Coincidence of High Nature Value farmlands with bird and butterfly diversity

Con en s lis s a ailable a ScienceDi ec Ag icul u e, Ecosys ems and En i onmen jou nal homepage: www.else ie .com/loca e/agee Coincidence o High Na u e Value a mlands wi h bi d and bu e ly di e si y Sanna Mäkeläinen a,⁎ , Annika Ha lio a , Ris o K Heikkinen b , I ina He zon c, , Mikko Kuussaa i b , Ka i Lepikkö a , And ea Maie a , Tuomas Seimola d , Juha Tiainen d , Anni A ponen e, a Facul y o Biological and En i onmen al Sciences, P. O. Box 65, FI-00014, Uni e si y o Helsinki, Finland b Finnish En i onmen Ins i u e (SYKE), P. O. Box 140, FI-00251, Helsinki, Finland c Depa men o Ag icul u al Sciences, P. O. Box 27, FI-00014, Uni e si y o Helsinki, Finland d Na u al Resou ces Ins i u e Finland (Luke), P. O. Box 2, FI-00791, Helsinki, Finland e Ecosys ems and En i onmen Resea ch P og amme, Facul y o Biological and En i onmen al Sciences, P. O. Box 65, FI-00014, Uni e si y o Helsinki, Finland Helsinki Ins i u e o Sus ainabili y Science, HELSUS, P. O. Box 65, FI-00014, Uni e si y o Helsinki, Finland ARTICLE INFO Keywo ds: Ag icul u al biodi e si y Bu e lies Fa mland bi ds High Na u e Value a mland indica o Species ichness ABSTRACT The amoun o High Na u e Value a mland (HNV ) is a commonly used en i onmen al indica o o assessing he pe o mance o he Common Ag icul u al Policy, o suppo sus ainable ag icul u e and moni o changes in ag icul u al land use in Eu ope. HNV comp ises ag icul u al a eas o semi-na u al s a e, low-in ensi y a ming and ine-scale landscape mosaics o di e en habi a ypes. Fo a success ul implemen a ion, he iden i ica ion o HNV should co ec ly e lec he a ia ion in biodi e si y alues be ween di e en ag icul u al landscapes. We examined how well he Finnish HNV indica o and he sub-indica o s cons i u ing i – ecalcula ed o he pu poses o his s udy o i e s udy egions – e lec he a ia ion in bi d and bu e ly species ichness and di e si y pa e ns a di e en spa ial scales. We ound ha bu e ly di e si y index was posi i ely associa ed wi h he HNV indica o a he ines scale o 0.5 km × 0.5 km squa es. Among he HNV sub-indica o s, ex- ensi e cul i a ion o g asslands was mos s ongly ela ed o he a mland bi d di e si y and he densi y o edge o he bu e ly di e si y. Thus, he HNV concep e lec s well he dis ibu ion o bu e lies in he Finnish ag icul u al landscapes bu insu icien ly he di e si y pa e ns o a mland bi ds. Impo an ly, semi-na u al ege a ion and long- e m pas u es – he backbone o he concep – p esen ly occu in small and highly ag- men ed pa ches in ag icul u al landscapes in Finland. The Pan-Eu opean concep o HNV has es ic ed appli- ca ion o a mland bi ds o his bo eal coun y and he na ional HNV concep may need o be e ised. 1. In oduc ion The e a e con as ing challenges in using ag icul u al en i onmen s: he impe a i e o in ensi ying ood p oduc ion o he g owing human popula ion, on one hand, and he need o p o ec ing biodi e si y and ecosys em se ices wi hin hem, on he o he hand (Foley e al., 2011). In he p essu e o hese con lic ing in e es s, many a mland bio opes o impo ance o biodi e si y, as well as hei bio a, ha e declined d as- ically du ing he las cen u y because o he in ensi ied p oduc ion (Ben on e al., 2003;Donald e al., 2001;S oa e e al., 2001). Fo ex- ample, he e a e well documen ed popula ion declines h oughou Eu ope in a mland bi ds (Bu le e al., 2010;G ego y e al., 2005) and bu e lies ( an Swaay e al., 2010,2015). Imp o ing he s a e o bio- di e si y in p esen -day ag icul u al landscapes is he e o e c ucial o hal ing he loss o biodi e si y (Bu cha e al., 2010). The concep o ‘High Na u e Value a mlands’ was de eloped in ecogni ion o he impo ance o ce ain cha ac e is ics o a mland o biodi e si y in Eu ope (Ande sen e al., 2003). Based on he High Na u e Value a mland (he ea e HNV ) concep , a sui e o a mland indica o s was de ined o assess he en i onmen al pe o mance o he Common Ag icul u al Policy (CAP) and he e ec i eness o Pilla 2 Ru al De elopmen P og ams (Lomba e al., 2014). Though he HNV is a landscape-le el concep , i also includes a c i e ion ha ocuses on species whose su i al depends on HNV a mlands (Ande sen e al., 2003). Howe e , only li le is known abou how well he HNV -based indica o s ul il an impo an c i e ion o biodi e si y indica o s: a plausible associa ion wi h he key unde lying biodi e si y elemen s, such as occu ences o declined and a e species (G ego y e al., 2005). h ps://doi.o g/10.1016/j.agee.2018.09.030 Recei ed 8 Ma ch 2018; Recei ed in e ised o m 12 July 2018; Accep ed 24 Sep embe 2018 ⁎ Co esponding au ho a : Finnish Museum o Na u al His o y Luomus, P. O. Box 17, FI-00014, Uni e si y o Helsinki, Finland. E-mail add ess: [email p o ec ed] (S. Mäkeläinen). Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 0167-8809/ © 2018 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/). T HNV e e s o a eas whe e ag icul u e is o en he p edominan land-use ype, and which a e cha ac e ized by ‘a high species and ha- bi a di e si y’ o ‘ he p esence o species o Eu opean conse a ion conce n o bo h’ (Ande sen e al., 2003). HNV needs o ul il a leas one o he ollowing c i e ia: (i) con ain a high p opo ion o semi- na u al ege a ion, (ii) comp ise a mosaic o ex ensi e and in ensi e ag icul u e and s uc u al elemen s such as ield ma gins and hedge- ows, o (iii) hos a e species, wi h bi d species as a ocal g oup (Pa acchini e al., 2008). Wi hin he CAP, HNV ea u es a e suppo ed only by he ag i-en i onmen -clima e measu e (AECM) unde Pilla 2. Howe e , he AECMs ha e shown mixed ou comes o a mland bio- di e si y and pa ial ine ec i eness (Ba á y e al., 2015;Kleijn e al., 2011;Tscha n ke e al., 2005). The iden i ica ion o HNV a he Eu opean le el is based on CORINE Land Co e ypes ha a e closely- ela ed o he semi-na u al ag icul u al elemen s, as well as on da a on bo h a ming sys ems and bi d species dis ibu ion (Ande sen e al., 2003;Pa acchini e al., 2008). Howe e , he e is much geog aphic a ia ion in he amoun , p esence and cha - ac e is ics o HNV ac oss Eu ope (Lomba e al., 2014;Schulman e al., 2005). Gi en his, he EU Membe S a es ha e been encou aged o de- elop hei own na ional-le el indica o s o measu e, moni o and epo he p opo ion o HNV (Benede i, 2017). The de elopmen o such in- dica o s has g ea ly a ied among he EU membe s a es and no common app oach cu en ly exis s o Eu opean-wide assessmen (Lomba e al., 2014). To de e mine he na ional-le el HNV indica o s wi h maximal po en iali y, i is impe a i e o examine how hey co espond o asso- cia ed a mland biodi e si y (see Aue e al., 2014;Benede i, 2017;Doxa e al., 2010;Janišo á e al., 2014;Mo elli e al., 2014). In Finland, he HNV indica o was de eloped in 2006 (Heliölä e al., 2009) o he manda o y assessmen o he Ru al De elopmen P og am. I ollowed he EU guidelines (EC, 2006), emphasizing he need o simple and cos -e icien annual calcula ion. A he hea o he HNV indica o in Finland is habi a a ailabili y and s uc u al com- plexi y, simila ly o elsewhe e in he EU. Ac ual dis ibu ion and di- e si y o a mland species is no conside ed. The na ional Finnish HNV indica o is calcula ed based on annual ield-le el in o ma ion, collec ed as pa o a m-le el ag icul u al epo ing. The indica o sums up sco es om h ee so-called s ong sub-indica o s (a eas o semi-na u al g asslands, pe manen pas u es and hose unde ele an AECM con ac s), and h ee weak sub-indica o s (edge densi y, ex- ensi e cul i a ion and li es ock a ming). All he sub-indica o s e lec he spa ial a ia ion in elemen impo an o biodi e si y in Finland (Heliölä e al., 2009), bu he h ee s ong sub-indica o s a e p esumed o be mo e c i ical o a mland biodi e si y han he h ee weak ones, and hus hey ecei e highe weigh s in he HNV indica o . O iginally, only a ms ha each he sco e o 20 o he na ional indica o we e classi ied as HNV (Heliölä e al., 2009) and, acco ding o he indica o , he HNV has declined om 10.2 o 8.7% o ag icul u al a ea in 2006–2014 (Biodi e si y. i, 2015). The o al indica o sco e has been con i med o co ela e wi h local habi a di e si y (Heliölä and He zon, 2012), bu i emains unknown how he indica o co esponds o spe- cies ichness and di e si y. The objec i e o his s udy was o use wo well-known indica o axa - a mland bi ds and diu nal bu e lies - he di e si y o which e lec s biologically aluable ag icul u al habi a s a di e en spa ial scales (G ego y e al., 2005; an Swaay e al., 2015). We in es iga ed whe he he amoun o HNV , ecalcula ed o ma ch he esolu ion o biological da a is posi i ely ela ed o he ichness and di e si y o species, and a which spa ial scales s onges ela ionships eme ge. We ela ed he HNV o wo aspec s o biodi e si y: (i) all a mland species eco ded in ou s udy a eas, and (ii) a subse g oup o declining species (bu e lies) o ed-lis ed species (bi ds) (Kuussaa i e al., 2007;Mikkola-Roos e al., 2010). Since he Finnish HNV indica o is based on sepa a e sub-in- dica o s ha can ela e in di e en ways o he di e si y o bi ds and bu e lies, we also examined whe he hese HNV sub-indica o s ela e o species di e si y measu es. 2. Me hods 2.1. Bi d and bu e ly di e si y da a Species dis ibu ion da a we e collec ed du ing esea ch pe iods 2 and 3 o MYTVAS moni o ing p og am (2000–2006 and 2007–2013) ha s udied he impac s o ag i-en i onmen al paymen s on biodi- e si y in Finland (Aakkula and Leppänen, 2014;Kuussaa i e al., 2008). A simila assessmen was made in a single yea o 2010 on he Åland islands (Sandholm e al., 2012;Tiainen e al., 2012). In his s udy, Finland was di ided in o i e geog aphical egions: Ka elia, Os obo hnia, sou he n Finland, sou hwes e n Finland and he Åland Islands. Fi s ly, his di ision was because he Åland islands is an au onomous egion o Finland which has an own ag icul u al policy, including he HNV assessmen and he ag i-en i onmen al p og am. Secondly, he s udy a eas in mainland Finland we e, o logis ical and esou ce easons, si ua ed wi hin 150-km dis ance om ou base si es (ci ies o Helsinki, Joensuu, Tu ku and Vaasa), om whe e he MYTVAS p og am was managed. Howe e , hese ou mainland a eas di e o hei biogeog aphic, clima ic and ag icul u al ea u es (Rikkinen, 1994). Fo example, he amoun o ag icul u al a ea and he size o con inuous a mland landscape dec ease and he p opo ion o dai y a ming and he agmen a ion o a mland landscape inc ease when mo ing om sou hwes e n Finland, o sou he n Finland, Os obo hnia, and inally o Ka elia. The Åland islands is a mix u e o all hese ea- u es because o i s loca ion in he Bal ic Sea, on one main and se e al smalle islands. In o al, 68 1-km 2 squa es we e placed in a mland landscapes in hese i e egions ( o he di ision o egions and he loca ion o hei squa es see Fig. 1). On con inen al Finland, hal o hem we e andomly selec ed, and each squa e had a andomly selec ed couple a a dis ance o 10–20 km. On he Åland islands all he squa es we e andomly se- lec ed excep o ha hey we e eques ed o be loca ed mo e han 5 km om each o he . Squa es had o consis a leas 20% o a mland. We used bi d e i o y da a ga he ed in 2005 om he squa es si ua ed on mainland Finland (52 squa es) and in 2011 om 10 squa es loca ed on he Åland Islands. Bi d e i o ies we e coun ed in all a mland o he 1- km 2 squa es, excluding o es . We used a mapping me hod in which he squa es a e su eyed h ee imes du ing May and he i s hal o June. The accumula ing obse a ions we e ans e ed o species maps on which he e i o ies we e in e p e ed, in he case o abundan species e y much based on eco ds o simul aneous obse a ions. The ea e , a ough mid-poin was es ima ed based on obse a ions in e i o ies in e p e ed and in oduced in a geo e e enced da abase. Only species de ined as a mland bi ds, excluding o es species, we e eco ded (Tiainen and Pakkala, 2001). Fo bu e ly censuses, he 1-km 2 squa es we e di ided in o ou equal squa es o 500 m × 500 m (25 ha in size) and wo o hem we e selec ed o ep esen he mos he e ogeneous and mos homogeneous landscapes om each 1-km 2 squa e; he selec ion was made wi h he aid o opog aphic maps and ae ial pho os, based on he numbe and size o ields and he amoun o e ges be ween hem o be ween hem and o es o a ms eads (Ek oos e al., 2010;Kuussaa i e al., 2004). A o al o en 50-m line ansec s we e placed wi hin bo h o hese squa es, so ha ansec s consis ed o one as homogeneous habi a ype as possible. They had also o be mo e han 50 m apa om each o he , and mos ly hey we e on di e en e ges ( his was no always possible). These line ansec s we e coun ed o bu e lies in 2010 on mainland o Finland (58 squa es) and in 2011 on he Åland Islands (10 squa es). We summed up he bu e lies coun ed in each 500-m × 500-m squa e o achie e a s anda d numbe o indi iduals pe 1000 m o coun ed ansec in each 1-km 2 squa e. We es ima ed he di e si y o bi ds and bu e lies sepa a ely. The di e si y measu es used we e species ichness (S), and he Shannon- Wiene di e si y index (H’), since hese depic di e en aspec s o species di e si y, species ichness weigh ing each species equally and H’ S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 225 gi ing mo e weigh o abundan species han o a e ones. The same measu es we e also coun ed o he subse s o species o conse a ion conce n. These subse s included 10 na ionally ed-lis ed bi d species ha ha e been on decline na ionally acco ding o IUCN c i e ia (see Mikkola-Roos e al., 2010) and 11 bu e ly species inhabi ing bo eal ag icul u al landscapes whose na ional popula ions ha e been declining in he ecen decades and which hus a e use ul indica o s o aluable a mland bio opes and biodi e si y (see Kuussaa i e al., 2007; o ull lis s o all eco ded species see Appendix A). Following hese species selec ion p ocedu es, he subse s o bi ds and bu e lies we e o he same o de o size and we e well ep esen ed wi hin ou da a. 2.2. High Na u e Value a mland indica o The Finnish HNV indica o (HNV i) was de i ed om he a m- le el da a in he egis y main ained by he In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y. The HNV i is compu ed by sum- ming he h ee s ong sub-indica o s (a eas o semi-na u al g asslands, pe manen pas u es and pa icula AECM con ac s), and he h ee weak sub-indica o s (edge densi y, ex ensi e cul i a ion and li es ock a ming) (Heliölä e al., 2009, o he s uc u e o Finnish HNV i see Table 1). Acco dingly, a ms wi h he highes sco es a e cha ac e ized by majo p opo ions o semi-na u al g asslands, pe manen pas u es o a eas wi h biodi e si y- ele an AECM con ac s o all o hem. In ad- di ion, a ms ha e agmen ed ield s uc u e wi h small pa cels, a low p opo ion o in ensi e c ops such as ce eals, and li es ock. The e- gional dis ibu ion o hese a m-le el HNV i sco es is p esen ed in Appendix B. O iginally, a ms wi h a h eshold sco e o 20 a e ega ded as HNV (see map in Appendix B). Howe e , he species da a o his s udy ha e been collec ed o andomized squa es ha ypically include pa s o se e al indi idual a ms. To ci cum en he spa ial misma ch in ou analysis, we calcu- la ed he HNV i alues o each o he squa es whe e species da a we e sampled ollowing he exac ly same p ocedu e o he same da a as used o he na ional HNV i. We used da a whe e each ield pa cel o a a m had an a ibu e o he c ops and cul i a ion ype and combined his in o ma ion wi h a geospa ial ec o da a o ield pa cels (Agency o Ru al A ai s, 2007;In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007). We c ea ed as e laye s (cell size o 25 m × 25 m) ha ep esen ed he HNV sub-indica o s (Heliölä e al., 2009) and co e ed he en i e sou he n pa o Finland. The i s h ee s ong sub-indica o s ha e been conside ed as he main c i e ia o HNV and hus he cons uc ion o he o al HNV i sco e is essen ially based on hem; in he calcula ion o he HNV i, his is enabled ia he wide po en ial ange o sub-indica o alues com- pa ed o he alues o h ee weak sub-indica o s (Table 1). The s ong sub-indica o s we e quan i ied as ollows: 1) he p opo ion o semi- na u al g asslands o u ilized ag icul u al a ea (UAA) o he squa e ( alue ange om 0 o 100), 2) he p opo ion o pe manen pas u es wi hin UAA ( ange 0–100), and 3) he p opo ion o ields wi h pa i- cula ag i-en i onmen al con ac s wi hin UAA ( ange 0–100). No e ha hese ca ego ies may o e lap, ha is, a semi-na u al g assland can be bo h a pe manen pas u e, and unde he AECM. Thus, a semi-na - u al g assland ha is bo h a pe manen pas u e and unde AECM is conside ed h ee imes as aluable as o he semi-na u al g asslands. In Finland, AECM-based con ac s o Managemen o adi ional u al bio opes, P omo ing na u al biodi e si y and landscape de elopmen and managemen , whe eas on he au onomous Åland Islands, con ac s o Na u al pas u es and wooded meadows we e included (Heliölä e al., 2009). The sub-indica o 4 desc ibes he agmen a ion o ields. He e, all ield edges we e con e ed o lines, bu no spa ial duplica es pe squa e we e allowed. The edge densi y (m/ha) was hen calcula ed by di iding he summed leng h o all ield edges by he combined ield a ea. The alues we e linea ly e-scaled om 0 o 30. The sub-indica o 5 is based on he p opo ion o UUA unde ex ensi e cul i a ion ( o lis o a mland ypes conside ed as ex ensi e cul i a ion see Appendix C) simpli ied o a scale o 0 o 10 poin s. Fo he six h sub-indica o , ei he 0 o 5 poin s we e appoin ed, e lec ing he absence o p esence o li- es ock a ms wi hin he squa e. The summed alue o he six indi idual sub-indica o s cons i u es he o al sco e o HNV i, wi h 345 as he maximal po en ial alue. This o al sco e was calcula ed o each 1-km 2 squa e and o i s su ounding landscape a di e en spa ial scales. The landscape con ex was ac- coun ed o by bu e ing squa es wi h 0.5, 1, 2, and 5-km adii, c ea ing bu e s in ela ion o cen oid o he squa e by he bu e polygons ool in A cGIS. While calcula ing HNV i alues o hese bu e ed a eas, he a ea o he squa e wi hin each bu e was always excluded (i.e. bu e s we e ounded squa e-shaped ings a ound a squa e). Landscape clas- si ica ion was conduc ed by A cGIS 10.3.1 (Es i, 2015) and he Fig. 1. Map o he loca ion o squa es (n = 68) o s udy he coincidence o High Na u e Value a mland wi h species di e si y in sou he n pa o Finland. The 1-km 2 squa es a e deno ed by small black squa es and hei su ounding landscape wi h a 5-km adius is deno ed by ci cles. Colou s o he ci cles indica e geo- g aphical egions whe eas g ey shaded a eas ep esen a mland a eas in Finland. S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 226 compu a ion o HNV i alues in R p og amming en i onmen (R co e eam, 2016). 2.3. S a is ical analyses We used linea and gene alized linea mixed models (LMMs and GLMMs, espec i ely) whe e he geog aphical egion was added as a andom e ec o he in e cep . To a oid con e gence p oblems in mixed models, explana o y a iables we e s anda dized wi h a mean o 0 and a s anda d de ia ion o 0.5 (G uebe e al., 2011). Fi s ly, we analyzed which spa ial scale o he HNV i was mos s ongly ela ed o he species ichness by GLMMs wi h Poisson e o s uc u e, whe e he numbe o species was he esponse a iable. The ela ionship o HNV i a di e en spa ial scales wi h he Shannon-Wiene di e si y index was also ana- lyzed using he same explana o y a iables. He e, we employed LMMs and he calcula ed alue o di e si y index as he esponse a iable. Secondly, we s udied he impo ance o indi idual HNV sub-indica o s on he conside ed di e si y measu es using GLMMs wi h Poisson dis- ibu ion in he models o he numbe o species and LMMs in he models o he Shannon-Wiene di e si y index. In gene al, ou models wi h Poisson e o s we e no o e dispe sed ( esidual de iance di ided by he esidual deg ees o eedom < 1), excep o models o ed- lis ed bi d and declining bu e ly species ichness, which showed sligh unde dispe sion (0.7–0.8). He e, we es ed o using nega i e binomial dis ibu ion assump ion ins ead o Poisson e o models, bu a com- pa ison wi h de i ed Akaike’s In o ma ion C i e ion (AIC) alues in- dica ed ha his change did no imp o e model i s. Spa ial au o- co ela ion in model esiduals was assessed isually by spline co elog ams wi h 95% boo s ap con idence in e als o GLMMs and by a iog ams o LMMs (Bjø ns ad and Falck, 2001). Signi ican spa ial au oco ela ion was no obse ed o al e na i ely, adding spa ial co - ela ion s uc u es did no imp o e he model i s. We used mul i-model in e ence and compa ed compe ing models wi hin each di e si y measu e and each species da a se based on hei AIC alues co ec ed o small sample sizes (AIC c ,Bu nham and Ande son, 2002). We checked o he collinea i y be ween explana o y a iables in he models using a iance in la ion ac o s (VIFs). We aimed o use he same candida e models o each di e si y measu e and each species da a se , and he null model was always included among candida e models. Howe e , i was no possible o cons uc a global model including all possible a iables due o high collinea i y be ween HNV i alues a consecu i e spa ial scales. Thus, he e he se o can- dida e models comp ised o sepa a e models cons uc ed o each spa- ial scale, in addi ion o geog aphical loca ion o he squa e and null model ( o ull lis o candida e models see Appendix D). To in es iga e he ela ionship o he HNV sub-indica o s wi h di e si y, we used a o emen ioned di e si y measu es as esponse a iables and log- ans o med sub-indica o alues as explana o y a iables. The i s h ee sub-indica o s (semi-na u al g asslands, pe manen pas u es and a eas wi h ag i-en i onmen al con ac s) we e summed up in o one alue ep esen ing a s ong indica o o HNV . The h ee o he a iables we e ield edge densi y, ex ensi e cul i a ion and ca ego ical a iable o he p esence o li es ock a ming. The de i ed VIFs we e lowe han 1.5 sugges ing o a de oid o collinea i y p oblems in he da a. In case whe e he e was no iceable model unce ain y and mo e han one model ecei ed s ong suppo , we conduc ed model-a e - aging o e he bes -app oxima ed models wi hin ΔAIC c < 2 by he ull- a e aging me hod (Symonds and Moussalli, 2011) and epo he e- sul s o 95% con idence in e als wi h ela i e a iable impo ances. Rela i e impo ance (Σwi) sums up he weigh s o all models whe e he a iable is p esen and hus desc ibes he p obabili y ha a a iable is componen o he bes model (Symonds and Moussalli, 2011). Lis s o bes -app oxima ed models a e p esen ed as supplemen a y ables, in appendices. All s a is ical analyses we e un by R [ e sion 3.3.1] wi h packages lme4 (Ba es e al., 2015) o mixed models and MuMIn (Ba on, 2014) o model a e aging. Table 1 S uc u e and a ionale o he Finnish High Na u e Value a mland (HNV ) indica o (a e Heliölä e al., 2009). Sub-indica o De ini ion Ra ionale Sou ce Spa ial scale Uni Pa ial sco e To al sco e = HNV i S ong Semi-na u al g asslands Na u al pas u es and meadows Cen al o HNV concep In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007 Field pa cel % o UAA 0–100 Pe manen pas u es Field pa cels used o cul i a e hay and eed plan s a leas du ing i e p eceding yea s, kep open by g azing o mowing G azed and open g asslands a e cen al o a mland biodi e si y (Ki inen e al., 2006;Luo o e al., 2004) In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007 Field pa cel % o UAA 0–100 Ag icul u al a eas wi h ag i- en i onmen al con ac s A eas unde AECM con ac s: Managemen o adi ional u al bio opes, P omo ing na u al biodi e si y and landscape de elopmen and managemen (mainland Finland) o Na u al pas u es and wooded meadows ( he Åland islands) T adi ional u al bio opes a e a na ional concep o semi- na u al g asslands unde adi ional managemen (Raa ikainen e al., 2017); measu e o “P omo ing na u al biodi e si y and landscape de elopmen and managemen ” a ge s landscape elemen s o ele ance o biodi e si y In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007 Field pa cel % o UAA 0–100 Weak max. 345 Ex ensi e cul i a ion Pa icula c op ypes ha desc ibe ex ensi e ( s. in ensi e) cul i a ion ( o ull lis see appendix C) Ex ensi e land use is cen al o HNV concep ; in Finland, inc easing he p opo ion o di e en g asslands, including allows, was shown o be key o biodi e si y (Kuussaa i e al., 2008;Toi onen e al., 2015) In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007 Field pa cel (% o UAA) /10 0–10 Edge densi y o ield pa cels Leng h o ield pa cel edges di ided by he ield a ea Re lec s he mosaic s uc u e o he a mland (Type 2 HNV); in Finland, agmen a ion o he ield s uc u e has posi i e biodi e si y e ec s (Piha e al., 2007;Vepsäläinen, 2007). Land pa cel egis e ; Agency o Ru al A ai s, 2007 Field m 3 /ha 0–30 Li es ock a ming Main p oduc ion sec o o a a m is ca le a ming, ho se managemen , o sheep o goa a ming Indica es (po en ially) pas u es; in Finland, also cul i a ed g asslands inc ease bi d di e si y (Ek oos e al., 2018; Vepsäläinen e al., 2010). In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y, 2007 Fa m le el p esence/ absence 5/0 S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 227 3. Resul s 3.1. Va ia ion in he High Na u e Value a mland indica o The o al indica o alues o HNV i anged om 8.8 o 130.6 among he 68 1-km 2 squa es, wi h squa es on he Åland Islands ecei ing he highes alues (Fig. 2,Table 2). The HNV i alues a b oade landscape scales a ound he 1-km 2 squa es showed simila egional a ia ion and he e we e only mino di e ences be ween he di e en scales (Table 2). Among he sub-indica o s, he s onges egional a ia ion was o he combined sco e o he h ee s ong HNV sub-indica o s; semi-na u al g asslands, pe manen pas u es and a eas wi h AECM con ac s (an a e age o 38.4 on he Åland Islands and an a e age o 4.3 on he mainland) (Table 2). Fo he 1-km landscape da a used o bi d di e si y (su ounding he 62 su eyed 1-km 2 squa es) bo h he com- bina ion o s ong HNV sub-indica o s and weake sub-indica o s showed simila a ia ion (Table 2). 3.2. High Na u e Value a mland indica o and bi d di e si y Species ichness o all a mland bi ds a ied be ween 11 and 31 (mean 19 ± SD 4.5) and ha o ed-lis ed bi ds be ween 0 and 5 (mean 2.3 ± SD 1.3), and he di e si y index o all bi ds be ween 1.7 and 3.1 (mean 2.5 ± SD 0.3) and ha o ed-lis ed bi ds be ween 0 and 1.5 (mean 0.6 ± SD 0.5), espec i ely. Models wi h landscape-scale HNV i alues (0.5–5 km) pe o med be e han models wi h 1-km 2 squa e- scale indica o alues in explaining he di e si y o all bi ds (see Appendix F.1). HNV i o 1-km 2 squa e was he only scale ha ecei ed any a iable impo ance wi h di e si y o ed-lis ed bi ds (Table 3). Fo he di e si y o ed-lis ed bi ds, models wi h longi udinal loca ion o he squa e showed a be e pe o mance han hose wi h HNV i a iables. Longi ude showed a clea nega i e ela ionship wi h he species ich- ness and di e si y index o ed-lis ed bi ds, deno ing ha communi ies o species o conse a ion conce n a e mo e di e se and hei abun- dances mo e e enly dis ibu ed in local communi ies occu ing in sou h-wes e nmos Finland. No clea ela ionship o any o he spa ial Fig. 2. Spa ially-explici p esen a ion o he dis ibu ion and magni ude o he ecalcula ed Finnish High Na u e Value a mland indica o (HNV i) o he 68 1-km 2 squa es. Size o he symbol (o ange ci cle) is p opo ional o o al sco e o he HNV i. To al sco e is a sum o he six High Na u e Value a mland sub-indica o s. Fo de ails o calcula ion see Me hods. G ey shaded a eas ep esen he sub-indica o s, ex- cep he densi y o ield edges. Table 2 Mean ( ± SD) o he o al High Na u e Value a mland indica o (HNV i) sco es a di e en scales (squa es, 0.5, 1, 2 and 5 km) and o he sepa a e sub-indica o s o he squa es and 1-km landscapes (n = 68) o each o he i e geog aphic egions (numbe o squa es pe egion is shown in pa en heses). The combined alue o he s ong HNV i alues included sub-indica o s o he semi-na u al g asslands, pe manen pas u es and a eas wi h AECM con ac s (see Me hods o de ails). Geog aphic egion Scale o HNV i Ka elia (n = 11) Os obo hnia (n = 15) Sou he n (n = 15) Sou hwes e n (n = 17) Åland (n = 10) Squa e 29.3 ± 8.5 33.1 ± 28.4 22.9 ± 14.1 21.4 ± 9.2 62.9 ± 28.7 0.5 km 29.9 ± 6.6 32.4 ± 23.3 28.3 ± 19.9 21.2 ± 8.5 73.4 ± 25.2 1 km 32.0 ± 6.3 31.9 ± 19.1 29.8 ± 18.8 23.9 ± 15.1 68.1 ± 19.2 2 km 32.9 ± 4.9 33.3 ± 13.2 31.2 ± 13.0 27.1 ± 14.0 68.3 ± 18.0 5 km 34.1 ± 4.0 32.5 ± 8.9 28.0 ± 4.8 25.8 ± 6.4 70.5 ± 18.1 Sub-indica o s S ong HNV i 1.5 ± 3.6 7.3 ± 22.8 4.2 ± 12.2 3.2 ± 7.7 36.2 ± 27.7 1 km 3.1 ± 3.9 0.9 ± 1.8 8.5 ± 19.1 4.2 ± 14.8 38.4 ± 19.3 Edge densi y 18.0 ± 5.3 16.5 ± 4.6 12.3 ± 3.0 12.2 ± 2.7 16.3 ± 4.8 1 km 18.6 ± 5.4 15.7 ± 3.1 13.6 ± 2.2 11.5 ± 1.2 19.6 ± 3.9 Ex ensi e cul i a ion 5.3 ± 2.6 5.0 ± 2.8 2.9 ± 2.3 2.9 ± 2.2 5.9 ± 2.9 1 km 5.9 ± 1.5 5.2 ± 1.4 3.1 ± 1.0 3.6 ± 1.8 5.2 ± 1.6 Li es ock a ming 4.5 ± 1.5 4.3 ± 1.8 3.5 ± 2.3 3.0 ± 2.5 4.5 ± 1.6 1 km 5.0 ± 0 4.2 ± 1.9 4.7 ± 1.3 4.3 ± 1.8 5.0 ± 0 S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 228 scales o he HNV i wi h he bi d di e si y measu es was ound (Table 3) and he explana o y powe o a iables was in gene al low (Fig. 3a, b). 3.3. High Na u e Value a mland indica o , loca ion and bu e ly di e si y Numbe s o all eco ded bu e ly species anged om 11 o 33 (mean 22.7 ± SD 4.6) and he numbe o he declining bu e lies om 0 o 5 (mean 1.3 ± SD 1.1). The di e si y index a e aged a 2.1 ( ± SD 0.3) among all species and 0.3 ( ± SD 0.4) among he declining species. Models wi h geog aphical loca ion only pe o med be e han models wi h HNV i alues o all di e si y measu es, excep o he di e si y index o all bu e lies (see Appendix F.2). The HNV i o squa e and ha o he 0.5-km scale we e impo an o bu e ly di e si y measu es, excep o he species ichness o declining bu e lies whe e he HNV i alues o g ea e landscape scales (2 o 5 km) ecei ed a iable im- po ance (Fig. 3a, b). The bu e ly di e si y index showed a posi i e ela ionship wi h he HNV i a 1-km 2 squa e scale (Fig. 4a, Table 3) and wi h longi ude bu a nega i e wi h la i ude. Also, he species ichness and communi y di e si y o declining bu e lies we e highe in sou he n han in no he n squa es. 3.4. High Na u e Value a mland sub-indica o s and species di e si y We conduc ed he analysis a 1-km landscape scale o all bi ds, and a he scale o squa e o ed-lis ed bi ds and all bu e ly di e si y measu es since hese scales showed mos impo ance by he mul i- model in e ence (Appendix F.3). Since g ea e landscape scales o HNV i (0.5, 2 and 5 km) showed some impo ance o di e si y measu es (S and H’ o declining bu e lies and H’ o bi ds), he associa ions wi h sub-indica o s we e explo ed sepa a ely (see suppo ing in o ma ion, Appendices F.4–F.5). The e was conside able unce ain y wi hin models o he bi d di- e si y measu es, since models wi h he lowes AIC c showed ela i ely low p obabili ies o being bes models (Akaike weigh o he highes app oxima ed models a ied om 0.22 o 0.66, see Appendix F.3). Howe e , all he sub-indica o s we e ep esen ed among models ha showed mos suppo . Fo he bi d di e si y index, ex ensi e cul i a ion (sub-indica o 5) had high a iable impo ance (Fig. 3d) and he con- idence in e als o model-a e aged es ima e indica ed a posi i e e- la ionship wi h di e si y index o all bi ds (Fig. 4b, Table 4). Li es ock a ming (sub-indica o 6) had a subs an ial a iable impo ance (0.66) o he di e si y index o all bi ds (Fig. 3d). The combined index o he h ee s ong sub-indica o s was also shown o ha e impo ance on bi d species ichness (Fig. 3c). Longi ude was nega i ely ela ed o ed-lis ed bi d ichness and di e si y index. O e all, all sub-indica o s excep li es ock a ming ela ed wi h he bu e ly di e si y (Fig. 3c, d, Table 4). Models wi h he loca ion a i- ables showed highes pe o mance o all o he bu e ly di e si y as- pec s excep o he di e si y index o all a mland species (Appendix F.3), o which he edge densi y o all ield pa cels (sub-indica o 4) wi h he la i ude was he bes -app oxima ed model wi h he highes weigh . The edge densi y had a high ela i e impo ance and a clea posi i e ela ionship wi h bu e ly di e si y index (Fig. 3d, 4c, Table 4). 4. Discussion We ound ha de i ed alues o he HNV i in Finland a ied clea ly be ween he geog aphical a eas, he Åland Islands ecei ing he highes sco es. Mo eo e , he e we e no able di e ences in how s ongly he bi d and bu e ly pa e ns we e ela ed o he HNV i measu ed a he di e en spa ial scales. Impo an ly, e en hough di e en spa ial HNV i scales had di e en impo ance on bi d and bu e ly di e si y measu es, only bu e ly di e si y index had a clea posi i e ela ion- ship wi h he o al HNV i a he smalles s udied scale, which is closes o he a m scale whe e he index is cu en ly being used. Among he HNV sub-indica o s, he ield edge densi y appea ed as he mos Table 3 The 95% con idence in e als o model-a e aged pa ame e es ima es and ela i e a iable impo ance alues (Σw i ) o High Na u e Value a mland indica o (HNV i) alues a di e en scales (HNV500 o 0.5 km, HNV1 o 1 km, HNV2 o 2 km and HNV5 o 5 km) and geog aphical loca ion explaining he species ichness (S) and di e si y (H’) in a mland bi ds and bu e lies. Con idence in e als ha do no include ze o a e bolded. Model-a e aging o each di e si y measu e was conduc ed on bes -app oxima ed model se s (see Tables 1 and 2 in Appendix F) by he ull-a e aging me hod (Symonds and Moussalli, 2011). G oup Di e si y HNV HNV500 HNV1 HNV2 HNV5 Longi ude La i ude Bi ds All S −0.109, 0.182 −0.116, 0.233 −0.109, 0.180 −0.109, 0.182 −0.164, 0.086 Σw i 0.22 0.34 0.22 0.22 0.44 Red-lis ed S −0.290, 0.196 −0.840, −0.069 Σw i 0.31 1.00 All H’ −0.076, 0.106 −0.093, 0.155 −0.064, 0.084 −0.125, 0.273 −0.132, 0.093 −0.066, 0.099 Σw i 0.11 0.25 0.08 0.38 0.17 0.18 Red-lis ed H’ −0.173, 0.127 −0.650, −0.176 Σw i 0.28 1.00 Bu e lies All S −0.059, 0.084 −0.145, 0.099 −0.247, 0.113 Σw i 0.24 0.26 0.50 Declining S −0.220, 0.172 −0.210, 0.166 −0.233, 0.296 −2.215, −0.854 Σw i 0.18 0.18 0.18 1.00 All H’ a 0.332, 1.058 0.121, 0.699 −0.640, −0.169 Σw i 1.00 1.00 1.00 Declining H’ −0.071, 0.093 −0.080, 0.114 −0.478, −0.172 Σw i 0.21 0.25 1.00 a Only one model was included wi hin ΔAIC c < 2, hus no model-a e aging was conduc ed o hese a iables. S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 229 impo an explana o y a iable o he di e si y o bu e lies and ex- ensi e cul i a ion o ha o bi ds. Ou esul s con i med ha he Åland Islands s and apa om mainland Finland in he amoun o HNV as demons a ed by h ee main HNV sub-indica o s. This is in line wi h he obse a ions ha meadow-like habi a s ha ha e dec eased s ongly on mainland Fig. 3. Rela i e impo ance o a iables explaining he ela ion be ween bi d and bu e ly species ichness (a,c) and di e si y index (b,d) wi h High Na u e Value a mland (HNV ) indica o a di e en scales (uppe panels), HNV sub-indica o s (lowe panels) and loca ion. Impo ance o di e en a iables is shown sepa a ely o g oups comp ising o all bi ds and bu e lies, and he sub-g oups o ed-lis ed bi ds and declining bu e lies. Rela i e impo ance is calcula ed by summing he weigh s o models whe e a a iable is p esen and hus desc ibes he p obabili y ha a a iable is included in he bes -app oxima ed models. I only one model is in he bes -app oxima ed se , a iable impo ance is ixed o 1. Ind = code o he six sub-indica o g oups. Fig. 4. Model p edic ions (black lines) and 95% con idence in- e als (dashed da k g ey lines) o bu e ly di e si y index wi h inc easing High Na u e Value a mland indica o (HNV i) a squa e scale (a), bi d di e si y index wi h log- ans o med ex- ensi e cul i a ion (b), and bu e ly di e si y index wi h log- ans o med edge densi y (c). P edic ed alues we e calcula ed ei he using he single bes -app oxima ed model ( o a) o by model-a e aged pa ame e es ima es ( o b and c). Va iables no shown in p edic ions we e se o hei mean s anda dized alues. S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 230 Finland s ill cons i u e a no able pa o ag icul u al land use on he Åland Islands and ha e also habi a quali y simila o ha o adi ional land use (Schulman e al., 2005). I is no ewo hy ha he ag i-en- i onmen al p og am o he Åland Islands g ea ly di e ed om he Finnish sys em in 2000–2006, wi h a g ea e emphasis on na u e alues (Schulman e al., 2005), and his has po en ially had e ec s on he landscape-le el habi a di e si y as well. Bi d communi ies show somewha di e en , bu s ill posi i e asso- cia ions wi h HNV in o he Eu opean coun ies (Aue e al., 2014;Doxa e al., 2010;Mo elli e al., 2014). Ou esul s indica ed ha a leas he o e all numbe o species does no inc ease wi h ising amoun in HNV in Finland. Howe e , while HNV i did no show any clea associa ion wi h bi d di e si y index a e model-a e aged pa ame e s, in he single models especially a 1-km and 5-km landscape scales he HNV i had a posi i e associa ion. A hese scales, hus, he bi d di e si y may bene i o landscape-le el habi a he e ogenei y (Ben on e al., 2003), when semi-na u al and o he uncul i a ed pa ches a e embedded in o an ag icul u al landscape. In con as , no ela ionship be ween he e- calcula ed HNV i and ed-lis ed bi ds was ound, mos likely caused by he low ep esen a ion o hese species wi hin ou s udy squa es and he low- ep esen a ion o a mland species in he Finnish ed lis . Bu abo e all, ou di e en esul o bi ds o igina es due o, i s ly, gene ally small sizes and amoun o semi-na u al a eas and pe manen pas u es o bo h in Finland in compa ison wi h coun ies o ea lie s udies (Aue e al., 2014;Doxa e al., 2010;Mo elli e al., 2014). Secondly, he Finnish HNV concep appea s o ake insu icien ly in o accoun some ele an ag icul u al elemen s o bi ds, o example, he ealized g azing p essu e o e iliza ion (Doxa e al., 2010;Tiainen and Seimola, 2014). The posi i e ela ionship o bu e ly di e si y especially wi h he occu ence o g asslands (e.g. Do e e al., 2011) and i s nega i e e- la ionship wi h ag icul u al in ensi ica ion ha e been con i med ea lie (Ek oos e al., 2010). Co esponding wi h hose s udies, ou esul sugges s ha HNV is a good indica o o he spa ial a ia ion in g assland bu e ly di e si y. Fu he mo e, Pöy y e al. (2004) ha e shown ha di e si y and e enness o bu e ly communi ies in sou h- wes Finland is g ea e on semi-na u al meadows unde long- e m con inuous g azing han on unmanaged o es o ed meadows. Ou e- sul s co obo a e his since HNV a eas wi h he highes sco es include managemen con ac s o semi-na u al meadows. Howe e , he species ichness and communi y di e si y o declining bu e ly species we e mo e s ongly ela ed o la i ude han o he dis ibu ion o aluable HNV , which is consis en wi h endange ed plan s and bu e lies being mo e abundan in sou hwes e n Finland (Heliölä e al., 2009). How- e e , i should be acknowledged ha he spa ial di e si y pa e ns o di e en axa do no necessa ily co ela e wi h each o he (Jonason e al., 2017), and HNV i may ail on some axa o his eason. Ou esul s highligh ed he impo ance o ex ensi e cul i a ion, one o he HNV sub-indica o s, on he di e si y index o all eco ded a mland bi d species. Indi idual HNV elemen s had also posi i e e- la ionship wi h specialis bi ds in Ge many (Aue e al., 2014). These indings a e suppo ed by o he Finnish s udies whe e low-in ensi y ag icul u e had a posi i e ela ionship wi h bi d di e si y, species ichness and abundance (Piha e al., 2007), non-cul i a ed a eas in open a mland in sou he n Finland had highe bi d densi ies han cul- i a ed a eas (Tiainen and Seimola, 2014) and se -aside ields doubled numbe s o bi d indi iduals and had 25 o 40% mo e species (He zon e al., 2011). Ex ensi e cul i a ion he e is de ined as all ield ypes co e ed by g ass, including in ensi ely managed g asslands such as silage and o a ional pas u es (Heliölä e al., 2009, Appendix B). Such a eas can ha e a posi i e e ec on a mland biodi e si y since hey can p o ide p o i able o aging and b eeding g ounds o bi ds (Vepsäläinen e al., 2010). The densi y o ield pa cels’ edges is a measu e o landscape con- igu a ion, ag icul u al landscapes wi h high edge densi y alues being ine-scale mosaics o habi a pa ches (Duelli, 1997;Hie ala-Koi u e al., 2004). Ou inding ha ield edge densi y is impo an o bu e ly di- e si y index ully suppo s he idea ha biodi e si y can be p omo ed by lea ing g ea e a eas ou side in ensi e cul i a ion (Ba á y e al., 2015). Al hough ield edges ha e di e en plan species composi ion Table 4 Rela ionships o he High Na u e Value a mland (HNV ) sub-indica o s wi h bi d and bu e ly di e si y (S o species ichness and H’ o di e si y index) in Finland as shown by he 95% con idence in e als o model-a e aged pa ame e es ima es and a ela i e a iable impo ance (Σw i ). Con idence in e als no including ze o a e bolded. Used scales we e he 1-km landscape scale o bi ds and he squa e scale o ed-lis ed bi ds and o bu e lies. Ind123 e e s o he combined sco e o he s ong HNV sub-indica o s, Ind4 is o edge densi y, Ind5 o ex ensi e cul i a ion and Ind6 o he p esence o li es ock a ming. G oup Di e si y Ind123 Ind4 Ind5 Ind6 Longi ude La i ude Bi ds All S −0.101, 0.269 −0.020, 0.312 −0.041, 0.026 −0.301, 0.076 −0.051, 0.043 Σw i 0.57 0.91 0.31 0.70 0.09 Red-lis ed S −0.436, 0.225 −0.798, −0.060 Σw i 0.44 1.00 All H’ −0.084, 0.129 −0.064, 0.099 0.074, 0.322 −0.069, 0.023 −0.266, 0.055 Σw i 0.22 0.24 1.00 0.66 0.77 Red-lis ed H’ −0.217, 0.134 −0.629, −0.166 Σw i 0.34 1.00 Bu e lies All S −0.063, 0.095 −0.143, 0.098 −0.255, 0.113 Σw i 0.27 0.25 0.52 Declining S −0.299, 0.221 −0.251, 0.327 −2.188, −0.823 Σw i 0.23 0.22 1.00 All H’ −0.073, 0.113 0.048, 0.305 −0.057, 0.074 −0.401, −0.141 Σw i 0.28 1.00 0.20 1.00 Declining H’ −0.099, 0.200 −0.496, −0.186 Σw i 0.46 1.00 S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 231 han semi-na u al g asslands (Toi onen e al., 2013), hey p o ide im- po an o aging habi a and dispe sal co ido s (Dela e e al., 2010; Ma shall and Moonen, 2002), especially i si ua ed in a landscape wi h high o es co e (Toi onen e al., 2017). In F ance, linea elemen s we e obse ed o p omo e bu e ly di e si y mo e han g asslands (Ouin and Bu el, 2002), and in Canada, bu e ly species ichness was highe in landscapes wi h smalle ields and pa ches han in landscapes wi h simple con igu a ion (Flick e al., 2012). In ou da a, only he weak HNV sub-indica o s (such as ex ensi e cul i a ion and edge densi y) conside ably a ied be ween geog aphical a eas and squa es. In con as , he combined sco e o he h ee s ong sub-indica o s (semi-na u al g asslands, pe manen pas u es and a eas wi h AECM con ac s) showed spo adically high alues, bu s ill o mos o he a eas he alues we e e y small. This spa ial imbalance be ween he s ong sub-indica o s and he weak sub-indica o s is e y likely e lec ed o he obse ed di e si y pa e ns. On one hand, he amoun o mos aluable HNV ea u es is oo small o con ibu e sig- ni ican ly o he di e si y o all axa s udied he e. On he o he hand, he weak HNV ea u es a e unde alued; a en ion o hei p ese a ion and de elopmen should be g ea e . Un o una ely, he e a e no biodi e si y da a on he scale o he o icial Finnish HNV i ha is a a m’s scale. The e o e, we had o e- calcula e he indica o o he landscape scale o ma ch i wi h ou biological da a. Howe e , since we used exac ly he same da a and p ocedu e ha is he basis o he HNV i na ionally, he indica i e e- la ionships be ween he indica o and he axa es ablished he e is likely o be highly consis en . Mo eo e , in Finland, i would be bene icial o link he spa ially mo e ex ensi e and long- e m da ase s o a mland bi d and bu e ly moni o ing o HNV a ming o be e iden i y in- dica o species o HNV . We p opose ha also he acili a ion o ed- lis ed o specialis species by HNV a ming should be examined close wi h la ge and mo e long- e m da a se s. Since he smalles spa ial scales o HNV did no show associa ion wi h bi d di e si y, he con- nec i i y o HNV a mlands in ela ion o species di e si y should be ollowed. Fu he mo e, ou esul s sugges ha some ea u es included in o weak HNV sub-indica o s can p omo e biodi e si y. Managemen and e en ion o such ea u es is no as cos ly as managing a eas co - esponding o he s ong HNV indica o s. This app oach was e lec ed in he new poli ical ool o Ecological Focus A eas in he cu en CAP (EU, 2013). Also, he inclusion o ‘weak’ sub-indica o ypes (including con en ional p oduc ion g asslands) in o he Finnish HNV i could be c i icized due o i s ela i ely mino ela ion o he adi ional species- ich a mland. In he end, he concep o HNV s i es o g asp a mland o op impo ance o biodi e si y in Eu ope, which is “cha ac e ized by long-es ablished, low-in ensi y and o en complex a ming sys ems” (Keenleyside e al., 2014) ha mos ly ceased o exis in he mode n landscape o Finland, excep pa s o he Åland Islands. Funding This wo k was suppo ed by he Kone Founda ion o S.M. (G an numbe : 085239), A.A. and A.H. T.S. and J.T. wo ked o he pape as a pa o he LIFE14 p ojec CCM/FI/000254 ( he EASME/Commission is no esponsible o any use ha may be made o he in o ma ion i con ains) and I.H. as pa o HNV-Link p ojec (Con ac numbe : 696391). Species da a collec ion was unded by he Finnish Minis y o Ag icul u e and Fo es y, he Minis y o En i onmen and Ålands landskaps y else. Decla a ions o in e es None. Acknowledgemen s We wan o hank all people who ook pa in ield wo k in coun ing bu e lies and a mland bi ds. In o ma ion Cen e o he Minis y o Ag icul u e and Fo es y is acknowledged o gi ing he ield pa cel and cul i a ion da a. Appendices A–F. Supplemen a y da a Supplemen a y ma e ial ela ed o his a icle can be ound, in he online e sion, a doi:h ps://doi.o g/10.1016/j.agee.2018.09.030. Re e ences Aakkula, J., Leppänen, J. (Eds.), 2014. Follow-up s udy on he impac s o ag i-en i on- men measu es (MYTVAS 3). Final epo . Agengy o Ru al A ai s Finland, 2007. Land pa cel egis e . h p://www.ma i. i/en/ abou - he-agency/Pages/in o ma ion-se ices.aspx (accessed 28 Sep embe 2018). Ande sen, E., Baldock, D., Benne , H., Beau oy, G., Bignal, E., B ouwe , F., Elbe sen, B., Eiden, G., Godeschalk, F., Jones, G., McC acken, D., Nieuwenhuizen, W., an Eupen, M., Hennekens, S., Ze as, G., 2003. De eloping a High Na u e Value Fa ming A ea Indica o . In e nal Repo o he Eu opean En i onmen Agency. IEEP. Aue, B., Diekoe e , T., Go schalk, T.K., Wol e s, V., Ho es, S., 2014. How High Na u e Value (HNV) a mland is ela ed o bi d di e si y in ag o-ecosys ems – owa ds a e sa ile ool o biodi e si y moni o ing and conse a ion planning. Ag ic. Ecosys . En i on. 194, 58–64. Ba on, K., 2014. MuMIn: Mul i-Model In e ence. R Package Ve sion 1.10.5. h p:// CRAN.R-p ojec .o g/package=MuMIn. Ba á y, P., Dicks, L.V., Kleijn, D., Su he land, W.J., 2015. The ole o ag i-en i onmen schemes in conse a ion and en i onmen al managemen . Conse . Biol. 29, 1006–1016. Ba es, D., Maechle , M., Bolke , B., Walke , S., 2015. Fi ing linea mixed-e ec s models using lme4. J. S a . So w. 67, 1–48. Benede i, J., 2017. T ends in High Na u e Value a mland s udies: a sys ema ic e iew. Eu . J. Ecol. 3, 19–32. h ps://doi.o g/10.1515/eje-2017-0012. Ben on, T., Vicke y, J., Wilson, J., 2003. Fa mland biodi e si y: is habi a he e ogenei y he key? T ends Ecol. E ol. 18, 182–188. Biodi e si y. i, 2015, FA8 High Na u e Value a mland. h ps://www.biodi e si y. i/en/ habi a s/ a mlands/ a8-high-na u e- alue-hn - a mland (accessed 1 Decembe 2016). Bjø ns ad, O., Falck, W., 2001. Nonpa ame ic spa ial co a iance unc ions: es ima ion and es ing. En i on. Ecol. S a . 8, 53–70. Bu nham, K.P., Ande son, D.R., 2002. Model Selec ion and Mul imodel In e ence. A P ac ical In o ma ion–Theo e ic App oach, second ed. Sp inge , New Yo k. Bu cha , S.H., Walpole, M., Collen, B., an S ien, A., Scha lemann, J.P., Almond, R.E., Baillie, J.E., Bomha d, B., B own, C., B uno, J., Ca pen e , K.E., Ca , G.M., Chanson, J., Chene y, A.M., Csi ke, J., Da idson, N.C., Den ene , F., Fos e , M., Galli, A., Galloway, J.N., Geno esi, P., G ego y, R.D., Hockings, M., Kapos, V., Lama que, J.F., Le e ing on, F., Loh, J., McGeoch, M.A., McRae, L., Minasyan, A., He nández Mo cillo, M., Old ield, T.E., Pauly, D., Quade , S., Re enga, C., Saue , J.R., Skolnik, B., Spea , D., S anwell-Smi h, D., S ua , S.N., Symes, A., Tie ney, M., Ty ell, T.D., Vié, J.C., Wa son, R., 2010. Global biodi e si y: indica o s o ecen declines. Science 328, 1164–1168. Bu le , S.J., Boccaccio, L., G ego y, R.D., Voříšek, P., No is, K., 2010. Quan i ying he impac o land-use change o Eu opean a mland bi d popula ions. Ag ic. Ecosys . En i on. 137, 348–357. Dela e, T., Pichancou , J., Bu el, F., Kindlmann, P., 2010. G assy ield ma gins as po- en ial co ido s o bu e lies in ag icul u al landscapes: a simula ion s udy. Ecol. Model. 221, 370–377. Donald, P.F., G een, R.E., Hea h, M.F., 2001. Ag icul u al in ensi ica ion and he collapse o Eu ope’s a mland bi d popula ions. P oc. R. Soc. Lond. B 268, 25–29. h ps://doi. o g/10.1098/ spb.2000.1325. Do e , J.W., Spence , S., Collins, S., Hadjigeo giou, I., Rescia, A., 2011. G assland bu - e lies and low in ensi y a ming in Eu ope. J. Insec Conse . 15, 129–137. Doxa, A., Bas, Y., Pa acchini, M.L., Poin e eau, P., Te es, J., Jigue , F., 2010. Low-in- ensi y ag icul u e inc eases a mland bi d abundances in F ance. J. Appl. Ecol. 47, 1348–1356. Duelli, P., 1997. Biodi e si y e alua ion in ag icul u al landscapes, an app oach a wo di e en scales. Ag ic. Ecosys . En i on. 62, 81–91. EC, 2006. Council decision o 20 Feb ua y 2006 on Communi y S a egic Guidelines o Ru al De elopmen (p og amming Pe iod 2007 o 2013) (2006/144/ EC). Ek oos, J., Heliölä, J., Kuussaa i, M., 2010. Homogeniza ion o lepidop e an communi ies in in ensi ely cul i a ed ag icul u al landscapes. J. Appl. Ecol. 47, 459–467. Ek oos, J., Tiainen, J., Seimola, T., He zon, I., 2018. Weak e ec s o a ming p ac ices co esponding o CAP g eening measu es on a mland bi d di e si y in bo eal ag i- cul u al landscapes. Submi ed. Unpublished esul s. . Es i, 2015. A cGIS 10.3.1. o Desk op. En i onmen al Sys ems Resea ch Ins i u e, Redlands, Cali o nia, USA. EU, 2013. Regula ion (EU) No 1307/2013 o he Eu opean Pa liamen and o he Council o 17 Decembe 2013 es ablishing ules o di ec paymen s o a me s unde suppo schemes wi hin he amewo k o he common ag icul u al policy and epealing Council Regula ion (EC) No 637/2008 and Council Regula ion (EC) No 73/2009. Flick, T., Feagan, S., Fah ig, L., 2012. E ec s o landscape s uc u e on bu e ly species ichness and abundance in ag icul u al landscapes in eas e n On a io, Canada. Ag ic. Ecosys . En i on. 156, 123–133. S. Mäkeläinen e al. Ag icul u e, Ecosys ems and En i onmen 269 (2019) 224–233 232