Reproductive rate of a top predator, the grey seal, as an indicator of the changes in the Baltic food web

Con en s lis s a ailable a ScienceDi ec Ecological Indica o s jou nal homepage: www.else ie .com/loca e/ecolind Rep oduc i e a e o a op p eda o , he g ey seal, as an indica o o he changes in he Bal ic ood web Kaa ina Kauhala a,⁎ , Samuli Ko pinen b , Maiju Leh iniemi b , Ja i Rai aniemi a a Na u al Resou ces Ins i u e Finland, Luke, I äinen Pi käka u 4 A, FI-20520 Tu ku, Finland b Ma ine Resea ch Cen e, Finnish En i onmen Ins i u e, La oka anonkaa i 11, FI-00790 Helsinki, Finland ARTICLE INFO Keywo ds: Bi h a e Clupeids Halichoe us g ypus Th eshold alue Zooplank on ABSTRACT Rep oduc i e a e o g ey seal (Halichoe us g ypus) emales has luc ua ed du ing he 2000s, al hough e- p oduc i e dis u bances, which occu ed ea lie , a e a e a p esen . He e we aimed o s udy especially he ood web ac o s possibly a ec ing he bi h a e o Bal ic g ey seals and whe he bi h a e can be used as an indica o o he changes in he Bal ic ood web. Ou esul s showed ha he bi h a e o g ey seals was signi ican ly ela ed o he ing (Clupea ha engus memb as) and sp a (Sp a us sp a us) quali y (weigh ) which, in u n we e in luenced by sp a and cod (Gadus mo hua) abundance, as well as zooplank on biomass and plank e size. This sugges s s ong ophic coupling o e h ee ophic le els. We hus conclude ha he bi h a e o g ey seals can be used as an indica o o he s a us o he Bal ic ood web. Based on his, we sugges a h eshold alue o good ood web s a us o a s able, non-g owing seal popula ion. 1. In oduc ion Rep oduc i e a e o mammals is usually a ec ed by hei body condi ion, as ep oduc ion is ene gy-demanding and he cos s o e- p oduc ion a e oo high o indi iduals in poo condi ion (e.g. Clu on- B ock e al., 1982; Boyd e al., 1995; McMahon and Hindell, 2003). Capi al b eede s, such as he g ey seal (Halichoe us g ypus) mainly ely on hei s o ed ene gy ese es du ing ep oduc ion and nu sing (Boyd, 2000). In seals, body condi ion a ec s he iming o pube y, im- plan a ion o emb yos and mo ali y a es o emb yos/ e uses and hus he ep oduc i e a e o emales (e.g. Boyd, 1984a; Guine e al., 1998; Bowen e al., 2006; P o i e al., 2007; Kauhala e al., 2017). Body condi ion (nu i ional s a us), and hus also ep oduc i e a e, may be a ec ed by en i onmen al condi ions, such as he quali y, quan i y o a ailabili y o ood esou ces (Lunn e al., 1994; C ocke e al., 2006; Biuw e al., 2007; Schick e al., 2013). Hence, ep oduc i e a e o Bal ic g ey seals could be used as an indica o o changes in he Bal ic ood web, such as he quali y o quan i y o hei main p ey (Ke shne e al., 2011, HELCOM, 2018a). Thei body condi ion and bi h a e has indeed ound o be ela ed o changes in he quali y o hei p e e ed p ey ish, he ing (Clupea ha engus) (Kauhala e al., 2016, 2017). The g ey seal is one o he ou ma ine mammal species in he Bal ic Sea and oday he mos nume ous o hem. I has been hun ed since he S one Age, a leas in he sou he n Bal ic (Hä könen e al., 2007), bu he popula ion c ashed om abou 100 000 in he beginning o he 20 h cen u y o only 2000–3000 in he 1970s, mainly due o he high hun ing p essu e (Jensen e al., 1969; Almk is , 1978; Kokko e al., 1999; Ha ding and Hä könen, 1999; Ha ding e al., 2007). En i on- men al pollu ion (PCBs and DDT) in he 1960s and 1970s hen hinde ed popula ion eco e y by causing s e ili y o emales (Be gman and Olsson, 1986; Be gman, 1999; Ha ding and Hä könen, 1999; Nyman e al., 2003). A e p o ec ion om hun ing (1982 in Finland and 1986 in Sweden) and due o he dec eased le el o en i onmen al pollu ion, especially PCBs and DDT, in he Bal ic Sea du ing ecen decades (Rou i, 2009) g ey seal numbe s ha e inc eased since he 1980s, and a p esen exceed 30 000 in he whole Bal ic Sea and 10 000 in he Finnish sea a ea only (Luke, 2017). Since he 1980s, ep oduc i e heal h o g ey seals g adually im- p o ed, and a p esen i ually no ep oduc i e dis u bances (i.e. s e ili y o emales) a e obse ed (Bäcklin e al., 2011; Kauhala e al., 2014). Rep oduc i e a e o g ey seal emales has, howe e , luc ua ed also du ing he 2000s (Kauhala e al., 2014, 2016), and one possible eason behind he luc ua ion is he quali y o he impo an p ey spe- cies, he he ing and sp a (Sp a us sp a us;MacKenzie and Kös e , 2004; Lunds öm e al., 2007, 2010; Kauhala e al., 2011, Gå dma k e al., 2012). He ing is he mos common p ey o Bal ic g ey seals composing abou 78% o hei iden i ied p ey ish, and occu ing in 70–85% o hei diges i e acks (e.g. Lunds öm e al., 2010; Scha - h ps://doi.o g/10.1016/j.ecolind.2019.03.022 Recei ed 20 No embe 2018; Recei ed in e ised o m 13 Ma ch 2019; Accep ed 17 Ma ch 2019 ⁎ Co esponding au ho . E-mail add esses: [email p o ec ed] (K. Kauhala), [email p o ec ed] (S. Ko pinen), [email p o ec ed] (M. Leh iniemi), [email p o ec ed] (J. Rai aniemi). Ecological Indica o s 102 (2019) 693–703 1470-160X/ © 2019 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/BY-NC-ND/4.0/). T Olsen e al., 2018). Sp a composed abou 15% o p ey ish (Lunds öm e al., 2007). The luc ua ions in he ish p ey a e caused, e.g. by zoo- plank on a ailabili y (Flinkman e al., 1992; Möllmann e al., 2004). Fu he mo e, he abundance o cod (Gadus mo hua), which p eys on sp a and he ing may ha e an impac on hei numbe s (e.g. Ruds am e al., 1994; Casini e al., 2008) and also hei quali y (weigh ) due o in a- and in e speci ic compe i ion o zooplank on be ween he ing and sp a (e.g. Casini e al., 2006). Also he ep oduc i e a e o seals may hus be connec ed o he abundance o cod. Cod is o mino im- po ance in g ey seal die (Scha -Olsen e al., 2018; T e in e al., 2019). To unde s and he ood web unc ioning in he Bal ic Sea, i is hus c ucial o analyze he ela ionships be ween in e eb a es (especially zooplank on), di e en ish species and op p eda o s, such as seals (Kuosa e al., 2017). Changes in he seal condi ion, such as ep oduc i e a e, and ul ima ely abundance, may e lec changes in he ood web o he Bal ic Sea (HELCOM, 2018a). In he p esen pape , we s udied especially which ood web ac o s migh a ec he bi h a e o g ey seals and whe he bi h a e can be used as an indica o o he changes in he ood web o he Bal ic Sea o no . We also discuss he possible h eshold alues o he indica o o good en i onmen al s a us. The e ec o he ing quali y on he bi h a e o g ey seals was s udied ea lie (Kauhala e al., 2016) bu since hen we ha e go mo e da a and he e ec s o he o he a iables (sp a , zooplank on and cod da a) we e no es ed ea lie . Wi h hese new pa ame e s and he new ocus, we p edic ed ha 1) bo h he ing and sp a quali y o quan i y ha e posi i e impac s on he bi h a e o g ey seals, and 2) biomass and mean size o zooplank on, as well as cod abundance may ha e an in- luence on he quali y o quan i y o he ing and sp a and, hus, on he bi h a e o g ey seals. 2. Ma e ial and me hods 2.1. Seal samples Samples o g ey seal emales (n = 447) included in his s udy o i- gina ed om he Finnish sea a ea om he Gul o Finland o he Bo hnian Bay (ICES SD 29–32) and mos o hem (414) we e killed in he Gul o Bo hnia (ICES SD 30 and 31; Fig. 1.). Samples used he e we e collec ed be ween 16 Ap il and 30 June (see below he calcula ion o bi h a e) in 2001–2017. Mos seals we e hus sho du ing he ime when hey agg ega e on sea ice in he Gul o Bo hnia bu hey ep esen ac ually a la ge popula ion which may sp ead o he sou he n a eas o eed a e he mol ing pe iod in la e sp ing. The e o e he seal e- p oduc i e da a is also ep esen a i e o he sou he n sea a ea and samples om di e en a eas we e pooled. Indeed, bi h a e did no di e be ween he sea a eas (logis ic eg ession: chi 2 = 0.092, p = 0.762). Samples om hun ed g ey seals a e ou inely collec ed and analyzed a Na u al Resou ces Ins i u e Finland (Luke) o en i on- men al moni o ing. Hun ing season is om 16 Ap il o 31 Decembe (excep in Åland o 31 Janua y). No seals we e killed o he pu pose o his s udy bu he samples we e ob ained du ing egula hun . No by- caugh seals we e included in he s udy. Samples o each emale seal included a leas u e us, o a ies and he lowe jaw. Seal species was con i med om he lowe jaw. Age de- e mina ion was done by coun ing he inc emen al lines in he ce- men um om ans e sal sec ions o lowe canine ee h (e.g. Mans ield, 1991). G ey seals gi e bi h o one pup in Feb ua y o Ma ch, o ula e du ing he nu sing pe iod wo o h ee weeks la e bu he emb yo does no implan in o he u e us un il mid-summe (Boyd e al., 1999; Kauhala e al., 2014). Du ing he p e-implan a ion pe iod all o ula ed emales ha e an ac i e co pus lu eum (CL) and emales which ga e bi h o a pup he p esen yea also ha e a co pus albicans (CA) because CL apidly a ophies a e pa u i ion and changes o CA (Boyd, 1983,1984b). Bo h he p esence o CA and emains o placen al sca s, which o en can be seen in he u e us a e pa u i ion, hus indica e ha he emale ga e bi h he p esen yea (Kauhala e al., 2014). The bi h a e was hus de e mined as he p opo ion o emales wi h CA, and i was con i med om he p esence o placen al sca s. We used only seals killed be o e he implan a ion pe iod, i.e. be o e he 1s o July because he ea e CA and placen al sca s may disappea (Boyd, 1984b). The bi h a e was calcula ed o emales which we e 7–25 yea s old (n = 447) because bo h younge and olde emales ha e lowe bi h a es (Bowen e al., 2006; Kauhala e al., 2014). Bi h a e could no be calcula ed o each yea because sample sizes a ied om 3 o 69 pe yea . The e o e he alues o bi h a e we e smoo hed wi h 3-yea mo ing a e age. (The i s and las alues we e he means o wo yea s). Sample sizes o each 3-yea pe iod a ied om 25 o 178, and we e 25 and 17 o he i s wo and he las wo yea s. O a ies we e emo ed and c oss sec ions ( hickness abou 2 mm) we e made by kni e. They we e examined by naked eye o he p esence o co pus lu eum o co pus albicans (Kauhala e al., 2014). Bo h aces o each sec ion we e examined o CL (yellowish, so in ex u e and oc- casionally wi h a hollow cen e ) and CA (whi ish in colo , sca -like in ex u e). In cases when CA was no clea , i was con i med om his- ological slides. The posi ions o CL and CA in he le o igh o a y we e eco ded. Seals usually o ula e om al e na e o a ies in succes- si e yea s (A kinson, 1997; Boyd e al., 1999). U e us was cu open and examined o signs o placen al sca s (Kauhala e al., 2014). We also eco ded hei place (le o igh ho n) in he u e us. A esh sca is like a black bel inside he u e ine ho n. A e some weeks only he ims o he sca can be seen bu he ho n in ques ion is s ill hicke a he si e o he placen al sca han he o he ho n. Some o ange spo s may be seen a he si e o he placen a (U. Siebe , pe s. com.). 2.2. Fish and plank on da a He ing da a (weigh and ca ch) we e ob ained om ICES epo om wo a eas: he Gul o Bo hnia (ICES SD 30 and 31) om 1980 o 2017 and he sou he n sea a eas (SD 25–29 and 32) om 1974 o 2017, he ea e sou he n a ea and he Gul o Bo hnia (ICES, 2018;Fig. 1). Fish s ocks in he sou he n sea a eas o Finland a e conside ed as pa o a wide s ock and hus pooled in ICES epo (ICES, 2018). Also he sp a da a we e om he sou he n a ea, i.e. om he single s ock ha li es in he Bal ic Sea, om 1974 o 2017 (ICES, 2018). Cod da a we e om ICES SD 25–32, al hough cod is e y a e in he Gul o Bo hnia (ICES, 2018) om 1974 o 2017. The e o e, cod da a a e ac ually om he sou he n a ea. Also sp a is spa se in he Gul o Bo hnia, and hus we used only he ing da a in he analyses om his a ea. The mean indi idual weigh s (he ea e weigh ) o ish we e used as Fig. 1. Map o he s udy a eas: he Gul o Bo hnia (ICES SD 30 and 31) and he sou he n a ea (ICES SD 25–29 and 32, excep he Gul o Riga) in he Bal ic Sea. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 694 indices o ish quali y and he ca ch sizes o comme cial ishe y o he ing and sp a (in numbe ) and o cod (in ons) we e used as indices o hei abundance (he ea e ca ch). Ca ch sizes o ish we e used as abundance indices because s ock numbe s we e no a ailable o he whole s udy pe iod, and i was es ed ea lie (Kauhala e al., 2017) ha ca ch and s ock numbe s o he ing co ela ed posi i ely du ing sho e pe iods bo h in he Gul o Bo hnia and Cen al Bal ic Sea. Al hough used a lo , ca ch pe e o (CPUE) is o en p oblema ic in assessing he s a us o ish s ocks, as i may no be p opo ional o he abundance o a ish popula ion (e.g. Maunde e al., 2006). The ca ches o he ish popula ions a e mo e o less ollowing quo as se o he Bal ic Sea coun ies ( om o al allowable ca ch), which a e based on ish s ock assessmen s. Fu he mo e, biomass would combine he e ec s o hese wo a iables bu we wan ed o es he e ec s o he ing and sp a quali y and abundance on he bi h a e o seals sepa a ely. We used he age g oups i e yea s and olde (5+) and 5–6 yea s o Fig. 2. Bi h a e o g ey seals ± SE (n = 447) om 2001 o 2017 (A; see also Kauhala e al., 2016), he ing weigh 5+ (B) and ca ch 5+ in numbe (C), sp a weigh 5+ (D) and ca ch 5+ in numbe (E), cod ca ch in ons (F) du ing he s udy pe iod 1974–2017, zooplank on biomass (G) and zooplank on size (µg we weigh ind −1 ; H) om 1979 o 2016. Bi h a e was om he whole Finnish sea a ea (ICES SD 29–32), ish da a we e om he sou he n a ea (ICES SD 25–29 and 32) and plank on da a om he Gul o Finland (ICES SD 32). K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 695 he ing because hese age g oups o he ing we e ound o be he mos impo an o body condi ion o g ey seals, whe eas no ela ion be ween younge he ing and body condi ion o seals was ound (Kauhala e al., 2017). The age dis ibu ion o he ing in g ey seal die is indeed skewed owa ds old indi iduals (mean = 6.3 yea s, median = 8 yea s, mean size 18.1 cm) and he mean age is highe han ha in he popula ion (3.3 yea s; Gå dma k e al., 2012). Fu he mo e, Gå dma k e al. (2012) es ima ed ha he ing mo ali y due o seal p eda ion is g ea es in he age g oup 8+. Weigh s o he ing om he Bo hnian Sea co ela ed posi i ely wi h hose om he Bo hnian Bay (age 5+: = 0.83, p < 0.001, age 5–6: = 0.90, p < 0.001). The e o e we pooled he - ing da a om hese wo a eas, i.e. we used he ing da a om he Gul o Bo hnia. We also used he age g oup 5+ o sp a because sp a is much smalle han he ing and i is unlikely ha seals would p ey on sp a younge han i e yea s. All age g oups a e included in cod da a. Da a o zooplank on biomass (a e age mg we weigh m-3 ) and he mean size o zooplank on indi iduals (µg we weigh ind -1 ; he ea e zooplank on size) in he communi y we e ob ained om once a yea moni o ing om 1979 o 2016 o he Finnish En i onmen Ins i u e (SYKE). Fo he sou he n sea a eas, we used he da a om he Gul o Finland because in o he a eas plank on da a we e oo spo adic. We used he plank on da a om he Bo hnian Sea and Bo hnian Bay o ep esen he Gul o Bo hnia. 2.3. S a is ical es s We examined he possible ends in he bi h a e o g ey seals, he ing and sp a weigh s, he ing, sp a and cod ca ches, and zoo- plank on biomass and size om cu es and using eg ession analysis wi h log ans o ma ions when needed o no malize he dis ibu ions. We checked he no mali y o dis ibu ions o esiduals by using one- sample Kolmogo o -Smi no es . We compa ed bi h a e om he whole Finnish sea a ea wi h ish da a om he wo a eas, he sou he n a ea and he Gul o Bo hnia, because ish s ocks in hese a eas di e . We i s calcula ed co ela ion ma ices be ween he abo e men- ioned a iables. Based on he co ela ions, we chose he independen a iables possibly a ec ing he bi h a e o g ey seals and he ish a iables, and calcula ed models wi h Gene al linea model (GLM, so wa e Sys a 13). We es ed he possible e ec s o di e en in- dependen a iables on he bi h a e bo h wi h and wi hou a ime lag o one yea . Only one yea ime lag was es ed because i is unlikely ha ish quali y o quan i y would a ec bi h a e a e wo o mo e yea s, and an ea lie s udy (Kauhala e al., 2017) showed co ela ion be ween ish quali y and body condi ion o g ey seals wi hou a ime lag. One yea ime lag is easonable because sca ce o poo quali y ood e- sou ces in summe o au umn may cause nu i ional s ess o emales and hus a ec implan a ion and su i al a es o emb yos and cause low bi h a e he nex sp ing. We used s epwise backwa ds p ocedu e excluding he non-signi ican independen a iables one a a ime, he one wi h he highes p- alue i s . Only a iables which signi ican ly inc eased he 2 – alues we e included in he models. The le el o signi icance was se o 0.05. Howe e , we also ook in o accoun he AICc– alues (AIC co ec ed o small sample sizes) in cases when he e was a a iable wi h p– alue be ween 0.10 and 0.05, and chose he inal model acco ding o he lowes AICc- alue. AIC ( he Akaike in o ma ion c i e ion) p o ides a means o selec he bes model, ela i e o o he models, and anks he eg ession models om he bes o he wo s . 3. Resul s Bi h a e o g ey seals in he Finnish sea a ea (SD 29–32; Fig. 1) inc eased om 2001 o 2004, hen dec eased un il 2007, emained low un il 2011 and inc eased he ea e (Fig. 2A, Appendix A). The inc ease in bi h a e le eled o du ing he las wo yea s. 3.1. Sou he n a eas: Long- e m ends in ish and plank on The e has been a long- e m decline in he ing weigh om 1974 o la e 1990s and he ea e i has emained on he same le els al hough he e we e sho e pe iods o inc eased g ow h du ing he ea ly 2000s and 2010s (Fig. 2B; Appendix A). He ing ca ch a ied du ing he s udy pe iod wi h no clea end, excep o sho e pe iods: he ca ch de- clined om 1975 o 1983, inc eased om 1983 o 2000 and a e a quick decline inc eased again om 2003 o 2017 (Fig. 2C). Sp a weigh declined om 1974 o 2017 (Fig. 2D, Appendix A), al hough he e we e sho e pe iods o inc ease o decline du ing he s udy pe iod: sp a weigh i s inc eased om 1974 o 1983, hen de- clined om 1990 o 1999 and inc eased om 1999 o 2017. Sp a ca ch was much la ge han ha o he ing and luc ua ed du ing he s udy pe iod dec easing om 1974 o 1983, sha ply inc easing om 1990 o 1999, and declining om 1999 o 2017 (Fig. 2E). The cod ca ch declined du ing he s udy pe iod (Fig. 2F), al hough he e was a sho pe iod o inc ease om 1974 o 1984. Zooplank on biomass has sligh ly dec eased in he Gul o Finland wi h a pe iod o highe alues in ea ly 1990s and he lowes biomass in ea ly 2000s (Fig. 2G). Also he zooplank on size has sligh ly dec eased bu he e ha e been a ew peak yea s in la e 1980s and some yea s in 2000s (Fig. 2H). 3.2. Sou he n a eas: ela ions be ween he ophic le els We i s calcula ed he co ela ion ma ices o h ee pe iods (1974–1983, 1983–1999 and 1999–2017) when he ing o sp a weigh ei he inc eased o dec eased bu we combined he i s wo pe iods o one pe iod om 1974 o 1999 (Appendix B1) because he esul s o hem we e almos equal. The esul s o he las pe iod om 1999 o 2017 di e ed somewha om hose o he o he pe iod and a e he e- o e p esen ed sepa a ely (Appendix B2). In he i s pe iod (be o e 1999) cod was abundan and had a s onge e ec on he ing and sp a popula ions. The co ela ion ma ices o he whole s udy pe iod we e also calcula ed (Appendix B3). The bi h a e o g ey seals co ela ed posi i ely wi h he ing and sp a weigh s (da a o bi h a e only om he second ime pe iod, a ime-lag o one yea ga e he mos signi ican esul s; Appendix B2). Sp a weigh co ela ed posi i ely wi h he ing weigh , whe eas sp a ca ch co ela ed nega i ely wi h bo h he ing and sp a weigh s du ing all s udy pe iods (Appendix B1, B2 and B3). Also he ing ca ch co e- la ed nega i ely wi h he ing weigh in all s udy pe iods, bu especially du ing he whole s udy pe iod he co ela ion was weake han ha be ween sp a ca ch and he ing weigh . Cod ca ch co ela ed nega- i ely wi h sp a ca ch and posi i ely wi h he ing and sp a weigh s, excep in he second pe iod when cod abundance was low. Zooplank on biomass and size co ela ed posi i ely wi h he ing and sp a weigh s du ing he whole s udy pe iod bu no signi ican co ela ions we e ound be ween plank on da a and he bi h a e o seals (wi h o wi hou a ime-lag). He ing and sp a weigh oge he bes explained he a ia ion in he bi h a e o g ey seals wi h he ime-lag o one yea (Fig. 3A, B, Table 1, see also Kauhala e al., 2016). Sp a and cod ca ches oge he explained well he a ia ion in he ing weigh du ing he i s and whole s udy pe iod (Fig. 4A, B), whe eas he ing ca ch alone bes ex- plained he a ia ion in he ing weigh om 1999 o 2017 (Fig. 4C). Sp a and cod ca ches oge he explained well he a ia ion in sp a weigh du ing he whole s udy pe iod (Fig. 5A, B). Sp a ca ch alone bes explained he a ia ion in sp a weigh du ing he sho e pe iods om 1974 o 1999 and om 1999 o 2017, whe eas cod ca ch ex- plained he a ia ion in sp a ca ch (Fig. 5C), excep in he second pe iod. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 696 3.3. The Gul o Bo hnia He ing weigh declined also in he Gul o Bo hnia ( = −5.4, 2 = 0.45, F = 29.6, p < 0.001; Fig. 6A), whe eas he ing ca ch in- c eased in he cou se o he s udy ( = 10.3, 2 = 0.75, F = 105.5, p < 0.001; Fig. 6B). Also zooplank on biomass inc eased du ing he s udy ( = 2.9, 2 = 0.21, F = 8.1, p = 0.008; Fig. 6C) in he Bo hnian Sea (bu no in he Bo hnian Bay) whe eas he e was no signi ican change in he zooplank on size in he Gul o Bo hnia. Sp a and cod a e no signi ican species in he Gul o Bo hnia and he e o e no ana- lyzed. Bi h a e o g ey seals and he ing weigh (wi hou a ime-lag) co ela ed posi i ely ( = 0.72, p = 0.001), whe eas he ing weigh and ca ch co ela ed nega i ely ( = −0.73, p < 0.001) in he Gul o Bo hnia om 1980 o 2015. He ing ca ch co ela ed posi i ely wi h zooplank on size ( = 0.37, p = 0.035) and zooplank on biomass ( = 0.39, p = 0.026). No signi ican co ela ion was ound be ween he ing weigh and zooplank on da a. He ing weigh pa ly explained he change in he bi h a e ( 2 = 0.52, F = 16.3, p = 0.001, Fig. 7A). The change in he ing weigh was well explained by he change in he ing ca ch ( 2 = 0.54, F = 42.0, p < 0.001; Fig. 7B). Zooplank on biomass om he Bo hnian Sea, in u n, explained some o he a ia ion in he ing ca ch ( 2 = 0.15, F = 5.5, p = 0.026; Fig. 7C). Fig. 3. Rela ionships be ween sp a weigh 5+ and bi h a e o g ey seals (A) and be ween he ing weigh 5+ and bi h a e (B). Bi h a e was om he whole Finnish sea a ea om 2001 o 2017 and ish da a we e om he sou he n a ea om 2000 o 2016 (i.e. a ime lag o one yea was included). Table 1 Signi ican models (GLM) o independen a iables a ec ing he bi h a e o g ey seals, he ing and sp a weigh and sp a ca ch in di e en ime pe iods. No model was ob ained o he ing ca ch. A ime-lag o one yea is included when ela ionships be ween bi h a e and he ing and sp a da a we e es ed. He ing and sp a ca ches a e in numbe s, cod ca ch is in ons. Bi h a e was om he whole Finnish sea a ea and ish da a we e om he sou he n a ea. Age g oups o he ing and sp a da a a e also gi en. Dependen a iable Pe iod Independen Model a iables 2 F p Bi h a e 2001–2017 Sp a weigh 5+ 2.4 0.65 13.2 0.001 He ing weigh 5+ 2.6 He ing weigh 5+ 1974–1999 Cod ca ch 4.5 0.79 44.2 < 0.001 Sp a ca ch 5+ −3.3 He ing weigh 5+ 1999–2017 He ing ca ch 5–6 −3.0 0.35 9.1 0.008 He ing weigh 5+ 1974–2017 Cod ca ch 6.5 0.84 110.6 < 0.001 Sp a ca ch 5+ −6.0 Sp a weigh 5+ 1974–1999 Sp a ca ch 5+ −6.1 0.61 37.1 < 0.001 Sp a weigh 5+ 1999–2017 Sp a ca ch 5+ −2.8 0.31 7.6 0.014 Sp a weigh 5+ 1974–2017 Sp a ca ch 5+ −5.6 0.70 48.2 < 0.001 Cod ca ch 2.5 Sp a ca ch 5+ 1974–1999 Cod ca ch −4.3 0.44 18.6 < 0.001 Sp a ca ch 5+ 1974–2017 Cod ca ch −5.5 0.42 30.5 0.001 Fig. 4. Rela ionships be ween sp a ca ch 5+ (in numbe ) and he ing weigh 5+ (A) and be ween cod ca ch (in ons) and he ing weigh 5+ (B) du ing he whole s udy pe iod, and be ween he ing ca ch 5-6 (in numbe ) and he ing weigh 5+ om 1999 o 2017 (C) om he sou he n a ea. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 697 4. Discussion 4.1. Rela ions be ween he ophic le els Ou i s p edic ion was ha he ing and sp a quali y (weigh ) o quan i y has an e ec on he bi h a e o g ey seals. We indeed ound ha bi h a e o g ey seals was posi i ely ela ed wi h bo h sp a and he ing weigh . The weigh pa ame e s o ish, on he o he hand, depended on he quan i y. I was shown ea lie (Kauhala e al., 2016, 2017) ha he ing weigh was ela ed o he nu i ional s a us o g ey seals and he e o e possibly also wi h hei bi h a e. Rep oduc i e a e is sensi i e o body condi ion in capi al b eede s, such as he g ey seal (e.g. Guine e al., 1998; Boyd, 2000). The esul s hus sugges ha changes in he ood web which a ec sp a and he ing quali y may be seen on he op o he ood web in he ep oduc i e a e o g ey seals, bo h wi h and wi hou a ime lag o one yea . Fig. 5. Rela ionships be ween sp a ca ch 5+ (in numbe s) and weigh 5+ (A), be ween cod ca ch (in ons) and sp a weigh 5+ (B), and be ween cod and sp a ca ches 5+ (C) du ing he whole s udy pe iod om he sou he n a ea. Fig. 6. T ends in he ing weigh 5+ (A) and he ing ca ch 5+ (in numbe ; B) om 1980 o 2017, and zooplank on biomass (C) in he Gul o Bo hnia om 1980 o 2016. Fig. 7. Rela ionships be ween he ing weigh 5+ and bi h a e o g ey seals om 2001 o 2017 (A), he ing ca ch 5+ (in numbe ) and he ing weigh 5+ (B) om 1980 o 2017, and zooplank on biomass and he ing ca ch 5+ (C) om 1980 o 2016. He ing da a we e om he Gul o Bo hnia, plank on da a om he Bo hnian Sea and bi h a e om he o al Finnish sea a ea. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 698 Ou s udy showed clea ly ha he abundances (ca ches in numbe s as abundance indices) o bo h he ing and sp a had a nega i e in lu- ence on he ing weigh , and sp a abundance had a nega i e impac on sp a weigh , sugges ing ha he e is densi y-dependen compe i ion o zooplank on wi hin and be ween he ing and sp a popula ions (Möllmann e al., 2004; Casini e al., 2006, 2010; Kuosa e al., 2017). Bo h species mainly consume copepods and cladoce ans bu sp a is a mo e e icien plank i o e: a high sp a densi ies, he ing g ow h is conside ably lowe han a low sp a le els (Flinkman e al., 1992; Möllmann e al., 2004, Casini e al., 2010). Ou models indica e, howe e , ha in he sou he n a eas, he ing weigh is in luenced e en mo e by in e speci ic compe i ion wi h sp a han by in aspeci ic compe i ion wi hin he he ing popula ion (Table 1). Fu he mo e, he ing abundance did no ha e a signi ican e ec on sp a weigh , sugges ing ha mainly in aspeci ic compe i ion o ood esou ces seems o a ec sp a weigh . I is hus sp a abundance which has he g ea es e ec on bo h he ing and sp a weigh (see also Möllmann e al., 2004; Casini e al., 2010), p obably because sp a is much mo e abundan han he ing (especially in SD 29) and mo e e icien han he ing as a plank i o e (Möllmann e al., 2004). Sp a ca ch inc eased exponen ially om 1990 o 2000 in he sou he n sea a eas, and he la ges sp a ca ches we e abou 5–6- old compa ed o he ca ch sizes o he ing (Fig. 2C and E; ICES, 2018). I is, howe e , possible ha in low sp a densi ies hyd o-clima ic condi ions, such as salini y, ha e an e - ec on he he ing popula ion (Casini e al., 2010). In he Gul o Bo hnia, sp a is spa se and he ing abundance alone had an impac on he ing weigh . Ou second p edic ion was ha zooplank on biomass and size, as well as cod abundance ha e an in luence on he ing and sp a weigh s, and, hus, indi ec ly may a ec g ey seal ep oduc ion. We ound in- deed posi i e co ela ions be ween he weigh s o hese ish species and zooplank on biomass and size du ing he whole s udy pe iod, sup- po ing ou second hypo hesis. Howe e , he in luence o zooplank on was no s ong, and ish popula ion densi ies seemed o be mo e im- po an o clupeid (he ing and sp a ) weigh . In he Gul o Bo hnia, zooplank on biomass was, howe e , a s ong p edic o o he ing ca ch. Kuosa e al. (2017) also ound ood limi a ion o he ing in he Bo hnian Sea. As we included cod abundance (ca ch in ons) in o ou models, we we e able o show ha cod had a nega i e e ec on sp a and he ing abundances and, he e o e, a posi i e impac on hei weigh s in he sou he n a ea especially du ing he i s ime pe iod when cod was abundan . The e ec o cod abundance on sp a abundance was ound also in se e al o he s udies (e.g. Ruds am e al., 1994; Casini e al., 2006, 2008; Ös e blom e al., 2007). Hence, cod abundance p obably has an indi ec in luence on he g ey seal ep oduc ion. The whole Bal ic ood web has, howe e , changed due o he decline o cod abundance because o a s ong p essu e by ishe ies and he low salini y and oxygen con en s o he Bal ic Sea (e.g. Casini e al., 2008). In he second ime pe iod in he 2000s, he op-down e ec o cod on clupeids has hus diminished (Casini e al., 2006, 2008; Kuosa e al., 2017) Fig. 8. Bi h a e o g ey seals in he Finnish sea a ea wi h he p esen h eshold alue o a g owing popula ion and he sugges ed alue o a s able popula ion wi h 95% con idence limi s (A), blubbe hickness o adul emales (wi h mon h as a co a ia e), modi ied igu e om Kauhala e al. (2017):Fig. 4C (B), he ing weigh wi h means and 95% cl in he sou he n a ea and in he Gul o Bo hnia o 2001–2017 (C) and sp a weigh wi h mean and 95% cl in he sou he n a ea o 2001–2017 (D), and zooplank on biomass in he Gul o Finland and in he Bo hnian Sea o 2001–2016 wi h a HELCOM h eshold alue (E). G een a eas e e o he ime pe iod when bi h a e was abo e he sugges ed h eshold alue and he ed a ea e e s o he ime pe iod when bi h a e was below he h eshold alue. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 699 which esul ed in an inc ease in sp a abundance (Casini e al., 2006, 2008, 2009; Ös e blom e al., 2007), and consequen ly, in a dec ease in he ing and sp a weigh s, which may be seen in he declining bi h a e o g ey seals in he 2000s (Kauhala e al., 2014). In ecen yea s, howe e , sp a abundance declined esul ing in an inc ease in he ing and sp a weigh s and p obably also in he bi h a e o g ey seals. The changes in he bi h a e ha e also in luenced he g ey seal popula ion g ow h a e (Kauhala e al., 2016). O he ish species a e also in ol ed in he Bal ic ood web dynamics. The s ickleback (Gas e os eus aculea us) popula ion has inc eased in ecen yea s, p obably because o inc eased nu ien s and sui able zooplank on in he Bo hnian Sea and possibly also due o he decline o he popula ions o p eda o y ish in mo e sou he n a eas (Ylipo imo, 2017). Compe i ion o ood esou ces be ween he h ee plank i o ous ish species (Ylipo imo, 2017) has p obably in luenced he loss o weigh o indi idual ish. In he Bo hnian Bay, also endace (Co egonus albula) may compe e wi h he ing o zooplank on and, hus, ha e an e ec on he ing weigh (e.g. Be genius e al., 2013). Un o una ely we did no ha e da a on endace o h ee-spined s ickleback o be es ed he e. 4.2. Rep oduc i e s a us o g ey seals as an indica o o changes in he Bal ic ood web Rep oduc i e s a us (bi h a e and p egnancy a e) o g ey seals, one o he op p eda o s in he Bal ic Sea, is used e.g. by HELCOM as an indi ec indica o o he en i onmen al s a us o he Bal ic Sea because ep oduc i e a e o seals is ela ed o hei body condi ion, which, in u n, is in luenced by ood esou ces (Bowen e al., 2006; Kauhala e al., 2017; HELCOM, 2018a). P egnancy a e is calcula ed om he au umn sample o 6–24-yea -old emales indica ing he p opo ion o emales p egnan a e he implan a ion pe iod in mid-summe . Bi h a e is calcula ed om he sp ing sample o 7–25-yea -old emales indica ing he p opo ion o emales ha ga e bi h he p esen sp ing. Indica o s should e lec changes in he ecosys em and be wa ning signals o de- le e ious changes in he en i onmen . The p esen s udy showed likely ela ions be ween zooplank on, clupeid quali y and bi h a e o g ey seals, sugges ing ha he bi h a e indeed e lec s changes in he ood web o e h ee ophic le els. The changes we e no only ophic cas- cades be ween zooplank on, ish and seals bu we e also in luenced by densi y-dependen compe i ion o ood esou ces be ween and among sp a and he ing popula ions, which, in u n, a e a ec ed by p eda ion by cod. Acco ding o HELCOM (2018a), a h eshold alue o good en- i onmen al s a us is eached when he mean p egnancy a e/bi h a e o a g owing seal popula ion is a leas 90% du ing he pas six yea s. The alue in he Finnish g ey seal popula ion is 87% (mean o 2012–2017) which is only sligh ly below his h eshold alue. I mus be emembe ed, howe e , ha in ecen yea s he g ow h o he Bal ic g ey seal popula ion has le elled o in he Finnish sea a ea (Luke, 2017) and, hence i is possible ha he popula ion is app oaching he ca ying capaci y o he en i onmen . In a dense seal popula ion he bi h a e o en declines and can a y be ween 50 and 90%, depending on he ood esou ces (e.g. Boyd e al., 1999). In he 2000s, he bi h a e o g ey seals a ied be ween 63% and 94% in he Finnish sea a ea (see also Kauhala e al., 2016). We he e o e sugges ha a lowe h eshold alue o a s able popula ion would s ill e lec a good s a e o he en i onmen . In Fig. 8, we ha e compa ed h eshold alues o zooplank on biomass as ag eed o he HELCOM zooplank on indica o (HELCOM, 2018b), yea s o high/low alues o sp a and he ing weigh and bi h a e o g ey seals, and he indica o h esholds o bi h a e. I can be seen om he compa ison ha , o ins ance, i he h eshold alue o bi h a e we e 78% (mean o he pe iod 2001–2017) o a s able seal popula ion, good s a e o bi h a e would coincide well wi h he good yea s o ood esou ces (Fig. 8). Con idence limi s (95%) o his h eshold alue would be 72.6–83.8%. The p esen h eshold alue (90%) o a g owing popula ion has no con idence limi s which makes i ai ly a bi a y. One exac alue in biological da a is no sensible. Nu i ional s a us ( he hickness o subcu aneous blubbe laye ) o sub-adul g ey seals in au umn is ano he seal heal h indica o used by HELCOM (2018c). The h eshold alue o blubbe hickness has been se a 40 mm o a g owing seal popula ion and a 25 mm o a popu- la ion a ca ying capaci y o he en i onmen (HELCOM, 2018c). The mean blubbe hickness o sub-adul g ey seals has been abo e 25 mm bu below 40 mm du ing he whole pe iod when we ha e ecei ed seal samples (since 2002; Kauhala e al., 2017), sugges ing ha he in- dica o , nu i ional s a us o sub-adul g ey seals, is no e y sensi i e o he changes in he ood web. On he o he hand, body condi ion o pups could be a good indica o because pups a e usually he i s o eac o poo en i onmen al con- di ions, and hei dec easing body condi ion could hus be he i s ala ming signal o dis u bances in he en i onmen , such as ood e- sou ces o ice condi ions (Kauhala e al., 2017). Body condi ion o adul emales may be he mos impo an o he popula ion because i a ec s implan a ion and mo ali y a es o emb yos/ e uses (e.g. Boyd, 1984a; Bowen e al., 2006) and hus he bi h a e o emales and g ow h a e o he popula ion. Blubbe hickness o adul emales a ied om 40 mm o 60 mm and was ela ed o he ing weigh du ing he s udy pe iod (Kauhala e al., 2017). In Fig. 8B we compa ed blubbe hickness wi h he ood web a iables o his s udy and show ha he a iable ollows well he changes in he ood web and is ela ed o he changes in he bi h a e. Thus, we sugges ha he mean blubbe hickness om 2002 o 2017 (49.5 mm, 95% cl: 46.5–52.5 mm) could well e lec a h eshold alue o adul emale g ey seals in a s able popula ion. 5. Conclusions We conclude ha he a iabili y in he bi h a e o g ey seals is p obably ela ed o changes in he Bal ic ood web, and ep oduc i e a e can hus be used as one indica o o he en i onmen al s a us o he Bal ic Sea. Howe e , as o he ac o s such as en i onmen al con- aminan s, especially PBCs and DDT, can cause se e e ep oduc i e dis u bances in seal emales (e.g. Be gman and Olsson, 1986; Be gman, 1999; Nyman e al., 2003), he possible oles o o he an h opogenic haza dous subs ances should also be s udied. Fu he mo e, he possible impac s o o he ish species, as well as he e ec o clima e change on he body condi ion and ep oduc i e a e o g ey seals should be in- es iga ed. We also sugges ha ou p oposal o a h eshold alue o he bi h a e o a s able seal popula ion would be conside ed. Acknowledgemen s We a e e y g a e ul o M. Isomu su, M. Kunnas an a, and P. Timonen who helped o examine seal samples. Hun e s sen us samples o hun ed seals. KK was inanced by BONUS Bal Heal h. The Bal Heal h p ojec has ecei ed unding om BONUS (A . 185), unded join ly by he EU, Inno a ion Fund Denma k (g an s 6180-00001B and 6180- 00002B), Fo schungszen um Jülich GmbH, Ge man Fede al Minis y o Educa ion and Resea ch (g an numbe FKZ 03F0767A), Academy o Finland (decision #311966) and Swedish Founda ion o S a egic En i onmen al Resea ch. K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 700 Appendix A Appendix B Appendix A T ends in di e en a iables in di e en s udy pe iods in he sou he n a ea, bi h a e om he whole Finnish sea a ea. The models we e calcula ed wi h Gene al linea model (GLM). Age g oups o he ing and sp a da a a e also gi en. Va iable Pe iod Model 2 F p Bi h a e 2004–2007 −7.8 0.97 61.3 0.016 2011–2015 5.1 0.90 25.6 0.015 He ing weigh 5+ 1975–1983 9.3 0.93 87.1 < 0.001 1983–1999 −17.6 0.95 308.5 < 0.001 1999–2017 2.2 0.22 4.7 0.045 1974–2017 −9.2 0.67 85.4 < 0.001 He ing ca ch 5+ 1975–1983 −5.5 0.81 29.8 0.001 1983–2000 2.4 0.26 5.6 0.031 2003–2017 5.2 0.68 27.3 0.038 Sp a weigh 5+ 1974–1983 5.7 0.80 32.3 < 0.001 1990–1999 −13.4 0.96 178.5 < 0.001 1999–2017 3.1 0.36 9.6 0.006 1974–2017 −6.6 0.51 43.9 < 0.001 Sp a ca ch 5+ 1974–1983 −6.9 0.86 47.3 < 0.001 1990–1999 9.6 0.92 91.4 < 0.001 1999–2017 −5.6 0.65 31.8 < 0.001 Cod ca ch 1974–1984 5.5 0.77 30.5 < 0.001 1985–2017 −11.8 0.82 138.8 < 0.001 1974–2017 −12.8 0.80 163.5 < 0.001 Plank on biomass 1974–2017 −2.2 0.12 4.8 0.035 Plank on size 1974–2017 −3.4 0.27 11.8 0.002 Appendix B1 Co ela ions be ween he weigh s o he ing and sp a , and he ca ch sizes o he ing, sp a and cod om 1974 o 1999 om he sou he n a ea. Hw = he ing weigh , Hc = he ing ca ch, Sw = sp a weigh , Sc = sp a ca ch, Cc = cod ca ch. * p < 0.05, ** p < 0.01, *** p < 0.001. Age g oups o he ing and sp a da a a e also gi en. Va iable Hw 5+ Hw 5–6 Hc 5+ Hc 5–6 Sw 5+ Sc 5+ Cc P Hw 5+ 1.00 Hw 5–6 0.99*** 1.00 Hc 5+ −0.41* −0.40* 1.00 Hc 5–6 −0.75*** −0.74*** 0.84*** 1.00 Sw 5+ 0.66*** 0.67*** −0.32 −0.47* 1.00 Sc 5+ −0.79*** −0.76*** 0.64*** 0.71** −0.78*** 1.00 Cc 0.84*** 0.80*** −0.20 −0.48* 0.45* −0.66*** 1.00 Appendix B2 Co ela ions be ween he weigh s o he ing and sp a , he ca ch sizes o he ing, sp a and cod om he sou he n a ea, and he bi h a e o 7–25-yea -old g ey seal emales om Finland. Da a o bi h a e is om 2001 o 2017, da a o ish om 2000 o 2016. Hw = he ing weigh , Hc = he ing ca ch, Sw = sp a weigh , Sc = sp a ca ch, Cc = cod ca ch. * p < 0.05, ** p < 0.01, *** p < 0.001, ⁰0.05 < p < 0.10. Age g oups o he ing and sp a da a a e also gi en. Va iable Hw 5+ Hw 5–6 Hc 5+ Hc 5–6 Sw 5+ Sc 5+ Cc Hw 5+ 1.00 Hw 5–6 0.96*** 1.00 Hc 5+ −0.44⁰ −0.51* 1.00 Hc 5–6 −0.59** −0.64** 0.84*** 1.00 Sw 5+ 0.64** 0.61** 0.14 −0.16 1.00 Sc 5+ −0.54** −0.42⁰ −0.16 0.13 −0.56* 1.00 Cc −0.35 −0.28 −0.50* −0.26 −0.47* 0.62** 1.00 Bi h a e 0.71** 0.65* −0.13 −0.27 0.69** −0.10 −0.26 K. Kauhala, e al. Ecological Indica o s 102 (2019) 693–703 701