Contents Inhoud - Zoogdierwinkel
  • No tags were found...

Contents Inhoud - Zoogdierwinkel

Volume Deel 56 - Number Nummer 1 – June 2013Contents InhoudLUTRAJournal of the Dutch Mammal SocietyVolume 56 – Number 1 – June 2013Editorial Redactioneel1 Moving mammalsKees J. CantersResearch Papers Artikelen5 ἀe l ong-term influence of grazing by livestock on common vole and raptors inman-made wetlands in the NetherlandsNico Beemster & J. ἀ eo Vulink23 ἀe r accoon dog (Nyctereutes procyonoides) in the Netherlands – its present status anda risk assessmentJaap L. Mulder45 Abundance of harbour porpoises (Phocoena phocoena) on the Dutch Continental Shelf,aerial surveys in July 2010-March 2011Steve C.V. Geelhoed, Meike Scheidat, Rob S.A. van Bemmelen & Geert Aarts59 Harbour porpoises (Phocoena phocoena) in the Marsdiep area, the Netherlands:new investigations in a historical study areaMichelle Boonstra, Yvonne Radstake, Kees Rebel, Geert Aarts & Kees (C.J.) CamphuysenPrinted by Drukkerij Wilco, AmersfoortLayout by Image Realize ISSN 0024-7634

EditorialMoving mammalsἀ e only personal, tangible experience I havehad that relates in any way to the 1997 Kyotoclimate treaty (official lifespan: 2002-2012,so now assigned to the dustbin of history…)are the ‘Kyoto cookies’ on the shelves of mylocal supermarket. I can’t remember whetherthe point about them was the ‘sustainability’of their ingredients or that they were non-fattening,or perhaps it was something else – forexample an especially tasty biscuit broughtback from Kyoto by someone attending a congressthere and recreated by a D utch baker.What was clear, though, was that ‘Kyoto’,despite all the scepticism about the need forand ‘realism’ of treaties like this (even backthen), was nonetheless passed down to theproverbial man in the street and imbued withmeaning by a s upermarket chain – n ot themost conventional approach to educating thepublic on nature and the environment.Over the past few decades the world, and particularlythe western world, has been confrontedwith a s eries of far-reaching developments,including the ICT revolution, ourproven vulnerability to terrorist actions anda series of crises in the financial and economicrealm and, partly as a r esult, largescale,far-reaching geopolitical changes at theglobal level. One of the effects of the currenteconomic-financial crises is that economicgrowth has fallen, in some cases to belowzero, while positive growth is often regardedas absolutely essential for continued prosperity.In concrete terms, though, it translates toless fossil fuel being burned, falling sales andconsumption of all manner of goods – apartfrom real estate, also luxury items like multipleforeign holiday trips every year – and aneasing of road congestion. And so in a roundaboutway the goals of green-minded NGOs,for years deemed utopian and therefore out ofthe question, are nonetheless being realised.So we should count our blessings!That the original Kyoto targets have not beensecured by a long chalk and that in later preliminarytalks on follow-up treaties it provedimpossible to even reach agreement on basictargets means, among other things, that theprocess of global climate change that began tounfold in the 1990s continues unabated and thatit is now questionable whether humanity stillhas time to achieve the kind of drastic changerequired to as yet turn the tide. Perhaps duringthe forthcoming negotiations the current positionof the US president – no longer guided, ashe was, by a desire for re-election when settingand implementing a climate agenda – will contributeto greater, and more structural, successwhen it comes to formulating and then actuallysecuring climate targets. And perhaps nowthat the BRIC nations (Brazil, Russia, India andChina) are also becoming more wealthy thesecountries will also show rather more responsibilityin this arena and translate that responsibilityinto concrete, practical action. It wouldcertainly be worth the effort.Editorial / Lutra 2013 56 (1): 1-3 1

ἀ e kind of insight yielded by carefully conductedresearch on an ‘exotic’ species is demonstratedby Mulder’s case study on Dutchoccurrence of the raccoon dog (Nyctereutesprocyonoides), published in this edition ofLutra. ἀ e basic procedure is straightforward:first, track down potential problems as carefullyas possible; next, based on the facts, examinewhat, if any, negative consequences are tobe anticipated; then, where applicable, identifythe potential for intervention; and, finally, takea decision on the basis of a c ost-benefit analysis.‘Control’ then generally emerges as oneof the least attractive options, and this is thecase with the raccoon dog, too. Indeed, controlis often impossible in practice (lack of physicalmeans, unfeasibly large-scale and thus tooexpensive or with too many side-effects) and,more importantly, alternative strategies areoften cheaper and more effective, such as compensationfor damage or intensification ofmanagement activities directed towards thespecies in question. In this context it’s of interestto note that a long-term study was startedthis year on the muskrat (Ondatra zibethicus),an exotic species now common throughoutmost of the Netherlands. ἀ e species is to bemonitored in 177 5x5 km grid squares, onethirdof these with ‘pest controllers’ workingat traditional intensity, one-third with effortsstepped up and one-third with efforts ratcheteddown. How does the muskrat populationdevelop in each case? How much damage isthere to dikes and the banks of watercourses?What are the costs of repair? ἀ ese are someof the questions to be answered in this study,which will probably vastly improve our knowledgeof this exotic species and provide practicalhandles for future policy.ἀ e movement of species and the appearanceof ‘new’ species in a particular countryare part and parcel of nature, and thereforealso of mammalian life, and are consequentlyan important issue for scientific researchers.Species are, in principle, always on the move,looking for a more favourable niche in termsof climate, food, cover or whatever. Besidesthe raccoon dog, this proves to hold true forall the other species featured in the articlesin this issue of Lutra: the harbour porpoises(Phocoena phocoena) coming to explore theMarsdiep tide-race, the colonisation of newpolders by the common vole (Microtus arvalis)and the temporal changes in the geographicdistribution of porpoises on the continentalshelf.Moving mammals – open to three interpretations,all of them covered or reflected in thisissue!Meanwhile, Kees Camphuijsen has steppeddown from Lutra’s editorial board. For oversix years Kees has made a highly valued contributionto the journal, not only throughhis lucid, succinct and always closely arguedcommentaries and the professional, accurateand stimulating support he gave to authors,but also as an author himself. In the period ofhis editorship Kees was (co-)author – and generallyfirst author – of six full papers and oneshort note. In earlier years Kees had alreadybeen a r egular contributor to Lutra and hehas stated his intention to remain faithful. Wewould like to take this opportunity to thankKees once more for all he has contributed.Kees J. CantersEditorial / Lutra 2013 56 (1): 1-3 3

The long-term influence of grazing by livestock oncommon vole and raptors in man-made wetlands inthe NetherlandsNico Beemster 1,2,* & J. ἀe o Vulink 1,2,**1Rijkswaterstaat, Programma GPO, P.O. Box 24057, NL-3502 MB Utrecht, the Netherlands2Animal Ecology Group, Centre for Ecological and Evolutionary Studies, University of Groningen,P.O. Box 14, NL-9750 AA Haren, the NetherlandsAbstract: Several studies have examined the effects of grazing by wild ungulates or livestock on the abundance ofsmall mammals; some studies have also examined the effects on the abundance of raptors feeding on small mammals.In most studies the abundance of small mammals was negatively affected by grazing, while raptors were foundto show a numerical response to the density of small mammals. However, most studies rely on census data from timespans of just 1-4 years. Because there are often large fluctuations in the numbers of small mammals, there is a needfor long-term studies. In this study we analyse the long-term effects (3-27 years) of grazing by livestock on vegetationdevelopment, vegetation structure, common vole (Microtus arvalis) index and density of vole-feeding raptors inman-made wetlands in the Netherlands. ἀ e man-made wet lands studied are characterised by a low level of physicalperturbation, and without additional management measures their vegetation, of short grasses, will soon be replacedby tall vegetation dominated by reed (Phragmites australis), wood small-reed (Calamagrostis epigejos) and shrubs.Grazing was initia ted while short grasses dominated the vegetation. ἀ e intens e grazing (summer grazing with astocking rate of more than 0.8 animals.ha - ¹), created a homogene ous short vegetation. Grazing with a low stockingrate (year-round grazing with a stocking rate less than 0.6 animals.ha - ¹ or summer grazing with a stocking rate ofless than 0.1 animals.ha - ¹) led to a heteroge neous vegeta ti on. However, after a few years, relati vely sharp boundariesdeveloped between short-grazed grassland and closed reed stands and the intermediate stage, characteri sed bya moderate reed height (between ca. 0.5-1.5 m), increasingly disappeared. In grazed areas relatively high densities ofcommon voles (vole indices 15-35 voles / 100 trap nights) were restric ted to parts with a moderate reed height. Suchareas only temporarily existed and their disappearance led to a decrease of vole abundance after some years. Volefeedingraptors showed a numerical response to changing vole densities. Vegetation structure also had an effect onraptor density. Maximum raptor densities were found at sub-maximum vole indices, where average reed height wassomewhat lower. ἀ e relevance of grazing as a tool for management of vole-feeding raptors in man-made wetlandsis highly depen dent on the potential of grazing to revert tall vegetation to an earlier successional stage. ἀ e regularoccurrence of high vole densities and their predators may be achieved by creating a cyclic variation in stocking rates.Years with relatively low stocking rates should be alternated with some years with higher livestock densities.Keywords: livestock grazing, long-term effects, commonvole, Microtus arvalis, vole-feeding raptors, reedheight, wetlands.*present address: Altenburg & W ymenga EcologicalConsultants, P.O. Box 32, NL-9269 ZR Feanwâlden,the Netherlands, e-mail:© 2013 Zoogdiervereniging. Lutra articles also on theinternet:**present address: Rijkswaterstaat, Programma Projectenen Onderhoud, P.O. Box 2232, NL-3500 GEUtrecht, the NetherlandsBeemster & Vulink / Lutra 2013 56 (1): 5-21 5

IntroductionAs a r esult of large hydraulic-engineeringworks intended to protect against flooding orto reclaim land for agriculture, about 30,000ha of new wetlands have been es tablishedin the Netherlands during the last century(Schultz 1992). Many of these man-madewetlands are of international conserva tionimportance because of the occurrence of largenumbers of birds. Many of these bird species,including vole-feeding raptors and owls, areattracted by early stages in vegetation succession(van Eerden 1984, Dijkstra et al. 1995).Since the level of perturbation , such as inundationby salt or freshwater, erosion by ice,or grazing by ungulates, is very low in thesewetlands, these early successional stages onlyexist for a s hort while. Livestock grazing isone of the management opti ons for stoppingor slowing down vegetation succession (e.g.Bakker 1989, Scherfose 1993), with the aimof maintaining early successional stages andtheir characte ristic plant and animal species.ἀ e effects of grazing on vegetation structureand habitat use by birds have attracteda great deal of attention (e.g. Larsson 1969,Soikkeli & S alo 1979, Holechek et al. 1982,van Wieren 1991, Duncan 1992, Vulink & vanEerden 1998). ἀ e effects of grazing by wildungulates (Keesing 1998, Smit et al. 2001) orlivestock (Grant et al. 1982, Bock et al. 1984,Heske & Campbell 1991, Hayward et al. 1997,Schmidt et al. 2005, Wheeler 2008, Johnson &Horn 2008, Bakker et al. 2009) on the abundanceof small mammals and their avian predatorshave been less well studied. In most studiesthe abundance of small mammals was foundto be negatively affected by grazing, while raptorswere found to show a numerical responseto small mammal density. However, except forthe work of Hayward et al. (1997) and Bakkeret al. (2009), these studies rely on censusdata from just 1-4 years. Because small mammalsoften show large fluctuations in numbers,there is a need for long-term studies (Haywardet al. 1997). Microtine rodents are known toshow large variation in population size; oftenthese fluctuations are cyclic with peaks every3-4 years (Krebs & M yers 1974, Hansson &Henttonen 1985). In the Netherlands populationsof common voles (Microtus arvalis), themain prey species for the majority of raptorspecies (Dijkstra et al. 1995), are weakly cyclic(van Wijngaarden 1957, Cavé 1968, Dijkstra &Zijlstra 1997). ἀ e numerical responses of volefeedingraptors to changing vole densities mayoccur either rapidly, without an obvious timelag (Korpimäki & Norrdahl 1989, Korpimäki& Norrdahl 1991, Korpimäki 1994) or with along delay (Keith et al. 1977, Erlinge et al. 1983).We examined the long-term effects of grazingby cattle and horses on the abundance ofcommon vole, vole-feeding raptors and onespecies of owl (hereafter referred to as raptors)in a long-term study (27 years). ἀ e effect ofvegetation development on the density of raptorsis analysed for the entire study period(1969-1995). ἀ e effects of (1) grazing on vegetationstructure, (2) vegetation structureon vole numbers, and (3) vegetation structureand vole numbers on raptor abundance,are analysed for the second part of the studyperiod (1983-1995).MethodsStudy areasἀ e study was conducted in two recentlyreclaimed areas in the Netherlands: Lauwersmeer(53°20’N, 6°10’E) and Oostvaardersplassen(52°26’N, 5°19’E) (figure 1). ἀ e Lauwersmeerpolder (9100 ha) was reclaimed fromthe Wadden Sea in 1969. ἀ e nature reserve(4500 ha) consists of former tidal flats (2100ha; hereafter referred to as flats), former accretionworks (300 ha) and shallow and deeperwaters (2100 ha). Soil types, varying fromloamy sand to clay-rich, are related to the elevationof the flats: the higher-lying flats beingmore sandy than the less-elevated ones. Afterempoldering, the soils of the flats gradually6 Beemster & Vulink / Lutra 2013 56 (1): 5-21

AWADDEN SEADIKELauwersoogNOostmahornLAKELAUWERSMEERsummer grazedstudy plotZoutkampEngwierumBLAKEIJSSELMEERgrazed former flatsungrazed former flatsstudy plot ungrazed0 1 2kmDIKEyear-round grazedstudy plotM A R S H Z O N Esummer grazedstudy plotB O R D E RZ O N E0 1 2kmFigure 1. O verview of the study areas in t he Netherlands: (A) L auwersmeer (a former estuary) and (B) O ostvaardersplassen(along lake IJsselmeer, the former Zuiderzee).desalinated (Joenje 1978, van Rooij & D rost1996). In the first few years after empoldering,an undistur bed vegetation successi on tookplace (Joenje 1978). In 1982, summer grazingwith cattle and horses was started on parts ofthe flats (1300 ha).ἀ e Oostvaardersplassen study area (5600ha) is located in southern Flevoland, a p olderreclaimed in 1968 from the freshwaterlake IJsselmeer. ἀ e nature reserve consistsof a central reed marsh (3600 ha) and a welldrainedborder zone (2000 ha). Soils in theOostvaardersplassen are clayey and thereforemore fertile than in Lauwersmeer. In the borderzone, grasslands (900 ha) were created onformer arable fields by sowing grass mixtures(for a d etailed description see Vulink & v anEerden 1998). Some areas in the border zonewere managed by summer grazing with cattleand horses between 1982 and 1993. From 1984onwards, year-round grazing with Heck cattle(Bos taurus, crossbred from primitive races)and konik horses (Equus ferus, a p rimitivebreed of horse originating from Poland) hastaken place. ἀ e area with year-round grazinggradually expanded in accordance withgrowth of the herds.Study plotsIn Lauwersmeer, common voles are morenu merous on high-lying flats (>0.65 m abovetarget level) than on less-elevated flats (

Table 1. Study plot characteristics of the summer-grazed and ungrazed study plots in Lauwersmeer, and the summer-grazedand year-round grazed study plots in Oostvaardersplassen.LauwersmeerOostvaardersplassenGrazing regime Summer grazing Ungrazed Summer grazing Year-round grazingStudy period 1983-1995 1989-1993 1991-1993 1989-1995Size (ha) 33 190 120 288Stocking rate (animal.ha - ¹) 0-1.1 0 1.4-1.7 0.3-0.9Annual grazing period 1982-1992; 7 June to 30 Sept.1993-1995; 1 May to 31 Oct.n.a. 1 May to 31 Oct. Year-roundabove target level). For example, in October1983 the density of burrows was 65.ha -1on high-lying flats versus 2 b urrows.ha -1 onless-elevated flats. ἀ e reason for this is thatlow-lying flats are regularly flooded for severaldays, whilst only parts of the high-lyingflats are occasionally inundated. ἀ e studyon the relationship between grazing, vegetationstructure, vole density and abundance ofraptors was carried out on high-lying flats ofthe nature reserve, some grazed and othersungrazed (figure 1A). For the most importantcharacteristics of the study plots, see table 1.In Oostvaardersplassen, grasses were sownin the summer-grazed study plot in 1989. ἀ evegetation was mown for a period of two years(twice a year), after which the area was summergrazed at high stocking rates for threeyears (for characteristics of the study plot, seetable 1). In the year-round grazed study plotin Oostvaardersplassen grasses were sown in1982. In the period 1984 to 1988 the density ofHeck cattle and konik horses was relatively lowand grassland was also mown; 1989 was thefirst year of grazing without additio nal mowing(for characteristics of the study plot, seetable 1). ἀ e locations of the summer-grazedand year-round grazed study plots in Oostvaardersplassenare shown in figure 1B.Vegetation developmentLong-term vegetation development on flatsin Lauwersmeer was derived from vegetationmaps (scale 1:5000 or 1:10,000) based on theinterpretation of satellite photographs (1972,1975) and aerial photographs (1980, 1984 and1989), combined with ground surveys thatidentified the different zones by using datafrom quadrates (Küchler & Zonneveld 1988).ἀ e vegetation composition in 1995 was basedon field visits. ἀ e original vegetation mapsdistinguished between 10 vegetation types.In this study, vegetation types are groupedinto three categories: vegetation of halophyticpioneers, dominated by glasswort (Salicorniaspp.), with herbaceous seepweed (Suaedamaritime), lesser sea-spurrey (Spergulariamarina) and greater sea-spurry (Spergulariamedia); vegetation of short grasses, dominatedby creeping bentgrass (Agrostis stolonifera)and common saltmarsh-grass (Puccinelliamaritima), with red fescue (Festuca rubra),marsh foxtail (Alopecurus geniculatus) andreed; tall vegetation, dominated by reed, withwood small-reed (Calamagrostis epigejos) andwillows (Salix spp.). ἀ e scientific nomenclatureof the plant species follows van derMeijden (1996).As succession proceeds vegetation dominatedby short grasses gradually changes intotall vegetation, dominated by reed on desalinatedformer flats in Lauwersmeer as well ason grassland in Oostvaardersplassen. Sincethere is a significant correlation between reedcover and reed height (Huijser et al. 1996), thelatter was used as an index for the vegetationstructure in the study plots. Reed height wasmeasured in autumn, using a measu ring rod.At each vole-trapping station, five measurementswere taken with a t otal of 100 or 1508 Beemster & Vulink / Lutra 2013 56 (1): 5-21

eadings per study plot. Before 1988, measurementsin Lauwersmeer were based on visualestimati ons (estimated to the nearest 5 cm).Developments in the vegetation structure ofshort grasses in Oostvaardersplassen in 1989-93 were described in terms of average swardheight. Sward height was measured by usinga polystyre ne disc (radius 50 cm, weight 320gram). ἀ e disc was gently lowered on to thesward, and the height of the vegetation wasread off on the measuring staff in the centre.At each vole trapping station, the height wasmeasured five times. ἀ e 100 to 150 readingsper study plot were averaged.Vole densitiesVole densities were measured two or threetimes a year (March, July and October). Smallmammals were caught according to the modifiedmethod of Hörnfeldt (1978). In each studyplot, trap lines were set up at 30 m i ntervalsand randomly assigned to each of the threetrapping periods per year. On each trap line10 trapping stations were situated at 10 mintervals, with five traps at each trapping station.Traps were controlled once a d ay, for aperiod of three days. Consequently, the numberof trap nights was 150 per trap line. Usually,two trap lines were set up per study plot.In the summer-grazed study plot in Lauwersmeer,the number of trap lines was three after1985, in the year-round study plot in Oostvaardersplassenthe number of trap lines wasfour in all years. In all study plots, commonvoles made up more than 95% of the smallmammals caught. ἀ e vole index was definedas the number of common voles caught per100 trap nights.Raptor densitiesCounts of raptors in Lauwersmeer were initiatedin 1969, the year of reclamation; counts inOostvaardersplassen were started in 1982, fifteenyears after empoldering was completed.In both study areas, counts were organisedabout once a month. In Lauwersmeer the raptorspecies (partly) feeding on common volesare hen harrier (Circus cyaneus), buzzard(Buteo buteo), roug h-leg ged buzzard (Buteolagopus), kestrel (Falco tinnun culus) andshort-eared owl (Asio ἀammeus). Hen harrierand rough-legged buzzard are mainly presentin winter, while buzzard, kestrel and shortearedowl are present throughout the year.Data of the two most common species (henharrier and kestrel) were selected for detailedanalysis in the study plots. In the study areas,the diets of these two species consisted ofmore than 90% of common voles (Masman etal. 1988, Dijkstra et al. 1995).Raptor densities were expressed as birddays.ha-1 per year. ἀ e number of bird-dayswas calculated for each interval between consecutivecounts as the average of these twocounts multiplied by the interval length indays. ἀ ese values were added up for each year(months from July till June). For the entireflats in Lauwersmeer all birds were included,whilst in the study plots only flight-huntingbirds were selected.ResultsThe effects of grazing on vegetationstructureOn Lauwersmeer flats with a natural successionof the vegetation, parallel to the processof desalination, halophytic pioneers weregradually replaced by short grasses, whichin their turn were replaced by tall vegetation(figure 2A). Flats where summer grazing wasinitiated were, on average, in an earlier successionalstage at the onset of grazing thanflats which remained ungrazed (cf. figures 2Aand 2B). After the initiation of grazing, thesuccessive decrease in the incidence of shortgrasses and of halophytic pioneers halted forsome years, this was in contrast to a s teadyBeemster & Vulink / Lutra 2013 56 (1): 5-21 9

incidence (%)100806040200100806040200NATURAL SUCCESSION1NATURAL SUCCESSIONhalophytic pioneersshort-growing grassestall vegetationsGRAZED5 10 15 20 25years after embankmentFigure 2. Vegetation succession on former tidal flats inthe Lauwersmeer nature reserve (year 1 = 1969): (A)flats with a natural succession of the vegetation (1030ha), (B) flats with summer grazing (since year 14; 890ha). ἀ e percentage of each vegetation type is presented(= % incidence) for years 4, 7, 12, 16, 21 and 27after empoldering.decrease on the ungrazed flats. ἀ ereafter, theproportion of halophytic pioneers and shortgrasses decreased again, while the proportionof tall vegetation increased. However, even inthe last year of the study period, the proportionof tall vegetation was much lower than onungrazed flats (figures 2A and 2B).In the summer-grazed study plot in Lauwersmeer,grazing transformed a v egetation ofmosaics of reed and short grasses into a largescaleopen vegetation. In the first five years ofgrazing, when the average stocking rate wasrelatively high (figure 3A1), the average reedheight was lower than 0.5 m (figure 3B1). Afterthe fifth year, when the stocking rate was verylow, the average reed height sharply increasedand gradually levelled off. In the last years ofthe study period, reed height was only slightlyABlower than on comparable flats in the ungrazedstudy plot at Lauwersmeer (1.4 m v ersus 1.6m). Gradually, the vegetation in the summergrazedstudy plot changed into a c losed reedstand, comparable to the ungrazed study plot.On less-elevated flats the vegetation remainedopen: the average reed height was lower than0.7 m for the entire period of grazing.In the summer-grazed study plot in Oostvaardersplassen,grazing at high stockingrates (cf. table 1) resulted in a homogeneousshort grassland with a v ery low reed height(less than 0.1 m) for the entire study period.Average sward height showed only a s lightincrease in summer (table 2A).In the year-round grazed study plot inOostvaardersplassen the average stockingrate strongly increased from about 0.3 to 0.9animals.ha -1 during the study period (table1). ἀ e average stocking rate increased from0.3 to 1.1 animals.ha -1 in summer, but wasrather stable in winter (in most years 0.3-0.6animals.ha -1 ; figure 3A2). Cattle and horsesdid not distribute equally over the study plot.In a p art of the study plot the animal densitywas extremely low from the first year ofgrazing without additional mowing. ἀ e areawith a r eed height higher than 1.5 m t hereforeincreased from 0% in year 1 to 30% inyear 7. ἀ e gradual increase in stocking rateduring the study period resulted in an expansionof heavily grazed grassland (reed height0-0.5 m) from 40% in year 1 to 60% in year7. In the same period the percentage of moderatelygrazed grassland (reed height 0.5-1.5 m) decreased from 60% to 10%. In thelast two years of the study period, the areamainly con sisted of intensively-grazed grassland(60%) and ungrazed closed reed stands(30%) with rather sharp boundaries betweenthe two. Maximum reed height in the reedstands in Oostvaardersplassen was higherthan in Lauwers meer, due to higher soil fertility.Along the vole-trap li nes, situated in thepart of the study plot with a moderate grazingintensity, average reed height increased from0.6 m in the first year of grazing to 0.8-1.0 m10 Beemster & Vulink / Lutra 2013 56 (1): 5-21

Table 2. A. Average sward height (cm, ±sd) in March, July and October in the summer-grazed and a part of theyear-round grazed study plot in Oostvaardersplassen. Measurements were made at vole-trap lines. B. Average voleindex (±sd) in March, July and October in the summer-grazed and a part of the year-round grazed study plot inOost vaar dersplassen (n refers to the number of trap lines).ASward heightSummer grazingYear-round grazingMarch 6.7±7.2 (1992-93; n=100) 5.1±8.4 (1991-93; n=150)July 18.8±15.5 (1991-92; n=100) 38.6±17.7 (1990-92; n=150)October 6.5±7.4 (1991-93; n=150) 19.5±14.1 (1989-93; n=250)BVole indexSummer grazingYear-round grazingMarch 0.0±0.0 (1992-93; n=4) 0.2±0.3 (1992-93; n=4)July 1.0±1.2 (1991-92; n=4) 2.5±2.9 (1991-92; n=4)October 0.7±0.6 (1991-93; n=6) 12.4±9.7 (1991-93; n=6)in years 3-4, and decreased to about 0.3-0.4 min years 5-7 (figure 3B2).ἀ ere was more seasonal variation in swardheight in the year-round grazed study plotthan in the summer-grazed study plot (table2A). In the year-round grazed study plot cattleand horses minimised sward height outsidethe closed reed stands during the winter,which led to short grassland in early spring.In March, the average sward height was significantlylower than in the summer-grazedstudy plot (Mann-Whitney U t est, P

1. Summer-grazed study plot Lauwersmeer 2. Year-round grazed study plotOostvaardersplassenAanimals.ha - height (m) index25201510504035302520151050Dbird-days.ha -114121086420hen harrierkestrel0 2 4 6 8 10 12 1443210hen harrierkestrel0 1 2 3 4 5 6 7years of grazing managementyears of grazing managementFigure 3. Response of vegetation structure to grazing and consequences for the vole index and raptor density in(1) the summer-grazed study plot in Lauwersmeer (33 ha; grazing by grazing) and (2) the year-round grazed studyplot in Oostvaardersplassen (288 ha; grazing by cattle and horses). Data are all expressed as a function of yearsof grazing management (Lauwersmeer: year 1 = 1982; Oostvaardersplassen year 1 = 1989). (A) Average stockingrate (in year-round grazed study plot in Oostvaardersplassen shown for summer (April-September) and winter(October-March)), (B) Average reed height in October , (C) vole index in October, vertical bars refer to vole peakyears(based on data for October from the well-drained border zone in the Lauwersmeer nature reserve (Dijkstraet al. 1995) and from the borderzone in Oostvaardersplassen (N. Beemster, unpublished data)), and (D) hen harrierand kestrel density, expressed as bird-days.ha -1 per year (1 July - 30 June) . Only hunting birds were included.12 Beemster & Vulink / Lutra 2013 56 (1): 5-21

Avole indexBbird-days.ha -1Cbird-days.ha -1year-round grazed OVPsummer-grazed LM40353025201510500 0.2 0.4 0.6 0.8 1 1.2 1.414hen harrier1210864200 0.2 0.4 0.6 0.8 1 1.2 1.49kestrel8765432100 0.2 0.4 0.6 0.8 1 1.2 1.4reed height (m)Figure 4. R elation of vole index (A) a nd raptor density(hen harrier (B), kestrel (C)) to reed height in thesummer-grazed study plot in L auwers meer (33 h a)and the year-round grazed study plot in Oostvaardersplassen(288 ha). Data are expressed as a function ofyears of grazing management (summer grazed studyplot in Lauwersmeer: year 1 = 1982; year-round grazedstudy plot in Oostvaardersplassen year 1 = 1989). (A)Vole index in October (summer-grazed study plot inLauwers meer: R 2 =0.85, P

ird-days.ha- harriersummer grazingnatural successionbuzzardrough-legged buzzardkestrelshort-eared owlGRAZED5 10 15 20 25years after embankmentFigure 5. D ensities of raptors, depending for a m ajorpart of their diet o n common voles, on former tidalflats in Lauwersmeer after empoldering (year 1 = 1969);closed dots, flats with a natural succession of the vegetation(1030 ha), open dots, flats with summer grazing(since year 14; 890 ha) . (A) hen harrier, (B) buzzard, (C)rough-legged buzzard, (D) kestrel, (E) short-eared owl.ABCDEafter empoldering, followed by a stabilisation.In the latter half of the study period raptornumbers gradually declined, with some speciesdeclining somewhat earlier than others(figure 5). Long-term developments in raptorabundance are related to a change in theextent of the area covered with short grasses(figure 2A) and consequently to a c hange infood supply of common voles (Dijkstra et al.1995). Buzzard numbers increased again morethan twenty years after empoldering, whena small breeding population became established.ἀ ese birds mainly find their food outsidethe ungrazed flats.At the time grazing was introduced, somespecies were still numerous (hen harrier andrough-legged buz zard), while others wereon the decline (buzzard and kestrel) or hadalmost disappeared (short-eared owl). Afterthe introduction of grazing, four out of fivespecies showed a temporary increase in number,contrasting with the general decline onthe ungrazed flats. ἀ e short-eared owl didnot respond to this change: the species hadalmost disappeared before the onset of grazing,and this measure did not result in thereturn of the species (figure 5E). ἀ e follo wingremarks are confined to the four species thatresponded to the introduction of grazing.For the entire period of grazing, three outof four species were sig nificantly more abundanton grazed flats than on the consistentlyungrazed flats (Wilcoxon signed rank test:n=14 years, buzzard P

years, P

grazed vegetation and closed reed standsdeveloped and there were almost no areaswith a moderate reed height. A nnual variationin vegetation structure was more pronouncedin the year-round grazed study plotthan in the summer-grazed study plots.Effects of grazing on vole densityIn grazed areas, relatively high densities ofcommon voles were restric ted to parts with aspecific vegetation structu re, characterised bya moderate reed height (between ca. 0.5 and1.5 m). Areas with such a reed height existedonly temporarily. Grazing may affect voleabundance in different ways: by trampling, bychanging the vegetation structure and/or byinfluencing the availability of food for voles.TramplingNormally, common voles live in burrowsin the upper soil and feed above ground. Inthe case of high densities of ungulates, tramplingmight well have a serious effect on voleabundan ce, by destroying burrows and compactingthe soil (cf. Heske & Campbell 1991).ἀ e risk of trampling is probably higher duringwet conditions in winter and thereforehigher in year-round grazed areas than insummer-grazed areas. On Lauwersmeer flats,characterised by high water tables in winterand occasionally in summer, common volesare forced to live above the ground for a substantialpart of the year. Above-ground, voleslive in nests of grass in the vegetation (ownobservations) and are probably vulnerable totrampling.Vegetation structureHerbivores may affect the vegetation structuredirectly by grazing and trampling, orindirectly by changing the abiotic environment(compacting and thereby salinatingthe soil, or influencing the level of nutrients(Scherfose 1993)). ἀ e vegetation structure isthought to influence vole density (Edge et al.1995, Peles & Barrett 1996). A uniform, low,vegetation has a negative effect on vole densitybecause it does not offer the voles enoughcover. ἀ is may have played a role in some ofthe grazing units. In the summer-grazed studyplot in Oostvaardersplassen, vegetation wasshort-grazed throughout the year and a l ackof cover may have been responsible for the lowvole densities. In a large part of the year-roundgrazed study plot, vegetation varied annuallybetween rough grassland in summer andshort grassland in winter and early spring. ἀ ispart of the study plot probably was a f avourablehabitat for voles in summer, but less so inwinter. Flats in Lauwer smeer, characterised byhigh groundwa ter tables, undoubtedly need amore structured vegeta tion for voles to survivein winter than well-drained environments. Alack of cover was probably the main reason forlow vole densities in years with a relatively highstocking rate.In the 14-year study period on former tidalflats in the summer-grazed study plot in Lauwersmeerthe vole population did not showthe characteristic 3-4 years cycle in density.ἀ e absence of this cycle might have been dueto large changes in vegetation structure overthe years and by high groundwater tables atthe flats during winter. ἀ roughout the wintermonths (October-March) the vole indexdeclined by 95-98% (n=2 winters).Availability of foodVoles (Microtus spp.) often fluctuate greatly innumber and food availability plays an importantrole in limiting their numbers (Hansson1979). ἀ roughout the year, the diet of thecommon vole in the study areas mainly consistsof the green parts of monoco tyledons,with the seeds of monoco tyledons and dicotyledonsplaying a r ole in summer (Hoog e-boom, unpublished data).Grazing affects food abundance and qualityfor voles in different ways. ἀ rough repeatedgrazing, grass maturation is prevented andgrowth stimulated, resulting in a higher foodquality. However, the prevention of grass16 Beemster & Vulink / Lutra 2013 56 (1): 5-21

matura tion suppresses the production ofseeds (McNaughton 1979).Where grassland in the study areas wasintensely grazed, seed production, as indicatedby inflorescence abundance, was verylow. On Lauwersmeer flats, seed productionin creeping bentgrass sharply decrea sed afterthe introduction of summer grazing (vanEerden et al. 1997). In grassland with a l owgrazing pressure, seed production appearedto be relatively high during the stage with amoderate reed height. At this stage commonvoles were relatively numerous. When grasslandchanged into a closed reed stand, grasscover- a nd hence seed production - sharplydecreased. It seems probable that the decreasein vole index at this stage can be explained bya decrease in the availability of food.In the year-round grazed study plot in Oostvaardersplassenand the summer-grazedstudy plot in Lauwersmeer hen harrier andkestrel showed a clear numerical response tovole abundance. ἀ e additional effect of vegetationcover was only detected in the summergrazed study plot in Lauwersmeer and notin the year-round grazed study plot in Oostvaardersplassen,but this was probably due tothe small sample sizes (number of years). InLauwersmeer the highest densities of raptorswere found at sub-maximum prey densities,where vegetation cover, as indicated by reedheight, was somewhat lower. We concludethat the numerical response of vole-feedingraptors to change in vole availability occurredvery rapidly. ἀ is holds for the hen harrier,which is mainly a w inter visitor, as wellas for the kestrel, which is a partial migrantand a c ommon breeding bird in the studyareas (Cavé 1968, Masman et al. 1988). ἀ eseresults are in accordance with the results ofKorpimäki (1994) and Korpimäki & N orrdahl(1989, 1991), who found that the densitiesof most avian predators in western Finlandtracked vole densities rapidly, withoutobvious time lags.Maximum raptor densities in the summergrazed study plot in Lauwersmeer weremuch higher than in the year-round grazedstudy plot in Oostvaardersplassen, despitelower maximum vole indices in Lauwersmeer.On Lauwersmeer flats, characterisedby high water tables in winter and occasionallyin summer, common voles are forced tolive above the ground for a substantial partof the year. Above-ground, voles live in nestsof grass in the vegetation (own observations)and are probably more vulnerable to predation.Additionally, Lauwersmeer flats have asomewhat more open vegetation, because of amore sandy soil.Effects of vole density and vegetationstructure on raptor densityConclusionsIn the study areas grazing regimes with relativelylow stocking rates created a suitablevegetation structure for common voles andsubsequently for vole-feeding raptors, for fiveyears at the most. In man-made wetlands,characterised by a l ow level of physical perturbation,the relevance of grazing as a managementtool for encouraging vole-feedingraptors will greatly depen d on the potential ofgrazing to revert tall vegetation to an earliersuccessional stage.Vegetation dominated by reeds can be successfullyreverted into a v egetation dominatedby short grasses by intense grazing(cf. van Deursen & Drost 1990, Vulink et al.2000, the present study). Subsequently, theregular occurrence of relatively high densitiesof common voles and vole-feeding raptorscan be achieved by creating a cyclic variationin stocking rate. Years with a r elative ly lowstocking rate should be alterna ted with someyears with a higher one. In the first years afterlowering the stocking rate, high densities ofcommon voles and vole-feeding raptors can beexpected. In Oostvaardersplassen, in 1997 theBeemster & Vulink / Lutra 2013 56 (1): 5-21 17

area with year-round grazing was extended tothe entire border zone and stocking rates in thisarea gradually increased to about 2.4 animals.ha - ¹ in recent years (in 2011 about 350 Heckcattle, 1150 konik horses and 3300 deer ( ἀ e area became shortgrazed for a large part of the year and densitiesof hen harrier and kestrel decreased to very lowlevels (less than 0.1 bird-days.ha - ¹). In spring2010 ten small exclosures (10x8 m) were erectedin the year-round grazed area of Oostvaardersplassenand vole trapping with life traps inOctober 2010 showed substantial densities ofcommon voles within the exclosures and novoles in the areas surrounding the exclosuresnearby (own observations). In the summergrazed study plot in Lauwersmeer the stockingrates remained very low and the reed stands arenowadays gradually being succeeded by willowspecies. Since 1995 densities of the hen harrierremained low (less than 2 bird-days.ha - ¹), andthe kestrel became an irregular visitor. Higherstocking rates are needed to revert tall vegetationto an earlier successional stage, before volefeeding raptors are able to profit from the abundanceof common voles again.In addition to vole-feeding raptors, red fox(Vulpes vulpes), mustelids (Mustela nivalis andMustela erminea) and herons (Botaurus stellaris,Egretta alba and Ardea cinerea) are also ableto profit from the abundance of common voles.In years with high stocking rates, raptor densitieswill be lower, and the area will be moresuitable for species that prefer a vegeta tion ofshort grasses, such as geese (Vulink et al. 2000,Vulink et al. 2010). If the preservation of volefeedingraptors is to be a goal of the managementof large-scale man-made wetlands, themanagement should include spatial and temporalvariation in grazing pressure, followingthe pattern of partial migration in natural systemssuch as the Serengeti (e.g. McNaughton &Banyikwa 1995).Acknowledgements: A great many people have contributedto the field work. We would especially like to thank:Dick Boshoff, Perry Cornelissen, Rudi Drent (†), Mennobartvan Eerden, Jan Hulscher, Léon Jobse, EduardKoopman, Reinko van der Laan (†), Klaas van Maanen,Stef van Rijn, Bram Smit, Joost Tinbergen, Paul Verhagen,Dirk Vogt, Kees Vos, Jan de Vries and Menno Zijlstra.Nico Dijkshoorn, Jan Griekspoor, Bert Versluis, SysKrap and their colleagues supported us by making theirequipment available to us. ἀ e diagrams were made byDick Visser. English language corrections were madeby Ineke Touber and Nicholas Parrott. Comments byRudi Drent (†), Herbert Prins, Serge Daan, Cor Dijkstra,Mennobart van Eerden, Maarten Platteeuw, LeoZwarts and two anonymous referees much improvedthe manu script.ReferencesBakker, J.P., J. de Leeuw & S .E. van Wieren 1984.Micropattern in grassland vegetation created andsustained by sheep-grazing. Vegetatio 55: 541-560.Bakker, J.P. 1989. Nature management by grazing andcutting. Kluwer Academic Publishers, Dordrecht,the Netherlands.Bakker, E.S., H. Olff, & J .M. Gleichman 2009. Contrastingeffects of large herbivore grazing onsmaller herbivores. Basic and Applied Ecology 10:141-150.Bock, C.E., J.H. Bock, W.R. Kenney & V.M. Hawthorne1984. Responses of birds, rodents, and vegetationto livestock exclosure in a s emidesert grasslandsite. Journal of Range Management 37: 239-242.Cavé, A.J. 1968. ἀ e breeding of the kestrel, Falco tinnunculusL., in the reclaimed area Oostelijk Flevoland.Netherlands Journal of Zoology 18: 313-407.Dijkstra, C., N. Beemster, M. Zijlstra, M.R. van Eerden& S. Daan 1995. Roofvogels in de Nederlandsewetlands. Flevobericht 381. Ministerie van Verkeeren Waterstaat, Rijkswater staat Directie IJsselmeergebied,Lelystad, the Netherlands.Dijkstra, C. & M . Zijlstra 1997. Reproduction of themarsh harrier Circus aeruginosus in recent landreclamations in the Netherlands. Ardea 85: 37-50.Duncan, P. 1992. Horses and grasses: ἀ e nutritionalecology of equids and their impact on the Camargue.Ecological Studies 87. Springer Verl ag, NewYork, USA.Edge, W.D., J.O. Wolff & R .L. Carey 1995. Density-18 Beemster & Vulink / Lutra 2013 56 (1): 5-21

dependent responses of gray-tailed voles to mowing.Journal of Wildlife Management 59: 245-251.Erlinge, S., G. Göransson, L. Hansson, G. Högstedt, O.Liberg, I.N. Nilsson, T. Nilsson, T. Von Schantz &M. Sylvén 1983. Predation as a re gulating factorin small rodent populations in southern Sweden.Oikos 40: 36-52.Grant, W.E., E.C. Birney, N.R. French & D.M. Swift1982. Structure and productivity of grasslandsmall mammal communities relating to grazinginducedchanges in vegetative cover. Journal ofMammalogy 63: 248-260.Hansson, L. 1979. Food as a l imiting factor for smallrodent numbers: Tests of two hypotheses. Oecologia37: 297-314.Hansson, L. & H. Henttonen 1985. Gradients in densityof small rodents: the importance of latitudeand snow cover. Oecologia 67: 394-402.Hayward, B., E.J. Heske & C .W. Painter 1997. Effectsof livestock grazing on small mammals at a desertcienaga. Journal of Wildlife Management 61: 123-129.Heske, E.J. & M. Campbell 1991. Effects of an 11-yearlivestock exclosure on rodent and ant numbers inthe Chihuahuan Desert, southeastern Arizona.Southwestern Naturalist 36: 89-93.Holechek, J.L., R. Valdez, S.D. Schemnitz, R.D. Pieper& C .A. Davis 1982. Manipulation of grazingto improve or main tain wildlife habi tat. WildlifeSociety Bulletin 10: 204-210.Hörnfeldt, B. 1978. Synchronous population fluctuationsin voles, small game, owls and Tularemia innorthern Sweden. Oecologia 32: 141-152.Huijser, M.P., J.ἀ . Vulink & M. Zijlstra 1996. Begrazingin de Oostvaardersplassen: Effecten op d evegetatie-structuur en het terreingebruik vangrote herbivoren en ganzen. Intern rapport 1996-5Li. Ministerie van Verkeer en Waterstaat, RijkswaterstaatDirectie IJsselmeergebied, Lelystad, theNetherlands.Joenje, W. 1978. Plant colonisation and succession onembanked sandflats. PhD thesis. University ofGroningen, Groningen, the Netherlands.Johnson, M.D. & C.M. Horn 2008. Effects of rotationalgrazing on rodents and raptors in a coastal grassland.Western North American Naturalist 68 (4):444-452.Keesing, F. 1998. Impacts of ungulates on the demographyand diversity of small mammals in centralKenya. Oecologia 116: 381-389.Keith, L.B., A.W. Todd, C.J. Brand, R.S. Adamcik &D.H. Rusch 1977. An analysis of predation duringa cyclic fluctuation of snowshoe hares. Proceedingsof the International Congress of Game Biologists13: 151-175.Korpimäki, E. 1994. Rapid or delayed tracking ofmulti-annual vole cycles by avian predators? Journalof Animal Ecology 63: 619-628.Korpimäki, E. & K. Norrdahl 1989. Predation of Tengmalm’sowls: numerical responses, functionalresponses and dampening impact on populationfluctuations of voles. Oikos 54: 154-164.Korpimäki, E. & K . Norrdahl 1991. Numerical andfunctional responses of kestrels, short-eared owls,and long-eared owls to vole densities. Ecology 72:814-826.Krebs, C.J., & J.H. Myers 1974. Population cycles insmall mammals. Advances in Ecological Research8: 267-399.Küchler, A.W. & I.S. Zonneveld (eds.) 1988. Vegetationmapping. Kluwer Academic Publishers,Dordrecht, the Netherlands.Larsson, T. 1969. Land use and bird fauna on shoremeadows in southern Sweden. Oikos 20: 136-155.Masman, D., S. Daan & C. Dijkstra 1988. Time allocationin the kestrel Falco tinnunculus and the principleof energy minimisation. Journal of AnimalEcology 57: 411-432.McNaughton, S.J. 1979. Grassland-herbivore dynamics.In: A.R.E. Sinclair & M . Norton-Griffiths(eds.). Serengeti; dynamics of an ecosystem: 46-81.Chicago Press, Chicago, USA.McNaughton, S.J. & F.F. Banyikwa 1995. Plant communitiesand herbivory. In: A.R.E. Sinclair & P.Arcese (eds.). Serengeti; dynamics, managementand conservation of an ecosystem: 49-70. ChicagoPress, Chicago, USA.Peles, J.D. & G.W. Barrett 1996. Effects of vegetativecover on the population dynamics of meadowvoles. Journal of Mammology 77: 857-869.Scherfose, V. 1993. Zum einfluss der Beweidung aufdas Gefasspflanzen-Artengefuge von Salz- undBrackmarschen. Zeitschrift für Ökologie undNaturschutz 2: 201-211.Beemster & Vulink / Lutra 2013 56 (1): 5-21 19

Schmidt, N.M., H. Olsen, M. Bildsøe, V. Sluydts & H.Leirs 2005. Effects of grazing intensity on smallmammal population ecology in wet meadows.Basic and Applied Ecology 6: 57-66.Schultz, E. 1992. Waterbeheersing van de Nederlandsedroogmakerijen. PhD thesis. Delft Technical University,the Netherlands / V an Zee tot Land 58.Rijkswaterstaat Directie Flevoland, Lelystad, theNetherlands.Smit, R., J. Bokdam, J. den Ouden, H. Olff, H.Schot-Opschoor & M . Schrijvers 2001. Effects ofintroduction and exclusion of large herbivores onsmall rodent communities. Plant Ecology155: 119-127.Soikkeli, M. & J . Salo 1979. ἀ e bird fauna on abandonedshore pastures. Ornis Fennica 56: 124-132.van der Meijden, R. 1996. Heukels flora van Nederland,22 nd ed. Wolters Noordhof, Groningen, theNetherlands.van Deursen, E.J.M. & H .J. Drost 1990. Defoliationand treading of reed Phragmitis australis. Journalof Applied Ecology 27: 284-297.van Eerden, M.R. 1984. Waterfowl movements inrelation to food stocks. In: P.R. Evans, J.D. Goss-Custard & W.G. Hale (eds.). Coastal waders andwildfowl in winter: 84-100. Cambridge UniversityPress, Cambridge, UK.van Eerden, M.R., B. Slager & L . Soldaat 1997. Maximisationof speed of autumn migration and fatteningrate by greylag geese (Anser anser L.) causeunderuse of natural food supply at a stopover site.In: M.R. van Eerden (ed.). Patchwork: Patch use,habitat exploitation and carrying capacity forwater birds in Dutch freshwater wetlands: 215-238.PhD thesis. University of Groningen, Groningen,the Netherlands / Van Zee tot Land 65. Ministryof Transport, Public Works and Water Management,Directorate IJsselmeergebied, Lelystad, theNetherlands.van Rooij, S.A.M. & H.J. Drost 1996. De Lauwersmeer,25 jaar zoet. Flevobericht 387. Ministerie van Verkeeren Waterstaat, Rijkswaterstaat Directie IJsselmeergebied,Lelystad, the Netherlands.van Wieren, S.E. 1991. ἀ e management of populationsof large mammals. In: I.F. Spellerberg, F.B.Goldsmith & M .G. Morris (eds.). ἀ e scientificmanagement of temperate communities for conservation:103-127. Blackwell Scien tific Publications,Oxford, UK.van Wijngaarden, A. 1957. ἀ e rise and disappearanceof continental vole plague zones in the Netherlands.Verslagen Landbouwkundige Onderzoekingen63: 1-21.Vulink, J.T. & M .R. van Eerden 1998. Hydrologicalconditions and herbivory as key operators for ecosystemdevelopment in Dutch artificial wetlands.In: M.F. Wallis De Vries, J.P. Bakker & S.E. vanWieren (eds.). Grazing and conservation management:212-252. Kluwer Academic Publishers,Dordrecht, the Netherlands.Vulink, J.T., H.J. Drost & L . Jans 2000. ἀ e influenceof different grazing regimes on Phragmites - andshrub vegetaton in the well-drained zone of aeutrophic wetland. Applied Vegetation Science 3(1): 73-80.Vulink, J.T., M.R. van Eerden & R .H. Drent 2010.Abundance of migratory and wintering geese inrelation to vegetation succession in man-madewetlands: the effects of grazing regimes. Ardea 98(3): 319-328.Wheeler, P. 2008. Effects of sheep grazing on abundanceand predators of field vole (Microtus arvalis)in upland Britain. Agriculture, Ecosystems andEnvironment 123: 49-55.SamenvattingDe lange termijn-effecten van begrazingmet landbouwhuisdieren op de talrijkheidvan de veldmuis en de daarvanafhankelijke roofvogels in door de mensgemaakte wetlands in NederlandVerschillende studies hebben het effect vanbegrazing door wilde herbivoren of landbouwhuisdierenop de talrijkheid van kleine zoogdierenbestudeerd; sommige studies gaan ookin op het effect van begrazing op de talrijkheidvan muizenetende roofvogels. In de meestestudies werd gevonden dat de dichtheid vankleine zoogdieren negatief beïnvloed werddoor begrazing, terwijl muizenetende roofvogelsmeestal een numerieke response lieten20 Beemster & Vulink / Lutra 2013 56 (1): 5-21

zien op de talrijkheids-index van de veldmuis(Microtus arvalis). Echter, de meeste studieswaren gebaseerd op gegevens uit een korteonderzoeksperiode (1-4 jaren). Omdat kleinezoogdieren vaak grote aantalsfluctuaties latenzien, is er behoefte aan langjarige studies.In deze studie analyseren we de lange termijn-effecten(3-27 jaren) van begrazing metlandbouwhuisdieren op vegetatieontwikkeling,vegetatiestructuur, talrijkheid van develdmuis en de daarvan afhankelijke muizenetenderoofvogels in recent door de mensgemaakte, grootschalige wetlands in Nederland(Lauwersmeer en Oostvaardersplassen).De bestudeerde wetlands worden gekenmerktdoor een laag niveau van dynamiek,waardoor vroege successiestadia (pioniervegetaties,grazige vegetaties) zonder aanvullendbeheer snel vervangen worden door latere successiestadia(riet, duinriet en w ilgenstruweel).Begrazing werd in de studiegebieden geïntroduceerdtoen de vegetatie nog overwegendbestond uit grazige vegetaties.Onder invloed van een hoge begrazingsdrukontstond een homogeen korte grazige vegetatie.Begrazing met een lage begrazingsdrukleidde tot een heterogene vegetatie van grazigevegetaties en riet. Na een aantal jaren ontstondenechter scherpe grenzen tussen kort afgegraasdegraslanden en g esloten rietvegetaties.Het intermediaire stadium, gekenmerkt doorgematigde riethoogtes (tussen 0,5 en 1 ,5 m)verdween meer en meer.Hoge dichtheden van veldmuizen inbegraasde gebieden waren beperkt tot gebiedsdelenmet gematigde riethoogtes. Gebiedenmet dergelijke riethoogtes bestonden slechtstijdelijk, waardoor de veldmuisdichtheid naenkele jaren afnam.Muizenetende roofvogels lieten een numeriekeresponse zien op de veranderendeveldmuisdichtheid. Behalve veldmuisdichtheidhad ook vegetatiestructuur effect op deroof vogeldichtheid. Maximale roofvogeldichthedenwerden gevonden bij een suboptimaleveldmuisdichtheid, waar de gemiddelderiethoogte wat lager was.De relevantie van begrazing als instrumentvoor het beheer van muizenetende roofvogelsin deze wetlands is afhankelijk van de potentievan herbivorie om latere successiestadia weerom te vormen in vroegere stadia. Het regelmatigevoorkomen van hoge dichtheden vanveldmuizen en d e daarvan afhankelijke predatorenkan bereikt worden door het instellenvan een cyclisch variërende begrazings druk.Jaren met een lage begrazingsdruk zoudenafgewisseld moeten worden met enige jarenmet een hogere begrazingsdruk.Received: 1 November 2010Accepted: 14 May 2013Beemster & Vulink / Lutra 2013 56 (1): 5-21 21

The raccoon dog (Nyctereutes procyonoides)in the Netherlands – its present status anda risk assessmentJaap L. MulderDe Holle Bilt 17, NL-3732 HM De Bilt, the Netherlands, e-mail: muldernatuurlijk@gmail.comAbstract: ἀ e raccoon dog (Nyctereutes procyonoides) was introduced from East Asia into the former USSRbetween 1928 and 1957. Since then it has colonised a large part of Europe and is considered an invasive alien species.An earlier paper (Mulder 2012) reviewed the current knowledge about its ecology. ἀi s paper deals with itspresent status in the Netherlands and provides an assessment of its ecological and human health risks. ἀ e colonisationof the Netherlands by the raccoon dog started from north-west Germany about 15 years ago. ἀ e patternof colonisation is blurred by the occurrence of individuals escaping from captivity. Up until 2013 ‘wild’ raccoondogs were probably recorded exclusively in the north-eastern part of the country. ἀi s is in accordance with thedistribution in Germany. It seems inevitable that the raccoon dog will colonise the whole territory of the Netherlandsin the future, maybe with the exception of the islands in the Wadden Sea. Its general impact on biodiversity isexpected to be small. Isolated populations of amphibians, however, may be at risk, as may ground breeding birds inmarshes. Raccoon dogs may increase the occurrence of diseases and parasites, of which Trichinella spiralis and thesmall fox tapeworm Echinococcus multilocularis probably constitute the most important health risks for humans.ἀ e options for effectively managing raccoon dogs are limited; only local and intensive measures of control or predationprevention may have the desired effect.Keywords: raccoon dog, Nyctereutes procyonoides, wasbeerhond, risk assessment, distribution, colonisation, theNetherlands, invasive species, management.IntroductionIn October 2007 the Dutch government publishedits policy on invasive species (Document20071012-dn-2007-2899.pdf). Accordingto the definition set out in this documentan invasive species is an organism whicharrives from elsewhere with the aid of humans(by transport or infrastructure) and which isa successful coloniser (by reproduction andpopulation growth). In accordance with theagreements in the Convention on BiologicalDiversity (Rio de Janeiro 1992) a s uccession© 2013 Zoogdiervereniging. Lutra articles also on theinternet: http://www.zoogdiervereniging.nlof policies should be applied to control invasivespecies: prevention of their arrival, eradicationwhen their populations are still small, and isolationand control management when populationshave grown too large to eradicate. ἀ eintensity of control measures depends on theimpact the invasive species is expected to haveon biodiversity and human health and safety.In the Netherlands the Invasive Species Team(TIE) of the Ministry of Economic Affairs hasthe task to advise the Minister on all issuesof invasive species. ἀ e TIE collects and publishesinformation, conducts risk analyses forinvasive species and recommends measuresfor the prevention, control and management ofsuch species. ἀ e risk assessment for the rac-Mulder / Lutra 2013 56 (1): 23-43 23

coon dog was published as an extensive report(Mulder 2011) and an earlier paper reviewedthe raccoon dog’s ecology in Europe (Mulder2012). ἀ e present paper deals with the historyand present situation (as of January 2013) of theraccoon dog in the Netherlands, and also containsa concise risk assessment. It ends by discussingthe management options.Two earlier publications have dealt with theraccoon dog in the Netherlands and the possiblerisks it poses (Oerlemans & Koene 2008,van Dijk & de Koning 2009). However, thesepublications were based on a limited selectionof the literature, and did not contain an evaluationand analysis of the raccoon dog observationsin the Netherlands.DistributionTo evaluate the history and present situationof the raccoon dog in the Netherlands, allrecords of raccoon dog sightings until 1 January2013 were collected and screened. ἀ isexercise drew on the available scientific literature,hunting journals and various databases(Alterra, Dutch Mammal Society,,, Yvette van Veldhuijsen (aprivate individual with a keen interest in raccoondogs), the Royal Dutch Hunters Societyand the AAP Foundation, a rescue centre andsanctuary for primates and other exotic animals).Efforts were made to collect previouslyunreported observations, through appeals inhunting journals and on the internet. Many ofthe original observers were contacted by telephoneor email and questioned. Obser verswere asked for their experience with wildlifein general, and with red foxes (Vulpes vulpes)and badgers (Meles meles) in particular. Detailsof the way of walking and other behaviour ofthe observed animal were asked for, as well asthe colours of the pelt and the relative length oflegs and tail. Most observations by hunters andnaturalists could be accepted. Many observationsby less experienced people were, however,too vague or incomplete to accept as certainor probable. If observers mentioned havingthought of a raccoon (Procyon lotor) when theyobserved the animal, this was taken as a positivesign. ἀ e vast majority of the records couldthus be validated in four categories: ‘certain’,‘probable’, ‘possible’ and ‘not likely’. All recordswhich could not be verified with additionalinformation, were placed in the ‘possible’ category.ἀ e resulting database has been submittedto the National Authority for Nature Data.Of course the validation of others’ observa-Raccoon dogs deposit their faeces in concentrated latrines along their routes. Photo: J.L.Mulder.24 Mulder / Lutra 2013 56 (1): 23-43

Figure 1. Tracks of raccoon dog (left) and red fox, theformer with more rounded prints and with the left andright feet spaced further apart.tions is ridden with the subjectivity of the validator.It is impossible to use objective criteria,and the dividing line between the categoriescannot be clearly defined. Dead animals poseno problem for identification. Sometimesphotographs were made of the animal or thetracks it left in mud or snow, making validationeasy for an expert. Footprints were onlyaccepted as proof if they were clearly roundand not elongated (as in the red fox), and ifthe prints of left and right feet were spacedapart instead of placed in almost one line (figure1). Pictures of prints showing the connectionbetween the caudal part of the two centraltoes were taken as definitive proof (figure2). ἀ is feature is only visible in sharp printsin mud or clay. Less sharp prints are similarto those of small domestic dogs, and can onlybe taken as produced by a raccoon dog if thepresence of a domestic dog can be absolutelyexcluded. Until 2013 there were, however, noFigure 2. Front paw of the raccoon dog, showing theconnection between the inner toes. Photo: Annemarievan Diepenbeek.records based exclusively on prints or othersigns (only in connection with an observationof an animal), except for one winter-recordof a den with a latrine at five metres distance;this observation was judged ‘probable’.Escaping raccoon dogsἀ e first raccoon dog in the Netherlands wasobserved in April 1981 in the south-east ofthe country, in the Province of Limburg (Vergoossen& Backbier 1993). ἀ e animal did notseem to be very shy, and may well have escapedfrom captivity or been deliberately set free.ἀ at is all the more likely since the first raccoondogs in north-eastern Germany (at a distanceof 600 km from Limburg) were recordedonly 15 years earlier, and it was another tenyears before the next raccoon dog was recordedin the Netherlands. ἀ is second animal wasMulder / Lutra 2013 56 (1): 23-43 25

Table 1. Known cases of escaping raccoon dogs. Sources: Yvette van Veldhuijsen (third and fourth case) and ownresearch.Date Place Number, sex, etc. How escaped What happened nextca 1997Speuld, Veluwe,children‘s farmHalfgrown male andfemaleUnknownRecovered from a fire wood shedin the next village, Putten (5 kmaway). After some weeks spentwith a private individual they werereturned to the children’s farm.2001 Gangelt (D), ZooHochwild Freigehege,on the Dutch/Adult male and female Unknown Male was killed on an adjacentroad on the night of the escapeGerman border, eastof Sittard.August 2002September2002Private house inEnschedeYearling male and female, Over a garden fencebrother and sister with horizontalwooden beamsPrivate house in Two yearling animalsIngen, Betuwe (Gld)ca 10 June2011Adult male and female,female pregnantWhite sterilized adultApril 2010 Children’s farmDondertman inEspelo, Holten,OverijsselAutumn 2010 It Schildhus AnimalRescue Asylum, female and normallyGoengarijp, Friesland(specialising in male. Had been previ-coloured neutered adultturtles)ously kept by private personin Utrecht province.Private person in Two males born in 2010.Egmond-Binnen, ἀ e third raccoon dogNoord-Holland who present did not escape.had already beenkeeping raccoondogs for ten yearsOne via a small tablethrough an openwindow. ἀ e othersome days later bybiting through meshwireChildren left thedoor of the pen openWere kept in a stableOver a fence, via thecollapsed roof of adog houseUnknown; no road kills shortlythereafter, but one in autumn 2003On 12 October 2002 a youngfemale was found as a roadkill nearWijk bij Duurstede. 9 km away onthe other side of the river Rhine.During the autopsy it was suspectedof being an escapee.Male was killed on the road within800 m and within a few days, on 26April 2010.ἀ e white female was shot on13 November, 5 km away. ἀ emale was probably spotted on 11November, 2-3 km awayOne was killed on the adjacentroad 800 m away on 17 June 2011.ἀ e other was seen on 20-21August in Heiloo, around 3 kmaway, in a chicken coopfound dead by a road east of the town of Groningen(in the north-eastern corner of theNetherlands) in February 1991, and may havebeen the first ‘wild’ raccoon dog in the Netherlands(Mulder & Broekhuizen 1992).Raccoon dogs are sometimes kept as pets,indoors as a d omestic dog or outdoors in akennel or behind fencing. ἀ ere is no commercialfarming of raccoon dogs in the Netherlands.ἀ ey are rather easy to keep, eat virtuallyanything, hardly make any noise anddeposit their faeces in the same spot. Advertisementsoffering young or adult raccoon dogsare not uncommon in animal journals or onthe internet. It is not known how many peoplekeep raccoon dogs in the Netherlands, but itmay be in the order of fifty or more. Although aone metre high fence is enough to contain raccoondogs (Stier 2006), it appears that escapinganimals are not rare, mostly through thenegligence of the owners. Seven such caseshave been documented (table 1), all regard-26 Mulder / Lutra 2013 56 (1): 23-43

ing two animals each. In four of these sevencases a road killed raccoon dog was found inthe vicinity shortly afterwards, and in one casean animal was shot by a hunter. ἀ is last animalcertainly was the escapee in question, for itwas a white sterilised female. Five of the recoveredescapees were found within 0.5 to 9 k mand within one month of the time of escape;one was killed on the road on the same nightit escaped.ἀ ese escaped raccoon dogs may blur thepattern of colonisation in the Netherlands. Isolatedobservations, far away from the majorityof the other observations, may indicate suchescapees. Proof, however, is rarely available.Sometimes it is obvious from the behaviour ofa raccoon dog that it is an escapee: such individualsshow no fear of people: it is possible tocome within a few metres, or even less, of them.When such behaviour had been reported, or aconnection with a known escape was clear, therecord was classified as an ‘escapee’.ObservationsAfter validation of all the observations of raccoondogs until 1 January 2013 the databasecontained 173 records (excluding the category‘not likely’), of which 77 were certain, 43 wereprobable and 53 were possible. In the threecategories a total of 11, 6 and 1 records respectivelywere obvious escapees. Figure 3 showsmaps of the Netherlands with all these raccoondog observations, separated into two periods.ἀ e majority of observations were made inthe north-eastern half of the country, in theProvinces of Groningen, Drenthe, Fries landand Overijssel, and in the Noordoostpolder,a distribution which accords very well withthe species’ distribution in Germany (figure4). Most of the records outside this area (indicatedwith a dashed line in figure 3) could beescaped animals. ἀ e Veluwe area (shownas an oval in figure 3), which is far from theGerman border, is interesting in this respect.Raccoon dogs started to be reported from theVeluwe relatively early (1993) and have continuedto be sighted. From their behaviour,at least three of the observed animals wereclearly escaped animals. In addition, severalpeople in the region were known to have raccoondogs as pets. I therefore consider all theraccoon dog records from the Veluwe (untilnow) as observations of escaped animals.Focussing on the north-eastern corner ofthe country only (with 55 certain, 24 probableand 24 possible observations) and excludingthe clear escapees, the pattern of raccoondog observations over time becomes clear(figure 5). After three isolated observations inthe early 1990s, there have been continuousobservations of raccoon dogs every year since2001. In north-east Germany the first raccoondogs were seen in 1964. ἀ at means that theraccoon dog has crossed the distance to theNetherlands in 37 years, at an average speedof 13 km/year. Since 2001 roughly 3 to 8 raccoondog observations have been recorded inthe Netherlands each year, with a maximumof 11-15 in 2006. Most probably however,many raccoon dogs go undetected. ἀ e Netherlandsapparently has reached the stage ofirregular, sporadic but steady observations ofraccoon dogs. In other countries this periodlasted 20-30 years before the populationstarted to grow exponentially (Mulder 2012).ἀ is is in keeping with the apparent lack ofreproduction in the Netherlands so far. Withtheir preference for badger setts the pups ofraccoon dogs or their signs (latrine) are likelyto be easily detected by badger watchers andhunters; up till now no such observations havebeen reported. Based on the temporal patternof raccoon dog colonisation in other countries,the exponential growth period may beexpected roughly to begin around the year2025. By 2035 the raccoon dog may be a commoninhabitant of the Netherlands.At present it is unclear how the (probablyrather lonely) raccoon dogs behave. Some animalsapparently settle for a while in a limitedarea and are spotted several times. In 2003a raccoon dog seems to have spent an entireMulder / Lutra 2013 56 (1): 23-43 27

A28 Mulder / Lutra 2013 56 (1): 23-43

BFigure 3. All records of raccoon dogs in the Netherlands. Left: until 1 January 2006. Right: between 1 January 2006and 1 January 2013. The observations from within the oval (Veluwe) are all considered escapees from captivity.The records north-east of the dashed line are mostly considered to be natural colonisers and are used as the basisof figure 5. The records south-west of the dashed line probably all relate to escaped animals, at least before 2006.Black dot: certain; black triangle: probable; open triangle: possible; cross: certainly or most probably an escapedanimal.Mulder / Lutra 2013 56 (1): 23-43 29

20062011Figure 4. Distribution of the raccoon dog in G ermany, by municipality, according to a repeated inquiry amonghunters, by the German project WILD. Black: at least one observation. Grey: no observations. White: no data.Upper map: 2006. Lower map: 2011 (adapted from: Mulder / Lutra 2013 56 (1): 23-43

201612possibleprobablecertain8401990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012Figure 5. All observations of the raccoon dog in the north-eastern part of the Netherlands (see figure 3) since 1990,grouped in three classes of certainty. Apparent escapees have been excluded.summer at a t ree nursery near Steggerda,Province of Friesland; it was last spotted inNovember. Another example is a raccoon dogwhich was seen on 29 May 2011 in Saeftinghe,in tidal salt marshes in the south-west of theNetherlands, far from any other raccoon dogobservation. Its tracks in the mud were foundat least eight times throughout the summer,for the last time on 25 August (see photo onpage ...). Despite many visits to the area, notracks were found thereafter (personal communication,M. Buise). Other raccoon dogsprobably roam the country. On a permanentfeeding station (under video surveillance)for wild boar (Sus scrofa), badgers and foxesin the Veluwe area, a raccoon dog showed uponly once, in December, which suggests that ithad not settled there.Considering the same raccoon dog observationsas in figure 5 (for the north-eastern partof the country, excluding escapees) but takinginto account only the ‘certain’ and ‘probable’categories (n=79), the methods of observationwere as follows: observed in the field(n=38), killed on the road (n=24), killed by aharvesting machine (n=7), shot (n=4), caughtalive and killed (n=2) and drowned (n=1). Inthree cases the details of observation werenot exactly recorded. Crops in which raccoondogs were killed, were beets (n=4) andmaize (n=3). Observations of live animals inthe field were often made in the headlightsof a c ar (n=11), during lamping (fox shootingwith a spotlight, n=6) or during daylighthunting (n=4). Once (and twice outside thenorth-eastern part of the country) a raccoondog was captured on photo by an automaticwildlife camera. Two dead raccoon dogs(one traffic victim, one drowned) appearedto be juvenile; they were collected on 12 and30 August, in Stiens (Province of Friesland)and Delden (Province of Overijssel) respectively.ἀ ey most probably are examples ofearly dispersers, from Germany (see Mulder2012). Alternatively, they may have been bornin the Netherlands; however, if reproductionwas already occurring here and there, morejuvenile traffic victims, and observations ofpups should be expected (see above). Of thefew animals which were sexed, ten were malesand two were females. Two raccoon dogs wereobserved together on four occasions (in theyears 2001, 2007, 2011 and 2012); the first pairFootprints of the raccoon dog o bserved in t he saltmarshes of Saeftinghe, summer of 2011. ἀ e pencil is10 cm long. ἀ e connection between the two front toesis clearly visible. Photo: Stefaan Thiers.Mulder / Lutra 2013 56 (1): 23-43 31

Many of the reliable records of raccoon dogs are traffic victims. Photo: J.L.Mulder.were found as traffic victims on consecutivenights, only 50 m apart, the other three werefield observations of two animals close to eachother.In conclusion, the raccoon dog is presentlyregularly but sporadically observed inthe north-eastern half of the Netherlands, inaccordance with its distribution in Germany.Every year between about three and eight certainor probable observations are reported.Elsewhere it is still very rare, and most of theraccoon dogs recorded in the south-westernhalf of the country probably escaped (or weredeliberately set free) from captivity.Risk assessment of the raccoondog in the NetherlandsProbability of arrival and establishmentἀ e analysis of the raccoon dog observationsin the Netherlands shows that the firstobservation of a raccoon dog, which probablydispersed from Germany, occurred in 1991.Considering the present colonisation speedof the raccoon dog in Germany, the lack ofpotential barriers (i.e. mountain ranges) andthe presence of suitable habitats in the Netherlands,it is likely that the raccoon dog willcontinue with its westward colonisation bydispersal and eventually become establishedin the whole of the Netherlands. Raccoondogs have a h igh dispersal and reproductionpotential which, in the Netherlands, willnot be adversely affected by predators or diseases(see below). ἀ e climate matches thatof regions where it has already successfullyinvaded large areas (for instance the northeastof Germany, with an average annual temperatureof 9.7 °C, ranging from a m ean of0.8° C in January to a mean of 18.2 °C in July),so the raccoon dog has the potential to be atruly invasive alien species.Almost all of the Dutch territory is suitable,or even very suitable, for raccoon dogs. Onlythe built-up areas will probably not be occupiedby the species. Research from elsewherein Europe indicates that resident raccoon dogs32 Mulder / Lutra 2013 56 (1): 23-43

avoid settlements and have not been observedto cross villages, even when these were locatedin the middle of their home range (Drygala etal. 2008). ἀ e numerous wetlands in the lowlyingnorth and west of the Netherlands, aswell as wet areas in the rest of the country,will probably harbour a dense population ofraccoon dogs in the future. Areas of higherground, that have badger populations andadjoin marshy areas with lots of dense cover(reeds, willow and alder thickets) will probablyconstitute the optimal habitat, with a maximumof two - ² locally in the mostvaried landscapes (Kauhala et al. 2006). Lessfavoured habitats will be large predominantlypine and fir forests and plantations (such asthe Veluwe) and large scale agricultural areas(such as most of the Flevopolders). ἀ e presenceof enough cover, in the form of densevegetation, is very important for the raccoondog. Average pre-breeding population densityover large suitable areas will probably bebetween 0.5 and 1.0 - ².Impact on native predatorsἀ e raccoon dog is an omnivorous, mediumsizedpredator whose ecology has sharedaspects with several native and non-nativepredators: badger, red fox, polecat (Mustelaputorius) and American mink (Mustelavison). Since the American mink is a n onnativespecies itself, and nothing is knownabout its relations with the raccoon dog, thespecies of interest here are badger, red fox andpolecat. Impacts can occur as a result of interferencecompetition or resource competition(Pianka 1978). Raccoon dogs rarely seem todirectly interfere with badgers (Mulder 2012).Raccoon dogs may kill badger cubs (Sidorovich2011) but the reverse, badgers killingraccoon dog pups, may be more common(Kowalczyk et al. 2008). Once an adult raccoondog was observed which had died ofwounds on its back most probably inflicted bya badger (Drygala 2009). Indirect interferenceof raccoon dogs with badgers is suspected tobe contributing to the decline of the badgerpopulation in central Belarus. Several deadbadgers were found in their setts during winter,which apparently had died of suffocation:raccoon dogs often hibernate in badger settsand block almost all the entrances (Sidorovich2011). In the mild winters in the Netherlandsthis will be less of a problem. Interferencewith foxes probably is much morecommon and both species occasionally killcubs of the other species. In north-east Germanyfox numbers (measured as the numberof foxes shot annually) decreased in the firstperiod after the arrival of the raccoon dog,but this effect disappeared later (Zoller 2006).An increase in the infection rate of sarcopticmange (see below), which is more common inraccoon dogs, might have been responsiblefor this temporary decline in the fox population.Drygala (2009) concludes that in Europecompetition between raccoon dog, red foxand badger might take place, but that it isunlikely that the competition is very severeor leads to a significant decrease of either species.In northern Belarus, an area with severewinters, the strong increase of the raccoon dogpopulation coincided with a strong decrease inpolecat numbers in two study areas, and witha decrease in pine marten (Martes martes),red fox and brown bear (Ursus arctos) in oneof the two study areas. ἀ e mechanism behindthe impact of raccoon dogs on other generalistpredators is thought to be the effective exploitationof available carrion by raccoon dogs inearly winter, resulting in a lack of food for theother predators in late winter; a classic exampleof resource competition (Sidorovich et al.2000). In the Netherlands such a competitionover carrion seems unlikely, except perhapswith the raven (Corvus corax). However, someresource competition between raccoon dogand polecat might be feasible, especially withregard to amphibians. Both polecat and raccoondog, although generalist predators, havea clear preference for amphibians.Mulder / Lutra 2013 56 (1): 23-43 33

Impact on prey speciesRaccoon dogs forage while slowly walking,mostly in dense vegetation. ἀ ey do not ‘hunt’like foxes, chasing their prey species. Uponencountering bird nests, they will eat the eggsand chicks, but only rarely the adult breedingbird (Mulder 2012). However, remains ofeggs in raccoon dog stomachs are rare in dietstudies. According to most authors, the addedimpact (on top of the impact of native predatorssuch as the red fox) of the raccoon dog onthe breeding success of ground nesting birdswill probably be negligible. However, solidresearch into the impact of the raccoon dogon its prey species is still lacking. ἀ e predictionsin this section are thus mainly based onexpert judgment.Bird colonies in wetlands (e.g., greylag goose(Anser anser), black-headed gull (Croicocephalusridibundus)) might be especially vulnerableto raccoon dog predation; as a raccoon dog canpossibly destroy many nests in a short time. Forthe Netherlands, with its many low-lying wetlandareas, including many Natura 2000 areas,the most vulnerable species will probably bethe purple heron (Ardea purpurea), the blacktern (Chlidonias niger) and the solitary bittern(Botaurus stellaris). Although the red fox hasalready (in recent decades) arrived in most ofthese wetlands, the raccoon dog may pose anadded threat because of its greater readiness toswim. ἀ e species mentioned above are possiblyat risk and measures to prevent predationby raccoon dogs may be necessary in thefuture.ἀ e raccoon dog’s preference for amphibiansmay lead to local declines of more or lessisolated populations of toads, frogs and possiblynewts, for instance in and around cattledrinking ponds in the dryer east and south ofthe Netherlands. Raccoon dogs also forage ongrass snakes (Natrix natrix) (Drygala 2009),and might be a threat to isolated populationsof this species as well. ἀ e common practiceof protecting amphibians from being killedon the road in spring, by erecting fences andcatching the animals in buckets during thenight, may in the future attract the unwelcomeattention of raccoon dogs, seeking to gather aneasy meal from the buckets (Puffpaff 2008).Viruses and parasitesἀ e raccoon dog can play a role in the transmissionof several infectious diseases, includingparasitic diseases, to other species and/orto humans. As far as is known there are nonew viruses or parasites imported by raccoondogs to Europe, however, they may act as areservoir for several pathogens already presentin Europe. Rabies, caused by the classicrabies virus, is an important disease in canids.Until the introduction of the raccoon dog, thered fox was by far the main reservoir and vectorof rabies in Europe. ἀ at situation has nowchanged. In Estonia about 50% of wildliferabies cases were raccoon dogs (WHO 2004as cited in Kauhala et al. 2007). During therabies epizootic in Finland in the late 1980s,and later in Poland and the Baltic states, theraccoon dog was the main vector and victimof rabies, accounting for 73% of all reportedrabies cases (Westerling 1991, Mól 2005, Kowalczyk2007, Zienius et al. 2007). Many countriesin Western Europe are free of rabies asa result of oral vaccination campaigns. Currentlythe front of rabies and rabies vaccinationcampaigns runs from eastern Polandto Croatia (figure 6). In the new situation inEurope, with two main rabies vectors, thevaccination campaigns may not be as effectiveas before; rabies might persist in the animalcommunity (due to incubation times of upto several months), even if the disease is notspreading in an individual vector species as aresult of its low density (Holmala & Kauhala2006, Kauhala & Kowalczyk 2011). However,Poland has in fact been largely freed fromrabies as a r esult of vaccination campaignsbetween 2000 and 2010, when raccoon dogswere already as common as foxes. Apparentlythe oral vaccination campaigns there were as34 Mulder / Lutra 2013 56 (1): 23-43

effective for raccoon dogs and foxes together,as they were for foxes alone.As in other canids, canine distemper causedby canine distemper virus (a paramyxo-virus),an important disease in domestic dogs, hasbeen observed in the raccoon dog. Recentlyit has become rather widespread in easternGermany (N. Stier, personal communication).An outbreak of canine distemper in 1991 nearTokyo killed about 70% of the local raccoondog population (Machida et al. 1993). Caninedistemper virus is not known to be zoonoticand thus is harmless for humans, but domesticdogs are susceptible; in fact, wild carnivoresaround the world may be more likely tobe infected by non-vaccinated domestic dogs,than vice versa (ἀ ree parasites are of importance in raccoondogs: the roundworm Trichinelle spiralis,the tapeworm Echinococcus multilocularisand Sarcoptic scabie var. vulpesi, a mite causingSarcoptic mange. In Finland 53 to 72%of the examined raccoon dogs were infectedwith Trichinella sp. (Mikkonen et al. 1995)and in eastern Germany 5.8% (ἀ iess 2004).In Finland an association between the densityof raccoon dogs and the incidence of infectionwith Trichinella larvae in the European lynx(Lynx lynx) has been demonstrated (Oksanenet al. 1998). ἀ e role of the raccoon dog as areservoir of Trichinella sp. seems remarkable:where raccoon dogs are common (Finland,Estonia) the prevalence of Trichinella in foxesis much higher than elsewhere (Oksanen etal. 1998, Oivanen et al. 2002). ἀ e percentageof infected wild boar in north-east Germanyalso increased in line with the numberof raccoon dogs shot each year (Pannwitz etal. 2010). Trichinellosis is also a human disease.However, as a result of the control measuresin the meat industry, it is rare in WesternEurope (J. van der Giessen, personal communication).Trichinellosis is caused by nematodesof the genus Trichinella and the diseaseresults from eating raw or undercooked meatfrom infected domestic pigs (mostly if theyFigure 6. A ll cases of sylvatic rabies (excluding batsand domestic animals) in Europe in 2012. M ap generatedwith access to outdoor pens) or game animals.Several carnivorous or omnivorous wildlifespecies (red fox, raccoon dogs, wild boar, rats,other rodents) can be carriers of Trichinella,and if pigs eat their bodies, they becomeinfected. In pigs, clinical symptoms are veryrare, but in humans clinical symptoms canoccur and are dependent on the dose of Trichinellalarvae ingested. In worst-case scenarios,the central nervous system and the myocardmay be affected, with potentially fatal consequences.ἀ e most important parasite in raccoondogs (and foxes) is the small fox tapewormEchinococcus multilocularis. ἀ is can cause asevere infection in humans, and if not diagnosedand treated properly the infection maylead to death. ἀ e parasite is very small (1.2to 3.7 mm; Faust & R ussell 1964) and hasno effect on the carnivore carrier, even if ithas a h igh burden of tapeworms. Eggs areexcreted with the faeces, thus contaminatingthe vegetation. Rodents, especially voles, eatthe grass and act as secondary hosts. Whenfoxes or other predators consume an infectedvole, the life cycle of the parasite is closed. Ifhumans ingest eggs of this parasite, the larvalstage develops in internal organs, primarilyin the liver. ἀ e incubation time is long, 5-15years, and the route of transmission is as yetMulder / Lutra 2013 56 (1): 23-43 35

unknown. Echinococcus eggs are very smalland may be ingested via water, contaminatedfood or direct contact with infected animals,such as a domestic dog.In Poland and eastern Germany severalstudies on E. multilocularis in raccoon dogshave been conducted (Machnika-Rowinskaet al. 2002, Tackmann et al. 2003, ἀ iess2004, Schwarz et al. 2011). ἀ e percentage ofinfected animals ranged from 0 to 10.7%, butthe studies have been limited in scope andnumber, so far.Intestinal parasites such as nematodes, cestodesand trematodes are common in raccoondogs, as in foxes, and are relatively harmless.Barbu (1972) found an emaciated raccoondog at the end of spring, with the exceptionalnumber of 1700 trematodes in its intestines.Sarcoptic mange (Sarcoptes scabie var.vulpesi, a mite) is rather common in raccoondogs (Stier 2006). ἀ e mite lives in the skin,causing bare patches, severe itching and, ifthe infection is severe, may cause death fromexposure. Sarcoptic mange sometimes comesin ‘waves’, decimating the populations of redfoxes (Lindström & Mörner 1985) and probablyalso of raccoon dogs. At other times it ispermanently present in low intensities in fox,raccoon dog and other wildlife populations,claiming few victims. Raccoon dogs seem tobe more frequently infected with mange thanfoxes (N. Stier, personal communication). Atpresent mange is rare among red foxes in theNetherlands, and the arrival of the raccoondog may boost its occurrence (cf. Stier 2006),especially in the fox population, but other speciesmay be affected as well. Humans may beinfected by the Sarcoptes scabei mite of dogs,and also of foxes, although the mite is not ableto finish its life cycle in humans.Human safety and health risksUp till now there are no records of raccoondogs being aggressive towards people. ἀ eyavoid contact and when cornered keep quietand can be grabbed easily. However, raccoondogs can be carriers of diseases and parasitesthat are harmful to people, i.e. the rabies virusand the fox tapeworm Echinococcus multilocularis.Kauhala & Kowalczyk (2011) considerthis to be the most severe risk of the raccoondog’s colonisation of Europe. However, rabieshas been eradicated in Western Europe and itis highly unlikely that it will return with thearrival of the raccoon dog (see above). ἀ erisk of rabies in Western Europe now mainlycomes from imported pets (dogs and cats) andfrom a few bat species (Lina & Hutson 2006).ἀ is said, the rabies control measures inEurope should be scrutinised and, where necessary,reviewed in order to remain successful(Kauhala & Kowalczyk 2011).ἀ e most important health risk constitutesthe small fox tapeworm Echinococcus multilocularis,of which the raccoon dog is a c arrier.Before the arrival of the raccoon dog, thered fox was the only vector of Echinococcus.It is unknown whether the raccoon dog willshow the same prevalence of E. multilocularisin the future as the red fox does now. Inthe Netherlands the distribution of the smallfox tapeworm is restricted to two areas, thenorth-east and the south-east corners of thecountry (Giessen et al. 2004a, Giessen et al.2004b, Opsteegh et al. 2013). It is expectedthat the distribution of E. multilocularis willslowly expand, as a result of the mobility andespecially the dispersal of the red fox. Sincethe indications are that average dispersal distancesof raccoon dogs are larger than those offoxes, it is to be expected that the distributionarea of E. multilocularis will expand slightlyfaster than with the fox as the sole vector.Since raccoon dogs will live alongside theexisting fox population, the density of potentialE. multilocularis-carriers will increaseand might double in the future. ἀ is impliesthat the infection risk in endemic areas mayincrease as well. Red foxes spread their faecesdiffusely over their whole territory, whileraccoon dogs defecate in just a f ew latrines.Raccoon dogs thus contaminate only a few36 Mulder / Lutra 2013 56 (1): 23-43

confined areas of the environment with E.multilocularis and pose a, potentially lowerrisk for human infection than foxes whichcontaminate a w ider environment. ἀ e onlyknown method to combat E. multilocularis,and diminish the risk for humans, is the(costly) permanent application of anthelminthicbaiting, targeting foxes and raccoon dogs.ἀ is involves distributing baits containing theworm-killer Praziquantel in the field (Hegglin& Deplazes 2008).Other risksἀ e raccoon dog is quite an isolated species inthe canid family (Mulder 2012), and hybridisationwith other dog species is unknown,even in captivity. ἀ ere is, therefore, no risk ofgenetic effects on native species. It is unlikelythat raccoon dogs will have a substantialimpact, directly or indirectly, on ecosystemsas a whole, e.g. by disrupting the existing foodwebs, maybe with the exception of some smallisland situations (Mulder 2012). To date, thereis no record of raccoon dogs having an economicor social impact in Europe. ἀ ey areshy and clumsy and avoid the vicinity of peopleand their infrastructures. Direct damageto property is not known, nor is there competitionwith economically important animals.ἀ ey do not climb and do not normally predateon pets or poultry. Raccoon dogs mighthave some economical impact (although thereare no data about this as yet) by eating fromcommercial crops of low hanging fruit (strawberries,blueberries, blackberries etc.) andmaize.Overall assessmentSeveral methods have been developed to makea numerical risk assessment for invasive species.Branquart (2007) and colleagues deviseda simplified hazard assessment for ecologicalimpacts: the Invasive Species EnvironmentalImpact Assessment protocol (ISEIA).ἀ e ISEIA protocol is originally designedfor species already established somewhere inEurope, and is therefore the most appropriatein the case of the raccoon dog. In this assessmenta number of aspects receive a score of 1,2 or 3 (for low, medium and high risk respectively)and the sum of scores leads to a classificationin one of three categories: A. the‘black list’, i.e. high environmental risk; B.the watch list’, i.e. a m oderate environmentalrisk and; C. species that are not considereda threat for native biodiversity and ecosystems.Table 2 shows how the raccoon dogscores according to this protocol. Accordingto this protocol, the raccoon dog receives ascore of 9 (more details in Mulder 2011), andfalls in category B, representing a ‘moderateenvironmental risk’. ἀ e main factors responsiblefor this score are its high dispersion andcolonisation potential. It should be noted thatthe ISEIA protocol does not take the humanhealth aspect into account.ManagementIn many countries the year round killing ofraccoon dogs is permitted (e.g., Sweden, Norway,Estonia, Latvia, Lithuania, Hungaryand Poland). However, in Finland, femaleswith pups are protected in May, June andJuly, and in Belarus hunting is allowed from1 October to the end of February (Kowalzcyk2007, Kauhala & S aeki 2008). In Denmarkhunting is not allowed unless there is anegative impact on game animals (Kowalczyk2007). In Germany the different federalregions (Bundesländer) have different rules;in most of them raccoon dogs can be huntedyear round. ἀ e exceptions are: Niedersachsenand Nordrhein-Westfalen, where adultraccoon dogs are protected from March toAugust, and Schleswig-Holstein where adultraccoon dogs are protected from March toJune; Hamburg, where adult and juvenile raccoondogs are protected from May to AugustMulder / Lutra 2013 56 (1): 23-43 37

Table 2. Scoring the ecological risks of the raccoon dog, according to the ISEIA-protocol.Aspect Sub-aspect Risk Score MaximumscoreDispersion potential High 3 3Colonisation of high value conservation habitat High 3 3Adverse impacts on native species Predation Medium 2 2Competition Low 1Disease Low 1Genetic interaction Low 1Alteration of ecosystem function Nutrient cycling Low 1 1Physical alteration Low 1Natural succession Low 1Food web Low 1Total score 9and Bremen and Saarland, where raccoondogs are protected year round (; viewed April 2013).Kowalczyk (2007) discussed the huntingintensity in Europe: “In Finland, the annualhunting bag varied between 75,000-130,000 in1998-2003 (Kauhala & S aeki 2004, Kauhala,personal communication), ca. 20,000 in Germany(S. Schwarz, personal communication),6,000-10,000 in Poland (data of Research Stationof Polish Hunting Society in Czempiń),4,000-5,000 in Estonia, 3,500-4,000 in Lithuania(L. Baltrūnaitė, personal communication),and 2,000 in Latvia. In other countriesraccoon dogs are hunted occasionally. (…)Locally, intensive trapping with box and wiretraps and hunting with dogs may be methodsof raccoon dog eradication. Eradication is,however, difficult, because raccoon dogs, likeother canids, tend to increase their litter sizewhen hunting pressure on them is high.”Figure 2 i n Mulder (2012) shows the bagrecord of the raccoon dog between 1994 and2011 for the whole of Germany. It peeked inthe hunting season 2007/08 with more than35,000 raccoon dogs shot. Since then theincrease in numbers seems to have stopped,most probably due to epizootics of canine distemperand sarcoptic mange in the north-eastof the country. In the mean time the expansionto western parts of Germany is still continuing.In Germany most raccoon dogs are beingshot more or less opportunistically, by hunterswaiting near a feeding ground for wild boar orroe deer (Capreolus capreolus). Raccoon dogsare also attracted to the bait in feeding pits forfoxes, and are shot there as well. Some huntersdo target raccoon dogs and use dogs to cornerthem or to find them in burrows, from whichthey are dug out. Many raccoon dogs are shotin autumn, when they flee from maize fieldsduring the harvest. When fishponds are beingdrained for the harvest, they work as magnetsfor raccoon dogs (and their hunters). Aswith the fox, box and wire cage traps work forjuvenile animals but are ineffective for adults(Stier & Joisten 2006).Stier (2006) argues that the high huntingbag in eastern Germany may look impressive,but that a two- to three-fold intensification ofthe hunting pressure would be needed to startreducing the breeding population of the raccoondog. Rabies, one major natural cause ofdeath has recently disappeared, increasing theraccoon dog’s expansion potential; however,another cause is slowly returning: the wolf.In his calculations Stier assumes (rather conservatively)a 300% annual potential increase(2 adults getting 6 pups), and a mortality byhunting of 50% and by other causes (traffic,diseases, old age, etc) of 50%. However, a twotothree-fold intensification of hunting is not38 Mulder / Lutra 2013 56 (1): 23-43

feasible and not realistic, and in view of thefavourable circumstances in habitat and foodavailability might still have a l imited effect.Stier (2006) advises to address specific possibleproblems, such as predation of bird colonies,on a l ocal scale. Very intensive controlbetween November (end of dispersal) andApril (start of reproduction) is needed, eachyear, to reduce the spring population in a targetarea (which should not be too large) inorder to achieve the desired results.One interesting example may illustrate theeffect of shooting raccoon dogs. In southernFinland raccoon dogs and other mediumsizedpredators (red fox, pine marten, Americanmink) were killed by shooting and capturingin a ‘removal area’ of 55 km², in orderto study the effects of these predators on thebreeding success of ducks. In a similar controlarea (48 km²) the numbers of medium-sizedpredators were not controlled. ἀ e experimentlasted five years. Raccoon dogs werekilled between 1 August to 31 April each yearby volunteer hunters. An index of raccoondog density was obtained using 50 scent stationseach spring. A total of 280 raccoon dogswere killed, i.e. 0.73 to 1.36 individuals/km²each year. Notwithstanding this effort, no significantdecrease in the raccoon dog indexwas observed. One reason behind this maybe that most raccoon dogs were killed in theautumn and were juveniles that would havedied anyway (Kauhala 2004). It is clearly noteasy to substantially reduce a p opulation ofraccoon dogs.Options for future managementIn all or most of the European countrieswhere raccoon dogs are living today, theyhave been hunted from the moment of theirarrival. Despite the high numbers of raccoondogs killed by hunters, their expansionhas not been halted anywhere and there areno indications that the population densityhas decreased as a result of hunting. To limitpopulation increase it is necessary to annuallyremove at least the numbers producedeach year in excess of the annual mortality.Despite the large hunting bag this seems to befar from the case in areas where the raccoondog has become common now (Stier 2006).ἀ e ‘usual’ shooting of raccoon dogs will, atthe most, decrease their expansion rate a little.Preventing the raccoon dog from establishingitself in the Netherlands, if possible atall, would at least require an effort and professionalorganisation similar to that establishedfor controlling muskrat (Ondatra zibethicus)and coypu (Myocastor coypus) (Broekhuizen2007).At the moment (February 2013) in theNetherlands the raccoon dog is placed in acategory of species that may be controlledaccording to Clause 67.1 of the Bill on Floraand Fauna (Appendix 1 of the Regulation ofManagement and Damage Control). ἀ is listcontains mostly non-native and feral species,but also some species which may be harmfulto crops or to flora and fauna. However, firearmscan only be used to control these specieswith the permission of the province. Suchpermission has only been granted in the Provinceof Friesland.All this means that two realistic managementoptions remain:A. Intensive hunting on a local scale (a fewkm²) in places where problems (might)arise, during the months with no dispersaland no reproduction (December –March).B. Prevention of predation by blockingaccess for raccoon dogs, for instance by(electric) fencing of colonies of breedingbirds or ponds with rare species ofamphibians. ἀ is method has provedto be effective in the case of the fox (J.L.Mulder, unpublished data).Acknowledgements: First of all I a m grateful to theTeam Invasieve Exoten, Ministerie van Economischezaken, Landbouw en I nnovatie, for commissioningthe risk assessment of the raccoon dog. Dr. Tom vanMulder / Lutra 2013 56 (1): 23-43 39

der Have made many helpful and stimulating commentsduring the process of preparing the report. Dr.Norman Stier from ἀ arandt (Germany) introducedme long ago to the raccoon dog, in his study area inMecklenburg-Vorpommern, and has since answeredmany of my questions. I h ave much appreciated thehelp of Dr. Joke van der G iessen, RIVM B ilthoven,in writing the paragraphs on infectious diseases andpublic health. Hans Vink provided me with relevantliterature. In collecting observations of raccoon dogsin the Netherlands I was assisted by a whole array ofpersons and institutions: Dr. Sim Broekhuizen, Alterra,the Dutch Mammal Society,,, Yvette van Veldhuijsen, Margriet Montizaan of theRoyal Dutch Hunters Society, and the AAP Foundation.I am very grateful for their kind cooperation, andfor their help in contacting many of the original raccoondog observers. Lastly I thank Martijn van Oene(Dutch Mammal Society) for making figure 3, a ndthree anonymous referees for helpful comments andcriticism.ReferencesBarbu, P. 1972. Beiträge zum Studium des Marderhundes,Nyctereutes procyonoides ussuriensis Matschie,1907, aus dem Donaudelta. SäugetierkundlicheMitteilungen 20: 375-405.Branquart, E. (ed.) 2007. Guidelines for environmentalimpact assessment and list classification of nonnativeorganisms in Belgium. URL:;viewed April 2013.Broekhuizen, S. 2007. Wordt de wasbeerhond eennieuwe muskusrat? Zoogdier 18 (2): 15-17.Drygala, F. 2009. Space use pattern, dispersal andsocial organisation of the raccoon dog (Nyctereutesprocyonoides Gray, 1834), an invasive, aliencanid in Central Europe. PhD thesis. TechnischeUniversität Dresden, Dresden, Germany.Drygala, F., N. Stier, H. Zoller, K. Bögelsack, H.M. Mix& M. Roth 2008. Habitat use of the raccoon dog(Nyctereutes procyonoides) in north-eastern Germany.Mammalian Biology 73: 371-378.Faust & Russell 1964. Craig and Faust’s clinical parasitology.Henry Kimpton, London, UK.Hegglin, D. & P. Deplazes 2008. Control strategy forEchinococcus multilocularis. Emerging InfectiousDiseases 14: 1626-1628.Holmala, K. & K. Kauhala 2006. Ecology of wildliferabies in Europe. Mammal Review 36: 17-36.Kauhala, K. 2004. Removal of medium-sized predatorsand the breeding success of ducks in Finland. FoliaZoologica 53: 367-378.Kauhala, K. & M. Saeki 2004. Raccoon dog Nyctereutesprocyonoides (Gray, 1834). In: C. Sillero-Zubiri,M. Hoffmann & D.W. Macdonald (eds.). Canids:foxes, wolves, jackals and dogs. Status survey andconservation action plan: 136-142. IUCN/SCCCanid Specialist Group, Cambridge, UK.Kauhala, K. & M . Saeki 2008. Nyctereutes procyonoides.In: IUCN 2010. IUCN Red List of ἀ reatenedSpecies. Version 2010.2. URL:;viewed April 2013.Kauhala, K. & R. Kowalczyk 2011. Invasion of the raccoondog Nyctereutes procyonoides in Europe: Historyof colonization, features behind its success,and threats to native fauna. Current Zoology 57(5): 584-598.Kauhala, K., K. Holmala & J. Schregel 2007. Seasonalactivity patterns and movements of the raccoondog, a vector of diseases and parasites, in southernFinland. Mammalian Biology 72: 342-353.Kauhala, K., K. Holmala, W. Lammers & J . Schregel2006. Home ranges and densities of medium-sizedcarnivores in south-east Finland, with special referenceto rabies spread. Acta ἀ eriologica 51: 1-13.Kowalczyk, R. 2007. NOBANIS - Invasive alien speciesfact sheet - Nyctereutes procyonoides. Online Databaseof the North European and Baltic Networkon Invasive Alien Species: 1-8. URL:;viewed April 2013.Kowalczyk, R., B. Jędrzejewska, A. Zalewski & W .Jędrzejewski 2008. Facilitative interactionsbetween the Eurasian badger (Meles meles), the redfox (Vulpes vulpes) and the invasive raccoon dog(Nyctereutes procyonoides) in Białowieża PrimevalForest, Poland. Canadian Journal of Zoology 86:1389-1396.Lina P.H. & A.M. Hutson 2006. Bat rabies in Europe: areview. Developments in Biologicals 125: 245-254.Lindström, E. & T. Mörner 1985. ἀ e spreading of sar-40 Mulder / Lutra 2013 56 (1): 23-43

coptic mange among Swedish foxes (Vulpes vulpesL.) in relation to fox population dynamics. Revued’Écologie (Terre et Vie) 40: 211-216.Machida, N., K. Kiryu, K. Oh-ishi, E. Kanda, N. Izumisawa& T. Nakamura 1993. Pathology and epidemiologyof canine distemper in raccoon dogsNyctereutes procyonoides. Journal of ComparativePathology 108: 383–392.Machnicka-Rowinska, B., B. Rocki, E. Dziemian & M.Kolodziej-Sobocinska 2002. Raccoon dog (Nyctereutesprocyonoides) - the new host of Echinococcusmultilocularis in Poland. Wiadomosci Parazytologiczne48: 65-68.Mikkonen, T., V. Haukisalmi, K. Kauhala & H. Wihlman1995. Trichinella spiralis in the raccoon dog(Nyctereutes procyonoides) in Finland. Bulletin ofthe Scandinavian Society for Parasitology 1995 (5):100.Mól, H. 2005. Wscieklinizna zwierat w 2004 r. Na tlepotrzeby jej badania w Polsce [Rabies in 2004 andthe need for research in Poland]. Zycie Weterynaryjne80: 655-658.Mulder, J.L. & S. Broekhuizen 1992. De wasbeerhondkomt. Zoogdier 3 (4): 34.Mulder, J.L. 2011. ἀ e raccoon dog in the Netherlands– a r isk assessment. Rapport Bureau Mulder-natuurlijk.URL:; viewed April 2013.Mulder, J.L. 2012. A review of the ecology of the raccoondog (Nyctereutes procyonoides) in Europe.Lutra 55 (2): 101-127.Oerlemans, M. & P . Koene 2008. Possible implicationsof the presence of the raccoon dog (Nyctereutesprocyonoides) in the Netherlands. Lutra 51 (2):123-131.Oivanen, L., C.M.O. Kapel, E. Pozio, G. La Rosa,T. Mikkonen & A . Sukura 2002. Associationsbetween Trichinella species and host species inFinland. Journal of Parasitology 88: 84–88.Oksanen, A., E. Lindgren & P. Tunkkari 1998. Epidemiologyof trichinellosis in lynx in Finland. Journalof Helminthology 72: 47–53.Opsteegh, M., M. Langelaar, C. van Dam, M. Maas, J.L. Mulder, M. Fonville, F. Franssen, K. Takumi & J.van der Giessen 2012. Onderzoek naar Echinococcusmultilocularis bij Nederlandse vossen: eindrapportage2012. RIVM Briefrapport 001/2013Z&O. RIVM, Bilthoven, the Netherlands.Pannwitz, G., A. Mayer-Scholl, A. Balicka-Ramisz &K. Nökler 2010. Increased prevalence of Trichinellaspp., Northeastern Germany, 2008. EmergingInfectious Diseases 16: 936-942.Pianka, E.R. 1978. Evolutionary ecology – second edition.Harper and Row, New York, US.Puffpaff, S. 2008. Naturschutzfachliche Kartierungund Bewertung der Gewässerstruktur desNationalparks Jasmund unter Berücksichtigungbestimmter Gewässer als Feuchtlebensräume derAnhang II Arten der Fauna-Flora-Habitat-Richtlinie.MSc thesis Hochschule Neubrandenburg,Neubrandenburg, Germany. URL:; viewed April2013.Schwarz, S., A. Sutor, C. Staubach, R. Mattis, K. Tackmann& F.J. Conraths 2011. Estimated prevalenceof Echinococcus multilocularis in raccoon dogsNyctereutes procyonoides in northern Brandenburg,Germany. Current Zoology 57 (5): 655-661.Sidorovich, V.E. 2011. Analysis of vertebrate predatorpreycommunity. Tesey, Minsk, Ukraine.Sidorovich, V.E., A.G. Polozov, G.O. Lauzhel & D .A.Krasko 2000. Dietary overlap among generalistcarnivores in relation to the impact of the introducedraccoon dog Nyctereutes procyonoides onnative predators in northern Belarus. Zeitschriftfür Säugetierkunde 65: 271-285.Stier, N. 2006. Rivale von Fuchs und Dachs? Marderhund:Ökologische Auswirkungen der Besiedlung.Neubürger auf dem Vormarsch. Sonderheft vonUnsere Jagd, Pirsch & Ni edersächsischer Jäger:24-29.Stier, N. & F. Joisten 2006. Mit Waffe und Bauhund.Bejagung des Marderhundes. Neubürger auf demVormarsch. Sonderheft von Unsere Jagd, Pirsch &Niedersächsischer Jäger: 30-35.Tackmann, K., J. Goretzki & F.J. Conraths 2003. DasNeozoon Marderhund als neue Entwirtpopulationfür Echinococcus multilocularis in OstDeutschland - ein Risiko? Bedrohung der biologischeVielfalt durch invasive gebietsfremde Arten.Schriftenreihe des Bundesministeriums für Verbraucherschutz,Ernährung und Landwirtschaft.Reihe A: Angewandte Wissenschaft 498: 176-181.Mulder / Lutra 2013 56 (1): 23-43 41

ἀ iess, A. 2004. Untersuchungen zur Helminthenfaunaund zum Vorkommen von Trichinella sp.beim Marderhund (Nyctereutes procyonoides) inBrandenburg. PhD thesis. Freie Universität Berlin,Berlin, Germany.van der Giessen, J.W.B., Y. Rombout & P. Teunis 2004a.Base line prevalence and spatial distribution ofEchinococcus multilocularis in a newly recognizedendemic area in the Netherlands. Veterinary Parasitology119: 27-35.van der Giessen, J.W.B., A. de Vries, M.L. Chu, V. Stortelder,J.L. Mulder, C. de Lezenne Coulander & P.Teunis 2004b. ἀ e prevalence of Echinococcus multilocularisin foxes in Limburg 2002-2003. Report330040001/2004. RIVM, Bilthoven, the Netherlands.URL:;viewed April 2013.van Dijk, M. & N . de Koning, N. 2009. De problematiekvan wilde exotische zoogieren in Nederland.Een inzicht in de problematiek, als basis voor hetopstellen van een beleidsvisie door de Zoogdiervereniging,geïllustreerd met een risicoanalysevan de Wasbeerhond (Nyctereutes procyonoides).Afstudeeronderzoek Diermanagement 594313.Zoogdiervereniging, Nijmegen, the Netherlands /Hogeschool van Hall Larenstein, Leeuwarden, theNetherlands.Vergoossen, W.G. & L . Backbier 1993. Waarnemingenvan de wasbeerhond in Limburg. NatuurhistorischMaandblad 82 (2): 36-41.Westerling, B. 1991. Raivotauti Suomessa ja sen torjuntavuosina 1988-90 [Rabies in Finland and its control1988–90]. Suomen Riista 37: 93-100. [with Englishsummary]Zienius, D., V. Sereika & R. Lelesius 2007. Rabies occurrencein red fox and raccoon dog population inLithuania. Ekologija 53: 59-64.Zoller, H. 2006. Koexistenz zwischen Enok und Reine ke.Neubürger auf dem Vormarsch. Sonderheft vonUnsere Jagd, Pirsch & Niedersächsischer Jäger: 26.SamenvattingDe wasbeerhond (Nyctereutes procyonoides)in Nederland – zijn huidige statusen een risico- analyseDe wasbeerhond heeft zijn natuurlijke verspreidingsgebiedin het verre oosten van Azië.Tussen 1928 en 1957 werden duizenden wasbeerhondenuitgezet in de voormalige SovjetUnie, voornamelijk ten westen van de Oeral.Van daaruit heeft de soort zich over een grootdeel van Europa verspreid. De soort wordtbeschouwd als een invasieve exoot, omdat hijdoor mensen is geïntroduceerd, zich succesvolvoortplant en zich verder verspreidt. Hetbeleid in Nederland met betrekking tot invasieveexoten bestaat uit het schatten van derisico’s voor de biodiversiteit met aandachtvoor de impact op dier- en volksgezondheiden economie. Een vorig artikel (Mulder 2012)bevatte een samenvatting van de huidige kennisvan de ecologie van de wasbeerhond inEuropa; het onderhavige artikel geeft een overzichtvan zijn huidige voorkomen in Nederlanden een analyse van de ecologische, veterinaireen sanitaire risico’s.Bijna vijftien jaar geleden werden in Nederlandde eerste uit Duitsland afkomstige wasbeerhondenwaargenomen. Het patroon van diekolonisatie wordt verstoord door het regelmatigvoorkomen van uit gevangenschap ontsnaptedieren. Tot 2013 werden ‘wilde’ wasbeerhondenwaarschijnlijk uitsluitend aangetroffen inhet noordoosten van Nederland: de provinciesGroningen, Friesland, Drenthe en O verijssel,plus de Noordoostpolder. Dit patroon sluit aanbij het huidige voorkomen in Duitsland. Er isin Nederland nog geen voortplanting geconstateerd.De huidige fase, met jaarlijks een beperktaantal waarnemingen, heeft in andere landen 20tot 30 jaar geduurd. Pas daarna begon de populatiesnel toe te nemen.Het is te verwachten dat de wasbeerhondheel Nederland gaat koloniseren, waarschijnlijkmet uitzondering van de Waddeneilanden.Zijn invloed op de biodiversiteit wordt in hetalgemeen ingeschat als beperkt. Alleen kleinegeïsoleerde populaties van amfibieën en grondbroedendevogels in moerasgebieden kunnengevoelig zijn voor predatie. De wasbeerhondheeft geen nieuwe ziektes meegebracht naarEuropa, maar zou wel het voorkomen van42 Mulder / Lutra 2013 56 (1): 23-43

eeds aanwezige ziektes en parasieten kunnenbevorderen: hondenziekte, Trichinella,schurft en de vossenlintworm Echinococcusmultilocularis. De wasbeerhond is ookeen potentiële verspreider van hondsdolheid,maar deze ziekte komt tegenwoordig na hetuitvoeren van grootschalige orale immunisatie-campagnesniet meer voor. De belangrijksteproblemen voor de volksgezondheid wordenwaarschijnlijk gevormd door Trichinellaen de vossenlintworm.De mogelijkheden om wasbeerhondeneffectief te bestrijden zijn beperkt. De bestrijdingdie vanaf het begin van de kolonisatie inDuitsland is toegepast, heeft de verdere verspreidingvan de wasbeerhond niet merkbaarverminderd en h eeft geen invloed gehad opde voorjaarsstand. Er zijn twee beheeroptiesvoor het tegengaan van mogelijke problemen:A. Intensieve bestrijding op een lokale schaal(maximaal enkele km 2 ) waar problemen,zoals predatie, zijn, of te verwachten zijn.Alleen effectief in de maanden zonder dispersievan jonge dieren en z onder reproductie:december tot en met maart.B. Preventie van predatie door de toegangvoor wasbeerhonden te verhinderen, bijvoorbeelddoor middel van (schrik-)hekwerkrond kwetsbare plaatsen als poelenmet zeldzame amfibieën of broedvogelkolonies.Received: 20 February 2013Accepted: 10 April 2013Mulder / Lutra 2013 56 (1): 23-43 43

Abundance of harbour porpoises ( Phocoenaphocoena) on the Dutch Continental Shelf,aerial surveys in July 2010-March 2011Steve C.V. Geelhoed 1 , Meike Scheidat 1 , Rob S.A. van Bemmelen 1 & Geert Aarts 1,21IMARES Wageningen UR, Institute for Marine Resource & Ecosystem Studies, Ecosystems Department,P.O. Box 167, NL-1790 AD Den Burg, Texel, the Netherlands, e-mail: steve.geelhoed@wur.nl2Wageningen UR, Department of Aquatic Ecology and Water Quality Management, Droevendaalsesteeg 3a,NL-6708 PB Wageningen, the NetherlandsAbstract: ἀ e harbour porpoise (Phocoena phocoena) is the most abundant marine mammal species in Dutchwaters. Nevertheless until 2010 abundance estimates for the entire Dutch Continental Shelf (DCS) were missing.Aerial surveys along designed track lines in July 2010, October/November 2010 and March 2011 provided densityand abundance estimates for the DCS. ἀ e highest abundance estimate was made in March 2011 (n=85,572);approximately three times higher than in July 2010 (n=25,998) and October/November 2010 (n=29,963). Distributionpatterns of porpoises differed between seasons, but a band of higher densities from the Brown Ridge to theBorkumer Reef was visible in all seasons. Calves were mainly seen in July, indicating that porpoises also reproducein Dutch waters. ἀ e total abundance estimate in March 2011 corresponds to 48% of the southern North Sea population,which implies that a large part of the North Sea population resides in Dutch waters during that season. Suchhigh densities may lead to increased conflict with human activities, making the instigation of local managementactions more imminent.Keywords: abundance, aerial survey, distance sampling, harbour porpoise, North Sea, Phocoena phocoena, populationsize.Introduction© 2013 Zoogdiervereniging. Lutra articles also on theinternet: http://www.zoogdiervereniging.nlἀ e harbour porpoise (Phocoena phocoena)is the most abundant marine mammal speciesin Dutch waters. After a sharp decline inthe first half of the 20th century, the occurrenceof harbour porpoises in Dutch watershas increased significantly in the last decades(Camphuysen 2011). ἀ is is probably a resultof a s outhward shift in distribution (SCANS2008). ἀ e reasons for this are not clear. Ashift in prey species is a likely cause (Camphuysen2004), although this remains a matterof debate (MacLeod et al. 2007).Systematically collected data on harbourporpoise abundance and distribution inDutch waters are scarce. Most data are a byproductof surveys aimed at seabirds, providinginformation on the relative occurrenceand distribution of porpoises. ἀ ese data werecollected during a l and-based sea watchingscheme, ship-based surveys (Camphuysen &Leopold 1994) and aerial surveys of the DutchContinental Shelf (DCS) (Baptist & W olf1993, Arts 2011).In the summer of 1994 and 2005, two largescalededicated cetacean surveys (SCANS andGeelhoed et al. / Lutra 2013 56 (1): 45-57 45

study is to present the results of three aerial surveysin July 2010, October/November 2010 andMarch 2011, and to estimate the distribution,density and abundance of the harbour porpoiseon the entire Dutch Continental Shelf.MethodsStudy area, survey design and dataacquisitionFigure 1. Map of the Dutch Continental Shelf showingthe study areas A-D, and some geographical namesused in the main text.SCANS II) resulted in abundance estimatesof harbour porpoises in European waters, butthe design of the survey areas did not coincidewith national borders (Hammond et al. 2002,SCANS 2008). Furthermore, the internationalsurvey took place in summer, while in Dutchwaters, most porpoises are present in winterand early spring (Camphuysen 2011, Scheidatet al. 2012).Since May 2008, dedicated aerial surveyswere conducted in the Dutch sector ofthe North Sea. However, in these surveys, thenorthernmost part was excluded (Scheidat etal. 2012). Hence, an un-biased estimate of porpoiseabundance and distribution for the entireDutch Continental Shelf (DCS) is still lacking.Such baseline data on porpoise abundance anddistribution is essential to monitor the effect ofhuman activities and to potentially assess theeffectiveness of conservation measures (e.g.Scheidat et al., in press). ἀ e objective of thisAerial surveys were carried out on the DutchContinental Shelf, divided into four areas(figure 1): A ( 9615 km²), B ( 16,892 km²), C(12,023 km²) and D (20,797 km²), which weresurveyed by aircraft along predesigned tracklines. ἀ e design of the track lines was parallelin ‘near shore’ areas C a nd D a nd zigzagin areas A a nd B t o ensure a r epresentativecoverage (figure 2). ἀ e direction of transectsin areas C and D followed depth gradientsin order to minimise potential variancein encounter rate within transect lines causedby depth (Buckland et al. 2001). ἀ e zigzagdesign of the offshore areas aimed at maximisingthe endurance of the plane and coveras large an area as possible. Additional tracklines were surveyed in two smaller areas W1and W2, within areas D a nd C, respectively,selected as potential areas for future offshorewind farms (figure 2).Surveys were conducted with a h ighwingedtwin-engine airplane, the Partenavia68, equipped with bubble windows (allowingobservations directly under the plane), flying atan altitude of 183 m (600 feet) with a speed of ca.186 km.h -1 (ca. 100 knots). Every four secondstime and the aircraft’s position were recordedautomatically onto a laptop connected to a GPS.Surveys were conducted by a team of three people.Details on environmental conditions wereentered in a database by the so-called navigatorat the beginning of each transect and wheneverconditions changed. Observations were madeby two dedicated observers each located at thebubble windows on the left and right side of46 Geelhoed et al. / Lutra 2013 56 (1): 45-57

The collected sightings were used to calculate densities and amaps. For the former only the surveys of the areas A, B, C anand W2 was omitted), while for the distribution estimates allallows for obtaining estimates of absolute densities, i.e. the nthe aircraft. For each observation, the observersacquired data that were entered in real time stratum v (i.e. area A, B, C a nd D) was estilandet al. 2001). Animal abundance in eachconfidence interval (C.I.) and coefficient of variation (C.V.; Binto a d atabase by the navigator. Observation stratum mated v using (i.e. area a HA, orvitz-ἀ B, C and ompson-like D) was estimated estimatorusing a Hordata included species (all cetaceans and seals),2001, Buckland(Bucklandet al.et2004)al. 2001,as follows:Buckland et al.declination angle measured with an inclinometerfrom the aircraft abeam to the individual2004) as follows:A ⎛⎛ nv v n ⎞⎞vor group, group size, presence of calves, behaviour,swimming direction relative to the tran-⎝⎝Nˆgs⎜⎜msv= + ⎟⎟svL ⎜⎜ ⎟⎟vˆ µ ˆgµm ⎠⎠sect, detection cue, whether the individual where A vis the area of the stratum, L vis theor group was above or below the sea surface length of transect line covered on effort inwhen abeam, and reaction to the survey plane. where good Aor v is moderate the area of conditions, the stratum, n gsvLis v is the the numberof sightings that occurred in good con-length of transectEnvironmental data included sea state (Beaufortscale), turbidity (assessed by visibility ofgsv is the number of sightings that occurred in goconditions, nditions in the stratum, n msvis the number ofobjects below the sea surface), cloud cover (in sightings that occurred in moderate conditionsoctaves), glare (area covered and strength) andsightingsin the stratum,that occurredˆµ gis theinestimatedmoderatetotalconditionseffectivein the stratumsubjective sighting conditions. ἀ ese sighting strip width (ESW) in good conditions, ˆµ mis theconditions represent each observer’s subjective estimated total effective strip width in moderateconditions and Sview of the likelihood that the observer wouldˆµ m(ESW) in good conditions, vis the mean observedis the estimated total effecsee a h arbour porpoise within the primary school size in the area should one be present, and these the mean ἀ e effective observed strip school width size is in the distance stratum. atconditions could be either good, moderate or which the The number effective of strip animals width detected is the distance outside at which thepoor. Furthermore, a category “not possible to the strip width equals the number of animalsobserve” is used. Sighting conditions could differbetween the left and right side of the plane. missed rected on for the the track proportion line by of a factor animals called missed g(0). on For the currenequals missed the inside number the of strip animals width. missed ἀ e ESW inside is cor-the strip width. TSurveys were conducted in weather conditionssafe for flying operations (no fog or rain,the track line by a factor called g(0). For the currentsurveyGerman studythe(Scheidatg(0) valueset al.obtained2005, Scheidatin a similaret al. 2008, Gilno chance of freezing rain, visibility >3 km) German study (Scheidat et al. 2005, Scheidat etand suitable for porpoise surveys (sea state ≤3 al. 2008, Gilles et al. 2009) were applied. ἀ eseBeaufort).0.37 g(0) for for good values good conditions are 0.37 and for 0.14 and good 0.14 for moderate conditions for moderate conditions, and conditions, resulting re i0.37 for good conditions and 0.14 for moderate conditions, resulting in effective s0.37 for good conditions and 0.14 for moderate respectively. conditions, 0.14 for moderate resulting conditions, in effective resulting strip widths in effec-of 76.5 and 27 mrespectively.strip widths of 76.5 and 27 m, respectively.respectively.Group abundance by stratum by stratum (areas (areas A, B, C A, and B, D) C was and estim D) wData analysisGroup Group abundance by stratum by stratum (areas A, (areas B, C and A, D) B, was C estimated by:Group abundance by stratum (areas A, B, and C and D) D) was was estimated by:ἀ e survey data were collected using distance N ˆ vsampling techniques (Buckland et al. 2001, vN ˆ N(group) v/ sˆv= N ˆby:ˆ(group)=v/vsv/svvN ˆ . . .vN ˆ(group)=v/ sv.Buckland et al. 2004). ἀ e collected . sightings Total Total animal animal and and group and group abundances group abundances of the of entire the of study entire the area study were area estim wwere used to calculate densities and abundanceestimates, and to produce distribu-NˆNˆ= =Total entire animal study and group area abundances were estimated of the entire by study area were estimated byTotal animal and group abundances of the entire study area were estimated bytion maps. For the former only the surveys NˆNˆ= NˆvN ˆ =vN ˆ(group)= ∑ Nˆ∑ Nˆ∑ NˆvN ˆ(group)= ∑ Nˆ(group)v (group)v (group)v v andv(group)vNˆ= vvof the∑ Nˆareas vN ˆ = andA, B, C and D were∑ Nˆand(group)used v (group)vv(i.e. the respectively. and Densities were estimated byextra survey effort and in W1 and W2 was omitted),while for Densities the distribution were estimated estimates by dividing all of the the abundance associated estimates stratum. by Mean the area group of the size associated stratrespectively. dividing Densities the abundance were were estimated estimates estimated by dividing by by the dividing area the abundance the abund estimrespectively. Densities were estimated by dividing the abundance estimates by were used. Line-transect distance samplingallows for obtaining estimates of abso-Mean group size size across across strata strata was estimated was estimated E by [ s]= Nby/ N(group)lute densities, i.e. the number of - ² Mean E ˆ[group s]= size Nˆacross / Nˆstrata (group) was estimated by.across strata was estimated byˆ ˆˆ E ˆˆ[s]ˆ=E N[ˆs/] N=ˆ(groNMean group size across strata was estimated by.Coefficients of variation of variation . (C.V.) (C.V.) and 95% and confidence 95% confidence intervalwith the associated 95% confidence interval Coefficients of variation (C.V.) (C.V.) and 95% and confidence 95% con-intervals (C.I.) were(C.I.) and Coefficients coefficient of variation (C.V.) (C.V.; Buck-and 95% bootstrap confidence (999 intervals (999 replicates) intervals (C.I.) within (C.I.) were within strata, estimated were strata, using estimated by transect using a non- by transect segments a non-parame segmen as thebootstrap (999 replicates) within strata, using transect segments as the sampling unbootstrap (999 replicates) within strata, using transect estimation segments of of ESW ESW was the was incorporated sampling using units. using a The parametric variance a parametric bootstrap due to boot proGeelhoed et al. / Lutra 2013 56 (1): 45-57 estimation of ESW was incorporated using a parametric bootstrap 47 procedure assumestimation of ESW was incorporated using a parametric normally distributed bootstrap random procedure random variables. assuming variables. More the details More ESW details on estimates this method on this to ca mnormally distributed random variables. More details on this method can be foundnormally distributed random variables. More details on this method can be found in Scheidat et al. (2008, 201

Figure 2. Survey effort in good or moderate sighting conditions on at least one side of the plane (on and off track line)with all sightings of harbour porpoises, including navigator sightings. Stars indicate groups with calves/neonates.parametric bootstrap (999 replicates) withinstrata, using transect segments as the samplingunits. ἀ e variance due to estimation of ESWwas incorporated using a parametric bootstrapprocedure assuming the ESW estimates to benormally distributed random variables. Moredetails on this method can be found in Scheidatet al. (2008, 2012).Distribution mapsDistribution maps are presented in two differentways: 1. Uncorrected sightings, and 2.Densities corrected for survey effort and subjectivesighting conditions per 1/9 ICES gridcell. Densities are represented spatially on a1/9 ICES grid. ἀ is grid has latitudinal rowsat intervals of 10’ and longitudinal columnsat intervals of 20’. Within the DCS, this correspondsto approximately 20x20 km grid cells,with areas ranging from 388 to 409 km 2 .Densities per 1/9 ICES grid cell were calculatedby dividing the total number of animalsobserved during good and moderate conditionsby the total surveyed area. ἀ e surveyedarea is the distance travelled multiplied by theeffective strip width (ESW). Grid cells withlow effort, such as grid cells extending outsidethe borders of the surveyed area (e.g. theWadden Sea), tend to be less reliable. ἀ erefore,grid cells with an effort less than one km 2were omitted from the maps (but used for theabundance estimates).ResultsEffort and sightings of harbour porpoisesἀ e July 2010 and March 2011 surveys hadthe best coverage with both around 6000 km48 Geelhoed et al. / Lutra 2013 56 (1): 45-57

Table 1. Total survey days, effort, sighting conditions (G – good, M – moderate, P – poor, X – not possible to observe)and harbour porpoise sightings during the three aerial surveys. Calves are included in the number of animals.Sighting conditions (%) Porpoise sightings (n)Survey Effort (km) G M P / X Sightings Individuals CalvesJuly 20106040 35 35 30 263 330 26(5,6; 8-11; 18-20 July)October/November 20104028 12 76 11 137 163 0(12-14 Oct; 19, 21, 24 Nov)March 20115945 29 62 9 684 743 2(18,19; 21 – 27 March)Total 16,013 1085 1236 28on effort (table 1). In these months, two setsof track lines in areas A-D plus the extra tracklines in W1 and W2 could be surveyed (figure2). Survey conditions were more adverse inOctober and November (shorter days, unstableweather) leading to lower survey effort,but one set of track lines could be completedin areas A-D and in the W1 and W2. In total,1085 sightings of 1236 harbour porpoises werecollected. Calves were sighted during the Julysurvey and the March survey, consisting 7.9%and 0.3% of the sighted individuals, respectively.During the October survey, one calf wasrecorded by the navigator (navigator sightingsare not included in the analysis of abundance).ἀ e ‘calves’ in March were big neonates.Average group size for all surveys combinedwas 1.14 animals. Average group size was1.09 (C.V. 0.31) animals in March, 1.25 (C.V.0.52) in July, and 1.19 (C.V. 0.37) in October/November. ἀ e largest group size observedwas a pod of eight animals in July. In all seasonsover 80% of the sightings consisted ofsingle animals. In summer and in autumn alarger proportion of the sightings consistedof two or more animals compared to earlyspring.Density and abundance of harbour porpoisesTable 2 gives an overview of density ( - ²) as well as abundance (number of animals)per survey area and survey period. ἀ e overalldensity was similar for the summer (July) andthe autumn (October/November) survey; 0.44and 0.51 - ² respectively. Densitywas about three times higher during the Marchsurvey with 1.44 - ².ἀ e total numbers of harbour porpoises onthe DCS (areas A-D) were estimated at 25,998(C.I.: 13,988-53,623) and 29,963 (C.I.: 16,098-59,011) animals in summer and autumnrespectively. ἀ e abundance in March comprised85,572 animals (C.I.: 49,324-165,443,table 2).Distribution of harbour porpoisesFigures 3-5 show densities of porpoises ( ²) per 1/9 ICES grid cell. In summer(July), higher densities were observed nearthe Brown Ridge, the Borkumer Reef, andaround the Botney Cut – Dogger Bank, nearthe UK border. ἀ e band of higher densitiesrunning from the Brown Ridge to the BorkumerReef was also visible in the autumnsurveys (October/November) and the springsurveys (March). In autumn the offshore densityin area B w as lower and porpoises weremore evenly distributed than in the other twosurvey periods. In spring, the overall densitywas much higher. In that period, the densitiesin areas B a nd D a lmost tripled, whilearea C showed an even stronger increase. ἀ ehighest density was found in area C, north ofthe Wadden Isles (figure 6). ἀ e high densityarea in the Dutch part of the Borkumer ReefGeelhoed et al. / Lutra 2013 56 (1): 45-57 49

Figure 3Figure 4Figure 5Figure 3. Summer density distribution of harbour porpoises( - ²) per 1/9 ICES grid cell, July 2010.Grid cells with low effort (

Table 2. Estimates of density and abundance of harbour porpoises from aerial surveys conducted on the DutchContinental Shelf in July 2010, October/November 2010 and March 2011. Estimates are given with the associated95% Confidence Interval (C.I.) and Coefficient of Variation (C.V.).July 2010Survey Density ( - ²) (95% C.I.) Abundance (n animals) (95% C.I.) C.V.AreaA 0.396 0.181 - 0.849 3806 1738 - 8165 0.404B 0.477 0.212 - 1.058 8055 3589 - 17,872 0.416C 0.336 0.046 - 0.890 4039 553 - 10,701 0.622D 0.484 0.208 - 1.056 10,098 4341 - 22,024 0.403Overall 0.438 0.236 - 0.903 25,998 13,988 - 53,623 0.336October / November 2010Survey Density ( - ²) (95% C.I.) Abundance (n animals) (95% C.I.) C.V.AreaA 0.391 0.117 - 0.872 3763 1124 - 8384 0.461B 0.573 0.298 - 1.157 9679 5035 - 19,543 0.352C 0.683 0.287 - 1.610 8216 3451 - 19,351 0.459D 0.398 0.212 - 0.733 8304 4431 - 15,296 0.317Overall 0.505 0.271 - 0.994 29,963 16,098 - 59,011 0.332March 2011Survey Density ( - ²) (95% C.I.) Abundance (n animals) (95% C.I.) C.V.AreaA 1.029 0.522 - 2.144 9890 5018 - 20,618 0.386B 0.908 0.521 - 1.791 15,331 8795 - 30,249 0.312C 2.982 1.645 - 5.806 35,850 19,772 - 69,808 0.325D 1.174 0.658 - 2.389 24,501 13,726 - 49,833 0.344Overall 1.441 0.830 - 2.786 85,572 49,324 - 165,443 0.316extended to the German part. Low densitieswere found in close proximity of the mainlandcoast and in the southern part of area D.DiscussionSeasonal patternsἀ e results show distinct differences in abundanceand distribution of harbour porpoisesbetween surveys; in summer and autumnsimilar numbers and densities occurred,while in March the numbers almost tripled.ἀ is pattern fits the general seasonal occurrenceas seen along the Dutch coast duringsystematic land-based observations of seabirdmigration and marine mammals (Camphuysen2011) and previous aerial surveys(Scheidat et al. 2012). ἀ ese land-based observationsshow that harbour porpoises are presentin coastal waters throughout the year.Peak numbers are observed from winter toearly spring (December-March), after whichthe numbers drop. Observations in June arerelatively scarce, but the numbers slightlyincrease from July onwards (Camphuysen2004, Camphuysen 2011). In the Belgian partof the North Sea, harbour porpoises are mostabundant from February to April, whereaslower numbers tend to occur offshore in therest of the year (Haelters et al. 2011). Frenchstrandings data also indicate a higher occurrenceof porpoises in the Channel during theGeelhoed et al. / Lutra 2013 56 (1): 45-57 51

density (animals/km²)65432ABCDoverall10July 10 Oct/Nov 10 March 11Figure 6. Density of harbour porpoises during the three surveys in the survey areas A-D and the entire study area.ἀ e black bars show the 95% Confidence Interval of the estimates.winter months (Jung et al. 2009). In the GermanNorth Sea bordering Dutch waters (areaC) the aerial surveys show the highest densitiesin spring. Further north along the Germancoast numbers peak in May and June (Gilles etal. 2009). Still further north, along the Danishwest coast, porpoise densities are highest fromApril to August, with a peak in August (datafor June-July are lacking however) (Teilmannet al. 2008). In the western North Sea porpoisenumbers peak in April off south-eastern England,and later further north (August; Evans etal. 2003). ἀ ese observations suggest a northwardsummer migration from the Channel,Belgium and ἀ e Netherlands to Danish andBritish waters and a southward migration inautumn.Comparison of densitiesIn 2010-2011, the estimated densities of harbourporpoises in Dutch waters ranged from0.34–2.98 - ² (table 2). ἀ ese densitiesare in the same order of magnitude (0.33-2.53 - ²) as obtained during comparablestudies in adjacent waters in Germany(ἀ omsen et al. 2006, Gilles et al. 2009), andin Belgium (Haelters et al. 2011), and in therelevant survey blocks during the large scaleSCANS II survey (SCANS 2008). Data fromGilles et al. (2009) reveal that the highest densitiesin the entire German North Sea ExclusiveEconomic Zone were reached in springwith an overall density of 1.34 - ²and a density of 0.85 - ² in the areaclosest to the Dutch border (East Frisia). ἀ eSCANS-II survey (SCANS 2008) showed aporpoise density for SCANS-block H i n thesouthern North Sea of 0.36 - ² inJune-July 2005. In the adjacent SCANS-blockB (the Channel) a d ensity of 0.33 was estimatedfor 2005. ἀ is density corresponds wellwith the 0.48 - ² estimated for theDutch survey area D i n July, which overlapswith SCANS-block B.52 Geelhoed et al. / Lutra 2013 56 (1): 45-57

Table 3. Comparison between density and abundance estimates obtained in the same areas andmonths (2008–2010) using results from the current study as well as from Scheidat et al. (2012).Estimates are given with the associated 95% Confidence Interval (C.I.) and Coefficient of Variation(C.V.).Area Period Density ( - ²)(95% C.I.)October-DecemberArea C Nov 2008 1.020(0.34 – 2.10)Oct/Nov 2010 0.683(0.29 - 1.61)Area D November/December 2009 1.511(0.91 – 3.08)October/November 2010 0.398(0.21 - 0.73)February-MarchArea B March 2010 0.660(0.28 – 1.45)March 2011 0.908(0.52 - 1.79)Area C March 2010 1.107(0.48 – 2.49)March 2011 2.982(1.65 - 5.81)Area D February/March 2009 1.468(0.78-2.70)March 2010 2.007(0.82 – 4.04)March 2011 1.174(0.66 - 2.39)Abundance(95% C.I.)12,227(4038 – 25,285)8216(3451 – 19,351)31,515(18,976 – 64,157)8304(4431 – 15,296)11,141(4692 – 24,560)15,331(8795 – 30,249)13,309(5819 – 29,918)35,850(19,772 – 69,808)30,534(16,265 – 56,161)41,878(17,145 – 84,302)24,501(13,726 – 49,833)C.V.0.420.460.320.320.420.310.440.330.330.390.34In March 2011, both Belgian and Germanwaters were surveyed simultaneously withthe Dutch surveys, thus allowing a d irectcomparison of the corresponding densities.ἀ e estimated densities of harbour porpoiseswere high, with 2.53 - ² and - ² in Belgian and German watersrespectively. ἀ ese densities correspond wellwith the maximum SCANS II density of - ² in area C, whereas the densityin area D of 1.17 - ² was lower.Inter-annual variabilityAerial surveys in the DCS using the samemethods as the present study have also beencarried out from May 2008 to March 2010.In February-May, August, November andDecember 2008-2010 10,557 km were coveredon effort during 16 survey days (Scheidat et al.2012). Within these surveys it was never feasibleto cover the complete DCS, but a comparisonper area is possible (table 3). Estimateddensities and abundances for the wintermonths (combined surveys from October toDecember) were similar for area C in 2008 and2010 (table 3). For area D the estimates in 2009seemed higher than in 2010. One reason mightbe that the difference in timing of the surveys(i.e. November/December versus October/November) could make a big difference in localdensity due to migration. When comparingMarch surveys, densities in the offshore area BGeelhoed et al. / Lutra 2013 56 (1): 45-57 53

were similar in 2010 and 2011, whereas area Cshowed an increase in density in 2011. Densityin area D has been fairly stable in 2009, 2010and 2011 - all estimates lie within the confidenceintervals of each other - but showed apossible decrease from 2010 to 2011. ἀ e totalabundances for areas B, C and D showed similarnumbers (66,328 animals in 2010 and75,682 animals in 2011), suggesting a shift indistribution between these two years, withmore animals present in the central-easternDCS (i.e. area C) in 2011 than in 2010.ReproductionIn recent years, stranding records of harbourporpoises along the Dutch and Belgian coastshowed increasing numbers of neonates in latesummer (e.g. Haelters & C amphuysen 2009).ἀ ese strandings are assumed to reflect birthsin coastal waters. Figure 2 shows the locationswhere calves were seen in Dutch waters, whichwere clearly not restricted to the coastal zone.ἀ e observations of (small) calves in July suggestthat harbour porpoises reproduce in Dutchwaters. Sexually mature female porpoises cangive birth to one calf each year or every secondyear (Gaskin et al. 1974). Given a gestationperiod of 10-11 months (Gaskin et al. 1974,Addink et al. 1995, Lockyer 2003), this meansthat mating will take place shortly after parturition.Hence, areas with calves are also importantreproduction areas. Based on the size ofthe foetus in by-caught porpoises, Börjesson& Read (2003) estimated the mean conceptiondate in the North Sea to be 25 July (± 20.3 days)and Gaskin et al. (1974) predicted a mean birthpeak from the end of May till the end of June.In the German Baltic and North Sea the majorityof births take place in May-July, with thefirst births in March (Hasselmeier et al. 2004).ἀ is is in line with the virtual lack of sightingsof calves in March 2011 and October/November2010, and the higher number of calves seenduring the July 2010 surveys.Management implicationsHarbour porpoises in the Atlantic can bedivided in several populations or managementunits (MUs). Evans et al. (2009) assessedthese for the north-eastern Atlantic. Basedon Danish telemetry data (Sveegaard et al.2011) and other available data (e.g. genetics),they concluded that the North Sea should bedivided into two MUs along an arbitrary linerunning NNW–SSE from northern Scotlandto Germany-Denmark. ἀ e Dutch porpoiseswould belong to the MU south of this line:the south-western North Sea and the easternChannel MU. ἀ e boundaries of thisMU are not well defined, but it lies withinthe SCANS II survey blocks V, U, H a nd B(SCANS 2008). An abundance estimate forthis MU is lacking, but given the smaller areait has to be less than the sum of the estimatesin the survey blocks V, U, H a nd B d uringSCANS II (SCANS, 2008). Based on SCANSII, the estimated number of porpoises in thismanagement unit is less than ca. 180 000 animals.Assuming that the population fromthe south-western North Sea and the easternChannel MU stays within this area throughoutthe year and that the population size didnot change much since 2005, at least 14% ofthis population was present in Dutch nationalwaters in July 2010. In March 2011 this proportionincreased to at least 48%. At thattime ἀ e Netherlands harbour a substantialproportion of the porpoise population in thesouthern North Sea and the eastern Channel.ἀ is emphasises the importance of livingup to the commitment the Netherlands havewithin the EU Habitats Directive to achieveand maintain a favourable conservation statusfor the harbour porpoise in Dutch waters.First steps in this process have been assessingtotal numbers occurring in Dutch waters(this paper) and the development of a DutchHarbour porpoise conservation plan (Camphuysen& Siemensma 2011).54 Geelhoed et al. / Lutra 2013 56 (1): 45-57

ConclusionsBy correcting for biases in the detection probability,this study provides the first un-biasedestimate of the harbour porpoise populationsize in Dutch waters and how it differs betweenthree seasons. ἀ e results show that there is astrong seasonal variation in density and thatthere are areas with higher densities of harbourporpoises in Dutch waters. In March2011, high densities were found in the wholeDCS, except for the southernmost parts anda narrow strip in close proximity of the mainlandcoast. In July, high densities were foundnear the Brown Ridge, Botney Cut-DoggerBank and Borkumer Reef. In October, the distributionseems more spatially homogeneous.Mother-calf pairs were mostly sighted in July,off the coast of the mainland (area D). Sincethese patterns may show (large) variability,repeated surveys will be necessary to ascertainif the established patterns are consistentwithin and between years. ἀ ese surveyscould be primarily aimed at surveying highdensity areas, but as the exact locations ofthese may vary over time, repeated DCS widesurveys should provide the necessary backgroundinformation.As harbour porpoises are a w ide rangingspecies, larger scale multi-season surveys incooperation with adjacent countries wouldallow a b etter understanding of the movementsand habitat use of porpoises in thesouthern North Sea. Using such data in combinationwith an overarching spatial modelincluding environmental parameters, wouldprovide information on distribution and ultimatelyon habitat preferences. ἀ is informationis necessary to develop adequate managementand protection measures in relationto offshore activities for harbour porpoises inthe future.Acknowledgements: First of all we would like to thankthe pilots from Ravenair and Sylt Air for safe and professionalflying, and the aerial observers Martin Baptist,Nicole Janinhoff, Mardik Leopold, Klaus Lucke,Hans Verdaat, and Richard Witte for their survey effort.ἀ ese surveys were conducted under the umbrella ofthe Shortlist Masterplan Wind programme commissionedby Rijkswaterstaat Waterdienst. Okka Jansen,Mardik Leopold and Richard Witte kindly providedinformation. Last but not least, Krissy Reeve improvedthe English of the manuscript.ReferencesArts, F. 2011. Trends en verspreiding van zeevogels enzeezoogdieren op h et Nederlands ContinentaalPlat 1991 - 2010. Rapport BM 11.19. RWS Waterdienst.Lelystad, the Netherlands.Addink, M.J.M., M. Garcia Hartman & C . Smeenk1995. ἀ e harbour porpoise Phocoena phocoena inDutch waters: life history, pathology and historicalrecords. International Whaling Commission,Scientific Committee Document SC/47/SM5: 1-8.Baptist, H.J.M. & P.A. Wolf 1993. Atlas van de vogelsvan het Nederlands Contintentaal Plat. RapportDGW-93.013. Rijkswaterstaat, Dienst Getijdewateren,Middelburg, the Netherlands.Börjesson, P. & A .J. Read 2003. Variation in timingof conception between populations of the harborporpoise. Journal of Mammalogy 84: 948–955.Buckland, S.S.T., D.R. Anderson, K.P. Burnham, J.L.Laake, D.L. Borchers & L. ἀ omas 2001. Introductionto distance sampling. Oxford UniversityPress, Oxford, UK.Buckland, S.S.T., D.R. Anderson, K.P. Burnham, J.L.Laake, D.L. Borchers & L. ἀ omas 2004. Advanceddistance sampling. Oxford University press, NewYork, USA.Camphuysen, C.J. 2004. ἀ e return of the harbourporpoise (Phocoena phocoena) in Dutch coastalwaters. Lutra 47: 113-122.Camphuysen, C.J. 2011. Recent trends and spatial patternsin nearshore sightings of harbour porpoises(Phocoena phocoena) in the Netherlands (SouthernBight, North Sea ), 1990-2010. Lutra 54: 39-47.Camphuysen, C.J. & M.F. Leopold 1994. Atlas of seabirdsin the southern North Sea. IBN Researchreport 94/6 / N IOZ Report 1994-8. Institute forForestry and Nature Research, Netherlands Institutefor Sea Research and Dutch Seabird Group,Geelhoed et al. / Lutra 2013 56 (1): 45-57 55

Texel, the Netherlands.Camphuysen, C.J. & M .L. Siemensma 2011. Conservationplan for the Harbour Porpoise Phocoenaphocoena in ἀ e Netherlands: towards a f avourableconservation status. NIOZ Report 2011-07.Royal Netherlands Institute for Sea Research,Texel, the Netherlands.Evans, P.G.H., P. Anderwald & M . Baines 2003. UKcetacean status review. Sea Watch Foundation,Oxford, UK.Evans, P., L.W. Andersen, A. Bjørge, M. Fontaine, A.Galatius, C.C. Kinze, C. Lockyer, C. De Luna, G.J.Pierce, S. Sveegaard, J. Teilmann, R. Tiedemann& M. Walton 2009. Harbour porpoise Phocoenaphocoena. Report of ASCOBANS/HELCOM smallcetacean population structure workshop, 8–10October 2007. UN Campus, Bonn, Germany.Gaskin, D.E., P.W. Arnold & B.A. Blair 1974. Phocoenaphocoena. Mammalian Species 42: 1-8.Gilles, A., M. Scheidat & U. Siebert 2009. Seasonal distributionof harbour porpoises and possible interferenceof offshore wind farms in the GermanNorth Sea. Marine Ecology Progress Series 383:295-307.Hammond, P.-S., P.P. Berggren, H.H. Benke, D.D.L.Borchers, A.A. Collet, M.M.P. Heide Jorgensen,S.S. Heimlich, A.R. Hiby, M.F. Leopold & N. Oien2002. Abundance of harbour porpoise and othercetaceans in the North Sea and adjacent waters.Journal of Applied Ecology 39: 361-376.Haelters, J. & C .J. Camphuysen 2009. ἀ e harbourporpoise (Phocoena phocoena L.) in the southernNorth Sea: abundance, threats, research andmanagement proposals. Report. Royal BelgiumInstitute of Natural Sciences (RBINS/MUMM) /the Royal Netherlands Institute for Sea Research(NIOZ); project funded by the International Fundfor Animal Welfare (IFAW) - Germany.Haelters, J., F. Kerckhof, T.G. Jacques & S . Degraer2011. Harbour porpoise Phocoena phocoena in theBelgian part of the North Sea: trends in abundanceand distribution. Belgian Journal of Zoology 141:75-84.Hasselmeier, I., K.F. Abt, D. Adelung & U. Siebert 2004.Stranding patterns of harbour porpoises (Phocoenaphocoena) in the German North and BalticSeas: when does the birth period occur? Journal ofCetacean Research and Management 6: 259-263.Jung, J-L., E. Stéphan, M. Louis, E. Alfonsi, C. Liret,F-G. Carpentier & S. Hassani 2009. Harbour porpoises(Phocoena phocoena) in north-westernFrance: aerial survey, opportunistic sightings andstrandings monitoring. Journal of the Marine BiologicalAssociation of the United Kingdom 89:1045-1050.Lockyer, C. 2003. Harbour porpoises (Phocoena phocoena)in the North Atlantic: Biological parameters.NAMMCO Scientific Publications 5: 71-90.MacLeod, C.D., G. J. Pierce & M.B. Santos 2007. Starvationand sandeel consumption in harbour porpoisesin the Scottish North Sea. Biological Letters3: 535–536.SCANS 2008. Small Cetaceans in the EuropeanAtlantic and North Sea. Final report to the EuropeanCommission under project LIFE04NAT/GB/000245. Available from SMRU, Gatty MarineLaboratory, University of St. Andrews, St.Andrews, Fife, UK.Scheidat, M., A. Gilles & U. Siebert 2005. Applying thecircle-back method to estimate g(0)- Experiencesand results from aerial surveys in German waters.In: F. ἀ omsen, F. Ugarte & P.G.H. Evans (eds.).Estimation of g(0) in line-transect surveys of cetaceans.ECS newsletter no 44-special issue: 18-24.Scheidat, M., A. Gilles, K. Kock & U . Siebert 2008.Harbour porpoise Phocoena phocoena abundancein the south-western Baltic Sea. Endangered SpeciesResearch 5: 215-223.Scheidat, M., H. Verdaat & G. Aarts 2012. Using aerialsurveys to estimate density and distribution ofharbour porpoises in Dutch waters. Journal of SeaResearch 69: 1-7.Scheidat, M., R. Leaper, M.J. van den Heuvel Greve &A.J. Winship, in press. Setting maximum mortalitylimits for harbour porpoises in Dutch waters toachieve conservation objectives. Open Journal ofMarine Science 3 (3).Sveegaard, S., J. Teilmann, P. Berggren, K.N. Mouritsen,D. Gillespie & J. Tougaard 2011. Acoustic surveysconfirm the high-density areas of harbourporpoises found by satellite tracking. ICES Journalof Marine Science 68: 929-936.Teilmann, J., S. Sveegaard, R. Dietz, I.K. Petersen, P.Berggren & G. Desportes 2008. High density areas56 Geelhoed et al. / Lutra 2013 56 (1): 45-57

for harbour porpoises in Danish waters. NERITechnical Report 657. National EnvironmentalResearch Institute, University of Aarhus, Denmark.ἀ omsen, F.F., M.M. Laczny & W.W. Piper 2006. Arecovery of harbour porpoises (Phocoena phocoena)in the southern North Sea? A c ase studyoff Eastern Frisia, Germany. Helgoland MarineResearch 60: 189-195.SamenvattingVoorkomen van bruinvissen (Phocoenaphocoena) op het Nederlands ContinentaalPlat: vliegtuigtellingen in de periodejuli 2010 - maart 2011De bruinvis (Phocoena phocoena) is de algemeenstezeezoogdiersoort in Nederlandsewateren. Desondanks waren er tot 2010 geenaantalsschattingen beschikbaar voor hetNederlands Continentaal Plat (NCP). In juli2010, oktober/november 2010 en m aart 2011werden vliegtuigtellingen langs vooraf ontworpentrack lines uitgevoerd waardoor hetmogelijk was dichtheden en a antalsschattingenvan bruinvissen op het NCP te berekenen.De hoogste aantallen werden in maart 2011(n=85.572) gevonden, ongeveer drie keer zoveel als in juli 2010 (n=25.998) en in oktober/november 2010 (n=29.963). Het verspreidingspatroonverschilde per telperiode, maar gedurendealle telperioden waren hogere dichthedenaanwezig in een strook tussen de BruineBank en de Borkumse Stenen. In juli werdenkalfjes gezien, hetgeen een indicatie vormtvoor het feit dat dat bruinvissen zich in Nederlandsewateren voortplanten. De aantalsschattingvoor maart correspondeert met 48% vande populatie in de zuidelijke Noordzee; eengroot deel van de Noordzeepopulatie verblijftdaarmee in die periode in Nederlandse wateren.De gevonden hoge dichtheden kunnenleiden tot een toename in conflicten met menselijkeactiviteiten. Het nemen van beleidsmaatregelenom deze conflicten te voorkomenof te mitigeren wordt hierdoor urgent.Received: 20 January 2013Accepted: 25 April 2013Geelhoed et al. / Lutra 2013 56 (1): 45-57 57

Harbour porpoises (Phocoena phocoena) in theMarsdiep area, the Netherlands: new investigationsin a historical study areaMichelle Boonstra 1,3 , Yvonne Radstake 2,4 , Kees Rebel 5 , Geert Aarts 1,2 &Kees (C.J.) Camphuysen 1,*1Royal Netherlands Institute for Sea Research (NIOZ), P.O. Box 59, NL-1790 AB Den Burg, Texel, the Netherlands,e-mail: kees.camphuysen@nioz.nl2IMARES Wageningen UR, P.O. Box 167, NL-1790 AD Den Burg, Texel, the Netherlands3University of Amsterdam, Biological Sciences, P.O. Box 94214, NL-1090 GE Amsterdam, the Netherlands4University of Utrecht, Environmental Biology, P.O. Box 80056, NL-3508 TB Utrecht, the Netherlands5Seawatchers Huisduinen, De Zeis 27, NL-8091 NK Wezep, the NetherlandsAbstract: ἀ e harbour porpoise (Phocoena phocoena) returned in Dutch waters in the late 20 th century after a nearabsenceof approximately three decades. Inspired by historical studies of harbour porpoises in the Marsdiep area(western Wadden Sea), mainly in the 1930s and 1940s, a study was initiated in the same area in 2011 to see if porpoisesare now as common as previously recorded and if future, more detailed investigations would be worthwhile.We investigated the current spatial and temporal distribution of harbour porpoises using a combination of visualobservations and acoustic recordings. In 2010 and 2011, harbour porpoises were most abundant around mid-March,and disappeared in April. ἀ ey were most frequently observed off Texel, with slightly lower numbers of visual detectionsalong the dike of Den Helder and at Huisduinen. Relatively high abundances were recorded around high tide atmost observation sites, but particularly so off Texel. ἀ e observed abundance of porpoises in spring in the Marsdieparea, and the apparent influence of tidal currents on spatial patterns in sightings frequencies, is such that the Marsdieparea could provide rather unique, future study opportunities for harbour porpoises in the wild.Keywords: Cetacea, Phocoena phocoena, Marsdiep, Wadden Sea, tidal rhythm, foraging habitats, sightings, recordings,history, Netherlands.Introductionἀ e harbour porpoise (Phocoena phocoena)returned as an abundant, indigenous speciesin Dutch waters in the late 20 th century,after a virtual absence of three decades (1960-1990; Camphuysen 2004, Camphuysen & Peet2006, Camphuysen 2011, Scheidat et al. 2012).Up to the late 1950s, harbour porpoises werecommon in Dutch coastal waters. ἀ e only© 2013 Zoogdiervereniging. Lutra articles also on theinternet: http://www.zoogdiervereniging.nlarea where studies were conducted prior tothe disappearance from Dutch waters wasthe Marsdiep, a sea inlet between Den Helderand Texel in the western Wadden Sea. Verwey(1975) reported sightings, documentedbehaviour and prey preferences and describedseasonal patterns of cetaceans in the Marsdieparea between 1931 and the early 1970s.ἀ e Marsdiep area is unique in the sensethat the entire area can be overseen from land,and that its complex bathymetry and oceanographyoffers a rare opportunity to study wildcetaceans, undisturbed, in a variety of marineBoonstra et al. / Lutra 2013 56 (1): 59-71 59

habitats. Due to its proximity to the RoyalNetherlands Institute of Sea Research (NIOZ),in and around the Marsdiep, several oceanographicand biological sampling programs areestablished. ἀ e historical and detailed dataseton the occurrence of harbour porpoises andcommon bottlenose dolphins (Tursiops truncatus)collected by the Zoological Institute justprior to and following the Second World Warserved as a reference (Verwey 1975).Following the return of harbour porpoisesin the early 21 st century, we felt that a resumptionof studies in the Marsdiep area (using modernobservation techniques) could yield muchneeded information on habitat characteristics,seasonality and behaviour of harbour porpoisesin Dutch coastal waters. We designed a feasibilitystudy which included a systematic porpoiseobservation programme. A t eam of observersmanned several, strategically chosen observationsites to assess spatial patterns in harbourporpoise abundance within the tidal inlet (visualobservations). ἀ e long-term seawatchingprogramme at Huisduinen was continued andintensified to provide a recent background ofseasonality in the area. Acoustic techniqueswere deployed to supplement the informationderived from sightings. ἀ e objective of thisstudy was to investigate if harbour porpoiseswithin the Marsdiep area are currently abundantenough to warrant a more complete ecologicalstudy in future years. We collected sightings(with corrections for observer effort andenvironmental conditions) during the periodof peak abundance in Dutch coastal waters andbeyond (February-April 2011), and documentedsite-specific patterns in abundance togetherwith aspects such as tidal phase, time of day,shipping traffic, presence and abundance ofseals, and the occurrence of multi-species feedingfrenzies of seabirds and marine mammals.Historical abundanceVerwey’s (1975) studies comprised descriptionsof sightings and of the behaviour of theanimals observed in the period 1931-1940 and1945-1973 along the shores of the Marsdiepnear Den Helder, with numbers being muchhigher in the earlier period. Jan Verwey andco-workers from the Zoological Station inDen Helder (now NIOZ, Texel) reconstructedthe seasonality in occurrence of harbour porpoises,and described interactions with commonbottlenose dolphins and their (presumedshared) prey (Verwey 1975). ἀ e animals wereobserved mostly during bicycle rides alongthe dike, from the ferry to Texel, or occasionallyfrom the institute’s research vessel, i.e.they were opportunistic sightings rather thansystematic observations. ἀ e last sightingsof significance were reported by Dudok vanHeel (1960), who recorded 40-50 “Phocaena”in Texelstroom, the sea inlet just to the southwestof the Marsdiep area. Since then, fewsightings were recorded until the late 1990s(Camphuysen 2004). ἀ e information on seasonaltrends was only based on data gatheredbetween 1934 and 1939, when porpoises werestill abundant. ἀ e more systematic observationswere made between the harbour of DenHelder and the lighthouse of Huisduinen(along a d ike), and were given as “monthlytotals”, where a month comprised circa twelveobservation days (during fine weather). Fromthese historical data and other publicationswe could summarise the historical abundanceof harbour porpoises in Dutch coastal water(with emphasis on the Marsdiep area) as follows:Before World War II, harbour porpoiseswere common in Dutch coastal waters (Weber1922, van Deinse 1925). Numbers in the Marsdieparea were very low from February/Marchto May, increased in early summer (May orJune-July) and rather high numbers were seenin winter (November-December to January-February; once in March; Verwey 1975). Verwey’sobservations contradict descriptions ofthe seasonal abundance of harbour porpoisesin nearshore waters along the mainland coastof the Netherlands by other authors, whoreferred to the summer as a p eriod with the60 Boonstra et al. / Lutra 2013 56 (1): 59-71

highest numbers (Heinsius 1914, van Deinse1925). IJsseling & Scheygrond (1943) notedthat most strandings in ἀ e Netherlandsoccurred in the summer months (notably inAugust), and many stranded animals wereapparently neonates (newborns; van Deinse1925). Harbour porpoises in the Marsdieparea were assumed to follow the tidal current(entering during flood, leaving during ebb;Verwey 1975). ἀ ese suggestions were supportedby only two observations. Some porpoiseswere observed in extremely shallowwaters in the Marsdiep area: in small creaks,and over sand banks at high tide that wereexposed at low tide. In 1939, the first signs ofa decline in numbers were noted (Viergever1955). During World War II, it was impossibleto collect data, but directly following the warnumbers seemed to have declined even further.Verwey (1975) stated that this decreasetook place around 1945, but the declinebecame more obvious in the 1960s. Factualdata are scarce, unfortunately, and assessmentsof (effort corrected) trends in relativeabundance cannot be reconstructed from thismaterial.Methodsἀ e Marsdiep forms an inlet between theWadden Sea and the North Sea, throughwhich half of the total body of water in theDutch Wadden Sea passes twice a day, equatingto one billion m 3 of water (Zimmerman1978). ἀ is movement has an average flow rateof 1 m.s -1 . ἀ is dynamic estuary has a highlyvariable bathymetry with depths rangingfrom 1 up to 45 meters (figure 1). ἀ e tide inthis area can be classified as a s emi-diurnalpattern with a cycle of 12.25 hours (Dron kers1964). ἀ e tide is asymmetric in this area witha relatively long-lasting high tide. ἀ ese hightides can consist of double-headed high tidesor a so called “agger”, i.e. a high tide in whichthe water rises to a c ertain level, recedesslightly, and then rises again. ἀ is phenomenonis mainly caused by non-linear effectslike friction and self-advection (Zimmerman1976).Observations 1981-2010A recent set of data, uploaded almost daily,was obtained from the database of the DutchSeabird group (NZG/CVZ; currently managedby Seawatchers, ahighly specialised and well trained group ofamateur ornithologists, conducted systematic,year-round observations from vantagepoints along the Dutch coast (Camphuysen1985, Camphuysen 2011). Although theirmain focus was the occurrence of true seabirdsand migratory wildfowl and waders,they have always recorded marine mammals.ἀ ese observations provide a long-term datasetfrom which long-term trends and seasonalpatterns can be analysed (Camphuysen2011). For this paper, we restrict our review todata collected at the seawatching site at Huisduinen(HD, figure 1) during 1981–2011. Harbourporpoises occur year-round in Dutchcoastal waters, but with a d istinct peak inJanuary-March (Camphuysen 2011). For theHuisduinen data, only this period has beenanalysed. Seawatchers recorded the date,weather, and duration of the counts (startandend-time), and reported sightings perhour of observation (n.hour -1 ).Observations in 2011Standard seawatches from observation siteHuisduinen (HD), following the exact samemethods as in previous seasons, were continuedto provide comparable data with earlieryears. Observation teams were stationedat three additional locations: at Texel on theWadden Sea dike near the NIOZ (ND), andnear the southern tip of the island near a navalbase (MB) (figure 1). In addition, observationswere conducted from the dike at Den HelderBoonstra et al. / Lutra 2013 56 (1): 59-71 61

Figure 1. Map of the study area including bathymetry and locations of C-PODs (green and yellow). Standard seawatcheswere conducted from Huisduinen (HD). Further observation sites included the Wadden Sea dike nearNIOZ at Texel (ND), the southern tip of Texel near a naval base (MB), and the dike at Den Helder (“Vlettenhelling”;VH). Green refers to the position of the first C-POD (#1482), attached to a green pole, yellow refers to theposition of the second C-POD (#1481) attached to a yellow pole.(Vlettenhelling; VH), some two km furthereast into the Wadden Sea relative to the Huisduinenseawatching site. Observations weremade between 20 February and 20 April 2011(516 hours of observation). However, datafrom April were excluded from the analysisbecause harbour porpoises were very scarcein that period (81 hours, 5 sightings). Observationswere generally done by three observers,with one observer always at Huisduinenand the other two at any of the other observationsites. Effort was logged in 15 minuteintervals, recording site, date, time, observer,meteorological conditions (wind directionand speed, cloud cover, visibility, precipitation),sea state, tide and any visual oceanographicfeatures, such as fronts and tidal ripples.For each sighting, real time, distance and62 Boonstra et al. / Lutra 2013 56 (1): 59-71

Table 1. O bserver effort (h) f or visual observations in t he Marsdiep area, February-April 2011. F or the moredetailed analysis of sightings, only the observer effort in bold was used: excluding data collected in April or whenvisibility was poor (see text); see figure 1 for site locations.Location Abbreviation Feb Mar Apr TotalHuisduinen HD 54 170 34 258Dike Den Helder VH 2 62 15 79NIOZ-dike Texel ND 10 61 13 84Naval base Texel MB 0 63 19 82Poor visibility 3 10 0 13Total effort (sum) 69 366 81 516angle were noted. To assess distance to theobserver, 7x50 reticle binoculars were used.ἀ e reticles were converted to meters followingBuckland et al. (1993).Details about group composition, swimmingdirection and association with birdsor oceanographic features were recorded foreach sighting. Apart from harbour porpoises,sightings of seals (Phocidae; not identified tospecies) and vessels were recorded, under theassumption that high traffic densities and highseal abundances (or the absence of shippingand/or seals as the other extreme) could influencethe behaviour and relative abundance ofporpoises within the study area.Weather permitting, six hours of observationper day were planned, with observereffort being equally distributed over the timeof day and over the tide, to account for obviousfactors that may affect the abundanceand behaviour of harbour porpoises. Duringrough weather (high sea state), scheduled visualobservations were cancelled. Data of observationsconducted in poor visibility (less than1000 m) were excluded from the analysis (effortand sightings; in total 13 h observation and sixsightings; table 1). Data collected in April, whenharbour porpoises appeared to be very scarce,were excluded from the analysis of factorsassociated with high abundances (a removalof 81 hours of observation, resulting into onlyfive sightings). Information on tide and waterlevel were obtained from the DONAR (DataOpslag Natte Rijkswaterstaatdatabase) madeavailable by the Dutch Ministry of Infrastructureand the Environment at Time to high-tide was classified into fourclasses, with break points at high and low tideand exactly in between low and high tide. ἀ iswas used to analyse the relation between porpoiseabundance, behaviour and tide.Passive acoustic monitoringBecause of their elusive nature, harbour porpoisesare easily overlooked, particularly inwindier conditions with rough seas. ἀ ereforewe used, in addition to our visual observations,passive acoustic monitoring devices, so-calledContinuous Porpoise Detectors (hereaftercalled C-PODs) at two sites; one situated closeto the NIOZ dike (53°00’N, 04°47’E) and theother one on the other side of the bay, closeto the naval base (52°99’N, 04°77’E; figure 1).C-PODs record frequencies of 20 to 160 kHzand pick up harbour porpoise clicks. ἀ e clicktrain filter developed by Tregenza (2011) wasused to extract click trains from backgroundnoise. Porpoise click frequency is very high,between 100 and 160 kHz, which means thatthe sound propagation range is relatively small(Au 2000). Hence, the range of a C-POD is onlyaround 250 meters horizontally, depending onBoonstra et al. / Lutra 2013 56 (1): 59-71 63

Observed harbour porpoises per hour of observation3.532.521.510.50198119821983sea conditions, but also on characteristics ofthe harbour porpoise clicks and orientationof the individual relative to the C-POD. ἀ ebeam width of the sound emitted by porpoisesis only about 12° (Au et al. 1999), hence a substantialnumber of porpoises in proximity ofthe C-POD may remain undetected.To increase buoyancy, a b uoy was gluedto the C-POD. ἀ e C-POD was attached toan anchor with a 1 m l ong rope, so that thedevice was not in contact with the sea floor, inorder to minimize the influence of sedimentnoise as much as possible. Once every fewweeks, the SD memory cards of the C-PODswere replaced. ἀ e data on the memory cardswas processed in the computer programmeCPOD.exe, V2.012 (Tregenza 2011). From 8February 2011 onwards, CPOD #1482 (“greenpole”) was deployed, and from the 17 February2011 onwards a s econd CPOD (#1481,“yellow pole”) was used (figure 1). Deploymentdepth during low tide was at approximately9 and 14 m, respectively. Data recordedbetween 22 February and 7 A pril were usedfor comparison with our visual observations.198419851986198719881989199019911992199319941995199619971998199920002001In the software programme the trains ofclicks were classified as low, moderate or highquality clicks on the basis of their probabilityto be a porpoise click (Tregenza 2011). In theanalysis, only click trains of high and moderatequality were used. Low quality click trainswere left out, because these were consideredto include many false positive detections. ἀ einput high pass filter was kept at the normalsetting (20 kHz) and the train filter was set ondetecting only NBHF cetaceans (thus ‘othercetacean’ was not ticked). For the data analysis,the range was set at 117 to 155 kHz. ἀ enumber of minutes per day during which porpoiseclicks where detected (porpoise positiveminutes – PPM) was used to describe the temporaltrend in porpoise occurrence.Results2002200320042005200620072008200920102011Figure 2. Harbour porpoises (n.h-1) in spring (January-March) at Huisduinen, based on systematic seawatchingdata, 1981-2011 (database NZG/Club van Zeetrekwaarnemers, and unpublished annual reportsVogelwerkgroep Den Helder, 1996-2005).Seawatching data Huisduinen 1981-2011During systematic seawatching observations inJan-Mar at Huisduinen between 1997 and 2011,64 Boonstra et al. / Lutra 2013 56 (1): 59-71

Observed harbour porpoises per hour of observation3.532.521.510.501Jan -7Jan* * * * *8Jan -14Jan15Jan -21Jan22Jan -28Jan29Jan -4Feb5Feb -11Feb12Feb -18Feb19Feb -25Feb26Feb -4Mar5Mar- 11Mar12Mar- 18Mar19Mar- 25Mar26Mar- 1Apr2Apr- 8Apr2010 2011Figure 3. Weekly observed harbour porpoises (n.h -1 ) at Huisduinen, January to mid-May, 2010 and 2011 ( * no observation effort9Apr- 15Apr16Apr- 22Apr23Apr- 29Apr30Apr- 6May7May- 13May1458 porpoises were recorded in 2400 hours ofobservation (mean 0.61 porpoises.h -1 ); whilenone were seen and recorded in earlier years(1981-96; figure 2). ἀ e overall mean sightingfrequencies in late winter did not exceed0.20.h -1 until 2004. In 2006, sightings peaked at3.25.h -1 , and from then on did not drop below1.h -1 , with the exception of 2008 and 2009,when 0.25.h -1 were recorded. Spring 2011 wasthe season with the second most observationsever, with slightly higher numbers than in2007 and 2010, but less than half the number ofporpoises per hour as in 2006. In the first fivemonths of both 2010 and 2011, sightings (n.h -1 )were most regular between late February andlate March (figure 3). In both years, numbersdeclined markedly in April.Marsdiep visual observations in 2011Between 20 February and the end of March2011, 605 harbour porpoises were recordedduring more than 400 hours of observation(poor visibility conditions excluded; 1.53.h -1 ).A first peak in sighting frequencies wasrecorded on 23 February, a second peak on 5-7March, and a third period of relatively highnumbers of sightings lasted from 12-30 March(figure 4). Very few animals were observed inApril, after a marked decline in sighting frequenciesduring the last week of March. Harbourporpoises were most abundant off thenaval base at Texel (MB, 2.4.h -1 ), followed bythe Wadden Sea dike at Texel (ND, 1.8.h -1 ),seawatching site Huisduinen (HD, 1.3.h -1 ),and the dike at Den Helder (VH, 0.75.h -1 ). ἀ emean pod size (±sd) was 1.31±0.59 per sighting(range=1-5).Particular phases of the tidal cycle producedmore sightings than others. Relativelyhigh abundances were recorded around hightide (late flood and early ebb; figure 5). Whena distinction is made between ebb and flood,higher abundances were found during flood.ἀ is pattern was found at most sites, but mostclearly at the naval base. Along the dike ofDen Helder near VH, however, a c ompletedifferent pattern was found: here the highestabundance was observed directly afterBoonstra et al. / Lutra 2013 56 (1): 59-71 65

10004.5Porpoise positive minutes per day900800700600500400300200100PPM-1482 (green)PPM-1481 (yellow)Visual observation43.532.521.510.5022/02/201125/02/201128/02/201103/03/201106/03/201109/03/201112/03/201115/03/201118/03/201121/03/201124/03/2011Observed harbour porpoises per hour of observation27/03/201130/03/201102/04/201105/04/20110Figure 4. Acoustic detections of harbour porpoises (porpoise positive minutes per day; PPM) based on continuousrecordings of two C-PODs in the Marsdiep area and visual detections during daytime (mean n.h -1 , indicatedby •) at all observation sites combined (animals per hour of observation). Estimates of porpoise sighting rates ondays during which no observations were possible (bad weather or otherwise), are based on linear interpolationsbetween the observation days.low tide, during early flood, after which itdecreased until the next early flood. ἀ e visualobservations included impressions of thebehaviour of porpoises. Some porpoises wereaccompanied with searching and plunge divinggulls, but by far most animals were seenaway from the most prominent feeding frenziesof seabirds.Acoustic observations in 2011ἀ e number of porpoise positive minutesrecorded per day by the C-PODs during thestudy period show a similar seasonal patternas that obtained by visual observations (figure4), but some disagreements in recordings(high peak in sightings with few acousticrecordings or vice versa) occurred on 23February (many sightings), 26-27 February(frequent acoustic recordings) and around6 March (many sightings). Daily activity indicesderived from both C-PODs (using daysthat visual observations were conducted inthe Marsdiep area) were positively correlated(Pearson’s correlation based on log PPMper day: r²= 0.67, df=43, t=5.9, P

Observed harbour porpoises per hour of observation4. Early floodHD Late floodHD Early ebbHD Late ebbVH Early floodVH Late floodVH Early ebbVH Late ebbMB Early floodMB Late floodMB Early ebbMB Late ebbObservation site and tidal phaseND Early floodND Late floodND Early ebbND Late ebbFigure 5. Visual detections of harbour porpoises (n.h -1 ) at each observation site relative to the tidal almost every location a r elation betweenabundance and tide was found, the effect wasmost clear off the naval base (MB), where theabundance was highest around high tide. Atthe constant effort-site Huisduinen, the abundanceof porpoises in spring 2011 (1.53.h -1 )was similar to 2007 and 2010, but higher thanin 2008 and 2009 (mean 1.6.h -1 for the sameobservation period at Huisduinen, 2006-2010). From the perspective of our feasibilitystudy, harbour porpoises were certainlynumerous enough in early spring to have theMarsdiep serve as a study area for harbourporpoises in future years.ἀ e higher sighting frequencies at the Texelside, particularly near the naval base, suggestthat porpoises have a preference for thisside of the Marsdiep. Alternatively, the morefrequent sightings may have been caused bylarger numbers of semi-resident animals(temporarily present on the spot; numerousre-sightings; movements within a p articulararea). Along the dike of Den Helder, more animalsmay have been just passing by and wouldhave a lower chance to be recorded over andover again. Directed movements (all animalsmoving in the same direction) are commonat Huisduinen, but on days with particularlyhigh numbers of sightings, local movements(producing double-counts) prevailed.Visual vs. acoustic observationsἀ e abundance of harbour porpoises in theMarsdiep area during our study period washighly variable (figure 4), and the causes offluctuations in numerical abundance are currentlynot well understood. ἀ e peak in sightingson 6-7 March consisted mainly of sightingsat one observation site. ἀ e abrupt endof sightings in April was, like in other years,very distinct, and seems to point at a g eneraldeparture towards deeper waters of theNorth Sea (cf. Camphuysen 2011). In contrast,recent coastal observations in 2013 suggest ahigh abundance of harbour porpoises in April( peaks in sightings were not reflectedin the acoustic data. Because acoustic recordingwere deployed during the same period inthe same area, we were able to compare theBoonstra et al. / Lutra 2013 56 (1): 59-71 67

acoustic recordings with indices of abundancerecorded during visual observations.ἀ e acoustic data was expressed in porpoisepositive minutes recorded per day. Porpoisesuse echolocation clicks to locate prey and tonavigate, but when travelling, less clicks perminute are produced than when searchingand hunting prey (Verfuβ et al. 2009). Hence,fluctuations in numbers of clicks does notnecessarily represent variations in numberof porpoises. However, PPM does not differentiatebetween variations in click frequency,and is therefore a m ore robust estimate oflocal occurrence. ἀ e relative strong and significantcorrelation between the visual sightingsand C-POD recordings indicates that theacoustic recordings can indeed be used as aproxy of local porpoise abundance.Both the acoustic and visual observationsreveal a s udden disappearance in April. ἀ ehistorical seasonal pattern in abundance(Verwey 1975) was compared with the seasonalityobserved today (Camphuysen 2011). Verweydid not describe a sudden disappearanceof porpoises in spring, nor a distinct peak seasonin late winter. In the old days, numbersgradually declined through May, June, andJuly. Interestingly, most recent observationsin 2013 suggest a h igh abundance in April.ἀ is may be a single year event caused by coldweather conditions in March, or may be a firstsign of a forward seasonal shift.Competition and preyἀ e decline in numbers of porpoises in latewinter described by Verwey (1975) may havebeen linked with the arrival of common bottlenosedolphins. Common bottlenose dolphinsentered the Wadden Sea in early springand in considerable numbers (ter Pelkwijk1937, Viergever 1940), chasing the migratingZuiderzee herring (Clupea harengus) (Verwey1975). ἀ is annual event coincided with aseasonal decline in porpoise numbers. Bottlenosedolphins can be aggressive towards andeven kill harbour porpoises (Ross & Wilson1996), and their arrival could have been themain reason for porpoises to avoid the areaduring this period. ἀ e bottlenose dolphinshave disappeared from the Marsdiep area anddid so far not return (Camphuysen & Peet2006). ἀ e abrupt decline in numbers in Aprilas observed during this study can thereforenot be explained by (aggressive) encounterswith larger cetaceans.According to Johnston et al. (2005), a strongrelation with the distribution of their mainprey species is an essential part of the harbourporpoise life-history. A link between theabundance of whiting (Merlangius merlangus)in the North Sea and the abundance ofporpoises in late winter, as proposed by Verwey(1975), may still be valid today. Whiting,and also gobies (Pomatoschistus spp. - acommon prey of immature porpoises in latewinter), form the major part of adult harbourporpoise’s diet in terms of mass and energy(M.F. Leopold, personal communication). ἀ emovements of harbour porpoises in this partof the North Sea can be related to the migrationof some important prey species and arelation between the presence of porpoises inthe Marsdiep area and the spring abundanceof gobies has been suggested. ἀ e annualmigration of gobies was described by Fonds(1973). ἀ e harbour porpoise is a r elativelysmall, endothermic predator with limitedenergy storage capacity, dependent on foragingthroughout the year without prolongedperiods of fasting (Kastelein et al. 1997, Bjørge2003). Future work in the Marsdiep area willneed to focus at the foraging ecology of porpoisesentering this tidal inlet to find a potentiallink with seasonal fluctuations of availableprey.Tide and foraging opportunitiesἀ e bathymetric composition of the site interactingwith the tidal current could induceupwelling or relative high vorticity, enhanc-68 Boonstra et al. / Lutra 2013 56 (1): 59-71

ing particle transports and food availabilityin the water column (Mann & L azier 1996).Such features could enhance the foragingopportunities for harbour porpoises (Jovallanos& Gaskin 1983, Johnston et al. 2005).If foraging opportunities off the naval basisdid increase around high tide, movements ofporpoises should be directed towards the areaduring flooding. Porpoises were in fact rarelyseen to join these feeding frenzies in the turbulentareas around the shallows off the navalbase. ἀ eir frequent occurrence in the areasuggests other foraging opportunities thatmust receive more attention in future studies.Our results suggest that individual porpoisesdo not stay in the Marsdiep for long periods,but rather move in and out on a more or lessdaily basis.ConclusionsIn conclusion, the Marsdiep and its springharbour porpoise population, provide excellentopportunities to study these elusive cetaceansin their natural habitat in the wild.During this pilot study, with relatively simpletechniques, a considerable amount of newinformation on temporal and spatial aspectsof the occurrence of porpoises in one of themost spectacular tidal inlets in the WaddenSea was collected. ἀ e study opportunities offthe naval basis at Texel are the most promising.ἀ e area is complex in its oceanographyand bathymetry, and serves as a f oragingarea for several piscivorous top-predators.Advanced observation techniques may haveto be deployed to pinpoint the exact positionsof porpoises in that area. Such exact locationsof sightings will then be combined with highresolution data on local bathymetry, currentsand vorticity, water properties such as salinity,turbidity and temperature, and hopefullyalso with the availability of suitable prey, inorder to achieve a better understanding of theecology of harbour porpoises in this dynamicsea area.Acknowledgements: We would like to thank JeroenHoekendijk for his company and help during observationsand the Royal Navy for kindly granted access totheir properties. ἀi s project was initiated by the RoyalNetherlands Institute for Sea Research (NIOZ) and theInstitute for Marine Resources and Ecosystem Studies(IMARES). ἀr ee anonymous referees and the editorialboard of Lutra kindly commented on an earlierdraft of this paper.ReferencesAu, W.W.L. 2000. Echolocation in dolphins. In:W.W.L. Au, A.N. Popper & R.R. Fay (eds.): Hearingby Whales and Dolphins: 364-408. Springer,Berlin, Germany / New York, USA.Au, W.W.L., R.A. Kastelein, T. Rippe & N.M. Schooneman1999. Transmission beam pattern and echolocationsignals of a h arbor porpoise (Phocoenaphocoena). Journal of the Acoustical Society ofAmerica 106: 3699-3705.Bjørge, A. 2003. ἀ e Harbour Porpoise (Phocoenaphocoena) in the North Atlantic: Variability inhabitat use, trophic ecology and contaminantexposure. In: T. Haug, G. Desportes, G.A. Víkingsson& L. Witting (eds.): Harbour Porpoises inthe North Atlantic: 223-228. NAMMCO ScientificPublications Volume 5. North Atlantic MarineMammal Commission, Tromsø, Norway.Buckland, S.T., D.R. Anderson, K.P. Burnham. & J.L.Laake 1993. Distance sampling: estimating abundanceof biological populations. Chapman andHall, New York, USA / London, UK.Camphuysen, C.J. 1985. Zeetrektellingen. In: M.F.H.Hustings, R.G.M. Kwak, P.F.M. Opdam &M.J.S.M. Reijnen (eds.): Vogelinventarisatie: 215-219. Pudoc, Wageningen, the Netherlands.Camphuysen, C.J. 2004. ἀ e return of the harbourporpoise (Phocoena phocoena) in Dutch coastalwaters. Lutra 47: 113-122.Camphuysen, C.J. & G . Peet 2006. Whales and dolphinsof the North Sea. Fontaine Uitgevers,Kortenhoef, the Netherlands.Camphuysen C.J. 2011. Recent trends and spatial patternsin nearshore sightings of harbour porpoises(Phocoena phocoena) in the Netherlands (South-Boonstra et al. / Lutra 2013 56 (1): 59-71 69

ern Bight, North Sea), 1990-2010. Lutra 54 (1):37-44.Dronkers, J.J. 1964. Tidal computations. North HollandPublishing, Amsterdam, the Netherlands.Dudok van Heel, W.H. 1960. Aanvulling op het walvisnieuws1958. Lutra 2 (1): 11-12.Fonds, M. 1973. Sand gobies in the Dutch Wadden Sea(Pomatoschistus, Gobiidae, Pisces). PhD thesis.Vrije Universiteit Amsterdam & NIOZ, Texel. E.J.Brill, Leiden, the Netherlands.Heinsius, H.W. 1914. Langs de Zuiderzee. De LevendeNatuur 18: 282-286.IJsseling, M.A. & A. Scheygrond 1943. De Zoogdierenvan Nederland II. ἀ ieme, Zutphen, the Netherlands.Johnston, D.W., A.J. Westgate & A .J. Read 2005. ἀ eeffects of fine-scale oceanographic features on thedistribution and movement of harbour porpoisesPhocoena phocoena in the Bay of fundy. MarineEcology Progress Series 295: 279-293.Jovallanos, C.L. & D .E. Gaskin 1983. Predicting themovements of juvenile Atlantic herring (Clupeaharengus harengus) in the SW Bay of Fundy usingcomputer simulation techniques. Canadian Journalof Fishery and Aquatic Science 40: 139–146.Kastelein, R.A., J. Hardeman & H . Boer 1997. Foodconsumption and body weight of Harbour Porpoises(Phocoena phocoena). In: A.J. Read, P.R.Wiepkema & P.E. Nachtigall (eds.): ἀ e biology ofthe Harbour Porpoise: 217-233. De Spil Publishers,Woerden, the Netherlands.Mann, K.H. & J.R.N. Lazier 1996. Dynamics of marineecosytems. Biological–physical interactions in theoceans. Blackwell Science, Malden, MA, USA.Ross, H.M. & B . Wilson 1996. Violent interactionsbetween bottlenose dolphins and harbour porpoises.Proceedings of the Royal Society LondonB 263: 283-286.Scheidat, M., H. Verdaat & G. Aarts 2012. Using aerialsurveys to estimate density and distribution ofharbour porpoises in Dutch waters. Journal of SeaResearch 69: 1-7.ter Pelkwijk, J.J. 1937. Dartele tuimelaars. Amoeba 16(4): 65-66.Tregenza, N. 2011 Cetacean monitoring. URL:; viewed May2013.van Deinse, A.B. 1925. De bruinvisch. De LevendeNatuur 29: 195-203.Verfuβ, U.K., L.A. Miller, P.K.D. Pilz & H.U. Schnitzler2009. Echolocation by two foraging harbour porpoises(Phocoena phocoena). ἀ e Journal of ExperimentalBiology 212: 823-834.Verwey, J. 1975. ἀ e cetaceans Phocoena phocoena andTursiops truncatus in the Marsdiep area (DutchWadden Sea) in the years 1931-1973, part 1 & 2 .Publicaties & Verslagen Nederlands Instituut voorOnderzoek der Zee 17a en 17b. NIOZ, Den Burg,the Netherlands.Viergever, J. 1940. Van bruinvissen, tuimelaars en dolfijnen.Amoeba 19 (12): 184-187.Viergever, J. 1955. Hoe staat het met de bruinvis? HetZeepaard 15: 90-91.Weber, M. 1922. Cetaceeën. In: H.D. Redeke (ed.):Flora en Fauna der Zuiderzee. Monografie van eenbrakwatergebied: 442-451. Nederlandsche DierkundigeVereeniging, Den Helder, the Netherlands.Zimmerman, J.T.F. 1976. Mixing and flushing of tidalembayments in the western Dutch Wadden SeaPart I: Distribution of salinity and calculationsof mixing time scales. Netherlands Journal of SeaResearch 10 (2): 149-191.Zimmerman, J.T.F. 1978. De Waddenzee, getijden engetijstromen. Natuur en Techniek 46 (4): 236-249.SamenvattingBruinvissen (Phocoena phocoena) in hetMarsdiep: nieuw onderzoek in een historischonderzoeksgebiedDe bruinvis (Phocoena phocoena) keerde naeen periode van afwezigheid van ongeveerdrie decennia aan het eind van de vorige eeuwterug in de Nederlandse kustwateren. Bruinvissenzijn hier momenteel het meest aanwezigin de winter en het vroege voorjaar (december-maart).In het Marsdiep, de doorgang vande Noordzee naar de westelijke Waddenzeetussen Den Helder en het eiland Texel, werdde bruinvis in het midden van de 20 ste eeuwbestudeerd. Dit onderzoek moest echter wor-70 Boonstra et al. / Lutra 2013 56 (1): 59-71

den gestopt wegens het verdwijnen van debruinvis. Na de terugkeer van bruinvissen inde Nederlandse kustwateren is er in 2011 eenhaalbaarheidsonderzoek uitgevoerd om nate gaan of het opnieuw de moeite waard zoukunnen zijn om bruinvissen in het Marsdiepte bestuderen. De aantallen bruinvissen die inhet voorjaar van 2011 werden gezien warengroot, maar weken niet wezenlijk af van water, op basis van zeetrektellingen bij Huisduinen,aan aantallen bruinvissen in de afgelopenvier seizoenen werd gezien. De meestedieren werden waargenomen rond middenmaarten z e verdwenen in april. In vergelijkingmet het historische voorkomen lijkt dezepiek in voorkomen twee tot drie maandenverschoven in de tijd. Waargenomen aantallenwaren het hoogst aan de Texelse kant vanhet Marsdiep, vooral in de monding van deMokbaai ter hoogte van de marinebasis opde zuidpunt van Texel. Op bijna alle observatiepostenwerden relatief veel waarnemingengedaan rondom hoog water, bij de marinebasiswas dat het meest uitgesproken. Dewaarnemingen suggereren dat, vooral tijdenshoogwater, deze plaats zuidelijk van deMokbaai een voor bruinvissen aantrekkelijkgebied vormt, wellicht om te foerageren of omhydrodynamische redenen. De door Verweywaargenomen verplaatsing van bruinvissen“met het getijde mee” kon tijdens ons onderzoekniet worden bevestigd. Wij concluderendat bij gelijkblijvende (of toenemende) aantallenbruinvissen in het gebied, het Marsdiepeen uniek en z eer geschikt studiegebied zoukunnen zijn om bruinvissen ongestoord en inhet wild te onderzoeken.Received: 18 March 2012Accepted:13 May 2013Boonstra et al. / Lutra 2013 56 (1): 59-71 71

More magazines by this user
Similar magazines