Niche overlap of mountain hare subspecies and the vulnerability of their ranges to invasion by the European hare; the (bad) luck of the Irish

ORIGINAL PAPER Niche o e lap o moun ain ha e subspecies and he ulne abili y o hei anges o in asion by he Eu opean ha e; he (bad) luck o he I ish An hony Ca a aggi .Ka ie Leach .F ancesco San illi . Jukka Rin ala .Pekka Helle .Juha Tiainen .F ancesco Bisi . Ad iano Ma inoli .W. Ian Mon gome y .Neil Reid Recei ed: 13 Feb ua y 2016 / Accep ed: 7 No embe 2016 / Published online: 17 No embe 2016 ÓThe Au ho (s) 2016. This a icle is published wi h open access a Sp inge link.com Abs ac Niche conse a ism is he endency o ela ed species o e ain ances al ole ances a e geog aphic sepa a ion. We used Ecological Niche Modelling and P incipal Componen s Analysis o bioclima ic and habi a a iables o desc ibe he ex en o he species niche, and deg ees o bioclima ic– habi a niche conse a ism wi hin he moun ain ha e (L. imidus) clade. Moun ain ha e niche space was con as ed wi h ha o he Eu opean ha e (L. eu opaeus), o shed ligh on species in e ac ions in con ac zones h oughou Eu ope. All i e subspecies o moun ain ha e had quan i iably dis inc niches. Fennoscandian (L. . syl a icus,L. . imidus) and highland (L. . sco icus,L. . a onis) subspecies, howe e , we e mos simila , exhibi ing g ea es appa en niche conse a ism. They inhabi und a, bo eal o es and uplands, and, hence a e p esumed mos simila o he ances al o m. The I ish ha e was dis inc , being consis en ly dis inguished om o he moun ain ha es in bo h 2D and n h dimensional (4D) niche space. The ecological dis inc i eness o he I ish ha e p o ides u he e idence ha i is an E olu ion- a ily Signi ican Uni , pa icula ly ulne able o Elec onic supplemen a y ma e ial The online e sion o his a icle (doi:10.1007/s10530-016-1330-z) con ains supple- men a y ma e ial, which is a ailable o au ho ized use s. A. Ca a aggi K. Leach W. I. Mon gome y N. Reid Que cus, School o Biological Sciences, Queen’s Uni e si y Bel as , Bel as BT9 7BL, UK A. Ca a aggi (&)K. Leach W. I. Mon gome y  N. Reid School o Biological Sciences, Queen’s Uni e si y Bel as , Bel as BT9 7BL, UK e-mail: [email p o ec ed] F. San illi Depa men o Ve e ina y Sciences, Uni e si a `di Pisa, Pisa, PI, I aly J. Rin ala J. Tiainen Na u al Resou ces Ins i u e Finland, P.O. Box 2, Viikinkaa i 4, 00791 Helsinki, Finland P. Helle Na u al Resou ces Ins i u e Finland, Uni e si y o Oulu, Paa o Ha aksen ie 3, 90014 Oulu, Finland F. Bisi A. Ma inoli Depa men o Theo e ical and Applied Sciences, Insub ia Uni e si y, Via J. H. Dunan 3, 21100 Va ese, I aly W. I. Mon gome y N. Reid Ins i u e o Global Food Secu i y (IGFS), Queen’s Uni e si y Bel as , Bel as BT9 5BN, UK 123 Biol In asions (2017) 19:655–674 DOI 10.1007/s10530-016-1330-z displacemen by in oduced Eu opean ha es wi h which i compe es and hyb idises. P ojec ions unde global clima e change sugges ha , by 2070, biocli- ma ic space o in asi e Eu opean ha es in I eland will expand (by 79%) bu con ac o endemic I ish ha es (by 75%), u he acili a ing hei eplacemen . The nea comple e species eplacemen o he hea h ha e (L. . syl a icus) in sou he n Sweden, whe e he Eu opean ha e has also been in oduced, may sugges a simila a e may be in s o e o he I ish ha e. Keywo ds En i onmen al Niche Modelling  In asion biology Lepus Niche conse a ism  P incipal Componen s Analysis Species Dis ibu ion Model In oduc ion The con o e sial concep , ‘niche conse a ism’, is he endency o eme gen species o e ain hei ances al ecological ai s such ha closely ela ed species may be mo e ecologically simila han would be expec ed based on hei phylogene ic di e gence (Wiens e al. 2010). A niche comp ises a mul i a ia e se o abio ic and bio ic condi ions which acili a e he pe sis ence o a species, and o which i is sui ably adap ed (Hu chinson 1957). Howe e , s udies in es iga ing he ela ionships be ween a species’ dis ibu ion and niche equen ly ail o app op ia ely de ine hei e ms (Sobe o `n 2007). The undamen al niche is uncon- s ained by limi ing bio ic ac o s such as ecological compe i ion, p eda ion, dispe sal abili y, and en i on- men al condi ions (Hu chinson 1957; Wiens and G aham 2005). The ealised niche is desc ibed as he undamen al niche cons ained by limi ing ac o s, i.e. he space occupied by, and he esou ces a ailable o, an o ganism (Hu chinson 1957; Sobe o `n 2007). The e a e, howe e , issues inhe en in he u ilisa ion o hese e ms due o conside a ions o spa ial esolu ion and bio ic in e ac ions (A au ´jo and Guisan 2006). Popu- la ion iabili y depends on a deg ee o en i onmen al s abili y o adap i e p edic abili y, and hence, clima e and habi a a e key ac o s (Sobe o ´n and Pe e son 2005;Ja ¨ka ¨la ¨niemi 2011). In he absence o gene low, popula ions may di e ge gene ically while occupying simila habi a o ances al species, and hence, he species niche is conse ed (Pe e son e al. 1999; Wiens 2004). Niche adap a ion may, he e o e, be spa io em- po ally s able and can be conse ed du ing allopa ic specia ion ia geog aphic isola ion. En i onmen al Niche Modelling (ENM) is inc eas- ingly used o es ima e en i onmen al sui abili y as a unc ion o geospa ial species occu ence ela i e o en i onmen al a iables (Phillips e al. 2006), hus cap u ing a species’ niche space. He e, we in es iga e species occu ences and o e lap in po en ia, wi h dispe sal cons ained by clima ic and habi a a iables (he ea e , e e ed o simply as he species ‘niche’). We aimed o desc ibe he deg ee o niche conse - a ism wi hin he moun ain ha e clade in Eu ope, cap u ing he niche o each subspecies. These a e con as ed agains he niche o he Eu opean ha e, hus shedding ligh on obse ed di e ences in species in e ac ions pos -con ac . We pay pa icula a en ion o he I ish ha e due o i s ecological equi alency o he Eu opean ha e (i.e. empe a e, lowland, g azing habi s). We hypo hesise ha , gi en p olonged, pos - glacial isola ion and ecological expansion: (1) he niche o he I ish ha e is mo e ecologically dis inc om o he moun ain ha e subspecies han hose subspecies a e om one ano he and, (2) he niches o he I ish and Eu opean ha e a e mo e simila o each o he han a e hose o o he moun ain ha e subspecies, ela i e o he Eu opean ha e (making he I ish ha e mo e ulne able o he impac o Eu opean ha e in asion). We also model he p edic ed shi in he bioclima ic space sui able o I ish and Eu opean ha es in I eland, unde p ojec ed global clima e change. We hypo hesised ha (3) he bioclima ic space a ailable o he I ish ha e is likely o become inc easingly unsui able, and, (4) he bioclima ic space a ailable o he Eu opean ha e is likely o become inc easingly sui able unde wa ming empe a u es due o hei con as ing o igins and di e en ially-adap ed physi- ology ( he o me ha ing A c ic, and he la e Middle Eas e n, ances y). Thus, we expec ha he ajec o y o global clima e change is likely o be bene icial o he in ade and de imen al o he na i e species. In he O de Lagomo pha, he genus Lepus (ha es and jack abbi s) is ep esen ed in Eu ope by i e ex an species, wo o which, he moun ain ha e (Lepus imidus, Linnaeus, 1758) and he Eu opean ha e (L. eu opaeus, Pallas 1837), a e widely dis ibu ed. These species a e eadily dis inguished pheno ypically, by ea and limb leng h (all sho e in moun ain ha es, apa om he hind ee ), head shape (con ex in 656 A. Ca a aggi e al. 123 Eu opean ha es), s ongly con as ing black ea - ips (p esen in Eu opean ha es), whi e-s iped muzzle (p esen in Eu opean ha es), en al ail su ace (black in Eu opean ha es), body mass (lowe in moun ain ha es) and pelage colou (da ke and mo e uni o m in moun ain ha es; Flux and Ange man 1990; Ca a aggi e al. 2016). The Eu opean ha e is a highly success ul in asi e species ha has been in oduced o a la ge numbe o coun ies wo ldwide (sensu Flux and Ange man, 1990). I is ypically pa apa ic wi h he moun ain ha e, being sepa a ed by ele a ion o habi a , wi h na ow con ac zones sugges ing each species has a dis inc niche sepa a ed by, o example, di e ences in habi a o clima e (Amo i e al. 2008). The in asion dynamics o he Eu opean ha e and i s in e ac ion wi h na i e moun ain ha e popula ions a e poo ly unde - s ood (Thulin 2003;Reid2011). Some con ac zones be ween he species a e la gely s able (e.g. in he Alps and Sco ish Highlands), hough i is p edic ed ha such ele a ionally-de ined con ac zones will shi upwa ds due o he e ec s o global clima e change (Leach e al. 2015a). O he , mo e ecen ly es ablished con ac zones a e highly uns able wi h he la ge Eu opean ha e ou compe ing and displacing he smalle moun ain ha e. Indeed, Eu opean ha es ha e displaced moun ain ha es o e much o sou he n Sweden (Jansson and Peh son 2007) and pa o sou he n Finland (Le a ¨nen e al. 2015) du ing he wen i h cen u y, and pa o I eland in he las ew decades (Reid and Mon gome y 2007;Reid2011; Ca a aggi e al. 2015,2016). Thus, pos -in oduc ion sympa y is a ypically ansien phenomenon (Thulin 2003). The moun ain ha e is a ci cumpola , a c o-alpine species complex, dis ibu ed om I eland in he wes , o Japan and Kamcha ka in he eas , and om he Alps in he sou h, o 75°N (Flux and Ange man 1990; Smi h and Johnson 2008b). The e a e i e ex an Eu opean moun ain ha e subspecies di e en ia ed by mo pho- physiological cha ac e is ics, beha iou , and ecology (Ange bjo ¨ n and Flux 1995). The e is gene ally low gene ic di e en ia ion be ween subspecies, indica i e o a pos -glacial panmic ic Eu opean popula ion, which subsequen ly unde wen agmen a ion, isola- ion, and di e gence (Hamill e al. 2006). The mos widesp ead subspecies, ecognised as he ypical o m (and hus p esumed simila o he ances al ype), is he no he n ha e (L. imidus imidus, Linnaeus 1758) which inhabi s und a (in he no h) and bo eal o es ( u he sou h) in he A c ic and Fennoscandia (Ange bjo ¨ n and Flux 1995). I s die a ies seasonally, wi h ha d woody ma e ial being consumed in win e , and g asses, sedges, and he bs, in la e summe and au umn (Flux and Ange man 1990; Helle 1995). The hea h ha e (L. . syl a icus, Nilsson 1831) occu s in sou he n Sweden and Go land (Winige 2014). The subspeci ic s a us o his axon is deba ed, wi h some ega ding i as a synonym o L. . imidus. Howe e , many o he s ecognise i as a dis inc subspecies based on win e pelage, which is blue-g ey a he han whi e (Linds o ¨m1980; Suchen unk e al. 1999; Thulin e al. 2003; Winige 2014). Bo h he no he n and hea h ha es a e hus Fennoscandian moun ain ha e subspecies and geog aphically dis inc om h ee isola ed moun ain ha e popula ions, wo o which a e ue highland moun ain ha es: he Sco ish ha e (L. . sco icus, Hizheime 1906) and he Alpine ha e (L. . a onis, Mille 1901). The o me is widesp ead h oughou mon ane habi a s in Sco land, occu ing up o 1300 m asl (Newey e al. 2011). The la e is gene ally ound on o es ed slopes (Bisi e al. 2013; Rehnus e al. 2013) up o 3500 m asl (Thulin 2003; Rehnus e al. 2013), h oughou he Alps (Ange bjo ¨ n and Flux 1995). The highland subspecies b owse ha d, woody plan ma e ial e.g. hea he Calluna ulga is (Flux and Ange man 1990). The I ish ha e (L. . hibe nicus, Bell 1837) is endemic o he island o I eland, whe e i has been isola ed o 30,000–60,000 yea s (Hughes e al. 2006). One es i- ma e placed he di e gence o I ish ha es om o he moun ain ha es (speci ically, Russian L. . imidus)a ca. 360,000 yea s be o e p esen (Hughes e al. 2006). This subspecies possesses a compa a i ely high num- be o unique gene ic o ms (mi ochond ial haplo- ypes) no sha ed by any o he subspecies ou side I eland (Hughes e al. 2006). I exhibi s conside able ecological plas ici y, being ound a all al i udes in I eland, bu is mos common in he lowlands (Whelan 1985; Reid e al. 2007). In con as o o he moun ain ha es i eeds p edominan ly on so , mos ly ag icul- u al g asses, e.g. yeg ass Lolium pe enne (S e ens and Roch o d 2004). Nea ly all moun ain ha e popu- la ions exhibi win e whi ening as camou lage du ing win e snow co e (e.g. Hewson 1958), wi h one excep ion. The I ish ha e has la gely los he ai , sa e o minimal whi ening o he ea ma gins and ee (Flux and Ange man 1990). Such is he gene ic, pheno ypic, beha iou al and ecological dis inc i eness Niche o e lap o moun ain ha e subspecies and he ulne abili y 657 123 o he I ish ha e, ha some con end i may wa an ull species s a us (Hughes e al. 2006). I is as di e gen om o he moun ain ha e subspecies as he moun ain ha e is om o he species such as he A c ic (L. a c icus, Ross 1819) o Alaskan (L. o hus, Me iam 1900) ha es (Paulo P odo ¨hl pe s. comm.), whose axonomic s a us and phylogene ic ela ionships wi h he moun ain ha e ha e been he subjec o deba e (e.g. Wu e al. 2005; MacDonald and Cook 2010). Me hods Da a sou ces and p epa a ion A o al o 238,813 eco ds o moun ain ha e sub- species and Eu opean ha e ound in Eu ope we e ob ained om a la ge numbe o sou ces, p incipally biodi e si y da a eco d cen es, academics and ecol- ogis s (Tables S1, S2 in Suppo ing In o ma ion). Da a we e collec ed ia a a ie y o me hods, combina ions o which di e ed be ween and wi hin egions, coun- ies and o ganisa ions, e.g. scien i ic su eys, hun ing bags, oppo unis ic sigh ings by he public, ecological su eys, oad casual ies. He ea e , we adop he e ms ‘‘(sub-)species’’ o e e o he moun ain ha e (includ- ing all subspecies) and he Eu opean ha e, o ‘‘sub- species’’ when e e ing o he moun ain ha e only. Reco ds we e ex ac ed du ing 2013–2014 and we e sub-sampled by da e (pos -1950, o ensu e consis ency wi h cu en bioclima ic da ase s), and geospa ial accu acy (B1 km esolu ion). Fu he mo e, while he e may be di icul ies inhe en in disc imina ing be ween sympa ic species, we we e unable o quan- i y obse e bias. Duplica e eco ds we e emo ed, as we e hose conside ed e oneous based on known dis ibu ions o each (sub-)species (i.e. alling beyond he bounda y o he In e na ional Union o Conse - a ion o Na u e ange polygon; Smi h and Johns on 2008a,b). Species-speci ic eco ds ha occu ed wi hin he known ange o ha species we e, he e o e, conside ed ‘ ue’ and e ained, while hose ha occu ed ou side he known ange we e conside ed ‘ alse’ and emo ed. The ange polygons o each moun ain ha e subspecies we e ex ac ed om he pa en IUCN ange polygon and sub-di ided in o geog aphically isola ed popula ions i.e. I eland, Sco - land, and he Alps, whils he Fennoscandian moun ain ha e subspecies anges we e delinea ed acco ding o Be geng en (1969). Due o a lack o su icien ly p ecise da a in no he n Fennoscandia and much o cen al Eu ope, No he n and Eu opean ha es appea ed e oneously ‘absen ’ om pa s o hei known ange. Fu he mo e, da a exhibi ed conside - able sample bias (Yackulic e al. 2013), wi h la ge numbe s o eco ds occu ing a ound u ban cen es, pa icula ly in he UK and Sweden. The e a e a numbe o me hods a ailable o accoun ing o sample bias (see Fou cade e al. 2014), including he u ilisa ion o a ge backg ound poin s (Phillips e al. 2009) o bias g ids (Eli h e al. 2010). Ta ge - backg ounds a e de ined as backg ound poin s d awn om occu ences o a ocal class (e.g. lagomo phs, he bi o ous mammals). Thus, backg ound da a will exhibi simila spa ial bias o ha o he modelled species (Phillips e al. 2009). Simila ly, a bias g id is a su ace scaled o ep esen su ey e o (Eli h e al. 2010), a quan i y unknown o almos all ([99.9%) o ou da a. Howe e , da a manipula ion (i.e. emo ing da a in o e -sampled egions) may be e ec i e in educing o emo ing bias (Phillips e al. 2009). Thus, in o de o educe sample selec ion bias, p esence eco ds we e hinned using Occu enceThinne e - sion 1.04 downloaded om www.phycoweb.ne / so wa e. Occu enceThinne uses p obabili y algo- i hms o emo e occu ence eco ds based on an associa ed ke nel densi y g id. The p obabili y ha an occu ence will be emo ed is p opo ional o occu - ence densi y desc ibed by he ke nel densi y g id (Ve b uggen e al. 2013). Due o he ex emely high densi y o occu ences in some egions (e.g. u ban a eas in he UK and sou he n Sweden), da a we e sequen ially hinned o app op ia e densi ies which we e in o med a p io i by densi ies o eco ds else- whe e in he species ange. A p io i hinning aimed o equalise he densi ies o occu ence eco ds on a landscape scale, and, hence, p oduce ecologically ele an models. A o al o 9075 eco ds we e used in modelling (see Table S2 o species speci ic p e- and pos - hinning occu ence coun s, Fig S1 o occu - ence dis ibu ion maps). 10,000 backg ound da a poin s (i.e. pseudo-absences) we e gene a ed an- domly wi hin he ange o each indi idual (sub-)spe- cies, analogous o he Res ic ed Backg ound app oach de ailed in Fou cade e al. (2014). Clima e da a we e downloaded om Wo ldClim (www.wo ldclim.o g) a 30 a c-second (ca. 1 km 2 ) esolu ion. Species eco ds we e associa ed wi h mean 658 A. Ca a aggi e al. 123 da a om 1950 o 2000 o cu en models only. Th ee aw- o ma (mean empe a u e, p ecipi a ion season- ali y and empe a u e seasonali y) and h ee composi e (Hilliness Index, No malised Di e ence Vege a ion Index (NDVI), and wa e balance) en i onmen al a iables we e used (Table S3). Eigh land co e a iables (coni e ous o es , c ops, mixed o es , moo land and hea hland, pas u e, pea bog, sc ub and spa se ege a ion; see Table S3 o ec o ilenames) we e ob ained om he CORINE Land Co e 2006 (EEA 2010). Shape ile and as e c ea ion and manipula ion we e ca ied ou using A cGIS 10.2.2 (ESRI 2011). En i onmen al Niche Modelling MAXENT is a popula p esence-only modelling ool (Phillips e al. 2006,2010), which uses a maximum en opy app oach, i.e. he p obabili y dis ibu ion which bes ep esen s he da a is he one wi h he la ges en opy. Despi e i s widesp ead use, MAXENT has been c i icised due o i s ulne abili y o o e i - ing and he use o logis ic ou pu o es ima e absolu e occu ence p obabili ies (e.g. Royle e al. 2012). Such limi a ions may be mi iga ed agains by ca e ul a p io i da a manipula ion, o close app oxima e he assump- ions o he model, e.g. ha occu ence da a ep esen unbiased independen samples, cons an p obabili y o de ec ion, and ha de ec abili y is independen o model a iables (Yackulic e al. 2013). Indeed, MAXENT has been shown o consis en ly ou pe o m o he compa able modelling echniques (e.g. Wisz e al. 2008; Ta kesh and Je schke 2012). While MAXENT is ela i ely obus agains collinea a iables (Rod iguez-Robles e al. 2010; Kuemme le e al. 2012), se e al clima ic a iables exhibi ed s ong collinea i y; explo a o y models sugges ed a s ong cumula i e in luence. Va iables wi h he g ea es pe mu a ion impo ance, i.e. mean empe a u e (col- linea wi h minimum empe a u e, maximum empe - a u e) and annual wa e balance (collinea wi h minimum p ecipi a ion, maximum p ecipi a ion, mean p ecipi a ion), we e e ained. Clima ic and en i on- men al a iables wi h a mean pe mu a ion impo ance o 2 (complex cul i a ion, human in luence index, inland ma sh, na u al g assland, adia ion, snow, u ban, numbe o mon hs wi h posi i e wa e balance) we e also emo ed. ENMs we e un using linea , quad a ic, p oduc and h eshold ea u es wi h clamping and ex apola ion disabled, o 50 eplica es. P esence eco ds we e spli andomly in o a 75% aining se and a 25% es se , wi h c oss- alida ion. Models o he I ish and Eu opean ha e we e p ojec ed unde global clima e change a ime-slices o he cu en pe iod (2010–2014), 2050s and 2070s. IPCC Fi h Assessmen Repo Coupled Model In e - compa ison P ojec Phase 5 (CMIP5) u u e clima ic da a o he Rep esen a i e Concen a ion Pa hway (RCP) 8.5 o 2050 (a e aged ac oss 2041–2060) and 2070 (a e age o 2061–2080) we e downloaded om Wo ldClim a 1 km 2 g id cell esolu ion. RCP 8.5 indica es a mean a e age global empe a u e inc ease o 2 °C by he 2050s and 3.7 °C by he 2070s. All a iables we e a e aged ac oss i e Global Ci cula ion Models (GCMs), CNRM-CM5, GFDL-CM3, GISS- E2-R, Had-GEM-ES and MIROC-ESM-CHEM, hus educing model e o (sensu Pie ce e al. 2009). O iginally desc ibed as ‘‘ex eme clima e change’’, his clima e scena io now appea s o bes i obse ed clima ological ends ( ollowing Leach e al. 2015a; Table S3). Changes in p edic ed ange ex en we e calcula ed using Max SSS, i.e. he sum o es speci ici y plus sensi i i y, which is e ec i e when using p esence only da a, and is no a ec ed by pseudo-absences (Liu e al. 2013). A majo ca ea o his app oach is ha CORINE habi a a iables we e kep cons an when p ojec ing in o u u e ime-slices as no obus p edic ions a e a ailable o how land co e will espond unde u u e clima ic condi ions. Howe e , his app oach is consis en wi h mos s udies ha p ojec species anges in o u u e condi ions (e.g. Ace edo e al. 2012). Model e alua ion Models we e e alua ed using he A ea Unde he Cu e (AUC; Fielding and Bell 1997) o he Recei e Ope a ing Cha ac e is ic (ROC) cu e, a model-accu- acy assessmen measu e ha is independen o p e alence (McPhe son e al. 2004). The classi ica ion o AUC alues ollows a commonly-used, ye a bi- a y anking sys em based on sugges ions by Swe s (1988), G eine e al. (2000). Values be ween 0.9 and 1.0 a e conside ed excellen , 0.9–0.8 good, 0.7 and 0.8 a e age and 0.7 poo . Howe e , whe e ROC cu es a e cons uc ed om p esence-only da a, he maxi- mum possible AUC is 1 (Wiley e al. 2003), and i is no possible o de e mine op imal pe o mance Niche o e lap o moun ain ha e subspecies and he ulne abili y 659 123 (Phillips e al. 2006). Ne e heless, ela i e pe o - mance may s ill be in e ed, gi en ha an AUC o 0.5 desc ibes andom p edic ion (Phillips e al. 2006). We also es ed he omission a e (p opo ion o ue occu ences misiden i ied), sensi i i y (p opo ion o p esences which a e co ec ly p edic ed), speci ici y (p opo ion o absences which a e co ec ly p e- dic ed), p opo ion co ec (p opo ion o he p esence and absence eco ds co ec ly iden i ied), and T ue Skill S a is ic (TSS), calcula ed using SDMTools package (Van de Wal e al. 2012) in R ( e sion 3.2.2). TSS is a p e alence-independen me ic de i ed om h eshold sensi i i y and speci ici y. Values ange om -1 o?1 and es he ag eemen be ween he expec ed and obse ed dis ibu ion, and whe he he ou come could be p edic ed due o chance (Allouche e al. 2006). A alue [0.4 was aken as indica ing ha he model was a good i (Landis and Koch 1977; sensu Leach e al. 2015a). Niche o e lap and equi alency The simila i y o (sub-)species con inuous-su ace p obabili y models (i.e. geog aphic niche o e lap) we e e alua ed using he niche o e lap me ic, I(Wa en e al. 2008). This me hod calcula es pai wise o e lap be ween models, p oducing alues be ween 0 (no o e lap be ween niche models) and 1 (iden ical niche models). Wa en’s Iis based on he p obabili y (p x,i ,p y,i ) o a species (X o Y) occu ing in a gi en cell (i); de ined by he ENM. In con as o Schoene ’s D (Schoene 1968), ano he commonly-used me ic, Wa en’s I ea s p x and p y as p obabili y dis ibu ions wi h no biological assump ions, and, hence, is mo e app op ia e o p esence-only analyses (Wa en e al. 2008). Niche o e lap me ics we e calcula ed o con empo a y and u u e clima e-p ojec ed models using he R package uzzysim (Ba bosa 2015). Niche equi alency es s we e used o assess whe he pai ed-species ENM o e lap alues (I) we e signi ican ly di e en om a one- ailed no malized null dis ibu ion o compa a i e o e lap alues. Null dis ibu ions we e gene a ed by compa ing ENMs o wo ocal species o andom subse s d awn om pooled p esences, whe e he numbe o ex ac ed (i.e. ‘null’) p esences we e equal o he numbe o obse ed p esences o each species. This was epea ed 100 imes o each species pai (Wa en e al. 2008). Ecological niches we e said o be non-equi alen i pai ed-species o e lap alues we e signi ican ly lowe han hose o he null dis ibu ion (P B0.05). Niche equi alency es s we e ca ied ou using ENMTools (Wa en e al. 2010) and using only con empo a y da a. Ecological dis ance P incipal Componen Analysis o occu ence eco ds and associa ed da a was used o educe bioclima ic and habi a a iables associa ed wi h all species eco ds o ou hypo he ical axes wi h eigen alues [1, desc ib- ing ecological niche space, using co e R unc ions. A mul i ac o ial Gene al Linea Model (GLM) was used o es ablish di e ences in P incipal Componen s (PC1 h ough PC4) be ween each (sub-)species wi h Bon- e oni pai wise pos hoc es o mul iple compa isons used o iden i y niche space di e ences. Biplo s o pai ed P incipal Componen Axes we e used o plo he p oximi y o each (sub-)species in 2D space. Fo each pai wise plo he mean Mahalanobis dis ance (De Maesschalck e al. 2000) was calcula ed be ween: (1) all pai wise compa isons o moun ain ha e subspecies excluding he ocal subspecies (i.e. he I ish ha e); and (2) all pai wise combina ions including he ocal subspecies. Mahalanobis dis ances we e calcula ed using he R package S a Ma ch (D’O azio 2015). The n-dimensional Euclidean dis ance be ween each pai o (sub-)species was also calcula ed ac oss all ou P incipal Componen s simul aneously, hus de i ing a single measu e o dis ance be ween (sub-)species in mul idimensional (4D) niche space. Euclidean dis- ances we e calcula ed using he R package pdis (Wong 2013). A es was used o es o signi icance o di e ences be ween he wo g oups (moun ain ha es including and excluding he I ish ha e). Resul s Model e alua ion All (sub-)species con inuous-sui abili y ENMs pe - o med well (AUC [0.7, TSS [0.4; Table 1). Tem- pe a u e seasonali y (43.1%) and mean annual empe a u e (26.3%) had he g ea es mean con ibu- ion ac oss all (sub-)species models (Table 2), bu hei con ibu ion o indi idual (sub-)species ENMs a ied subs an ially. Fo example, empe a u e 660 A. Ca a aggi e al. 123 seasonali y was he single mos impo an a iable o he I ish ha e (97.2%), ye was ela i ely unimpo an o he no he n ha e (0.5%). The p edic ed p obabili ies o moun ain ha e (sub- )species p esence closely app oxima ed he ac ual ange ex en o each (sub-)species (Fig. 1). Niche space o he Eu opean ha e was p edic ed no hwa d beyond i s no he n (in asi e) ange edge in Sweden ex ending wes in o sou he n No way, sou hwa d beyond i s sou he ly (na u al) ange edge in no h- eas e n Ibe ia and in all di ec ions a ound i s cu en in asi e ange in No he n I eland (Fig. 1 ). Ecological (dis)simila i ies Geog aphic niche o e lap measu es de i ed om con inuous-su ace p obabili y models desc ibed po en ial o e lap be ween six (sub-)species pai s (IC0.4; Table 3). Almos all pai wise compa isons be ween ha e (sub-)species and ENMs gene a ed using andomly selec ed backg ound poin s did no di e om null dis ibu ions, and, hence, hei niches can be said o be simila . Only ou pai wise compa isons be ween we e ound o be signi ican ly di e en (i.e. less simila han expec ed by chance; P B0.05), hough he ela ionship was unidi ec ional a he han ecip ocal: he Alpine ha e was dis inc om he Sco ish ha e and he Eu opean ha e; he I ish ha e was dis inc om he Eu opean ha e; and he No he n ha e was dis inc om he I ish ha e (Table 3). Thus, he ecological niche o he Alpine ha e, o example, was mo e dis inc om ha o he Eu opean ha e han would ha e been expec ed by chance, bu no ice e sa. Ou esul s sugges ha while he ecological niches o ha e (sub-)species in Eu ope a e simila , hey a e no iden ical. Ecological niche space om occu ence poin da a was desc ibed by P incipal Componen Axis 1 (PC1) cap u ing 24% o bioclima ic and habi a a ia ion, desc ibing mean annual empe a u e (0.82; linea combina ion coe icien , o loading), No malized Di e ence Vege a ion Index (0.89), pas u e (0.55), p ecipi a ion seasonali y (-0.59), and empe a u e seasonali y (-0.72), PC2 cap u ed 17% o he a ia- ion, desc ibing annual wa e balance (0.80), hilliness (0.80), and spa se ege a ion (0.59), PC3 cap u ed 8% o a ia ion, desc ibing coni e ous o es (0.79) and sc ub (0.53), and PC4 also cap u ed 8% o a ia ion, desc ibing pea bogs (0.84; Table 4). All P incipal Componen alues a ied signi i- can ly be ween (sub-)species (Table S4). The biplo o PC1 and PC2 (accoun ing o 41% o cumula i e a ia ion) sugges ed ha he niches o Fennoscandian moun ain ha e subspecies we e mo e simila o one ano he han hey we e o any o he ha e (sub-)species (Fig. 2). Bo h highland moun ain ha e subspecies we e also mo e simila o one ano he han hey we e o any o he ha e (sub-)species. The Sco ish ha e occupied a simila , ye sligh ly mo e p oduc i e en i onmen , sugges ed by a mo e posi i e alue on Table 1 En i onmen al Niche Model e alua ion me ics o six Eu opean ha e (sub-)species, using 75% aining and 25% es da a (50 eplica ions) (Sub-)species Da a AUC Omission a e Sensi i i y Speci ici y P opo ion co ec TSS Alpine ha e T aining 0.74 0.19 0.81 0.66 0.66 0.47 Tes 0.74 0.19 0.81 0.66 0.66 0.47 Hea h ha e T aining 0.73 0.20 0.80 0.67 0.67 0.47 Tes 0.73 0.21 0.79 0.67 0.67 0.47 I ish ha e T aining 0.72 0.27 0.73 0.70 0.70 0.43 Tes 0.73 0.24 0.76 0.70 0.70 0.46 No he n ha e T aining 0.74 0.20 0.80 0.68 0.68 0.48 Tes 0.74 0.21 0.79 0.68 0.68 0.47 Sco ish ha e T aining 0.73 0.27 0.73 0.72 0.72 0.45 Tes 0.73 0.26 0.74 0.72 0.72 0.46 Eu opean ha e T aining 0.74 0.16 0.84 0.64 0.64 0.48 Tes 0.73 0.17 0.83 0.64 0.64 0.47 AUC A ea Unde he Cu e o he Recei e Ope a ing Cha ac e is ic cu e, TSS T ue Skill S a is ic Niche o e lap o moun ain ha e subspecies and he ulne abili y 661 123 PC1, indica ing highe NDVI. Va ia ion in I ish ha e niche space no only did no o e lap wi h any o he moun ain ha e subspecies, bu i s cen oid was u he away om o he moun ain ha e subspecies han i was om he Eu opean ha e, which was associa ed wi h ag icul u al c ops. The I ish ha e was associa ed wi h empe a e, highly p oduc i e pas u es (Fig. 2). O he pai wise compa isons be ween emaining P incipal Componen s showed ewe dis inc di e - ences (Fig. S2), as hey accoun ed o less a ia ion, Table 2 Compa ison o en i onmen al esponse cu es o each a iables used in En i onmen al Niche Modelling and hei es i- ma ed ela i e con ibu ion o o each model. Va iables a e anked in descending o de o hei a e aged con ibu ion ac oss all six (sub-) species. x-axis =me ics o he ocal a iable; y-axis =p obabili y o sui able condi ions Va iable Alpine ha e Hea h ha e I ish ha e No he n ha e Sco ish ha e Eu opean ha e x  Tempe a u e seasonali y 36.1% 18.9% 97.2% 0.5% 79.5% 26.6% 43.1% Mean empe a u e 17.5% 49.7% 0.0% 51.6% 15.8% 23.2% 26.3% NDVI 6.4% 2.8% 0. 4% 29.1% 2.9% 13.9% 9.3% Hilliness Index 29.6% 4.1% 0.1% 0.9% 0.1% 7.1% 7.0% Wa e balance 2.7% 10.5% 0.1% 5.0% 0.4% 3.5% 3.7% P ecipi a ion seasonali y 2.0% 0.9% 1.9% 1.3% 0.0% 11.9% 3.0% Pas u e 0.4% 5.4% 0.1% 0.7% 0.0% 2.6% 1.5% Pea bog 1.9% 1.5% 0.1% 1.3% 0.0% 2.9% 1.3% C ops 1.3% 0.2% 0.1% 3.4% 0.2% 2.3% 1.3% Spa se ege a ion 0.6% 3.0% 0.0% 1.9% 0.0% 1.0% 1.1% Coni e ous o es 0.3% 1.7% 0.1% 1.1% 0.1% 1.7% 0.8% Moo land & hea hland 0.9% 0.8% 0.0% 0.4% 0.6% 1.1% 0.6% Mixed Fo es 0.1% 0.3% 0.0% 0.6% 0.4% 1.9% 0.6% Sc ub 0.1% 0.2% 0.0% 2.1% 0.0% 0.4% 0.5% 662 A. Ca a aggi e al. 123 (a) (b) (c) (d) (e) ( ) Fig. 1 P edic ed bioclima ic and habi a sui abili y om En i onmen al Niche Models o aAlpine ha e, bHea h ha e, cI ish ha e, (d) No he n ha e, eSco ish ha e, and Eu opean ha e. Shaded a eas indica e he (sub-) species ange ex en as de i ed om IUCN polygons o known dis ibu ions (Be gen- g en 1969; Winige 2014; Ca a aggi e al. 2015) Niche o e lap o moun ain ha e subspecies and he ulne abili y 663 123 E olu iona ily Signi ican Uni (i.e. a se o popula- ions which a e gene ically and mo phologically dis- inc om simila species; Ryde 1986), i no a ull species, o iginally desc ibed by Bell (1837) as L. hibe nicus, becomes inc easingly pe suasi e. Fou een his o ical in oduc ions o Eu opean ha es occu ed h oughou I eland be ween 1848 and 1890 (Reid 2011), wi h mos ailing o become es ablished (Reid and Mon gome y 2007). Ou ENMs p edic ha mos o I eland is (and p esumably was) unsui able o he Eu opean ha e, p o iding a po en ial explana ion as o why mos in oduc ions ailed. A p esen , he e is a ela i ely ange- es ic ed popula ion o in oduced Eu opean ha es in No he n I eland (Ca a aggi e al. 2015), he only egion o I eland cu en ly p edic ed by ou ENMs as being sui able o he species. Thei ange expanded h ee- old be ween 2005 and 2012/2013 (Ca a aggi e al. 2015) wi h a co e ange popula ed solely by he in ade being es ablished ecen ly (Ca a aggi e al. 2016). Mo eo e , he e is a high deg ee o mul i-gene a ional, bidi ec ional hyb idisa ion in a eas o sympa y (Hughes e al. 2009; P odo ¨hl e al. 2013). Ce ainly, he I ish ha e sha es much in common wi h he Eu opean ha e, hough he la e has a g ea e associa ion wi h a able land, ypical o i s mo e eas e ly and cen al Eu opean dis ibu ion (Smi h e al. 2005). Indeed, p e ious s udies ha e demons a ed ha bo h species exhibi compa able niche b ead hs and almos comple e niche o e lap in I eland (Reid and Mon gome y 2007; Ca a aggi e al. 2015). Niche o e lap me ics in he p esen s udy a e, on he ace o i , con adic o y. Howe e , i mus be emembe ed ha he Eu opean ha e niche desc ibed he ein is e lec i e o i s en i e ange, whe eas niche simila i ies desc ibed by p e i- ous s udies we e con ined o I eland. Niche o e lap es ima es de i ed om ENMs a e associa ed wi h geog aphic o e lap o gene a ed p obabili y su aces (Wa en e al. 2008). Ou es ima es, he e o e, mus be placed in con ex o he small ange o he Eu opean ha e in I eland a p esen , and u u e p ojec ions o Eu opean ha e ange inc ease and I ish ha e ange dec ease. Clima ological p ojec ions unde u u e clima e change sugges I eland is likely o ge wa me and d ie , hough wi h hea ie win e ain all (Holden and B e e on 2003), a ou ing an inc ease in a able ag icul u e (Holden e al. 2003). ENMs sugges such changes will esul in a d ama ic educ ion in he sui able bioclima ic en elope o he I ish ha e. As such, while he I ish ha e exhibi s conside able ecological plas ici y and adap abili y, he species may s uggle unde inc eased clima e ins abili y. Remaining sui able a eas a e likely o be in he cool, we wes o I eland in habi a s subop imal o he Eu opean ha e, such as pea bogs o he no h-wes uplands. Con e sely, he bioclima ic en elope sui - able o he Eu opean ha e is likely o expand in I eland in u u e, wi h condi ions and any inc ease in a able c opland becoming inc easingly a ou able. Indeed, i has been sugges ed ha in asi e in e ac ions may be mo e likely in a eas wi h g ea e han a e age clima ic ins abili y (Leach e al. 2015b). The Eu opean ha e may, he e o e pose a di ec h ea o he ecological in eg i y o he I ish ha e in he sho - e m (i.e. o e he nex 30 yea s o so) only, a e which a eas o sympa y a e likely o be empo ally ansien . Ou ENMs p edic ha he wes o I eland will emain sui able o he I ish ha e and unsui able o he Eu opean ha e by 2070. I should be cau ioned, howe e , ha no moun ain in I eland is high enough o ha e a pe manen snowline, no any habi a subop imal o he Eu opean ha e expansi e enough, in e ms o indi idual pa ch size, o p o ide e uge o he I ish ha e, should pos -Eu opean ha e in oduc ion popula ion dynamics mi o hose o Sweden. Conse- quen ly, only occupied o sho e islands in he no h- wes a e likely o p o ide e uge o he I ish ha e in he long- e m. Con ol o in asi e species is equen ly ecom- mended and inc easingly common (e.g. Cou champ e al. 2003). Mos success ul e adica ions ha e occu ed on islands, whe e ecolonisa ion is less likely (e.g. Imbe e al. 2000; Howald e al. 2007). In ade s may be emo ed ia biological (e.g. he in oduc ion o a p eda o , compe i o o pa hogen), physical (e.g. shoo ing, apping) o chemical (poi- soning) me hods (Cou champ e al. 2003). The use o any one o combina ion o con ol me hods equi es ca e ul e alua ion due o he po en ial o signi ican unin ended consequences on non- a ge species and he wide ecosys em. Con ol o e ad- ica ion o ange- es ic ed non-na i e species is eminen ly easible gi en he applica ion o app o- p ia e echniques, and obse a ion o con ol/ e- mo al c i e ia (e.g. Bom o d and O’B ien 1995). Once in asi e species become widesp ead, howe e , e adica ion may be imp ac ical and applied man- agemen becomes inc easingly di icul and 670 A. Ca a aggi e al. 123 expensi e. This is ce ainly he case wi h ega ds o he Eu opean ha e popula ions in Sweden. The Eu opean ha e was in oduced o Sweden in he la e 19 h cen u y (Lonnbe g 1905), and apidly expanded ac oss sou he n Sweden, comple ely displacing he hea h ha e om [9,000 km 2 by 1999, and es ab- lishing a conside able zone o sympa y which also ex ends in o he sou he ly ex en o he no he n ha e (sensu Be geng en 1969; Jansson and Peh son 2007). The si ua ion in Sco land is mo e nuanced, as he Eu opean ha e is conside ed a p io i y species in G ea B i ain, despi e i s alien o igin, as is he na i e moun ain ha e. Ou niche o e lap p ojec ions sug- ges ha in e speci ic compe i ion be ween Eu opean and Sco ish ha es will become inc easingly com- mon. We may expec , he e o e, he Sco ish ha e o come unde inc easing p essu e om wha we can easonably assume o be an ecologically dominan compe i o , leading o spa ial displacemen . I mus be emembe ed, howe e , ha ou p ojec ions did no accoun o changes in land use. Gi en he di e en habi a p e e ences o Sco ish and Eu o- pean ha es (e.g. Hewson 1962, Schai-B aun e al. 2015), compe i ion and displacemen may be less se e e han ou models sugges and may be media ed by habi a managemen o he bene i o he Sco ish ha e (e.g. he main enance o hea he moo land and o he upland habi a s). Gi en ha he cu en Eu opean ha e popula ion in I eland may ha e been in oduced as ecen ly as he 1970s (Ca a aggi e al. 2015) wi h i s ex en and numbe s expanding apidly, policy make s and con- se a ionis s in I eland would do well o ake heed o he I ish ha e’s likely u u e p ospec s by using Sweden as a case s udy example. The au ho i ies o bo h poli ical ju isdic ions o I eland a e signa o ies o he Con en ion on Biological Di e si y (UNEP 1992), he Be n Con en ion (1979), he Eu opean Habi a s Di ec i e (EEC 43/1992), and he EU Regula ion 1143/2014 on In asi e Alien Species (O icial Jou nal o he Eu opean Union [OJ] 2014), and hence, a e obliged o add ess in asi e species issues. Lessons om he hea h and No he n ha e in Sweden highligh he p eca ious, uns able na u e o moun ain and Eu opean ha e in e speci ic popula ion dynamics, and sugges ha con inued inac ion om au ho i ies in No he n I eland will only se e o acili a e he con inued expansion o he in ade , o he de imen o he endemic. Acknowledgemen s This p ojec was unded by he Na u al He i age Resea ch Pa ne ship (NHRP) (P ojec QU12-07 Accoun Numbe R3326BSC) be ween he No he n I eland En i onmen Agency (NIEA) and Que cus, Queen’s Uni e si y Bel as (QUB). Species p esence da a and pe mission o use in his publica ion we e ob ained om a la ge numbe o biological eco d cen es and academics which a e lis ed in ull in Table S1 in Suppo ing In o ma ion. We a e g a e ul o he Edi o and e iewe s o hei ins uc i e guidance and commen s which subs an ially imp o ed he manusc ip . AC was he p ima y au ho , KL c ea ed se e al clima ic a iables, de eloped he a e aged u u e clima ic da a and con ibu ed o he manusc ip , FS, JR, PH, JT, FB and AM p o ided species p esence eco ds and con ibu ed o he manusc ip , while WIM and NR supe ised he wo k, concei ed he idea, and edi ed he manusc ip . Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p:// c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un e- s ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Com- mons license, and indica e i changes we e made. Re e ences Ace edo P, Jime ´nez-Val e de A, Melo-Fe ei a J, Real R, Al es PC (2012) Pa apa ic species and he implica ions o cli- ma e change s udies: a case s udy on ha es in Eu ope. Glob Change Biol 18:1509–1519 Allouche O, Tsoa A, Kadmon R (2006) Assessing he accu acy o species dis ibu ion models: p e alence, kappa and he ue skill s a is ic (TSS). J Appl Ecol 43:1223–1232 Amo i G, Con oli L, Nappi A (2008) Mammalia II. E inaceo- mo pha, So icomo pha, Lagomo pha, Roden ia. Fauna d’I alia, ol XLIV. Edizioni Calde ini de Il Sole 24 ORE Business Media S l, Milano Ange bjo ¨ n A, Flux JEC (1995) Lepus imidus. Mamm Species 495:1–11 A au ´jo MB, Guisan A (2006) Fi e (o so) challenges o species dis ibu ion modelling. Glob Ecol Biogeog 33:1677–1688 A ibas O, Ca anza S (2004) Mo phological and gene ic e i- dence o he ull species s a us o Ibe olace a cy eni ma inez icai (A ibas, 1996). Zoo axa 634:1–24 Ba bosa AM (2015) uzzySim: applying uzzy logic o bina y simila i y indices in ecology. Me hods Ecol E ol 6:853–858 Be geng en A (1969) On gene ics, e olu ion and his o y o dis ibu ion o he hea h-ha e, a dis inc popula ion o he a c ic ha e, Lepus imidus. S enaka jaga e o bunde , S ockholm Be n (1979) Con en ion on he conse a ion o Eu opean wildli e and na u al habi a s. Eu T ea y Se 104:1–16 Bisi F, Noda i M, Oli ei a NMDS, Ossi F, Masse oni E, P e- a oni DG, Wau e s LA, Ma inoli A (2013) Habi a selec- ion and ac i i y pa e ns in Alpine moun ain ha e (Lepus imidus a onis). Mamm Biol 78:28–33 Niche o e lap o moun ain ha e subspecies and he ulne abili y 671 123 Bisi F, Wau e s LA, P ea oni DG, Ma inoli A (2015) In e - speci ic compe i ion media ed by clima e change: which in e ac ion be ween b own and moun ain ha e in he Alps? Mamm Biol 80:424–430 Bom o d M, O’B ien P (1995) E adica ion o con ol o e - eb a e pes s? Wildl Soc Bull 23:249–255 Ca a aggi A, Mon gome y WI, Reid N (2015) Range expansion and compa a i e habi a use o insula , congene ic lago- mo phs: in asi e Eu opean ha es Lepus eu opaeus and endemic I ish ha es Lepus imidus hibe nicus. Biol In a- sions 17:687–698 Ca a aggi A, Zacca oni M, Riga F, Schai-B aun SC, Dick JTA, Mon gome y WI, Reid N (2016) An in asi e-na i e mammalian species eplacemen p ocess cap u ed by came a ap su ey Random Encoun e Models. Remo e Sens Ecol Conse 2:45–58 Cou champ F, Chapuis J-L, Pascal M (2003) Mammal in ade s on islands: impac , con ol and con ol impac . Biol Re 78:347–383 D’O azio M (2015) S a Ma ch: s a is ical ma ching. R package e sion 1.2.3. h p://CRAN.R-p ojec .o g/package= S a Ma ch. Accessed 3 Aug 2015 De Maesschalck R, Jouan-Rimbaud D, Massa D (2000) The Mahalanobis dis ance. Chemome In ell Lab 50:1–18 EEA (2010) Co ine Land Co e 2006 as e da a. Eu opean En i onmen Agency. h p://www.eea.eu opa.eu/da a-and- maps/da a/co ine-land-co e -2006- as e -3. Accessed 17 Jan 2015 EEC (1992) Habi a s Di ec i e. J Eu Comm L 206/7:1–44 Eli h J, Kea ney M, Phillips S (2010) The a o modeling ange- shi ing species. Me hods Ecol E ol 1:330–342 ESRI (2011) A cGIS desk op: elease 10. En i onmen al Sys- ems Resea ch Ins i u e, Redlands Fielding AH, Bell JF (1997) A e iew o me hods o he assessmen o p edic ion e o s in conse a ion p esence/ absence models. En i on Conse 24:38–49 Flux JEC, Ange man R (1990) The ha es and jack abbi s. In: Chapman J, Flux JEC (eds) Rabbi s, ha es and pikas: s a us su ey and conse a ion ac ion plan. IUCN, Gland, pp 61–94 Fou cade Y, Engle JO, Ro ¨dde D, Secondi J (2014) Mapping species dis ibu ions wi h MAXENT using a geog aphi- cally biased sample o p esence da a: a pe o mance assessmen o me hods o co ec ing sampling bias. PloS ONE 9:e97122 G eine M, P ei e D, Smi h RD (2000) P inciples and p ac ical applica ion o he ecei e -ope a ing cha ac e is ic analysis o diagnos ic es s. P e Ve Med 45:23–41 Hamill RM, Doyle D, Duke EJ (2006) Spa ial pa e ns o gene ic di e si y ac oss Eu opean subspecies o he moun ain ha e, Lepus imidus L. He edi y 97:355–365 Helle E (1995) Me sa ¨ja ¨nis (Lepus imidus). In: Linde ´n H, Ha io M, Wikman M (eds) Riis an ja ¨ljille, Riis a - ja kala alouden u kimuslai os. Edi a, Helsinki, pp 18–21 (Finnish wi h English summa y) Hewson R (1958) Moul s and win e whi ening in he moun ain ha e Lepus imidus sco icus, Hilzheime . P oc Zool Soc Lond 131:99–108 Hewson R (1962) Food and eeding habi s o he moun ain ha e Lepus imidus sco icus, Hilzheime . P oc Zool Soc Lond 139:515–526 Holden N, B e e on AJ (2003) Impac o clima e change on I ish ag icul u e. Pages. In: Sweeney J, B e e on T, By ne C, Cha l on R, Emblow C, Fealy R, Holden N, Jones M, Donnelly A, Moo e S, Pu se P, By ne K, Fa ell E, Mayes E, Minchin D, Wilson J (eds) Clima e change: scena ios and impac s o I eland. En i onmen al P o ec ion Agency, Wex o d, pp 33–81 Holden N, B e e on AJ, Fealy R, Sweeney J (2003) Possible change in I ish clima e and i s impac on ba ley and po a o yields. Ag ic Fo Me e ol 116:181–196 Howald G, Donlan CJ, Gal a ´n JP, Russell JC, Pa kes J, Sama- niego A, Wang Y, Vei ch D, Geno esi P, Pascal M, Saunde s A, Te shy B (2007) In asi e oden e adica ion on islands. Conse Biol 21:1258–1268 Hughes M, Mon gome y WI, P odo ¨hl P (2006) Popula ion gene ic s uc u e and sys ema ics o he I ish Ha e. Repo p epa ed by he Na u al He i age Resea ch Pa ne ship, Que cus o he No he n I eland En i onmen & He i age Se ice, No he n I eland, UK Hughes M, Reid N, Mon gome y I, P odo ¨hl P (2009) Ve i ica- ion o hyb idisa ion be ween in oduced Eu opean and na i e I ish ha es. Repo p epa ed by he Na u al He i age Resea ch Pa ne ship, Que cus o he No he n I eland En i onmen Agency, No he n I eland, UK Hu chinson G (1957) A ea ise on limnology. Wiley, New Yo k Imbe M, Ha ison M, Ha ison J (2000) In e ac ions be ween pe els, a s and abbi s on Whale Island, and e ec s o a and abbi e adica ion. N Z J Ecol 24:153–160 Ja ¨ka ¨la ¨niemi A (2011) Na ow clima e and habi a en elope a ec he su i al o elic popula ions o a no he n A nica angus i olia. En i on Exp Bo 72:415–421 Jansson G, Peh son A ˚(2007) The ecen expansion o he b own ha e (Lepus eu opaeus) in Sweden wi h possible implica- ions o he moun ain ha e (L. imidus). Eu J Wildl Res 53:125–130 Kuemme le T, Hickle T, Olo sson J, Schu ge s G, Radelo VC (2012) Recons uc ing ange dynamics and ange ag- men a ion o eu opean bison o he las 8000 yea s. Di e s Dis ib 18:47–59 Landis JR, Koch GG (1977) The measu emen o obse e ag eemen o ca ego ical da a. Biome ics 33:159–174 Leach K, Kelly R, Came on A, Mon gome y WI, Reid N (2015a) Expe ly alida ed models and phylogene ically- con olled analysis sugges s esponses o clima e change a e ela ed o species ai s in he o de Lagomo pha. PLoS ONE 10:e0122267 Leach K, Mon gome y WI, Reid N (2015b) Biogeog aphy, mac oecology and species’ ai s media e compe i i e in e - ac ions in he o de Lagomo pha. Mamm Re 45:88–102 Le a ¨nen R, Kunnas an a M, Pohjoisma ¨ki J (2015) Abundance and dis ibu ion o ha e hyb ids in Finland. In: Ange bjo ¨ n A, Dale ´n L, Elmhagen B, We delin L (eds) P oceedings o he 7 h Eu opean cong ess o mammalogy. S ockholm Uni e si y, S ockholm, p 54 Linds o ¨m E (1980) The ed ox in a small game communi y o he sou h aiga egion in Sweden. In: Zimen E (ed) The ed ox: symposium on beha iou and ecology. Sp inge , Do d ech , pp 177–184 Liu C, Whi e M, Newell G (2013) Selec ing h esholds o he p edic ion o species occu ence wi h p esence-only da a. J Biogeog 40:778–789 672 A. Ca a aggi e al. 123 Lonnbe g E (1905) On hyb id ha es be ween Lepus imidus L. and Lepus eu opaeus Pall. om Sou he n Sweden. P oc Zool Soc Lond 1:278–287 MacDonald SO, Cook JA (2010) Recen mammals o Alaska. Uni e si y o Alaska P ess, Fai banks, pp 123–125 McPhe son JM, Je z W, Roge s DJ (2004) The e ec s o spe- cies’ ange sizes on he accu acy o dis ibu ion models: ecological phenomenon o s a is ical a e ac ? J Appl Ecol 41:811–823 Newey S, Po s J, Baines D, Cas illo U, Duncan M, Ha ison A, Ramsay S, Thi good S, Iason G (2011) De elopmen o a eliable me hod o es ima ing moun ain ha e numbe s. SNH Commissioned Repo , ol 444, pp 1–29 OJ (2014) Regula ion (EU) No 1143/2014 o he Eu opean Pa liamen and o he Council o 22 Oc obe 2014 on he p e en ion and managemen o he in oduc ion and sp ead o in asi e alien species. J Eu Union 35–55 Pe e son A, Sobe o ´nJ,Sa ´nchez-Co de o V (1999) Conse - a ism o ecological niches in e olu iona y ime. Science 285:1265–1267 Phillips SJ, Ande son RP, Schapi e RE (2006) Maximum en opy modeling o species geog aphic dis ibu ions. Ecol Model 190:231–259 Phillips SJ, Dudı ´k M, Eli h J, G aham CH, Lehmann A, Lea hwick J, Fe ie S (2009) Sample selec ion bias and p esence-only dis ibu ion models: implica ions o backg ound and pseudo-absence da a. Ecol Appl 19: 181–197 Phillips SJ, Dudik M, Schapi e R (2010) Maxen So wa e, e . 3.3.3e. h ps://www.cs.p ince on.edu/*schapi e/maxen /. Accessed 3 Feb 2014 Pie ce DW, Ba ne TP, San e BD, Gleckle PJ (2009) Selec ing global clima e models o egional clima e change s udies. PNAS 106:8441–8446 P odo ¨hl PA, Hughes MA, Hynes RA, Mon gome y WI, Reid N (2013) Molecula e idence o bidi ec ional hyb idisa ion be ween he endemic Lepus imidus hibe nicus and he in asi e Lepus eu opaeus in I eland. In: Mon gome y WI (ed) P oceedings o he 11 h in e na ional mammalogical cong ess, p 77 Rehnus M, Ma coni L, Hackla `nde K, Filli F (2013) Seasonal changes in habi a use and eeding s a egy o he moun ain ha e (Lepus imidus) in he Cen al Alps. I al J Mammal 24:161–165 Reid N (2011) Eu opean ha e (Lepus eu opaeus) in asion ecology: implica ion o he conse a ion o he endemic I ish ha e (Lepus imidus hibe nicus). Biol In asions 13:559–569 Reid N, Mon gome y WI (2007) Is na u alisa ion o he b own ha e in I eland a h ea o he endemic I ish ha e? Biol En i on 107:129–138 Reid N, Dinge kus K, Mon gome y WI, Ma nell F, Lynn D, Kings on N, Mcdonald RA (2006) S a us o ha es in I e- land. I ish Wildli e Manuals 30. Na ional Pa ks and Wildli e Se ice, Depa men o En i onmen , He i age and Local Go e nmen , Dublin Reid N, Sweeney O, Wilson C, P es on SJ, Mon gome y WI, McDonald RA (2007) De elopmen s in ha e su ey me hodology—AS applied o he NI I ish ha e su ey 2007. Repo p epa ed by Que cus o he En i onmen and He i age Se ice (DOE, N.I.) Rod iguez-Robles JA, Jezko a T, Leal M (2010) Clima ic s a- bili y and gene ic di e gence in he opical insula liza d Anolis k ugi, he Pue o Rican ‘Laga ijo Ja dine o de la Mon ana ˜’. Mol Ecol 19:1860–1876 Royle JA, Chandle RB, Yackulic C, Nichols JD (2012) Like- lihood analysis o species occu ence p obabili y om p esence-only da a o modelling species dis ibu ions. Me hods Ecol E ol 3:545–554 Ryde OA (1986) Species conse a ion and sys ema ics: he dilemma o subspecies. T ends Ecol E ol 1:9–10 Schai-B aun SC, Reichlin TS, Ru T, Klansek E, Ta a uch F, A nold W, Hackla ¨nde K (2015) The Eu opean ha e (Lepus eu opaeus): a picky he bi o e sea ching o plan pa s ich in a . PLoS ONE 10:e0134278 Schoene TW (1968) Anolis liza ds o Bimini: esou ce pa i- ioning in a complex auna. Ecology 49:704–726 Smi h AT, Johns on CH (2008a) Lepus eu opaeus. The IUCN ed lis o h ea ened species 2008: e.T41280A10430693. doi:10.2305/IUCN.UK.2008.RLTS.T41280A10430693. en. Accessed 18 Dec 2013 Smi h AT, Johns on CH (2008b) Lepus imidus. The IUCN Red Lis o Th ea ened Species 2008: e.T11791A3306541. doi:10.2305/IUCN.UK.2008.RLTS.T11791A3306541.en. Accessed 18 Dec 2013 Smi h RK, Jennings NV, Ha is D (2005) A quan i a i e analysis o he abundance and demog aphy o Eu opean ha es Le- pus eu opaeus in ela ion o habi a ype, in ensi y o ag icul u e and clima e. Mamm Re 35:1–24 Sobe o ´n J, Pe e son AT (2005) In e p e a ion o models o undamen al ecological niches and species’ dis ibu ional a eas. Biodi e s In o m 2:1–10 S e ens TC, Roch o d JM (2004) The die and impac o he I ish ha e (Lepus imidus hibe nicus, Bell 1837) in a young plan a ion. Biol En i on 104:89–94 Suchen unk F, Pols e K, Giacome i M, Ra i P, Thulin C-G, Ruhle C, Vasile AG, Slo a-Bachmay L (1999) Spa ial pa i ioning o allozyme a iabili y in Eu opean moun ain ha es (Lepus imidus): gene pool di e gence ac oss a dis- junc dis ibu ional ange? Mamm Biol 64:308–318 Swe s JA (1988) Measu ing he accu acy o diagnos ic sys ems. Science 240:1285–1293 Sy ja ¨la ¨P, Nylund M, Heinikainen S (2005) Eu opean b own ha e synd ome in ee-li ing moun ain ha es (Lepus imi- dus) and Eu opean b own ha es (Lepus eu opaeus)in Finland 1990–2002. J Wildl Dis 41:42–47 Ta kesh M, Je schke G (2012) Compa ison o six co ela i e models in p edic i e ege a ion mapping on a local scale. En i on Ecol S a 19:437–457 Thulin C-G (2003) The dis ibu ion o moun ain ha es Lepus imidus in Eu ope: a challenge om b own ha es L. eu o- paeus? Mamm Re 33:29–42 Thulin C-G, Jaa ola M, Tegels o ¨m H (2003) The occu ence o moun ain ha e mi ochond ial DNA in wild b own ha es. Mol Ecol 6:463–467 Tiainen J, Pankakoski E (1995) Rusakko (Lepus eu opaeus). In: Linde ´n H, Ha io M, Wikman M (eds) Riis an ja ¨ljille, Riis a - ja kala alouden u kimuslai os. Edi a, Helsinki, pp 22–25 ( Finnish wi h English summa y) UNEP (1992) The con en ion on biological di e si y. Uni ed Na ions En i onmen al P og am. h ps://www.cbd.in /. Accessed 14 Feb 2015 Niche o e lap o moun ain ha e subspecies and he ulne abili y 673 123 Van de Wal J, Falconi L, Januchowski D, Shoo L, S o lie C (2012) SDMTools: Species Dis ibu ion Modelling Tools: ools o p ocessing da a associa ed wi h species dis ibu- ion modelling exe cises. R package e sion 1.1-13. h p:// c an. -p ojec .o g/web/packages/. Accessed 6 Aug 2015 Ve b uggen H, Tybe ghein L, Bel on GGS, Mineu F, Jue e - bock A, Hoa au G, Gu gel CFD, De Cle ck O (2013) Imp o ing ans e abili y o in oduced species’ dis ibu- ion models: new ools o o ecas he sp ead o a highly in asi e seaweed. PLoS ONE 8:e68337 Wa en DL, Glo RE, Tu elli M (2008) En i onmen al niche equi alency e sus conse a ism: quan i a i e app oaches o niche e olu ion. E olu ion 62:2868–2883 Wa en DL, Glo RE, Tu elli M (2010) ENMTools: a oolbox o compa a i e s udies o en i onmen al niche models. Ecog aphy 33:607–611 Whelan J (1985) The popula ion and dis ibu ion o he moun- ain ha e (Lepus imidus L.) on a mland. I ish Na J 21:532–534 Wiens JJ (2004) Specia ion and ecology e isi ed: phylogene ic niche conse a ism and he o igin o species. E olu ion 58:193–197 Wiens JJ, G aham CH (2005) Niche conse a ism: in eg a ing e olu ion, ecology, and conse a ion biology. Ann Re Ecol E ol Sys 36:519–539 Wiens JJ, Acke ly DD, Allen AP, Anacke BL, Buckley LB, Co nell HV, Damschen EI, Da ies TJ, G y nes JA, Ha - ison SP, Hawkins BA, Hol RD, McCain CM, S ephens PR (2010) Niche conse a ism as an eme ging p inciple in ecology and conse a ion biology. Ecol Le 13:1310–1324 Wiley EO, McNyse KM, Pe e son AT, Robins CR, S ewa AM (2003) Niche modeling and geog aphic ange p edic ions in he ma ine en i onmen using a machine-lea ning algo i hm. Oceanog aphy 16:120–127 Winige A (2014) The appa en popula ion c ash in hea h- ha es Lepus imidus syl a icus o sou he n Sweden—do complex ecological p ocesses lea e de ec able inge p in s in long- e m hun ing bag eco ds? Unpublished mas e s hesis, Swedish Uni e si y o Ag icul u al Sciences Wisz MS, Hijmans RJ, Li J, Pe e son AT, G aham CH, Guisan A, Eli h J, Dudı ´k M, Fe ie S, Hue mann S, Lea hwick JR, Lehmann A, Lohmann L, Loiselle BA, Manion G, Mo i z C, Nakamu a M, Nakazawa Y, O e on JM, Phillips SJ, Richa dson KS, Scache i-Pe ei a R, Schapi e RE, Sobe o ´n J, Williams SE, Zimme mann NE (2008) E ec s o sample size on he pe o mance o species dis ibu ion models. Di e s Dis ib 14:763–773 Wong J (2013) pdis : pa i ioned dis ance unc ion. R package e sion 1.2. h p://CRAN.R-p ojec .o g/package=pdis . Accessed 6 Aug 2015 Wu CH, Wu JP, Bunch TD, Li QW, Wang YX, Zhang YP (2005) Molecula phylogene ics and biogeog aphy o Lepus in Eas e n Asia based on mi ochond ial DNA sequences. Mol Phylogene E ol 37:45–61 Yackulic CB, Chandle R, Zipkin EF, Royle JA, Nichols JD, Campbell G an EH, Ve an S (2013) P esence-only mod- elling using MAXENT: when can we us he in e ences? Me hods Ecol E ol 4:236–243 674 A. Ca a aggi e al. 123