is relatively low because of the low levels of shipping traffic there; the highest relative risk areas are found shipping shipping traffic there; the highest relative risk areas are found shipping traffic there; the highest relative risk areas are found in Dixon Entrance where ship traffic more concentrated. concentrated. in Dixon inDixon DixonEntrance Entrancewhere whereship shiptraffic trafficisis more in concentrated. risk assessments can predict where ship strikes While the While the risk assessments can predict where shipstrikes strikes While the risk assessments can predict where ship are likely to tooccur, occur,they theyare cannot predict how many are most likely cannot predict how are most likely to occur, they are cannot predict how many are most likely to occur, they are cannot predict how many strikes are actually occurring. One technical development strikes are actually occurring. strikes are actually occurring. One technical development strikes are actually occurring. One technical development that will assist these ongoing efforts isis aa more consistent use that will assist these ongoing thatwill willassist assistthese theseongoing ongoingefforts effortsis moreconsistent consistentuse use that a more of the AIS system coastwide. While AIS coverage in BC is of the AIS system of the AIS system coastwide. While AIS coverage in BC of the AIS system coastwide. While AIS coverage in BC isis currently sparse, the system is expected to come into currently currently sparse, the system is expected to come into currently sparse, the system is expected to come into widespread use in the near future. At that point, the risk widespread widespread use use in in the the near near future. future. At At that that point, point, the the risk risk widespread metric could be recalculated in absolute, rather than metric could be recalculated in absolute, rather than relative metric could be recalculated in absolute, rather than relative metric could be recalculated in absolute, rather than relative relative units. Efforts will stall, still however, thepoint point of of units. Efforts will still stall, however, at the point units. Efforts Efforts will still still stall,stall, however, at atthe the point of units. will however, at evaluating whether current mortality rates can be deemed evaluating whether current mortality evaluating whether current mortality rates can be deemed evaluating whether current mortality rates can be deemed acceptable. A considerable hurdle for setting mortality limits acceptable. A setting mortality acceptable.A Aconsiderable considerablehurdle hurdlefor forsetting settingmortality mortalitylimits limits acceptable. is the inability to state Canada’s current management is the inability Canada’s current is the inability to state Canada’s current management is the inability to state Canada’s current management objectives in quantitative terms and whether quantitative objectives in quantitative terms objectives in in quantitative quantitative terms terms and and whether whether quantitative quantitative objectives objectives will be N N (i.e. the degree objectives based on N or N (i.e. the degree of of objectives will be based on N or N (i.e. the degree best min best min best or N min objectives will be based on Nbest min (i.e. the degree of uncertainty uncertainty that will be tolerated). In BC, this is especially uncertainty that will be tolerated). In BC, this is especially uncertainty that will be tolerated). In BC, this is especially problematic fin whales because of the large uncertainty problematic for for fin fin whales whales because because of of the the large large uncertainty uncertainty problematic for fin of the large uncertainty problematic whales because associated existing abundance associatedwith withexisting existingabundance abundanceestimates estimates(Williams (Williamsand and associated with existing abundance estimates (Williams and associated Thomas, 2007). For the two baleen whale species, limits for Thomas, For the two baleen whale species, limits for Thomas, 2007). For the two baleen whale species, limits Thomas, 2007). For the two baleen whale species, limits for an area, rather than a population, have been calculated an area, rather than a population, have calculated for an area, rather than a population, have been calculated an area, rather than a population, have been calculated because currently because it currently unknown unknown what what fraction fraction of of the the because itit is isis currently currently unknown what fraction of the because populations likely to be in the study populationswas waslikely likelyto tobe bein inthe thestudy studyarea areaat atthe thetime timeof of populations was likely to be in the study area at the time of populations the survey, which will way, the survey, which will differ among species. In a related way, the survey, which will differ among species. In a related the survey, which will differ among species. In a related way, ship strike mortality shipstrike strikemortality mortalitymay mayapply applyto tokiller killerwhales whalesyear-round year-roundin in ship strike mortality may apply to killer whales year-round in ship this region but only for a limited period for the other this region but only for a limited period for the other species. this region but only for a limited period for the other species. this region but only for a limited period for the other species. Until information stock boundaries seasonal patterns Untilinformation informationon onstock stockboundaries boundariesand andseasonal seasonalpatterns patterns Until information on stock boundaries and seasonal patterns Until in abundance available, the range of mortality in abundance becomes available, the range of mortality in abundance becomes available, the range of mortality in abundance becomes available, the range of mortality limits presented limits presented presented are are necessarily necessarily simplistic, simplistic, but but aaa useful useful limits presented are necessarily simplistic, but useful limits starting starting point for discussion. Based on objectives from the starting point for discussion. Based on objectives from the starting point for discussion. Based on objectives from the different different management management approaches approaches reviewed reviewed (Table (Table 1), 1), different management approaches reviewed (Table 1), different potential would potential limits to anthropogenic mortality would vary by an potential limits to anthropogenic mortality vary by an potential limits to anthropogenic mortality would vary by an order order of magnitude for both fin and humpback whales (Table order of magnitude for both fin and humpback whales (Table order of magnitude for both fin and humpback whales (Table 1). Regardless of the management 1). Regardless Regardless of of the the management management approach approach and and objectives objectives 1). Regardless of the management approach and objectives 1). that Canada eventually specifies, mortality that Canada Canada eventually eventually specifies, specifies, mortality mortality limits limits will will be be that Canada eventually specifies, mortality limits will be that relatively low for these species, both because populations relatively low for these species, both because populations are relatively low for these species, both because populations are relatively low for these species, both because populations are small and uncertainty in abundance estimates is large (Table smalland anduncertainty uncertaintyinin in abundance estimates islarge large (Table small abundance estimates is large (Table 1). small and uncertainty abundance estimates is (Table J. CETACEAN RES. MANAGE. 11(1):1–8, 2009 J. MANAGE. 11(1):1–8, 2009 J. CETACEAN CETACEAN RES. MANAGE. MANAGE. 11(1):1–­­ 11(1):1–8, 2009 CETACEAN RES. 8, 2010 1). It remains to be seen whether ship strikes are causing 1). It be seen whether ship strikes are It remains to to be are causing 1). It remains remains to beseen seenwhether whether ship strikesprecautionary are causing mortality rates that exceed all butship the strikes most mortality rates that exceed all but the most precautionary mortality rates that exceed all but the most precautionary limits to anthropogenic mortality, but a cursory review of the limits mortality, but aa cursory review of limits to to anthropogenic anthropogenic mortality, butthat cursory revieware of the the primary and grey literature reveals ship strikes far primary and grey literature reveals that ship strikes are primary and grey literature reveals that ship strikes are far far more common in the region than expected. more more common common in in the the region region than than expected. expected. Estimating ship-strike mortality Estimating Estimating ship-strike ship-strike mortality mortality Fin whales Fin Fin whales whales Many British Columbians first became aware of the threat that Many British Columbians first aware of threat that Manystrikes British Columbians firstinbecame became aware of the the threat ship that ship pose to fin whales June 1999, when a cruise ship strikes pose to fin whales in June 1999, when a cruise ship ship strikes pose to fin whales in June 1999, when a cruise ship arrived in the port of Vancouver with a fin whale carcass draped arrived in the of with aa fin carcass arrived the port portbow. of Vancouver Vancouver fin whale whale carcass draped draped over itsinbulbous Althoughwith mortality rate estimates based over its bulbous bow. Although mortality rate estimates overanecdotal its bulbous bow. Although mortality estimates based based on information received throughrate self-reporting and on information received self-reporting and on anecdotal anecdotal received through through and compiled in aninformation informal monitoring scheme self-reporting (Table 2; average compiled in an informal monitoring scheme (Table 2; average compiled in an informal monitoring scheme (Table 2; average of one animal per year in BC-Washington waters) are no doubt of animal per in BC-Washington waters) no doubt of one one lower animalthan per year year BC-Washington waters) are are nowould doubt much total inmortality rates, estimated rates much lower than total mortality rates, estimated rates would muchbelower totaltomortality rates, estimated ratesinwould still high than enough trigger management action other still enough to trigger action still be be high high (e.g. enough trigger management management action in in other other jurisdictions NewtoZealand; Table 1). Nevertheless, high jurisdictions (e.g. New Zealand; Table 1). Nevertheless, high jurisdictions (e.g. New Zealand; Table 1). Nevertheless, high priority must be placed on identifying the degree to which priority on the to which priority must must be beofplaced placed on identifying identifying the degree degree which under-reporting ship strikes is occurring for thistospecies. under-reporting of strikes occurring for this under-reporting of ship ship strikesforis isfin occurring for accompanied this species. species. Existing abundance estimates whales are Existing abundance estimates for fin whales are accompanied Existing abundance estimates for fin whales are accompanied by such large CVs (Williams and Thomas, 2007) that only the by large and Thomas, 2007) only by such such large CVs CVs (Williams (Williams Thomas,problems 2007) that that only the the most catastrophic populationand declines would be most catastrophic population declines problems would be most catastrophic declinesestimates, problemscoupled wouldwith be detected. This lack ofpopulation robust abundance detected. This lack abundance estimates, coupled with detected. Thispropensity lack of of robust robust coupled with an apparent for abundance fin whales estimates, to be struck by ships an propensity for fin to be by an apparent apparent fin whales whales be struck struck by ships ships (Douglas et propensity al., 2008;for Laist et al., to 2001), suggests that (Douglas et al., 2008; Laist et al., 2001), suggests that (Douglas et al., 2008; Laist et al., 2001), suggests thata understanding ship-strike impacts on fin whales should be understanding ship-strike impacts on fin whales understanding ship-strike impacts et onal., fin2008). whales should should be be aa priority for future work (Panigada priority priority for for future future work work (Panigada (Panigada et et al., al., 2008). 2008). Humpback whales Humpback whales Humpback with whales Collisions humpback whales are reported frequently Collisions with humpback whales are reported frequently Collisions with concern. humpback whales enough to raise Three of theare fivereported reported frequently collisions enough to raise concern. Three of the five reported collisions enough2)tooccurred raise concern. Three of the fivewhich reported (Table in the ‘high-risk area’, maycollisions reflect a (Table 2) in area’, which may reflect (Table 2) occurred occurred in the the ‘high-risk ‘high-risk which may reflect aa true tendency for ship strikes to occurarea’, in areas where humpback true tendency for ship strikes to occur in areas where humpback true tendency for ship occur inmay areas where humpback whales aggregate and strikes where to shipping intensify in narrow whales where shipping intensify whales aggregate aggregate and and(i.e. where shipping may may intensify in initnarrow narrow coastal passageways ‘bottlenecks’). Alternatively, could coastal passageways (i.e. ‘bottlenecks’). Alternatively, itit could coastal passageways could reflect simply the (i.e. high‘bottlenecks’). probability Alternatively, that whalewatchers, reflect the probability that whalewatchers, reflect simply simply the high high will probability that whalewatchers, researchers and naturalists detect and report such events researchers and naturalists will detect report such events researchers andtoo naturalists detecttoand and report such whales events because they would bewill drawn places where because they too would be drawn to places where whales because they would drawnsparse to places whales aggregate. Onetoo pattern seenbe in these data is where a tendency for aggregate. pattern seen these sparse data is for aggregate. One One pattern seen in in is aa tendency tendency for humpback whale collisions to these resultsparse in an data uncertain fate of the humpback whale to in an fate humpback whale collisions collisions to result resultthat in additional an uncertain uncertain fate of of the the animal. A priority is thus to ensure resources are animal. priority is thus that resources are animal. A A to priority thus to to ensure ensure that additional additional resources are allocated allowislong-term monitoring of struck individual allocated to allow long-term monitoring of struck individual allocatedtotoassess allowpost-strike long-termsurvivorship. monitoring of struck individual animals However, it is clear animals to survivorship. However, it is clear animals to assess assess post-strike post-strike clear that under-reporting would survivorship. have to be However, severe forit isannual that under-reporting would have to be severe for annual that under-reporting would have to be severe for annual mortality to be approaching anything but the most precautionary mortality approaching anything but most precautionary mortality to to be beobjectives approaching anything butItthe the precautionary conservation for this species. is most possible that 10-20 conservation objectives for this species. It is that 10-20 conservation objectives for this species. It is possible possible (Table 1) humpback whales could be killed each yearthat by10-20 ships (Table 1) humpback whales could each year (Table humpback whales could be be killed killed year by by ships ships in BC 1)and this level of mortality could each go unnoticed or in this level of could go or in BC BC and andbut this level data of mortality mortality could go unnoticed unnoticed or unreported, existing do not allow the plausibility of this unreported, but existing data do not allow the plausibility of this unreported, but existing data do not allow the plausibility of this scenario to be evaluated. scenario scenario to to be be evaluated. evaluated. Killer whales Killer whales Killernumber whalesof collisions reported between resident killer The The number of reported between resident killer The number of collisions collisions reported between residentpaid killer whales and vessels was surprising given the attention in whales and vessels was surprising given the attention paid in whales and vessels was guidelines. surprising given the attention in BC to whalewatching However, residentpaid killer BC guidelines. However, resident killer BC to to whalewatching whalewatching resident killer whales are censused in guidelines. most years However, by Fisheries and Oceans whales in years Fisheries and whales are are censused censused in most most(Ford yearsetby byal., Fisheries and Oceans Oceans (DFO) Canada researchers 2000) and heavily (DFO) Canada researchers (Ford et al., 2000) and heavily (DFO) Canada researchers (Ford et al., 2000) and scrutinised by commercial whalewatchers, making heavily it less scrutinised by whalewatchers, making itit less scrutinised by commercial commercial whalewatchers, making less likely that vessel strikes go unreported for killer whales than likely that vessel go for whales than likely vessel strikes strikes go unreported unreported for killer killer whalessocial than for fin that or humpback whales. The small size and highly for humpback whales. The small and social for fin fin or or small size sizemeans and highly highly nature ofhumpback BC killerwhales. whale The populations that social these nature of BC killer whale populations means that these nature of BC populations populations are killer unablewhale to absorb anythingmeans beyondthat verythese low populations are absorb beyond very populations are unable unable to tomortality absorb anything anything beyond very low low levels of anthropogenic (Table 1) (Williams and levels of mortality (Table and levels of anthropogenic anthropogenic mortality (Table 1) 1; (Williams Williams and mortality (Table 1) (Williams Lusseau, 2006). Any limit to anthropogenic mortality Lusseau, 2006). Any limit to anthropogenic mortality Lusseau, 2006). Any limit to anthropogenic mortality established for these small populations would be low, established established for for these these small small populations populations would would be be low, low, 7 77 regardless of the conservation approach (Table 1) and the regardless approach (Table 1) and the regardless of of the the conservation conservation (Table the minimum mortality or serious approach injury rates due1)toand vessel minimum mortality or serious injury rates due to minimum mortality or serious injury rates due to vessel vessel collisions based on anecdotal information and self-reporting collisions on information and self-reporting collisions based based on anecdotal anecdotal information self-reporting approach or exceed these limits already.and Fortunately, BC approach or exceed these limits already. Fortunately, BC approach or populations exceed these already. Fortunately, BC killer whale are limits very well studied, and variation killer whale are very well and variation killer whale populations populations areship verystrikes well studied, studied, anddetectable, variation in mortality resulting from would be in resulting from ship would be in mortality mortality ship strikes strikes would be detectable, detectable, provided thatresulting DFO’s from Cetacean Research Program and the provided that DFO’s Cetacean Research Program and the provided that DFO’s Cetacean Research Program the Center for Whale Research (Washington State)andhave Center Whale (Washington State) have Center for for Whale toResearch Research (Washington have adequate resources continue their long-termState) monitoring adequate resources to continue their monitoring adequate resources to whales. continueHowever, their long-term long-term monitoringa study of resident killer clearly attributing study of resident killer whales. However, clearly attributing a study of resident killer whales. However, clearly attributing proportion of mortality to ship-strikes, or any anthropogenica proportion of mortality to ship-strikes, or any anthropogenic proportion of mortality to ship-strikes, or anyefforts. anthropogenic cause, remains an obstacle for conservation For this cause, an for efforts. For this cause, remains remains an obstacle obstacle for conservation conservation Forkiller this reason, increasing the recovery and necropsyefforts. rates of reason, increasing the recovery and necropsy rates of killer reason,carcasses increasingis the recovery necropsy rates of killer whale a priority forand future research supporting whale carcasses priority for future supporting whale carcasses is is aathis priority future research research supporting the conservation of speciesfor (Raverty and Gaydos, 2004). the conservation of this species (Raverty and Gaydos, the conservation of this species (Raverty and Gaydos, 2004). 2004). Utility of the approach Utility of Utility of the the approach approach The approach described here represents an early attempt to The approach described here represents an to The approach here density represents an early early attempt attempt to overlay whale described and shipping to calculate the spatial overlay whale and shipping density to calculate the spatial overlay whale and shipping density to calculate the spatial distribution of relative risk, which has been identified by the IWC distribution of risk, which has identified by distribution of relative relativeas risk, has been been identified by the the IWC IWC Scientific Committee anwhich important step in understanding ship Scientific Committee as an important step in understanding ship Scientific as an important inGAM-based understanding ship strikes. TheCommittee approach adopted, namelystep to use spatial strikes. approach namely GAM-based spatial strikes. The The approach adopted, adopted, namely to to use use GAM-based models to estimate whale distribution (Hedley et al., spatial 1999; models whale distribution (Hedley et 1999; models to toetestimate estimate whale (Hedley et al., al.,data 1999; Williams al., 2006) and distribution overlay spatially explicit on Williams et al., 2006) and overlay spatially explicit data Williams et al., 2006) and overlay explicit data on on marine vessel traffic intensity, providesspatially a reasonable, quantitative marine vessel intensity, provides aa reasonable, quantitative marine vessel traffic traffic intensity, provides reasonable, quantitative and objective method to identify areas in which animals are and method to identify areas which animals are and objective objective method to tohuman identify areas in inThere which animals particularly vulnerable activities. is also valueare in particularly vulnerable to human activities. There is also value particularlyavulnerable human activities. also value in in reporting range oftomortality limits, There whenis conservation reporting of limits, conservation reporting aare a range range of mortality mortality limits, when whenterms conservation objectives not framed in easily quantifiable (Wade et objectives are in easily quantifiable (Wade et objectives are not not framed framed quantifiableforterms terms (Wade et al., 2008). For example, oneinofeasily the motivations this study was al., 2008). For example, one of the motivations for this study was al., 2008). For example, one of the motivations for this study was to assess the likely impacts on whales resulting from the to likely impacts on resulting from to assess assess the the likelyPort impacts on whales whales from the the expansion of the of Prince Rupertresulting to accommodate expansion of the Port of Prince Rupert to accommodate expansionbulk of the Port of Prince Rupert increased and container shipping. Giventothataccommodate most of the increased bulk shipping. Given that of the increased bulk and andtocontainer container Given that most most the traffic is expected travel inshipping. an east-west direction, thisofport traffic is expected to travel in an east-west direction, this port traffic is expected to travel in an east-west direction, this port development might lead to greater risk to fin whales than development might lead greater risk to whales than development lead to toOn greater risk hand, to fin finfin, whales than humpback or might killer whales. the other humpback humpback or killer whales. On the other hand, fin, humpback humpback or killer whales. the other fin, humpback and killer whales would all beOn impacted byhand, the construction of a and would all be by of and killer killertowhales whales alland be impacted impacted by the the construction construction of aa pipeline Kitimatwould (Fig. 1) the concomitant rise in petroleum pipeline to Kitimat (Fig. 1) and the concomitant rise in petroleum pipeline to Kitimat (Fig. 1) and the concomitant rise in petroleum tanker traffic in narrow passages along the central coast (Fig. 1: tanker traffic in passages along the coast (Fig. 1: tanker trafficmainland in narrow narrownorth passages along the central central coast 1: the coastal of Port Hardy and south of(Fig. Prince the coastal mainland north of Port Hardy and south of the coastalGiven mainland of Port Hardy and south of Prince Prince Rupert). the north difficulty in adequately monitoring oil Rupert). Given the difficulty in adequately monitoring oil Rupert). in Given difficulty adequately monitoring oil pollution most the regions of BC, in shipping intensity is one of the pollution in most regions of BC, shipping intensity is one of the pollution in most regions of BC, shipping intensity is one of the best available proxy indices for ship-source oil pollution (O’Hara best proxy indices for ship-source oil (O’Hara best available available ship-source oil pollution pollution (O’Hara and Morgan, proxy 2006).indices In thefor same way, movement patterns for and Morgan, 2006). In the same way, movement patterns and Morgan, 2006). In thealso same way, patterns for for large vessels will probably serve as amovement proxy for catastrophic large vessels will probably also serve as for catastrophic large vessels will probably also serveloaded as aa proxy proxy catastrophic oil spill risk. In August 2007, a barge with for a fuel truck and oil risk. 2007, aa barge loaded with and oil spill spill risk. In In August August loaded with aa fuel fuelbetruck truck and other equipment tipped2007, over inbarge the area identified to the area other equipment tipped over in the area identified to be the area other equipment tipped over in the area identified to be the area of highest risk for interactions between killer whales and ships of risk for between whales ships of highest highest riskaccident for interactions interactions between killer killer10,000L whales and and ships (Fig. 2). The spilled approximately of diesel (Fig. 2). The accident spilled approximately 10,000L of diesel (Fig.and 2).aThe accident spilled approximately 10,000L of diesel fuel similar volume of other hydrocarbons. It was estimated fuel aa similar of other hydrocarbons. ItIt was estimated fuel and and similar volume volume hydrocarbons. was estimated that approximately 25% of ofother the northern resident killer whale that approximately 25% of the northern resident killer whale that approximately 25% of the northern resident killer population was seen in the vicinity of the spill and may havewhale been population was seen in of population in the the vicinity vicinity of the the spill spill and and may may have have been been exposed to was fuel seen (Williams et al., 2009). exposed to et exposed to fuel fuel (Williams (Williams et al., al., 2009). 2009). Quantitative risk assessments such as those presented Quantitative risk such as those Quantitative riskforassessments assessments those presented presented here can be useful identifying such areas as of overlap between here can be useful for identifying areas of overlap between here can be useful for identifying areas of overlap between intense or high-risk human activities and relatively large intense human and large intense or orofhigh-risk high-risk human activities activities and relatively relatively large fractions wildlife populations. This framework might be fractions of wildlife This might be fractions wildlife populations. populations. This framework framework be useful forofevaluating various least-cost scenariosmight to plan useful for various least-cost scenarios to plan usefulshipping for evaluating evaluating various least-cost plan new routes that minimise threatscenarios to whalestowhile new shipping that threat to while new minimising shipping routes routes that minimise minimise threat to whales whales while also disruption to industry. It would certainly be also minimising disruption to industry. It would certainly be also minimising disruption to industry.early It would certainly be easier to consider whale distribution in the planning easier consider whale early in easier to tobefore considerenvironmental whale distribution distribution earlyassessments in the the planning planning stages impact are stages environmental impact are stages before before environmental impact assessments assessments are completed, permits attained, business/operation plans are completed, permits attained, business/operation plans are completed, permits attained, business/operation plans developed and infrastructure is built. There is a need are for developed infrastructure is There is for developedtoand and infrastructure is built. built. There is aa need need for research inform policy as soon as possible, before research to policy as as before research traffic to inform inform policy as soon soon as possible, possible, shipping patterns become established, becausebefore once shipping patterns become established, because once shipping traffic traffic become because once entrenched and patterns integrated into established, business plans, shipping entrenched and integrated into business plans, shipping entrenched and integrated into business plans, shipping 8 WILLIAMS & O’HARA: MODELLING SHIP STRIKE RISK IN BRITISH COLUMBIA routes become difficult to modify. For humpback and killer whales, several channels along the Inside Passage emerge from the analyses as candidates for places where ships might be requested to travel at low speed, or to avoid altogether where feasible. Future risk assessments along these lines can inform management of protected areas and lead to efficient resource allocation for emergency preparation and response measures. If If there thereisisananaccident, accident, industry responsible thethe industry responsible for for accident likelybenefit benefitfrom from such such emergency the the accident willwilllikely preparation as this this will will lead lead to to aa more moreefficient efficientresponse. response. management objectives objectives for marine mammal As Canadian management stocks are being developed and articulated in quantitative terms (Hammill and Stenson, 2007; Johnston et et al., al., 2000; Williams et et al., al., 2008), it is time to assess the population-level consequences of ship strikes and non-fishery mortality in similarly quantitative terms. Fisheries and Oceans Canada is developing a regional marine mammal response network to respond to cetacean strandings, particularly for those species that are listed under Canada’s Species at Risk Act. The spatial statistical modelling methods presented presented here here provide a useful, visual tool for managers to identify potential problem areas, to manage shipping activities accordingly in as efficient a manner as possible, to allocate funds in priority regions for research, for identifying priority beaches to monitor for carcass detection and possible recovery and to mitigate impacts wherever possible. ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS The line transect data came from a study designed with Len Thomas Dougdata Sandilands, and were collected RW in The line and transect came from a study designed by with Len partnership with Sandilands, Raincoast Conservation Society. Len Thomas and Doug and were collected by RW in Thomas alsowith provided substantial assistance with data partnership Raincoast Conservation Society. Len analysis. attempts at estimating potentialwith limits to Thomas The also first provided substantial assistance data anthropogenic mortality were on collaborations with analysis. The first attempts at based estimating potential limits to Anna Hall and mortality Arliss Winship, with additional advice from anthropogenic were based on collaborations with Erin Moira Brown, Meganwith Ferguson, Philadvice Hammond, AnnaAshe, Hall and Arliss Winship, additional from Dave Johnston, Bruce Mate, Andrew Trites Phil and Paul Wade. Erin Ashe, Moira Brown, Megan Ferguson, Hammond, We Jackie Stephen Doug Davethank Johnston, BruceHildering, Mate, Andrew Trites Raverty, and Paul Wade. Sandilands LisaHildering, Spaven forStephen sharing Raverty, information on We thank and Jackie Doug vessel strikesand in the All for analyses were supported on by Sandilands Lisaregion. Spaven sharing information avessel grantstrikes to RW theAllNational Wildlife in thefrom region. analysesFish were and supported by Foundation’s National thank a grant to RW fromWhale the Conservation National FishFund. and We Wildlife the IWC and National Sarah Haney Whale Institute) for Foundation’s Whale(Canadian Conservation Fund. We thank travel funds attend the (Canadian 2009 Annual IWC Scientific the IWC and to Sarah Haney Whale Institute) for Committee and the the 2009 members of the bycatch subtravel fundsmeeting to attend Annual IWC Scientific committee valuable PDOofbenefitted fromsubinCommittee for meeting andfeedback. the members the bycatch kind support the Department of Fisheries andfrom Oceans committee forfrom valuable feedback. PDO benefitted inat thesupport Institute of the Ocean Sciences,ofSidney, BC.and WeOceans thank kind from Department Fisheries Russell Leaper ofand one Sciences, anonymous reviewer for thank their at the Institute Ocean Sidney, BC. We Russell Leaper and one reviewer for their comments on an earlier draftanonymous of the manuscript. comments on an earlier draft of the manuscript. REFERENCES Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake, J.L., Borchers, REFERENCES D.L. and Thomas, L. 2001. Introduction to Distance Sampling: Buckland, S.T., Anderson,ofD.R., Burnham, K.P., Laake, J.L., University Borchers, Estimating Abundance Biological Populations. Oxford D.L. and Thomas, L. 2001. Introduction to Distance Sampling: Press, Oxford, UK. vi+xv+432pp. Estimating of Biological Populations. Oxford University Douglas, A.B.,Abundance Calambokidis, J., Raverty, S., Jeffries, S.J., Lambourn, Press,and Oxford, UK. S.A. vi+xv+432pp. D.M. Norman, 2008. Incidence of ship strikes of large whales Douglas, A.B., Calambokidis, J., Raverty, S., Jeffries, S.J., Lambourn, in Washington State. J. Mar. Biol. Assoc. UK. D.M. and S.A. 2008. ship strikes large whales Fisheries and Norman, Oceans Canada. 2007.Incidence Recoveryof strategy for the of Transient Killer in Washington Biol. Assoc. UK. draft]. Species at Risk Act Whale (OrcinusState. orca)J.inMar. Canada [Consultation Fisheries andstrategy Oceansseries, Canada. 2007. Recovery strategy for the Transient Killer recovery Fisheries and Oceans, Canada, Vancouver. 42pp. Whale and (Orcinus orca) in Canada draft].for Species at Risk Act Fisheries Oceans Canada. 2008.[Consultation Recovery strategy the Northern and recovery strategy series, Fisheries and Oceans, Canada, Vancouver. 42pp. Southern Resident Killer Whales (Orcinus orca) in Canada. Species at Fisheries and Oceans Canada. 2008. Recovery strategy for the Northern and Risk Act recovery strategy series, Fisheries and Oceans, Canada, Ottawa. Southern Resident Killer Whales (Orcinus orca) in Canada. Species at Risk ix+81pp. Act recovery strategy series, Fisheries and Oceans, Canada, Ottawa. ix+81pp. Ford, 2000. Killer Killer Whales: Whales: The The Ford, J.K.B., J.K.B., Ellis, Ellis, G.M. G.M. and and Balcomb, Balcomb, K.C. K.C. 2000. Natural History and Genealogy of Orcinus orca in British Columbia and Washington. 2nd ed. UBC Press, Vancouver. 102pp. View publication stats Fujiwara, and Caswell, H. 2001. the endangered NaturalM. History and Genealogy of Demography Orcinus orca of in British Columbianorth and Atlantic right2nd whale. 414(November Washington. ed. Nature UBC Press, Vancouver.2001): 102pp.537-41. Fujiwara, M.ofand Caswell, H. SARA 2001. Demography of the endangered north Government Canada. 2006. (Species At Risk Act). Final recovery Atlantic for right whale. Nature 414(November 2001): 537-41. B. physalus, strategy blue, fin and sei whales (Balaenoptera musculus, Government of Canada. 2006. SARA (Species At Risk Act).ofFinal recovery and B. borealis) in Pacific Canadian waters. Government Canada. strategyM.O. for blue, and seiG.B. whales (Balaenoptera B. physalus, Hammill, and fin Stenson, 2007. Applicationmusculus, of the precautionary and B. borealis) in Pacific Canadian of Canada. approach and conservation referencewaters. pointsGovernment to management of Atlantic Hammill, M.O. and Stenson, G.B. 2007. Application of the precautionary seals. ICES J. Mar. Sci. 64: 702-06. approach and conservation reference points to management of Atlantic Hedley, Buckland, seals.S., ICES J. Mar.S.T. Sci.and 64: Borchers, 702-06. D.L. 1999. Spatial modelling from line transect data. J. Cetacean Manage. 1(3): 255-64. Hedley, S., Buckland, S.T. and Res. Borchers, D.L. 1999. Spatial modelling International Whaling 1996. Report1(3): of 255-64. the Scientific from line transect data.Commission. J. Cetacean Res. Manage. Committee. Whaling Rep. int. Whal. Commn 46:50-97. International Commission. 1996. Report of the Scientific Jensen, A.S. and G.K.Commn 2003. 46:50-97. Large whale ship strike database. Committee. Rep.Silber, int. Whal. NOAAA.S. Tech.and Mem. NMFS-OPR: 37pp. [Available at: http://www.nmfs. Jensen, Silber, G.K. 2003. Large whale ship strike database. NOAA Tech. Mem. NMFS-OPR: 37pp. [Available at: noaa.gov/pr/overview/publicat.html] http://www.nmfs.noaa.gov/pr/overview/publicat.html] Johnston, D.W., Meisenheimer, P. and Lavigne, D.M. 2000. An evaluation of Johnston, D.W.,objectives Meisenheimer, P. and Lavigne, D.M.harp 2000. Anhunt, evaluation management for Canada’s commercial seal 1996of management objectives for Canada’s commercial harp seal hunt, 1998. Conserv. Biol. 14: 729-37. 1996-1998. Conserv.A.R., Biol.Mead, 14: 729-37. Laist, D.W., Knowlton, J.G., Collet, A.S. and Podesta, M. 2001. Laist, D.W., Knowlton, A.R., Mead, J.G., Collet, A.S. and Podesta, M. 2001. Collisions between ships and whales. Mar. Mammal Sci. 17(1): 35-75. Collisions between ships and whales. Mar. Mammal Sci. 17(1): 35-75. O’Hara, P.D. and Morgan, K.H. 2006. Do low rates of oiled carcass O’Hara, P.D. and Morgan, K.H. 2006. Do low rates of oiled carcass recovery recovery in in beached beached bird bird surveys surveys indicate indicate low low rates rates of of ship-source ship-source oil oil spills? spills? Mar. Mar. Ornithol Ornithol34: 34:133-40. 133-40. Panigada, Panigada, S., S., Donovan, Donovan, G. G. and and Hammond, Hammond, P. P. 2008. 2008. Work Work programme programme and and protocols to assess human induced mortality on fin whales and ship strikes with large whales and smaller cetaceans in the ACCOBAMS area. SC/60/BC7presented presentedto to Scientific Committee, June Paper SC/60/BC7 thethe IWCIWC Scientific Committee, June 2008, 2008, Santiago, Chile (unpublished). 5pp.available [Paper available from the Santiago, Chile (unpublished). 5pp. [Paper from the Office of Office of this Journal]. this Journal]. Panigada, S., S., Pesante, Pesante, G., G., Zanardelli, Zanardelli, M., M., Capoulade, Capoulade, F., F., Gannier, Gannier, A. A. and and Panigada, Weinrich, M.T. M.T. 2007. 2007. Mediterranean Mediterranean fin fin whales whales at at risk risk from from fatal fatal ship Weinrich, ship strikes. Mar. Mar. Poll. Poll.Bull. Bull.52(10): 52(10):1287-98. 1287-98. strikes. Raverty, S.A. and Gaydos, J.K. 2004. Killer whale necropsy and disease Raverty, S.A. and Gaydos, J.K. 2004. Killer whale necropsy and disease testing protocol. [Available at testing protocol. [Available at http://www.vetmed.ucdavis.edu/whc/ http://www.vetmed.ucdavis.edu/whc/seadoc/pdfs/orcanecropsyprotocol. seadoc/pdfs/orcanecropsyprotocol.pdf.] pdf.] Thomas, S., Marques, Marques, F.F.C., F.F.C., Thomas, L., L., Laake, Laake, J.L., J.L., Rexstad, Rexstad, E.A., E.A., Strindberg, Strindberg, S., Buckland, S.T., Borchers, Borchers,D.L., D.L.,Anderson, Anderson,D.R., D.R., Burnham, K.P., Burt, Buckland, S.T., Burnham, K.P., Burt, L., L., Hedley, S.L., Pollard, J.H.,Bishop, Bishop,J.R.B. J.R.B.and and Marques, Marques, T.A. T.A. 2006. Hedley, S.L., Pollard, J.H., Distance 6.0. Release Beta 5. Research Unit for Wildlife Population Assessment, University of St Andrews, UK. [Available from: http://www.ruwpa.st-and.ac.uk/distance/]. http://www.ruwpa.st-and.ac.uk/distance/]. Thomas, L., L., Williams, Williams, R. R. and and Sandilands, Sandilands, D. D. 2007. 2007. Designing Designing line line transect transect Thomas, surveys for for complex complex survey survey regions. regions. J. J. Cetacean Cetacean Res. Res. Manage. Manage. 9(1): 9(1): 1-13. 1-13. surveys Vanderlaan, A. and Taggart, C.T. 2007. Vessel collisions with whales: the Vanderlaan, A. and Taggart, C.T. 2007. Vessel collisions with whales: the probability of lethal injury based on vessel speed. Mar. Mammal Sci. probability of lethal injury based on vessel speed. Mar. Mammal Sci. 23(1): 144-56. 23(1):P.R. 144-56. Wade, and Angliss, R.P. 1997. Guidelines for assessing marine Wade, P.R. and Angliss, Guidelines for assessing mammal stocks: Report R.P. of the1997. GAMMMS Workshop April 3-5,marine 1996, mammal stocks: Report of theTechnical GAMMMS Workshop April 3-5, 1996, Seattle, Washington. NOAA Memorandum NMFS-OPR-12. Seattle, Washington. Technical Memorandum NMFS-OPR-12. 93pp. [Available fromNOAA http://www.nmfs.noaa.gov/publications.htm ]. 93pp.P.R., [Available fromand http://www.nmfs.noaa.gov/publications.htm]. Wade, Bass, C.L. Kasuya, T. 2008. A comparison of methods for Wade, P.R., Bass, C.L. andforKasuya, T. 2008. A comparison methods for calculating thresholds evaluating levels of catch of of Japan’s Dall’s porpoise (Phocoenoides hand-harpoon Paper SC/60/SM24 calculating thresholds fordalli) evaluating levels ofhunt. catch of Japan’s Dall’s presented(Phocoenoides to the IWC Scientific Committee, June Santiago, Chile porpoise dalli) hand-harpoon hunt.2008, Paper SC/60/SM24 (unpublished). [Paper available at the Office this Santiago, Journal]. Chile presented to the8pp. IWC Scientific Committee, June of 2008, Williams, R. and Lusseau, D. 2006. A killer whale social is vulnerable (unpublished). 8pp. [Paper available at the Office of network this Journal]. to targeted removals. Biology Letters 2: 497-500.Williams, R. is and Thomas, Williams, R. and Lusseau, D. 2006. A killer whale social network vulnerable L. 2007. Distribution and abundance of marine mammals in the coastal to targeted removals. Biology Letters 2: 497-500.Williams, R. and Thomas, waters of British Columbia, Canada. J. Cetacean Res. Manage. 9(1): 15-28. L. 2007. Distribution and abundance of marine mammals in the coastal Williams, R. and Thomas, L. 2009. Cost-effective abundance estimation of waters of British Columbia, Canada.ofJ.small-boat Cetacean Res. Manage. 9(1):whales 15-28. rare animals: testing performance surveys for killer Williams, R. Columbia. and Thomas, L. Conserv. 2009. Cost-effective abundance estimation of in British Biol. 142: 1542-47. rare animals: performance of small-boat surveys for killer whales Williams, R., testing Hall, A. and Winship, A. 2008. Potential limits to in British Columbia. Biol.ofConserv. 142: 1542-47. anthropogenic mortality small cetaceans in coastal waters of British Williams, R.,Can. Hall, A. and Winship, A. 2008. Potential limits to Columbia. J. Fish. Aquat. Sci. 65: 1867-78. Williams, R., Hedley, S.L. andofHammond, P.S. 2006. Modeling distribution and anthropogenic mortality small cetaceans in coastal waters of British abundance of Antarctic baleen whales using ships of opportunity. Ecology and Columbia. Can. J. Fish. Aquat. Sci. 65: 1867-78. Society R., 11(1): [online: Williams, Hedley, S.L.http://www.ecologyandsociety.org/vol11/iss1/art1/]. and Hammond, P.S. 2006. Modeling distribution and Williams, R.,ofLusseau, and Hammond, 2009. The role of social abundance AntarcticD.baleen whales usingP.S. ships of opportunity. Ecology aggregations and protected areas in killer whale conservation: the mixed and Society 11(1): [online: http://www.ecologyandsociety.org/vol11/iss1/art1/]. blessingR., of critical habitat. Biol. Conserv. P.S. 142:2009. 709-19. Williams, Lusseau, D. and Hammond, The role of social Wood, S.N. 2006. Generalized Additive Models: an introduction with R. aggregations and protected areas in killer whale conservation: the mixed Chapman and Hall, Boca Raton, Florida. 391pp. blessing of critical habitat. Biol. Conserv. 142: 709-19. Wood, S.N. 2006. Generalized Date received: December 2008. Additive Models: an introduction with R. Chapman Hall, Boca Raton, Florida. 391pp. Date accepted: and September 2009. Date received: December 2008. Date accepted: September 2009.