Using fire to enhance rewilding when agricultural policies fail

Using fi e o enhance ewilding when ag icul u al policies ail João C. Campos a, ⁎, Julia Be nha d b , Nú ia Aquilué c,d ,LluísB o ons d,e, , Jesús Domínguez g ,ÂngelaLomba a,h , B uno Ma cos a,h ,Fe nandoMa ínez-F ei ía a , F ancisco Mo ei a a,i , Sil ana Pais a,j , João P. Hon ado a,h ,Ad iánRegos a,g a InBIO/CIBIO - Cen o de In es igação em Biodi e sidade e Recu sos Gené icos, Campus Ag á io de Vai ão, Rua Pad e A mando Quin as, n° 7, 4485-661 Vai ão, Po ugal b Uni e si ä Leipzig, Ri e s aße 30, 36 04109 Leipzig, Ge many c Cen e d'É ude de la Fo ê , Uni e si é du Québec à Mon éal, Mon éal, Canada d InFo es J u (CTFC-CREAF), C a. An iga S Llo enç de Mo unys km. 2, 25280 Solsona, Spain e CREAF, 08193 Ce danyola del Vallès, Spain CSIC, 08193 Ce danyola del Vallès, Spain g Depa amen o de Zooloxía, Xené ica e An opoloxía Física, Uni e sidade de San iago de Compos ela, 15782 San iago de Compos ela, Spain h Depa amen o de Biologia, Faculdade de Ciências da Uni e sidade do Po o, Rua do Campo Aleg e, S/N, Edi ício FC4, 4169-007 Po o, Po ugal i InBIO/CIBIO - Cen o de In es igação em Biodi e sidade e Recu sos Gené icos, Labo a ó io Associado, Ins i u o Supe io de Ag onomia, Uni e sidade de Lisboa, Tapada da Ajuda,1349-017 Lisboa, Po ugal j p oMe heus, Resea ch Uni in Ma e ials, Ene gy and En i onmen o Sus ainabili y, Ins i u o Poli écnico o Viana do Cas elo, 4990-706 Pon e de Lima, Po ugal HIGHLIGHTS •HNV is a long- e m oppo uni y o fi e supp ession and biodi e si y conse a- ion. •HNV managemen would benefi he majo i y o species (mo e han 60%). •Rewilding benefi s up o 20% o species, including endange ed/endemic axa. •Up o 33% o species p ofi om open habi a s c ea ed by fi e. •Fi e-media ed ewilding is a na u e- based solu ion when ag icul u al policies ail. GRAPHICAL ABSTRACT abs ac a icle in o A icle his o y: Recei ed 22 July 2020 Recei ed in e ised o m 13 Sep embe 2020 Accep ed 2 Oc obe 2020 A ailable online 14 Oc obe 2020 Edi o : Paulo Pe ei a Keywo ds: Biodi e si y conse a ion High Na u e Value a mlands Land abandonmen Rewilding has been p oposed as an oppo uni y o biodi e si y conse a ion in abandoned landscapes. How- e e , ewilding is challenged by he inc easing fi e isk associa ed wi h mo e flammable landscapes, and he loss o open-habi a specialis species. Con as ingly, suppo ing High Na u e Value a mlands (HNV ) has been also highligh ed as a aluable op ion, bu he e ec i e implemen a ion o ag icul u al policies o en ails leading o unce ain scena ios whe ein he e ec s o wildfi e managemen emain la gely unexplo ed. He ein, we simu- la ed fi e-landscape dynamics o e alua e how fi e supp ession scena ios a ec fi e egime and biodi e si y (102 species o e eb a es) unde ewilding and HNV policies in he u u e (2050), in a ansna ional biosphe e e- se e (Ge ês-Xu és Moun ains, Po ugal-Spain). Rewilding and HNV scena ios we e modula ed by h ee di e - en le els o fi e supp ession e ec i eness. Then, we quan ified scena io e ec s on fi e egime (bu ned and supp essed a eas) and biodi e si y (habi a sui abili y change o 2050). Simula ions confi m HNV as a long- e m oppo uni y o fi e supp ession (up o 30,000 ha o addi ional supp essed a eas be ween 2031 and 2050 Science o he To al En i onmen 755 (2021) 142897 ⁎Co esponding au ho . E-mail add esses: [email protected] (J.C. Campos), nu ia.aquilue@c c.ca (N. Aquilué), lluis.b o ons@c c.ca (L. B o ons), [email protected] (J. Domínguez), angelalomba@ c.up.p (Â. Lomba), b uno.ma [email protected] (B. Ma cos), ma inez- ei [email protected] (F. Ma ínez-F ei ía), mo ei [email protected] (F. Mo ei a), jhon ado@ c.up.p (J.P. Hon ado), ad ian. [email protected] ,ad ian. [email protected] (A. Regos). h ps://doi.o g/10.1016/j.sci o en .2020.142897 0048-9697/© 2020 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). Con en s lis s a ailable a ScienceDi ec Science o he To al En i onmen jou nal homepage: www.else ie .com/loca e/sci o en Land-use Sou he n Eu ope Wildfi es in compa ison o ewilding scena io) and o conse a ion (benefi ing a ound 60% o species). Rewilding benefi s some species (20%), including c i ically endange ed, ulne able and endemic axa, while se e al species (33%) also p ofi om open habi a s c ea ed by fi e. Al hough HNV emains he bes scena io, ewilding ein o ced by low fi e supp ession managemen may p o ide a na u e-based solu ion when socie al suppo h ough ag icul u al policies ails. © 2020 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY license (h p:// c ea i ecommons.o g/licenses/by/4.0/). 1. In oduc ion Land abandonmen in u al landscapes is one o he mos impo - an d i e s o egional land-use change (Es oque e al., 2019), and has been sugges ed as an oppo uni y o biodi e si y conse a ion and he eins a emen o na u al ecological p ocesses (e.g. Quei oz e al., 2014;Me ckx and Pe ei a, 2015). Rewilding app oaches ely on his assump ion as hey in end o ees ablish na u al sel - sus ained ecosys ems wi h educed human in e en ion (Gillson e al., 2011). In Eu ope, ewilding ini ia i es (e.g. Rewilding Eu ope Ne wo k; h ps:// ewildingeu ope.com) migh benefi biodi e si y by p o iding enewed sui able habi a s and dispe sal co ido s, alle- ia ing o es habi a agmen a ion (Pe ino e al., 2019). This man- agemen s a egy migh be pa icula ly beneficial o a di e si y o o es -adap ed species, including emblema ic axa o conse a ion conce nsuchaswol sandbea s(Na a o and Pe ei a, 2015). Addi- ionally, ewilded landscapes migh po en ia e a ious ecosys em unc ions and se ices (Pe ino e al., 2019), such as he suppo o soil, nu ien s and wa e cycles, he p o isioning o biomass and cli- ma e egula ion (Na a o and Pe ei a, 2015). Howe e , ewilding holds some cons ain s ha may limi i s suc- cess ul implemen a ion. The inhe en homogeniza ion o ewilded landscapes leads o loss and agmen a ion o open habi a s, mainly due o sh ub enc oachmen and o es expansion (Mo ei a e al., 2011). S udies also indica ed ha he g adual cessa ion o adi ional a ming a eas, many o which known o suppo “High Na u e Value a mlands”(he ea e HNV ), is a majo cause o local biodi e si y losses, accele a ing popula ion declines o species adap ed o we g ass- lands, pas u es and o he ex ensi e ag icul u al a eas (F anks e al., 2018;Lomba e al., 2020). Fu he mo e, sh ub enc oachmen caused by land abandonmen con ibu es o uel accumula ions and highe connec i i y o flammable ege a ion, inc easing he po en ial inci- dence, se e i y and impac s o la ge wildfi es (see Mo ei a e al., 2011 and e e ences he ein). These changes in he fi e- ege a ion dynamics could cause se e e impac s on local ecosys em p ocesses and associa ed biodi e si y, pa icula ly in egions whe e landscapes and bio a ha e endu ed a long- e m ela ion wi h human ac i i ies and fi e egimes. The Medi e anean egions o sou he n Eu ope a e pa adigma ic cases o human- and fi e-media ed landscapes (Loep e e al., 2010; Fe nandes e al., 2011;Viedma e al., 2015). The ma ked clima e season- ali y o Medi e anean egions, cha ac e ized by wa m-d y summe s and cold-humid win e s, p o ides op imal condi ions o equen and in ense fi e e en s, mos ly concen a ed in he summe season. The numbe o ex eme fi e e en s has escala ed in sou he n Eu ope du ing he las decades (San-Miguel-Ayanz e al., 2013;Tedim e al., 2013; Williams, 2013), mainly due o in ensified e ec s o clima e change (wa me and d ie clima ic condi ions, and mo e equen hea h wa es) and inc eased uel accumula ion and connec i i y caused by u al land abandonmen (Fe nandes e al., 2016a;Tu co e al., 2019). Fo hese easons, sou he n Eu opean coun ies ha e in es ed as e- sou ces in fi e supp ession (Fe nandes e al., 2016b). Howe e , exhaus- i e fi e supp ession migh be coun e -p oduc i e in fi e-p one ecosys ems, since ege a ion will become dense and p o ide mo e uel o la ge , mo e in ense and mo ese e e fi es —a p ocess desc ibed as he “fi efigh ing ap”(Mo ei a e al., 2020). Consequen ly, al e na- i e managemen s a egies ha e been ecommended, highligh ing he u gency o a wide- anging policy shi owa ds fi e mi iga ion and adap a ion (Fe nandes, 2013;Mo i z e al., 2014;Mo ei a e al., 2020). The in ica e links be ween land abandonmen and fi e egimes complexi y he selec ion o app op ia e al e na i e scena ios, and sub- sequen ly decision-making in fi e managemen and planning (Mo ei a e al., 2011;McLauchlan e al., 2020). Conside ing he di e en p one- ness o land co e (LC) ypes o bu n, ag icul u al policies implemen ed unde he Eu opean Union's Common Ag icul u al Policy ha e been claimed as an oppo uni y o educe fi e isk by p omo ing HNV in de - imen o mo e flammable LC ypes, such as dense sc ublands (Mo ei a and Pe'e , 2018). Al e na i e fi e managemen s a egies ha e been e alua ed in he las decades o inc ease fi e supp ession e ec i eness unde he ongoing land abandonmen ends. Fo ins ance, since wild- fi es migh educe uel load and connec i i y be ween flammable ege- a ion, se e al s udies ha e sugges ed ha allowing unplanned fi es o bu n (unde sa e condi ions) would help supp essing u u e la ge o es fi es unde ad e se clima ic condi ions(e.g. Hou man e al., 2013;Regos e al., 2014;Duane e al., 2019). Allowing low-in ensi y wildfi es o bu n migh also con ibu e o biodi e si y conse a ion, an idea ad oca ed by he “py odi e si y hypo hesis”, a p ominen concep in fi e ecology (Kelly and B o ons, 2017;Kelly e al., 2017). This hypo hesis elies on he assump ion ha a high le el o landscape he e ogenei y, ela ed o spa ial- empo al a - ia ions o esou ces and condi ions media ed by he in ensi y and se e - i y o fi e egimes, benefi s biodi e si y (see Kellyand B o ons, 2017 and e e ence he ein). This is s ill a con o e sial hypo hesis since s udies ha e ound bo h posi i e (e.g. Ponisio e al., 2016;Tingley e al., 2016) and nega i e (e.g. Taylo e al., 2012;Be y e al., 2015) e ec s o py odi e si y on di e en biological communi ies. The incong uen ou comes o py odi e si y unde line he complex ela ions be ween biodi e si y, fi e egime, and landscape s uc u e and composi ion (McLauchlan e al., 2020). In o ma i e assessmen s should hus con- side how land-use change and fi e can a ec bo h he di e si y, a ail- abili y and connec i i y o sui able habi a s o mul iple species (Kelly and B o ons, 2017), p e e ably om di e en axonomic g oups. He e, we assess he impac s o al e na i e landscape ajec o ies and fi e supp ession managemen s a egies on u u e fi e egimes and on biodi e si y conse a ion. We ocused on changes in bu ned and supp essed a eas and habi a sui abili y o 102 e eb a e species om h ee axonomic g oups (amphibians, bi ds, and ep iles) in a ansbounda y p o ec ed moun ain egion be ween Po ugal and Spain, ep esen a i e o he ypical sou he n Eu opean moun ain landscapes unde going land abandonmen o a leas hal a cen u y (Regos e al., 2015;Lasan a e al., 2017). We aimed o answe he ol- lowingques ions: 1)How would di e en land-use and fi esupp ession managemen scena ios con ibu e o u u e fi e mi iga ion (i.e. ewe a eas bu ned associa ed wi h highe supp ession e ficiency)?; 2) How would hose scena ios a ec biodi e si y?; 3) Which is he bes man- agemen scena io o p omo ing fi e mi iga ion and biodi e si y conse - a ion?; and 4) Wi h ag icul u al policies ailing o cope wi h u al abandonmen , could a ewilding ajec o y in eg a ed wi h fi e supp es- sion policies con ibu e o enhance biodi e si y conse a ion? We fi s modelled he fi e- ege a ion dynamics om 2010 o 2050, unde wo con as ing land-use managemen scena ios ( ewilding e sus HNV ) coupled wi h h ee fi e managemen s a egies di e ing in fi e J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 2 supp ession le els. Then, based on u u e p ojec ions o fi e- ege a ion and on species dis ibu ion modelling, we quan ified he po en ial im- pac s o hese con as ingpolicy op ions on fi e egime, landscape s uc- u e, and hei subsequen e ec s on biodi e si y. 2. Me hods 2.1. S udy a ea The s udy a ea is he T ansbounda y Biosphe e Rese e Ge ês-Xu és (BR-GX; c.a. 2670 km 2 ) in he NW Ibe ian Peninsula (Fig. 1). The ese e encompasses ou Na u a 2000 si es and wo na ional p o ec ed a eas: he Peneda-Ge ês Na ional Pa k (PGNP; Po ugal) and he Baixa Limia- Se a do Xu és Na u al Pa k (BLXNP; Spain). This moun ain egion is o- pog aphically complex, wi h al i udes a ying be ween 15 and 1545 m and is composed by a di e sified o ien a ion elie , deep alleys, plains and s eep slopes. The clima e is p edominan ly A lan ic, cha ac e ized by high p ecipi a ion le els (a ound 1000 o 3000 mm/yea ), mos ly concen a ed in au umn and win e seasons (Oc obe –Ma ch), al- hough he a ied elie and wide- anging al i udes a o d di e se mi- c oclima ic condi ions. The ese e is loca ed be ween he Medi e anean and Eu osibe ian biogeog aphical egions, holding high biodi e si y le els and se ing as e uge o se e al emblema ic species o conse a ion conce n. The landscape is domina ed by sc ublands and agmen ed o es s o deciduous ees (mainly oaks) and coni e s (mainly pines). The ese e is expe iencing a long- e m u al exodus (ha ing a cu en low popula ion o 29.4 inhabi an s/km 2 ), which has p og essi ely con ibu ed o sc ub and woodland expansion a he ex- pense o c opland and pas u eland (Regos e al., 2015). The e ec s o land abandonmen a e pa icula ly in ensified in he moun ainous and ha dly accessible a eas o he ese e, ollowing con empo a y endencies obse ed ac oss o he Medi e anean egions (e.g., Ga cía- Ruiz and Lana-Renaul , 2011; an Leeuwen e al., 2019). S ill, he a ea is subjec ed o a high equency o human-induced wildfi es (e.g. adi- ional use o fi e; Chas-Amil e al., 2015), leading o a fi e egime cha ac- e ized by ecu en fi e e en s and conside able o al bu ned a eas (195,000 ha bu ned be ween 1983 and 2010) wi h a fi e e u n in e al o 37 yea s. Despi e inc eased in es men s in fi e supp ession, hese e - o s ha e been p o en unsuccess ul o alle ia e he local fi e egime and he impac s o la ge wildfi es, a pa e n ha has been obse ed ac oss o he moun ain u al a eas in sou he n Eu ope, such as in Po ugal (e.g. Fe nandes e al., 2016b) and Spain (e.g. Loep e e al., 2010;Pausas and Fe nández-Muñoz, 2012;Viedma e al., 2015), o F ance (San Roman Sanz e al., 2013), I aly (U sino and Romano, 2014;Bajocco e al., 2019) and G eece (Xys akis e al., 2017). 2.2. Scena io design Two land-use scena ios we e es ablished (see Fig. 2 and Table 1 o de ails conce ning annual con e sion a es): (1) a ewilding scena io, ep esen ing he ongoing end o land abandonmen epo ed in ou s udy a ea (see Regos e al., 2015), as well as in o he moun ain egions ac oss Sou he n Eu ope unde he cu en EU Common Ag icul u al Pol- icy (CAP) (see Lasan a e al., 2017 and e e ences he ein); and (2) an HNV scena io, eflec ing he adop ion o policies ha e ec i ely main- ain and expand HNV ag icul u al a eas (i.e. a g eene pa h o he CAP; Pe'e e al., 2014, 2020). Fo each land-use scena io, h ee fi e manage- men scena ios di e ing in supp ession policies we e compa ed (Fig. 2): (1) a scena io o le ing fi es bu n in o es and sc ubland a eas while fi efigh ing esou ces a e p io i ized owa ds human asse s p o ec ion (i.e. fi efigh ing e o s mos ly ocused on ag icul u al mosaics, in which human asse s a e mos ly loca ed; low fi e supp ession-LS; i.e. Fig. 1. Loca ion o he T ansbounda y Biosphe e Rese e Ge ês-Xu és in heIbe ian Peninsula ( hick black con ou s) and o hemain p o ec ed a eas o Peneda-Ge êsNa ional Pa k and he Baixa Limia-Se a do Xu és Na u al Pa k (yellow con ou s). J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 3 using fi e o p omo e biodi e si y; see e iew in Kelly and B o ons, 2017); (2) a scena io ep esen a i e o cu en fi e supp ession ap- p oaches and capabili ies (cu en fi e supp ession-CS; see model pa- ame iza ion in Supplemen a y ma e ial in Pais e al., 2020); and (3) a scena io wi h highe in es men in fi e supp ession (i.e. inc eased fi efigh ing capaci y; high fi e supp ession-HS). 2.3. Fi e-landscape modelling and e ec s o managemen scena ios on fi e egime Fi e-media ed landscape dynamics we e add essed using a landscape dynamic modelling app oach based on he REMAINS model (see Pais e al., 2020 conce ning REMAINS model). B iefly, he model inco po a es landscape change p ocesses d i ing fi e- ege a ion dynamics, gene a ing dis inc scena ios o landscape dynamics based on p e iously designed condi ions. The model simula es fi e igni ion, sp ead and ex inc ion, cali- b a ed unde business-as-usual scena ios wi h he cu en fi efigh ing e - ec i eness specifically de e mined o each adminis a i e egion o he s udy a ea (please see Supplemen a y ma e ial in Pais e al., 2020), na u- alsuccessionandpos -fi e egene a ion. Annual bu n a eas a e quan i- fied o define a bu n a ea dis ibu ion and o consequen ly inco po a e his o ical fi e egime in o he model. In addi ion, he fi e-succession model is coupled wi h a land-use model o implemen land-use policies a he landscape le el. Pa icula ly, he model used (1) dynamic a iables in o ming abou he main LC ype, ime since las change o LC ype and which ansi ions has ecen ly aken place; and (2) s a ic a iables, such as oad densi y, ele a ion, slope and aspec . The a iables we e imple- men ed in he model a 30-m esolu ion o he ese e. Then, he model es ima ed he fi e igni ion p obabili y and simula ed he fi e sp ead based on h ee majo d i e s. Fi s ly, he land-use changes (e.g. u al abandonmen , ag icul u al in ensifica ion and o es con e sions) a e simula ed by using a demand-alloca ion app oach whe e he quan i y o change by ime s ep is based on his o ical land-use ends, which al- lows emula ing landscape dynamics unde cu en land-use policies. Sec- ondly, wildfi e and fi e supp ession, in which he fi e egime was calib a ed by p e iously speci ying a a ge annual bu ned a ea based on fi e s a is ical da a be ween 1983 and 2010 a he municipal le el o he Spanish Minis y o Ag icul u e and Fishe ies, Food and En i onmen (MAPAMA, 2018) o Spain, and on fi e pe ime e s o he pe iod 1990–2010 o he Ins i u e o Na u e Conse a ion and Fo es s (h ps:// www.icn .p /) o Po ugal. Thi dly, he ege a ion dynamics, including na u al succession and pos -fi e egene a ion, which was compu ed by iden i ying pos -fi e egene a ion cu es o each pixel o he s udy a ea Fig. 2. Schema ic ep esen a ion o he s udy app oach. The land-use and fi e supp ession managemen scena ios implemen ed in his s udy a e p esen ed, as well as he esul ing al e na i e scena ios es ed o imp o e fi e mi iga ion and biodi e si y conse a ion in he s udy a ea. Table 1 Land-use and fi e supp ession scena ios implemen ed in he modelling simula ions. The land-use scena io, he annual ege a ion con e sion a es unde each land-use policy and he fi e supp ession le els a e p esen ed o each scena io. Scena io ac onym Land-use scena io C opland o sc ubland (ha) Sc ubland o c opland (ha) Fi e supp ession le el RWild_LS RWild 400 0 Low RWild_CS RWild 400 0 Cu en RWild_HS RWild 400 0 High HNV _LS HNV 0 400 Low HNV _CS HNV 0 400 Cu en HNV _HS HNV 0 400 High J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 4 by using a No malized Di e ence Vege a ion Index ime se ies acco ding o Gu man and Igna o (1998).Finalfi e pe ime e s esul ed om he spa io- empo al in e ac ions be ween he landscape i sel and he imple- men a ion o fi efigh ing s a egies. 2.4. Species dis ibu ion modelling and e ec s o managemen scena ios on biodi e si y 2.4.1. Species da a and en i onmen al a iables In o de o assess he impac s o al e na i e managemen scena - ios on biodi e si y, we buil species dis ibu ion models (SDMs) o 102 e eb a es (79 bi ds, 15 ep iles and 8 amphibians; Appendix A) using a se o en i onmen al a iables. To do so, we ga he ed p es- ence/absence da a om he PGNP (Pimen a and San a ém, 1996; Soa es e al., 2005)and heBLXNP(Domínguez e al., 2005, 2012) a lases. Bi ds da a we e a ailable a 2 km esolu ion o he PGNP ( o 1990–1995) and a 1 km ( o 1998–2000) and 2 km esolu ion ( o 2010) o he BLXNP. The da a o amphibians and ep iles we e a ailable a 1 km esolu ion o bo h PGPN ( o 1998–2003) and BLXNP ( o 2010). SDMs we e hen conduc ed on o a 1-km g id co - e ing he en i e s udy a ea, using a minimum o 30 p esences pe species (Appendix A). En i onmen al a iables we e selec ed in acco dance wi h he empo al pe iods o he a lases da a. Al i ude, slope and aspec we e acqui ed a 30 m esolu ion om a Global Digi al Ele a ion Model o he Ad anced Spacebo ne The mal Emission and Reflec ion Radiome e (h ps://as e web.jpl.nasa.go /gdem.asp). LC a iables, conc e ely pe cen age o e e g een o es s, deciduous o es s, ag i- cul u al a eas, ocky a eas wi h spa se ege a ion and sh ublands, we e ob ained om Landsa -de i ed da a o 2010, and om REMAINS simula ions o he pe iod be ween 2020 and 2050, unde he p e iously defined six managemen scena ios (RWild_LS, RWild_CS, RWild_HS, HNV _LS, HNV _CS and HNV _HS). All en i on- men al a iables we e consequen ly esampled in o 1 km o spa ial esolu ion. 2.4.2. Model fi ing and e alua ion We buil ou SDMs using six modelling echniques (Gene alized Linea Models, Gene alized Addi i e Models, Random Fo es s, A ifi- cial Neu al Ne wo ks, Gene alized Models o Boos ed Reg ession, and Mul i a ia e Adap i e Reg ession Splines) a ailable in he R package ‘biomod2’(h p:// - o ge. -p ojec .o g/p ojec s/biomod/). The models we e fi using 80% o he da a, while he emaining 20% we e used o model e alua ion h ough he a ea unde he cu e (AUC) o a ecei e ope a ing cha ac e is ics (ROC; Fielding and Bell, 1997). Subsequen ly, we buil consensus p edic ions based on weigh ed a e ages o single-model p ojec ions o each species (Ma mion e al., 2009;A aújo e al., 2011). Ensemble models we e de i ed using a minimum o 10 model eplica es (see Appendix A conce ning model eplica es and AUC quali y h esholds o each species). These models we e hen p ojec ed o 2010 (one p ojec- ion) and 2050 (using 40 fi e-landscape model uns pe manage- men scena io, hus o aling 240 p ojec ions pe species). Finally, consensus p edic ions we e eclassified in o p esence/absence maps h ough ROC op imized h esholds a ailable in ‘biomod2’ (Thuille e al., 2009). 2.4.3. E ec s o managemen scena ios on biodi e si y To e alua e he e ec s o land-use and fi e supp ession managemen scena ios on biodi e si y, we calcula ed he p edic ed habi a sui abili y change be ween 2010 and 2050 o each species. We fi ed mul iple lin- ea eg ession models using he species habi a sui abili y changes as dependen a iables, and fi e supp ession le el (low/cu /high) and land-use managemen ( ewilding/HNV ) as independen a iables. The e ec s o land-use and fi esupp ession managemen on speciesdis- ibu ions we e assessed based on model significance s a is ics (Appen- dix B). Finally, habi a sui abili y changes (be ween 2010 and 2050) we e e alua ed acco ding o (i) species g oups significan ly a ec ed by land-use and fi e supp ession managemen ; and (ii) species' egional IUCN s a us (Appendix A). 2011−2030 2031−2050 Bu ned A ea Oppo unis ic Fi e Supp esion RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS 0 50 100 150 200 250 0 20 40 60 Scena ios Thousands hec a es Land−use managemen HNV Rewilding Fig. 3. Es ima ed bu ned and supp essed a eas ( ollowing an oppo unis ic fi e supp ession,i.e. adop ing open mosaics o igina ed om ag icul u al ac i i ies as oppo uni ies o s op fi es) o wo u u e pe iods (2011–2030 and 2031–2050), acco ding o di e en land-use and fi e managemen scena ios. Poin s and ba s ep esen he mean and he mean s anda d e o , espec i ely. J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 5 3. Resul s 3.1. E ec s o managemen scena ios on fi e egime Acco ding o ou simula ions, u u e fi es would bu n a ound 175,000 (RWild_CS) and 150,000 (HNV _CS) hec a es be ween 2031 and 2050 unde he cu en fi e supp ession sys em (mo e han 8500 and 7500 ha/yea , espec i ely). Scena ios wi h he highes fi e sup- p ession le els would la gely educe bu ned a eas (Fig. 3). Howe e , fi efigh ing oppo uni ies om open a eas gene a ed by ag icul u al ac- i i ies would become mos impac ul unde low supp ession scena ios (see fi e supp ession in Fig. 3), wi h significan di e ences (up o 30,000 ha; i.e. 1500 ha/yea ) be ween ewilding and HNV scena ios in he long e m (2031–2050) as a mland a eas g adually inc ease (see “c opland”in Appendix C). 3.2. E ec s o managemen scena ios on biodi e si y Ou ensemble models p esen ed high p edic i e accu acy (AUC = 0.92 ± 0.06; TSS = 0.72 ± 0.14; Appendix A) and e ealed di e en e- sponses o biodi e si y o land-use and fi e supp ession managemen (Fig. 4 and Appendix D). Land-use managemen significan ly a ec ed mo e species (83%) han fi e supp ession managemen (48%) (Fig. 4 and Appendix B). HNV scena ios would inc ease habi a sui abili y o mo e han hal o he species (N = 64), whe eas ewilding scena ios a e p edic ed o in- c ease sui able habi a s o 20% o species (N = 21). Howe e , se e al h ea ened and endemic species a e significan ly benefi ed by ewilding scena ios (Fig. 5), including he only c i ically endange ed species (hab- i a gains o mo e han 50%), and mos o he endange ed (3 ou o 5 species; Appendix E) and ulne able species (6 ou o 11 species; 15 34 53 0 20 40 High FS SE Low FS SE No FS e ec % species A 21 64 17 0 20 40 60 RWild SE HNV SE No land−use e ec % species B RWild SE HNV SE No land−use e ec High FS SE Low FS SE No FS e ec RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS −50 −25 0 25 −50 −25 0 25 −50 −25 0 25 Scena ios % o change om 2010 Land−use managemen Rewilding HNV C Fig. 4. E ec s o managemen scena ios on biodi e si y. (A) Pe cen age o speciesbenefi ed by fi e supp ession (FS) managemen scena ios (2050), independen ly o he land-use scena io and (B) pe cen age o species benefi ed by land-use managemen scena ios, independen ly o he fi e managemen scena ios (2050). (C) To al habi a sui abili y change (2010–2050) acco ding o species g oups significan ly a ec ed by (1) land-use managemen (significan e ec s - SE; species wi h no significan e ec s a e plo ed as “No FS/land-use e ec s”) and (2) fi e supp ession (FS) managemen scena ios, i.e. species significan ly a ec ed by high (High FS SE) and low (Low FS SE) fi e supp ession endencies, and wi h no significan e ec s (No FS e ec ). Poin s and ba s a e ep esen ing he mean and he mean s anda d e o , espec i ely. J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 6 Appendix F). Ibe ian endemics would be pa icula ly benefi ed by HNV unde low fi e supp ession le els (15% o habi a gain). S ill, se e al en- demic species also p ofi om ewilding scena ios, conc e ely hose sig- nifican ly a ec ed by high supp ession s a egies and hose no a ec ed by fi e supp ession managemen (Fig. 5). Conce ning he e ec s o fi e supp ession managemen , a high p o- po ion o species (33%) is p edic ed o benefi om low fi e supp es- sion (i.e., fi e-media ed landscapes), while only 15% p ofi om high fi e supp ession, bo h in de imen o he cu en fi e supp ession sys- em (Fig. 4). 4. Discussion 4.1. E ec s o managemen s a egies on fi e mi iga ion The cu en endency o land abandonmen would inc ease he im- pac s o u u e fi es (Fig. 3), which migh be explained by enc oachmen o ege a ion domina ed by mo e flammable species in a eas p e iously occupied by a mlands and spa se ege a ion (Appendix D). The imple- men a ion o an HNV scena io could ep esen a mo e e ficien (and cos -e ec i e) fi e mi iga ion s a egy in he long- e m (Fig. 3), as also demons a ed in o he egions o he Ibe ian Peninsula (Aquilué e al., 2020). Ou fi e- ege a ion simula ions ein o ce hese assump ions, by demons a ing po en ial ad an ages o le ing unplanned fi es bu n o inc ease fi e supp ession oppo uni ies de i ed om new open a eas (Fig. 3). Le ing unplanned fi es bu n unde a o able wea he condi ions in landscapes wi h highe a ailabili y o semi-open a eas migh con ibu e o inc ease fi e-supp ession oppo uni ies by educing uel accumula ion and p omo ing accessibili y o fi efigh ing (Duane e al., 2019). 4.2. E ec s o managemen s a egies on biodi e si y Th ee main findings should be highligh ed omou esul s: 1) he e is a high a iabili y o biodi e si y esponses o di e en combina ions o land-use and fi e supp ession managemen s a egies; 2) HNV man- agemen would benefi he majo i y o species (mo e han 60%) in de - imen o ewilding scena ios (Fig. 4); and 3) a subs an ial numbe o species (mo e han a hi d) would p ofi om scena ios wi h mo e fi e in he landscape (i.e., low fi e supp ession le els; Fig. 4). The majo - i y o he s udied species p ofi ed om HNV scena ios, ein o cing ha main aining some ag icul u al a eas con ibu es o biodi e si y a he landscape le el (Halada e al., 2011;Lomba e al., 2020). The e ec i e implemen a ion o HNV policies could ha e an impac ul epe cussion on biodi e si y conse a ion in ou s udy a ea, gi en he small ex en o ag icul u al a eas in compa ison o o he LC ypes such as sh ublands (c.a. 60%) and o es s (c.a. 20%). Land abandonmen and subsequen ege a ion enc oachmen leads o habi a losses o a mland specialis s and o he species adap ed o open and semi-open a eas (Appendices B and D).Howe e , o he species would benefi om hese condi ions, es- pecially when associa ed wi h high fi e supp ession s a egies (Fig. 4; Appendices B and D). These species a e mos ly ep esen ed by o es High FS SE Low FS SE No FS e ec RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS −50 0 50 Managemen scena ios % o change om 2010 IUCN s a us CR EN VU SE − Signi ican e ec Fi e managemen e ec s (% habi a change) on h ea ened species (IUCN c i e ia) A High FS SE Low FS SE No FS e ec RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS RWild_LS RWild_CS RWild_HS HNV _LS HNV _CS HNV _HS −75 −50 −25 0 Managemen scena ios % o change om 2010 Endemic species SE − Signi ican e ec Fi e managemen e ec s (% habi a change) on Ibe ian endemic species B Fig. 5. E ec s o managemen scena ios on h ea ened and endemic biodi e si y. (A) Changes (%) o habi a sui abili y o 2050 (in compa ison o 2010) o h ea ened ( egional IUCN c i e ia) and (B) endemic species om he Ibe ian Peninsula. The species a e plo ed acco ding o he fi e supp ession (FS) managemen scena ios significan e ec s (SE), i.e. species significan ly a ec ed by high (High FS SE) and low (Low FS SE) fi e supp ession endencies, and wi h no significan e ec s (No FS e ec ). Poin s and ba s ep esen he mean and he mean s anda d e o , espec i ely. J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 7 bi d specialis s and amphibians (Appendices B and D), which a e mo e dependen on we e and coole habi a s wi h dense ege a ion. The impo ance o fi e lies in he c ea ion and main enance o open habi a s, pa icula ly in a eas suscep ible o g adual a mland abandon- men ha would p oduce mo e homogeneous and flammable land- scapes. In ac , he posi i e esponse o a la ge g oup o species o lowe fi e supp ession le els (Fig. 4; Appendices B and D) migh be p e- cisely ela ed o he c ea ion o open spaces(and agmen a ion o wood and sc ub ma ices) p omo ed by wildfi es. New open habi a s can be po en ially ecolonized om su ounding habi a pa ches, a pa e n al- eady acknowledged o ep iles and ea ly successional bi d species (Regos e al., 2016;Che gui e al., 2019). S ill, p e ious s udies ha e obse ed dis inc esponses o biodi e si y communi ies o fi e acco d- ing o di e en pos -fi e eco e y and ecoloniza ion ajec o ies in Medi e anean bu n a eas (B os e al., 2011;Ma eos e al., 2011; San os e al., 2016). All hese con as ing pa e ns sugges a con ex - dependen esponse (i.e. habi a dominance a he landscape le el and habi a p e e ence o he species) o biodi e si y and bio ic communi- ies o fi e-landscape dynamics (Pausas and Pa , 2018). 4.3. Implica ions o biodi e si y conse a ion and landscape managemen 4.3.1. Op imal managemen scena io: HNV wi h high fi e supp ession The implemen a ion o HNV combined wi h high fi e supp ession s a egies, al hough demanding a ca e ul s a egic planning o a oid ex- eme fi e haza ds de i ed om fi efigh ing ap e ec s, would ep e- sen an op imal managemen app oach ha could con ibu e o bo h fi e mi iga ion and biodi e si y conse a ion (Fig. 4). The his o ical human in e en ions ha shaped Medi e anean landscapes o e millennia sus ained high le els o biodi e si y. The p omo ion o HNV would benefi se e al species adap ed o hese semi-na u al habi a s and dependen on he con inua ion o specific a ming sys ems (Fig. 4 and Appendix D), including many axa o conse a ion conce n cu - en ly h ea ened by land abandonmen p ocesses (F anks e al., 2018; Pais e al., 2020;Fig. 5). The con inuous abandonmen o hese p ac ices will cause nega i e impac s o se e al species (Fig. 4 and Appendix D), which highligh s he need o s able u al policies and long- e m socio- economic p ojec s o sus ain ag icul u al ac i i ies in u al moun ain a eas. None heless, he successand con inua ion o HNV demands ailo ed socio-economic and poli ical s a egies explo ing al e na i e and mo e e ficien solu ions han economic incen i es (Lomba e al., 2020). In- c easing capaci y building, socie al coope a ion and aising awa eness o he alues o biodi e si y and ecosys em se ices in hese egions a e pa amoun o sus aining he na u al and cul u al he i age associ- a ed wi h hese adi ional ac i i ies. S ill, gi en he cu en ly domina - ing endencies o land abandonmen and he di ficul ies o ein o ce he cu en fi e supp ession sys em (due o he high cos s in ol ed and he low accessibili y in s eep moun ainous a eas), HNV should be consid- e ed as complemen a y o o he al e na i e managemen scena ios o assu e cos -e ec i eness and adequa e implemen a ion. 4.3.2. Al e na i e managemen scena io: using fi e o enhance ewilding Ou simula ions suppo ewilding as a beneficial scena io o o es - adap ed species and axonomic g oups conside ed mos ulne able o en i onmen al and landscape dis u bances (e.g. amphibians). Mo e- o e , ewilding scena ios would p omo e s abiliza ion o sui able a eas and conside able inc eases o a ailable habi a s o se e al h ea ened and endemic species inside he ese e (Fig. 5). The habi a gains a e pa icula ly no iceable o c i ically endange ed (Appendix E) and ul- ne able axa (Appendix F), highligh ing he po en ial o ewilding poli- cies o p o ec ing unique biodi e si y o conse a ion conce n. Acco ding o hese esul s, and gi en he o e all ine ec i eness o cu en Eu opean ag icul u al policies (Pe'e e al., 2014, 2019), ou s udy sugges s an al e na i e and p ac icable scena io cen e ed on fi e-media ed ewilding, consis en wi h he con inuous end o depopula ion in he s udy a ea. In such a scena io, he po en ial haza ds o ewilding would be bu e ed by low fi e supp ession s a egies, bo h by con olling he e ec s o unplanned fi es and by using s a egic p e- sc ibed fi es. This scena io would con ibu e o an oppo unis ic fi e mi iga ion and o po en ial inc eases in py odi e si y o suppo ing biodi e si y conse a ion (Kelly e al., 2015, 2017). Al e na i e scena ios in eg a ing bo h ewilding and HNV manage- men a he landscape le el should also be conside ed (Me ckx and Pe ei a,2015;Lomba e al., 2020),since hei combinedimplemen a ion would po en ially con ibu e o achie e di e en conse a ion goals (e.g. conse a ion o biodi e si y adap ed o ag icul u al and o es ed a eas) and o secu e a wide ange o ecosys em se ices (Pais e al., 2020). Fu u e s a egic planning and decision-making in he egion, howe e , should conside he p ospec i e ade-o s be ween fi e mi i- ga ion, ecosys em se ices and conse a ion o di e en ace s o biodi e si y. 5. Conclusions This s udy con ibu es o he inc easing e idence o ag icul u al policies as essen ial ools o ensu e biodi e si y while educing fi e haza d, an aspec ha has been equen ly neglec ed when assessing he beneficial e ec s o ag icul u al policies. Also, ou s udy sugges s using fi e o enhance ewilding as an al e na i e managemen s a - egy in ou s udy a ea —an issue ha decision make s and manage s should conside when implemen ing ewilding ini ia i es in o he fi e-p one egions. Addi ionally, ou s udy highligh s he need o enewed poli ical and socio-economic e o s explo ing di e en so- lu ions o economic incen i es and/o managemen s a egies in eg a ing bo h ewilding and HNV . In his con ex , ou s udy dem- ons a es how an e ec i e implemen a ion o Eu opean ag icul u al policies could benefi biodi e si y ( h ough he c ea ion o new open habi a s o endange ed species) while p o iding u he fi e- supp ession oppo uni ies. Ou s udy also shows how fi e supp es- sion policies can help he implemen a ion o ewilding ini ia i es in o he abandoned, fi e-p one moun ain a eas ac oss Sou he n Eu ope. I also goes beyond he business-as-usual scena ios and p o- ides plausible u u e pa hways whe ein ewilding modula ed by fi e supp ession can eme ge as na u e-based solu ion i he new EU Common Ag icul u al Policy con inues o ail a e e sing u al aban- donmen ends. F om a me hodological pe spec i e, ou s udy high- ligh s he need o aking in o accoun he landscape dynamics and fi e- ege a ion eedbacks when modelling u u e managemen sce- na ios in moun ain landscapes o Sou he n Eu ope —highly dynamic a eas his o ically d i en by he spa ial in e ac ion be ween fi e- ela ed p ocesses (such as fi e and pos -fi e eco e y) and human in- e en ions (land-use and fi e supp ession policies). CRediT au ho ship con ibu ion s a emen Concei ed and designed he expe imen s: JCC, AR. Expe imen s and da a analyses: JCC, JB, AR. Con ibu ed da a/analysis ools: NA, JD, BM, FM-F. D a p epa a ion: JCC. Visualiza ion: JCC. W i ing, e iew and edi ing: all au ho s. Decla a ion o compe ing in e es The au ho s decla e ha hey ha e no known compe ing financial in e es s o pe sonal ela ionships ha could ha e appea ed o influ- ence he wo k epo ed in his pape . Acknowledgemen s The au ho s hank Ad ián Lamosa To es, Xosé Pa da ila and Albe o Gil o hei help du ing fieldwo k in Xu és; and Ra ael Vázquez o p o- iding addi ional da a o amphibians and ep iles. J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 8 Funding This esea ch was suppo ed by Po uguese na ional unds h ough FCT - Founda ion o Science and Technology, I.P., unde he Fi ESma p ojec (PCIF/MOG/0083/2017), and by p ojec INMODES (CGL2017-89999-C2-2-R) unded by he Spanish Minis y o Science and Inno a ion. AL, FM and FM-F we e suppo ed by FCT (DL57/2016/CP1440/ CT0001, IF/01053/2015; and DL57/2016/CP1440/CT0010, espec- i ely). AR was suppo ed by he Xun a de Galicia (ED481B2016/ 084-0) and IACOBUS p og am (INTERREG V-A España–Po ugal, POCTEP 2014–2020). Appendix A–F. Supplemen a y da a Supplemen a y da a o his a icle can be ound online a h ps://doi. o g/10.1016/j.sci o en .2020.142897. Re e ences Aquilué, N., Fo in, M.-J., Messie , C., B o ons, L., 2020. The po en ial o ag icul u al con e - sion o shape o es fi e egimes in Medi e anean landscapes. Ecosys ems 23, 34–51. h ps://doi.o g/10.1007/s10021-019-00385-7. A aújo, M.B., Alagado , D., Cabeza, M., Nogués-B a o, D., Thuille , W., 2011. Clima e change h ea ens Eu opean conse a ion a eas. Ecol. Le . 14, 484–492. h ps://doi. o g/10.1111/j.1461-0248.2011.01610.x. Bajocco, S., Fe a a, C., Guglie a, D., Rico a, C., 2019. Fi een yea s o changes in fi e igni- ion equency in Sa dinia (I aly): a ich-ge - iche p ocess. Ecol. Indic. 104, 543–548. h ps://doi.o g/10.1016/j.ecolind.2019.05.030. Be y, L.E., Lindenmaye , D.B., D iscoll, D.A., 2015. La ge unbu n a eas, no small unbu n pa ches, a e needed o conse e a ian di e si y in fi e-p one landscapes. J. Appl. Ecol. 52, 486–495. h ps://doi.o g/10.1111/1365-2664.12387. B os, V., Mo eno-Rueda, G., San os, X., 2011. Does pos fi e managemen a ec he eco - e y o Medi e anean communi ies? The case s udy o e es ial gas opods. Fo . Ecol. Manag. 261, 611–619. h ps://doi.o g/10.1016/j. o eco.2010.11.014. Chas-Amil, M.L., P es emon, J.P., McClean, C.J., Touza, J., 2015. Human-igni ed wildfi e pa - e ns and esponses o policy shi s. Appl. Geog . 56, 164–176. h ps://doi.o g/ 10.1016/j.apgeog.2014.11.025. Che gui, B., Fahd, S., San os, X., 2019. A e ep ile esponses o fi e shaped by o es ype and ege a ion s uc u e? Insigh s om he Medi e anean basin. Fo . Ecol. Manag. 437, 340–347. h ps://doi.o g/10.1016/j. o eco.2019.01.046. Domínguez, J., A enas, M., Tapia, L., 2005. Guía de las a es del Pa que na u al Baixa Limia- Se a do Xu és y LIC Baixa Limia. Conselle ía de Medio Ambien e, Xun a de Galicia. Domínguez, J., Lamosa, A., Pa da ila, X., Ma ınez-F ei ıa, F., Regos, A., Gil, A., Vidal, M., 2012. A las de los e eb ados e es es ep oduc o es en el Pa que Na u al Baixa Limia-Se a do Xu és y ZEPVN-LIC Baixa Limia. p. 423. Duane, A., Aquilué, N., Canelles, Q., Mo án-O doñez, A., De Cáce es, M., B o ons, L., 2019. Adap ing p esc ibed bu ns o u u e clima e change in Medi e anean landscapes. Sci. To al En i on. 677, 68–83. h ps://doi.o g/10.1016/j.sci o en .2019.04.348. Es oque, R.C., Gomi, K., Togawa, T., Ooba, M., Hijioka, Y., Akiyama, C.M., Ku oda, K., 2019. Scena io-based land abandonmen p ojec ions: me hod, applica ion and implica ions. Sci. To al En i on. 692, 903–916. h ps://doi.o g/10.1016/j. sci o en .2019.07.204. Fe nandes, P.M., 2013. Fi e-sma managemen o o es landscapes in he Medi e anean basin unde global change. Landsc. U ban Plan. 110, 175–182. h ps://doi.o g/ 10.1016/j.landu bplan.2012.10.014. Fe nandes, P.M., Rego, F.C., Rigolo , E., 2011. The FIRE PARADOX p ojec : owa ds science- based fi e managemen in Eu ope. Fo . Ecol. Manag. 261, 2177–2178. h ps://doi.o g/ 10.1016/j. o eco.2010.12.024. Fe nandes, P.M., Mon ei o-Hen iques, T., Guioma , N., Lou ei o, C., Ba os, A.M., 2016a. Bo om-up a iables go e n la ge-fi e size in Po ugal. Ecosys ems 19, 1362–1375. h ps://doi.o g/10.1007/s10021-016-0010-2. Fe nandes, P.M., Pacheco, A.P., Almeida, R., Cla o, J., 2016b. The ole o fi e-supp ession o ce in limi ing he sp ead o ex emely la ge o es fi es in Po ugal. Eu . J. Fo . Res. 135, 253–262. h ps://doi.o g/10.1007/s10342-015-0933-8. Fielding, A.H., Bell, J.F., 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. h ps://doi. o g/10.1017/S0376892997000088. F anks, S.E., Roodbe gen, M., Teunissen, W., Ca ing on Co on, A., Pea ce-Higgins, J.W., 2018. E alua ing he e ec i eness o conse a ion measu es o Eu opean g assland-b eeding wade s. Ecology and E olu ion 8, 10555–10568. h ps://doi.o g/ 10.1002/ece3.4532. Ga cía-Ruiz, J.M., Lana-Renaul , N., 2011. Hyd ological and e osi e consequences o a mland abandonmen in Eu ope, wi h special e e ence o he Medi e anean egion—a e iew. Ag ic. Ecosys . En i on. 140, 317–338. h ps://doi.o g/ 10.1016/j.agee.2011.01.003. Gillson, L., Ladle, R.J., A aújo, M.B., 2011. Baselines, pa e ns and p ocess. Conse a ion Biogeog aphy. Wiley-Blackwell, Ox o d, pp. 31–44. Gu man, G., Igna o , A., 1998. The de i a ion o he g een ege a ion ac ion om NOAA/ AVHRR da a o use in nume ical wea he p edic ion models. In . J. Remo e Sens. 19, 1533–1543. h ps://doi.o g/10.1080/014311698215333. Halada, L., E ans, D., Romão, C., Pe e sen, J.E., 2011. Which habi a s o Eu opean impo - ance depend on ag icul u al p ac ices? Biodi e s. Conse . 20, 2365–2378. h ps:// doi.o g/10.1007/s10531-011-9989-z. Hou man, R.M., Mon gome y, C.A., Gagnon, A.R., Calkin, D.E., Die e ich, T.G.,McG ego , S., C owley, M., 2013. Allowing a wildfi e o bu n: es ima ing he e ec on u u e fi e supp ession cos s. In . J. Wildland Fi e 22, 871–882. h ps://doi.o g/10.1071/ WF12157. Kelly, L.T., B o ons, L., 2017. Using fi e o p omo e biodi e si y. Science 355, 1264–1265. h ps://doi.o g/10.1126/science.aam7672. Kelly, L.T., Benne , A.F., Cla ke, M.F., McCa hy, M.A., 2015. Op imal fi e his o ies o biodi e si y conse a ion. Conse . Biol. 29, 473–481. h ps://doi.o g/10.1111/ cobi.12384. Kelly, L.T., B o ons, L., McCa hy, M.A., 2017. Pu ing py odi e si y o wo k o animal con- se a ion. Conse . Biol. 31, 952–955. h ps://doi.o g/10.1111/cobi.12861. Lasan a, T., A náez, J., Pascual, N., Ruiz-Flaño, P., E ea, M.P., Lana-Renaul , N., 2017. Space– ime p ocess and d i e s o land abandonmen in Eu ope. Ca ena 149, 810–823. h ps://doi.o g/10.1016/j.ca ena.2016.02.024. Loep e, L., Ma inez-Vilal a, J., Oli e es, J., Piñol, J., Llo e , F., 2010. Feedbacks be- ween uel educ ion and landscape homogenisa ion de e mine fi e egimes in h ee Medi e anean a eas. Fo . Ecol. Manag. 259, 2366–2374. h ps://doi.o g/ 10.1016/j. o eco.2010.03.009. Lomba, A., Mo ei a, F., Klimek, S., Jongman, R.H., Sulli an, C., Mo an, J., McC acken, D.I., 2020. Back o he u u e: e hinking socioecological sys ems unde lying high na u e alue a mlands. F on . Ecol. En i on. 18, 36–42. h ps://doi.o g/10.1002/ ee.2116. MAPAMA, 2018. Es adís icas de incendios o es ales. Consul ado en Ma zo de 2018. h p://www.mapama.gob.es/es/desa ollo- u al/es adis icas/Incendios_de aul .aspx. Ma mion, M., Pa iainen, M., Luo o, M., Heikkinen, R.K., Thuille , W., 2009. E alua ion o consensus me hods in p edic i e species dis ibu ion modelling. Di e s. Dis ib. 15, 59–69. h ps://doi.o g/10.1111/j.1472-4642.2008.00491.x. Ma eos, E., San os, X., Pujade-Villa , J., 2011. Taxonomic and unc ional esponses o fi e and pos -fi e managemen o a Medi e anean Hymenop e a communi y. En i on. Manag. 48, 1000. h ps://doi.o g/10.1007/s00267-011-9750-0. McLauchlan, K.K., Higue a, P.E., Miesel, J., Roge s, B.M., Schwei ze , J., Shuman, J.K., Balch, J.K., 2020. Fi e as a undamen al ecological p ocess: esea ch ad ances and on ie s. J. Ecol. 108, 2047–2069. h ps://doi.o g/10.1111/1365-2745.13403. Me ckx, T., Pe ei a, H.M., 2015. Reshaping ag i-en i onmen al subsidies: om ma ginal a ming o la ge-scale ewilding. Basic and Applied Ecology 16, 95–103. h ps://doi. o g/10.1016/j.baae.2014.12.003. Mo ei a, F., Pe'e , G., 2018. Ag icul u al policy can educe wildfi es. Science 359, 1001. h ps://doi.o g/10.1126/science.aa 8120. Mo ei a,F.,Viedma,O.,A ianou sou,M.,Cu ,T.,Kou sias,N.,Rigolo ,E.,Bilgili,E., 2011. Landscape - wildfi e in e ac ions in sou he n Eu ope: implica ions o landscape managemen . J. En i on. Manag. 92, 2389–2402. h ps://doi.o g/ 10.1016/j.jen man.2011.06.028. Mo ei a, F., Ascoli, D., Sa o d, H., Adams, M.A., Mo eno, J.M., Pe ei a, J.M., Cu , T., 2020. Wildfi e managemen in Medi e anean- ype egions: pa adigm change needed. En- i on. Res. Le . 15, 011001. h ps://doi.o g/10.1088/1748-9326/ab541e. Mo i z, M.A., Ba llo i, E., B ads ock, R.A., Gill, A.M., Handme , J., Hessbu g, P.F., Sypha d, A.D., 2014. Lea ning o coexis wi h wildfi e. Na u e 515, 58–66. h ps://doi.o g/ 10.1038/na u e13946. Na a o, L.M., Pe ei a, H.M., 2015. Rewilding abandoned landscapes in Eu ope. Rewilding Eu opean Landscapes. Sp inge , Cham h ps://doi.o g/10.1007/s10021-012-9558-7. Pais, S., Aqulué, N., Campos, J., Sil, Â., Ma cos, B., Ma ínez-F ei ía, F., ... Regos, A., 2020. Moun ain a mland p o ec ion and fi e-sma managemen join ly educe fi e haza d and enhance biodi e si y and ca bon seques a ion. Ecosys em Se ices 44, 101143. h ps://doi.o g/10.1016/j.ecose .2020.101143. Pausas, J.G., Fe nández-Muñoz, S., 2012. Fi e egime changes in he Wes e n Medi e a- nean Basin: om uel-limi ed o d ough -d i en fi e egime. Clim. Chang. 110, 215–226. h ps://doi.o g/10.1016/j.landusepol.2019.01.018. Pausas, J.G., Pa , C.L., 2018. Towa ds an unde s anding o he e olu iona y ole o fi e in animals. E ol. Ecol. 32, 113–125. h ps://doi.o g/10.1007/s10682-018-9927-6. Pe’e , G., Dicks, L.V., Viscon i, P., A le az, R., Báldi, A., Ben on, T.G., Sco , A.V., 2014. EU ag- icul u al e o m ails on biodi e si y. Science 344, 1090–1092. h ps://doi.o g/ 10.1126/science.1253425. Pe'e , G., Zinng ebe, Y., Mo ei a, F., Si ami, C., Schindle , S., Mülle , R., Hansjü gens, B., 2019. A g eene pa h o he EU Common Ag icul u al Policy. Science 365, 449–451. h ps://doi.o g/10.1126/science.aax3146. Pe’e , G., Bonn, A., B uelheide, H., Dieke , P., Eisenhaue , N., Feind , P.H., Ma qua d, E., 2020. Ac ion needed o he EU Common Ag icul u al Policy o add ess sus ainabili y challenges. People and Na u e 2, 305–316. h ps://doi.o g/10.1002/pan3.10080. Pe ino, A., Pe ei a, H.M., Na a o, L.M., Fe nández, N., Bullock, J.M., Ceaușu, S., ... Wheele , H.C., 2019. Rewilding complex ecosys ems. Science 80 (364), eaa 5570. h ps://doi. o g/10.1126/science.aa 5570. Pimen a, M., San a ém, M.L., 1996. A las das A es do Pa que Nacional da Peneda-Ge ês. Ponisio, L.C., Wilkin, K., M’Gonigle, L.K., Kulhanek, K., Cook, L., Tho p, R., K emen, C., 2016. Py odi e si y bege s plan -pollina o communi y di e si y. Glob. Chang. Biol. 22, 1794–1808. h ps://doi.o g/10.1111/gcb.13236. Quei oz, C., Beilin, R., Folke, C., Lindbo g, R., 2014. Fa mland abandonmen : h ea o op- po uni y o biodi e si y conse a ion? A global e iew. F on . Ecol. En i on. 12, 288–296. h ps://doi.o g/10.1890/120348. Regos, A., Aquilué, N., Re ana, J., De Cáce es, M., B o ons, L., 2014. Using unplanned fi es o help supp essing u u e la ge fi es in medi e anean o es s. PLoS One 9, 1–10. h ps://doi.o g/10.1371/jou nal.pone.0094906. J.C. Campos, J. Be nha d , N. Aquilué e al. Science o he To al En i onmen 755 (2021) 142897 9