Elgstudiene på Vega. Erfaringer fra 30 år i havgapet

Abstract

Solberg, E.J., Heim, M., Herfindal, I., Rolandsen, C.M., Haanes, H., Lee, A.M., Tremblay, J.-P., Markussen, S.S., Ofstad, E.G., Kvalnes, T. & Sæther, B.-E. 2024. Elgstudiene på Vega – Erfaringer fra 30 år i havgapet. NINA Rapport 2258. Norsk institutt for naturforskning. Elgstudiene på Vega består av en rekke forskningsprosjekter om elgens økologi, evolusjon og forvaltning som vi gjennomførte på øya Vega i Nordland i perioden 1992-2021. Studiene ble opprinnelig igangsatt av Norsk institutt for naturforskning (NINA) med hensikt 1) å studere de demografiske konsekvensene av varierende kjønns- og aldersstruktur i en høsta elgbestand, men ble siden utvidet med fire nye delprosjekter. Disse ble gjennomført i samarbeid med NTNU i Trondheim, og omhandlet 2) elgens demografi og livshistoriestrategi, 3) elgens områdebruk og 4) elgens genetikk og evolusjon. Sistnevnte prosjekt var delt inn i to påfølgende delprosjekter (4a og 4b). I denne rapporten beskriver vi elgens bestandsutvikling på Vega i perioden 1985-2021 og oppsummerer resultatene fra de ulike forskningsprosjektene. Prosjektene var basert på individbaserte data over elgens kroppsvekst, reproduksjon, overlevelse, atferd og slektskapsforhold. Disse ble framskaffet gjennom å radiomerke (VHF- og GPS-sendere), veie, måle, prøveta, og siden følge opp tilnærmet alle elgene som overlevde fram til ca. 8 måneders alder (kalvens første vinter) og alle innvandrende nye individer til bestanden. I studieperioden (1992-2021) merket vi totalt 324 elger på Vega, hvorav de fleste ble påskutt og bedøvd fra helikopter ved bruk av pilgevær med bedøvelsesmiddel. I tillegg merket vi 50 elger på samme måte på naboøya Ylvingen (i Vega kommune) i perioden 2009-2013. Også elger felt under jakt ble veid og prøvetatt. Prøvene inkluderte DNA fra vevsprøver som vi siden benyttet til å bestemme slektskapsforhold (f.eks. foreldre-avkom), bestandsstørrelse og bestandsstruktur tilbake til 1985, da de 2 første elgene i moderne tid koloniserte Vega fra fastlandet. Siden 1985 har elgbestanden på Vega variert mye i størrelse og sammensetning. Bestanden økte kraftig fram til 1994 da det før jakt var drøye 50 elger på øya. Deretter ble bestanden regulert til omkring 30 dyr etter jakt (vinterbestand) og holdt på omtrent samme nivå de neste 6 årene. De siste 20 årene har bestanden økt, og nådde en foreløpig topp i 2019, med 55 individer etter jakt. Elgjakt ble først tillatt i Vega kommune i 1989 (2 dyr) og har siden vært brukt av forskning (1992-1999) og forvaltning (2000-) til å regulere bestanden til ønsket nivå og sammensetning. Totalt ble det felt 635 elger på Vega i perioden 1989-2021, hvorav de fleste var kalv (52 %). Jaktuttaket var høyest i 2015, med 34 dyr. I tillegg har vi registrert 46 elger (eldre enn 4 måneder) som har omkommet av andre årsaker enn jakt, hvorav 3 i forbindelse med merking (en fjerde omkom i 2022). Vi har også registrert 10 elger som utvandret fra Vega i studieperioden. De fleste svømte til fastlandet eller øyer mellom fastlandet og Vega. I motsatt retning ble 13 okser og 5 elgkyr registrert som innvandrere til Vega. Elgbestanden på Vega er preget av en høy rekrutteringsrate (f.eks. andel kalv i førjakt-bestanden) som følge av høye fruktbarhetsrater, og periodevis svært lav okseandel. Dette gjaldt særlig på siste halvdel av 1990-tallet da okseandelen ble redusert mye som en del av forskningsprosjektet på øya. Siden har okseandelen økt, samtidig som det har skjedd en nedgang i elgkyrnes aldersspesifikke ruktbarhetsrater. Sistnevnte henger sannsynligvis sammen med en tilsvarende nedgang i kalv- og åringsvekter og kan skyldes økende grad av matbegrensning som følge av dagens høyere bestandstetthet. I tillegg kan økende temperaturer både sommer og vinter ha ført til synkende kvalitet på elgens beiteplanter og høyere energikostnader i forbindelse med termoregulering (regulering av kroppstemperatur). I det første delprosjektet på Vega (1992-1999), undersøkte vi hvordan skjevheter i kjønns- og aldersstruktur påvirker elgens kroppsvekst og reproduksjon i to påfølgende perioder. Først reduserte vi oksenes gjennomsnittsalder (men ikke andelen okser) og deretter reduserte vi andelen okser (men ikke oksealder). Til tross for betydelige endringer i kjønns- og aldersstruktur fant vi ingen endringer i andelen elgkyr som ble bedekt. Det var heller ingen endringer i oksenes kroppsvekst som følge av at flere deltok i brunsten ved tidligere alder. Vi fant imidlertid betydelige forsinkelser i kalvenes fødselstidspunkt, som i neste omgang førte til lavere kalvevekt på høst og vinter. Siden eldre okser viste seg å produsere en overvekt av oksekalver, fant vi også en nær sammenheng mellom oksenes gjennomsnittsalder i bestanden og andelen oksekalver produsert hvert år. En lav andel okser kan dermed ha flere uheldige konsekvenser for elgbestanden. I de påfølgende årene (2000-2021) var bestanden gjenstand for ordinær jakt og uten større manipulasjoner av bestandsstrukturen. I denne fasen gjennomførte vi flere studier av elgens demografi og livshistoriestrategi, og trakk da veksler på stadig bedre slektskapsdata fra de genetiske analysene. Foruten effekten av fødselstidspunkt, fant vi i disse studiene at kalvens vekt også påvirkes av mors alder, kalvens kjønn, og antallet søsken samme år (0 eller 1). Denne vektvariasjonen vedvarte gjennom livet slik at store kalver ble store individer også som voksne. Kroppsvekten er viktig i elgens liv ettersom tunge elgkyr gjerne starter å reprodusere i ung alder og dermed kan produsere mange kalver i løpet av et langt liv. På Vega er elgene i svært god kondisjon (høye vekter) og er jevnt over svært produktive. For eksempel kom mer enn 50 % av elgkyrne med kalv som toåring (bedekt som åringsdyr), hvorav 7 % med tvillingkalv. De mest produktive elgkyrne på Vega (15-18 kalv i løpet av livet) startet å reprodusere som toåringer, og holdt seg i live til godt voksen alder (11-15 år). Interessant nok var det også et negativt forhold mellom alder ved første reproduksjon og framtidig reproduktiv suksess (antall kalver), og mellom alder ved første reproduksjon og sannsynligheten for å overleve jakta. Førstnevnte viser at det er store kvalitetsforskjeller mellom elgkyr, mens sistnevnte mest sannsynlig skyldes at jegerne velger å ikke felle kua fra kalven. Elgkyr som starter å reprodusere som toåring har derfor større sannsynlighet for å overleve jakta enn elgkyr som starter som 3- eller 4-åring. Variasjonen i reproduktiv suksess er større blant okser enn blant kyr. På Vega oppnådde kun en tredjedel av oksene å bli far, og selv blant disse var det stor variasjon i antall avkom. De mest produktive oksene paret seg med opptil 8 elgkyr i løpet av brunstsesongen, og den mest suksessfulle oksen ble far til hele 44 kalver. I gjennomsnitt ble hver okse far til 3,2 kalver. Oksenes overlevelse var den viktigste årsaken til variasjonen i antallet avkom. Svært mange okser ble skutt alt som halvannetåring og hadde dermed få muligheter til å bli far. I tillegg er oksenes vekt og gevirstørrelse av betydning for å oppnå parringer. Disse trekkene øker med oksens alder og følgelig hadde eldre okser større suksess enn yngre okser under brunsten. Selv små okser i sin aldersgruppe kan dermed oppleve å bli far dersom de får leve til fullvoksen alder (5 år +). Dette introduserer en viss grad av tilfeldigheter med hensyn til hvilke okser som bidrar til neste generasjon, og disse tilfeldighetene blir mer framtredende når andelen okser i bestanden er lav. Under slike forhold kan selv genetisk svake okser bli far til en stor andel av kalvene. På starten av 2000-tallet og 2010-tallet gjennomførte vi også flere studier av elgens arealbruk på Vega. Basert på data fra elgens radiohalsbånd kan vi avdekke hvor elgen oppholdt seg på ulike deler av døgnet og året, og vi kan studere hvordan elgen bruker landskapet på Vega i forhold til ressursene som finnes der (eks. mat). Denne typen studier ble spesielt interessante etter 2006, da vi for første gang utstyrte elgen med GPS-sendere. Dette er sendere som kan registrere et stort antall nøyaktige posisjoner (± 5 m) med jevne mellomrom. Basert på denne typen romlige data estimerte vi størrelsen på områdene som elgen bruker i løpet av året eller i ulike sesonger (hjemmeområdet) og vi analyserte elgens habitatbruk og habitatseleksjon. Habitatbruken viser hvordan elgen fordeler tiden sin i ulike arealtyper (eks. ulike typer skog, innmark, lynghei etc.), mens habitatseleksjon angir hvorvidt elgen bruker ulike habitattyper mer eller mindre enn forventet ut fra habitattypenes relative areal i leveområdet. Studiene viser at enkelte habitattyper er spesielt mye brukt i forhold til forekomst, for eksempel næringsrike lauvskoger. Habitatbruken varierte også mellom kjønn, mellom sommer og vinter, og i en viss grad mellom aldersgrupper. I tillegg fant vi en tydelig effekt av menneskelig aktivitet på hvor elgen oppholder seg. Dette var særlig tydelig i kalvingsperioden og i de første ukene etter kalving. I denne perioden oppholder elgkua seg lengst mulig unna folk eller i andre ‘sikre’ områder, til tross for at disse områdene ofte kan tilby relativt lite mat. I de siste to delprosjektene gjennomførte vi en omfattende kartlegging av elgens genetiske slektskapsforhold på Vega, og undersøkte hvordan innavl og jakt påvirker elgens genetiske variasjon og evolusjonære utvikling. Øyas relative isolerte beliggenhet, kombinert med et begrenset antall innvandrende elg, gjør bestanden sårbar for innavl og innavlsdepresjon. I tillegg er det ingen andre større rovdyr enn mennesker på øya, hvilket gjør at jakta og de valgene som gjøres av forvaltning og jegere kan ha stor effekt på bestandens evolusjonære utvikling. Resultatene fra disse studiene viser at elgbestanden på Vega i perioder har vært betydelig innavlet som følge av lav bestandsstørrelse og hyppige parringer mellom nære slektninger. Særlig stor var graden av innavl på 2000-tallet, da gjennomsnittlig innavlsgrad i bestanden tilsvarer hva vi kan forvente dersom halvsøsken parrer seg med hverandre. Dette skyldes en relativt lav innvandring av dyr fra fastlandet i denne perioden, og dermed en akkumulering av innavl over generasjoner (innavlede slektninger som parer seg med hverandre). Motsatt fant vi betydelig reduksjon i graden av innavl i årene etter at nye individer innvandret fra fastlandet. Graden av innavl hadde flere konsekvenser for elgens overlevelse og reproduktive suksess (fitness). Dette gav seg uttrykk i forsinket kalvingstidspunkt, redusert kalvevekt og redusert tvillingkalvproduksjon hos de mest innavla individene. Reduksjonen var likevel ikke veldig stor, og sammenlignet med verdiene vi finner i de fleste fastlandsbestander av elg, var kalvevekter og produksjonen av tvillingkalver fortsatt høy hos innavla individer på Vega. Graden av innavl på Vega var dessuten lavere enn hva vi kunne forventet dersom elgkyrne valgte sine partnere tilfeldig fra bestanden. På grunn av de mulig uheldige effektene av innavl, har mange arter utviklet strategier for å unngå dette, for eksempel ved at unge hanndyr utvandrer, eller ved at slektninger gjenkjennes og unngås. På Vega så vi at elgen ofte unngikk nære slektninger som partnere, men kun når det eksisterte gode alternativer. I perioder med lav bestandstetthet og lav okseandel var det få gode alternativer og parringer mellom nære slektninger forekom mer hyppig. I de genetiske studiene undersøkte vi også hvordan ulike individer bidro til fremtidige generasjoner og hvordan jakta påvirket dette. Blant elgkyrne var det seleksjon for tidlig kalving fordi tidligkalvende kyr i gjennomsnitt fødte flere rekrutter. I tillegg var små kyr mer suksessfulle fordi kalvene hadde mindre sjanse til å bli skutt enn kalvene til større kyr. Blant oksene var det seleksjon for et seint fødselstidspunkt. Oksekalver som var født sent hadde lavere vekt og lavere sannsynlighet for å bli skutt. Dette viser at jakt kan påvirke hvilke individer som fører sine egenskaper videre til etterfølgende generasjoner. Både vekta til kalvene og fødselstidspunktet hadde en betydelig arvbar komponent. Av den grunn kan begge egenskapene endre seg over tid, for eksempel som en tilpasning til endrede miljøbetingelser eller som en konsekvens av et selektivt uttak av individer under jakta. Selv om usikkerheten er stor, antyder våre analyser en framtidig reduksjon i kroppsvekten til kukalvene på Vega som en følge av større jakttrykk på store ungdyr, og en tilsvarende utvikling for oksekalv fordi tidligfødte kalver er mer utsatt for å bli skutt. En forutsetning er at jakta på Vega fører til samme type seleksjon også i årene som kommer. I likhet med mennesker, har elger ulike ‘personlighetstrekk’, og flere av disse kan gjøre elgen mer eller mindre utsatt for å bli skutt under jakt. På Vega undersøkte vi spesifikt hvordan elgens bruk av åpne (eks. innmark) og lukkede (eks. skog) habitattyper varierte mellom individer og hvordan denne egenskapen påvirkes av jakt. Jevnt over fant vi at elgene var relativt konsistente i sine preferanser for disse habitattypene og i tillegg var det markante forskjeller i fruktbarhet og dødelighet mellom individer som tilbrakte mye tid i skogen (sky individer) eller på innmarka (dristige individer). Som forventet fant vi at dristige elger med større sannsynliget ble felt under jakta eller de fikk med større sannsynlighet fraskutt en kalv. På den annen side var dristige individer også mer fruktbare (høyere tvillingandel) og produserte i gjennomsnitt større kalver. Dette er sannsynligvis fordi bruken av åpne arealer (innmark) fører til inntak av mere næringsrik mat. I løpet av livet var det derfor ingen betydelig forskjell i gjennomsnittlig kalveproduksjon fra sky og dristige elger, med den følge at frekvensen av individer med de ulike atferdstrekkene vil forbli på samme nivå over tid. En forutsetning er at jakttrykket ikke endrer seg mye og at de ernæringsmessige konsekvensene av å bruke de ulike habitattypene forblir som nå. Elgstudiene på Vega har bidratt med mye data og kunnskap med relevans for norsk og internasjonal hjorteviltforvaltning og -forskning. Fra et faglig perspektiv er det ønskelig å fortsette merkestudiene i årene som kommer. Studiene er imidlertid kostnadskrevende og til tider praktisk vanskelig å gjennomføre, og det er derfor usikkert hvorvidt vi får til å gjennomføre disse studiene med samme metodikk. I samarbeid med Vega kommune og lokal elgforvaltning, er dette noe vi håper å avklare i nær framtid.

Solberg, E.J., Heim, M., Herfindal, I., Rolandsen, C.M., Haanes, H., Lee, A.M., Tremblay, J.-P., Markussen, S.S., Ofstad, E.G., Kvalnes, T. & Sæther, B.-E. 2024. The moose studies on Vega – Experiences from the period 1992-2021. NINA Rapport 2258. The moose studies on Vega consist of a series of research projects on moose ecology, evolution, and management that we conducted on the island of Vega in Nordland in the period 1992-2021. Initially, the studies were started by the Norwegian Institute for Nature Research (NINA) with the purpose of studying the demographic consequences of varying sex and age structure in a harvested moose population and were later expanded with four new projects. These were conducted in collaboration with the Norwegian University of Science and Technology (NTNU) in Trondheim and dealt with 2) moose demography and life history strategies, 3) moose land use and 4) moose genetics and evolution. The latter project was divided into two subsequent subprojects (4a and 4b). In this report, we describe the development of the moose population on Vega in the period 1985-2021 and summarize the results from the different research projects. The projects were based on individual-based data on moose body growth, reproduction, survival, behavior, genetics and kinship. These were obtained by radio-tagging (collars with VHF and GPS transmitters), weighing, measuring, sampling (tissue, feces, blood) and tracking virtually all moose that survived to about eight months of age (the calf's first winter) and all new individuals immigrating to the population. During the study period (1992-2021), we tagged a total of 324 moose on Vega, most of them from helicopters using a dart gun and anesthetic. In addition, we tagged 50 moose from helicopters on the neighboring island of Ylvingen (in Vega municipality) in the period 2009-2013. Moose killed during hunting (age > 4 months) were also weighed and sampled. From tissue samples we extracted DNA that we later used to determine kinship (e.g., parent-offspring), and to estimate population size and structure back to 1985, when the first two moose in modern times colonized Vega from the mainland. Since 1985, the moose population on Vega has varied greatly in size and composition. The population increased until 1994, when there were just over 50 moose on the island before hunting. The population was then regulated to around 30 animals after hunting (winter population) and remained at about the same level for the next 6 years. Over the past 22 years, the population has increased, reaching a preliminary peak in 2019, with 55 individuals after hunting. Moose hunting was first permitted in Vega municipality in 1989 (2 animals) and has since been used by research (1992-1999) and management (2000-2021) to regulate the population to the desired level and composition. A total of 635 moose were harvested in Vega during the period 1989- 2021, most of which were calves (52%). The harvest was highest in 2015, with 34 animals. In addition, we recorded 46 moose (older than 4 months) that died for reasons other than hunting, of which three in connection with tagging (a fourth died in 2022). We also recorded 10 moose that emigrated from Vega during the study period, most of which swam to the mainland or islands in between the mainland and Vega. In the opposite direction, 13 bulls and five cow moose were recorded as immigrants to Vega. The moose population on Vega experience high recruitment rates (e.g., pre-harvest proportion calves) because of high fecundity rates, and at times a very low proportion of bulls. The latter was particularly true in the late half of the 1990s, when the proportion of bulls was greatly reduced as part of the research project on the island. Since then, the proportion of bulls has increased, and the age-specific fecundity rates have declined. The latter is probably related to a corresponding decline in age-specific body masses, potentially due to density dependent food limitation. In addition, increasing temperatures in both summer and winter may have led to a declinie in quality of moose forage and increased energy costs due to thermoregulation. In the first project on Vega, we investigated how biased sex and age structure affect moose body growth and reproduction in two consecutive periods. First, we reduced the average age of bulls (but not the proportion of bulls) and then we reduced the proportion of bulls (but not bull age). Despite significant changes in sex and age structure, we found no changes in the proportion of impreganted moose cows and no changes in bulls' body growth due to early start of breeding. However, we found significant delays in the date of birth, which in turn led to lower body mass of calves in fall and winter. Since older bulls produced more male calves, we also found a close correlation between the average age of bulls in the herd and the proportion of male calves produced. A low proportion of bulls can thus have several unfortunate consequences from a harvesting point of view. In the following years (2000-2021), the population was subject to normal hunting and without major manipulations of the population structure. During this phase, we conducted several studies of moose demography and life history strategies, drawing on increasingly better kinship data based on genetic analyses. Besides an effect of time of birth, we found in these studies, that calf body mass is also affected by maternal age, calf sex, and the number of siblings in the same year (0 or one). This variation persisted throughout life, so that large calves became large individuals even as adults. Body mass is important in the moose's life, as heavy moose cows tend to start reproducing at a young age and can thus produce many calves over a long life. On Vega, the moose are in good condition (high age-specific body mass) and are generally very productive. More than 50% of the moose cows gave birth to calf as two-year-olds (covered as yearlings), of which 7% had twin calves. The most productive moose cows (15-18 calves in their lifetime) on Vega started to reproduce as two-year-olds and stayed alive until well into adulthood (11-15 years). Interestingly, there was also a negative relationship between age of first reproduction and future reproductive success (number of calves), and between age of first reproduction and the probability of surviving the hunting season. The former shows that there are large quality differences between cow moose, while the latter is most likely because hunters rarely shoot the cow from the calf. Moose cows that start reproducing as two-year-olds are therefore more likely to survive the hunt than moose cows that start reproducing as three- or four-year-olds. The variation in reproductive success is greater among bulls than among cows. On Vega, only one third of the bulls succeeded in becoming a father, and even among these there was great variation in the number of offspring. The most productive bulls mated with up to 8 cows during the rutting season, and the most successful bull fathered as many as 44 calves. On average, each bull sired 3.2 calves. The survival was the main reason for the number of offspring produced by bulls. Many bulls were shot at the age of 18 months and thus had little chance of becoming a father. In addition, the weight and antler size of bulls are important for achieving matings. These traits increase with the age of the bull, and consequently, older bulls were more successful than younger bulls during the rut. Even small bulls in their age group can thus experience sirehood if they are allowed to live to a high adult age. This introduces a degree of randomness in terms of which bulls contribute to the next generation, and this randomness becomes more prominent when the proportion of bulls in the population is low. Under such conditions, even genetically weak bulls can sire a large proportion of calves. In the early 2000s and 2010s, we also conducted several studies of moose land use on Vega (spatial dynamics). Using radio collars, we recorded the location of moose at different times of the day and year, and, combined with high resolution resource maps, we studied both habitat use and selection. This type of study became particularly interesting after 2006, when we equipped moose on the island with GPS radiocollars for the first time. These can record many precise positions (± 5 m), which are then stored in the collar and can be sent to NINA via satellite or the cellphone network. Based on these data, we estimated moose home range size in different seasons or the full year, and we determined the most used and preferred habitats by moose on the island. Habitat use shows how the moose distribute themselves across different types of forest, agricultural land, heathland etc. (i.e., habitat types), while the level of habitat selection indicates the use of different habitat types relative to their availability. These studies showed that some habitat types were strongly selected, such as nutrient-rich deciduous forests. Habitat use also varied between males and females, between summer and winter, and to some extent between age groups. In addition, we found a clear effect of human activity on the spatial distribution of moose. This was particularly evident at calving and in the subsequent few weeks. During this period, cow moose stayed as far away from people as possible, or in other 'safe' areas, even though these areas often offered little food. In the last two sub-projects, we conducted a comprehensive mapping of the genetic relationships between moose on Vega and investigated how inbreeding and hunting affect the genetic variation and evolutionary development. The relative isolation of the island, combined with a limited number of immigrants, make the population vulnerable to inbreeding and inbreeding depression. In addition, there are no large predators other than humans on the island, which means that hunting and the choices made by the management and hunters can have a major effect on the evolutionary development of the population. The results show that the moose on Vega was significantly inbred in certain periods because of frequent mating between close relatives. The degree of inbreeding was particularly high in the 2000s, when the average degree of inbreeding in the population corresponds to what we can expect from mating between half-siblings. This corresponds with relatively low immigration of moose from the mainland during this period, leading to an accumulation of inbreeding over generations (i.e., mating between inbred related individuals). Conversely, we found a significant reduction in the degree of inbreeding in the population in years after new individuals immigrated from the mainland. The degree of inbreeding had several consequences for the fitness of moose. This was expressed in delayed calving, reduced calf body mass and reduced twin calf production in the most inbred individuals. However, the reduction was not very large, and calf mass and twin calf production in inbred individuals on Vega were still rather high compared to what is found in most mainland (and outbred) moose populations. The degree of inbreeding on Vega was also lower than expected if moose cows chose their partners by random. Because of the potential adverse effects, animals have often developed strategies to avoid inbreeding (inbreeding avoidance), for example by forcing young males to emigrate, or by recognizing and avoiding relatives. On Vega, we observed that moose often avoided close relatives as mates, but mainly when genetically more distant alternatives were available. In periods of low population density and bull:cow ratio, there were few good alternatives and mating between close relatives occurred more frequently. In these studies, we also investigated how different individuals contributed to future generations and how hunting affected the outcome. Among moose cows, we found selection for early calving because early calving cows gave birth to on average more recruits. In addition, small cows were more successful because their calves were less likely to be shot than calves of larger cows. Among bulls, there was selection for a later birth, as late born male calves had lower weight and lower probability of being shot. This shows that hunting can affect which individuals that pass on their genes (and traits) to subsequent generations. Both the body mass of calves and time of birth had a significant heritable component. Consequently, both traits may evolve, for example as an adaptation to changing environmental conditions or selective harvesting. Although there is considerable uncertainty, our analyses suggest a future reduction in the body weight of female calves on Vega because of higher hunting pressure on calves from larger females, and a similar development for male calves because early born calves are more likely to be shot. A prerequisite is that similar hunting selection will occur on Vega also in the years to come. In addition to characteristics such as size and calving time, hunting can affect the prevalence of various behavioral traits in a moose population. Like humans, moose have different personalities, which to a varying degree can make them susceptible to be shot during hunting. On Vega, we specifically investigated how the use of open (e.g., infields) and closed (e.g., forest) habitat types varied between individuals over time and how these behavioural traits affected their individual fitness. Overall, we found that moose were relatively consistent in their preferences for these habitat types, and there were marked differences in fecundity and mortality between shy individuals that spent more time in the forest and bold individuals that spent more time on infields. As expected, bold individuals were more likely to be killed during the hunt or they were more likely to have a calf killed. However, bold individuals were also more fecund (higher twinning rate) and produced larger calves than shy individuals, on average. This may be because moose can consume more nutritious food on infields than in the forest. As a result, there was no significant difference in the individual fitness of shy and bold moose, with the consequence that these behavioral traits are likely to remain in the population with the same frequency. A prerequisite is that the hunting pressure will be similar and that the nutritional benefits of using the different habitat types remains as now. The moose studies on Vega have provided substantial data and knowledge of great relevance to Norwegian deer management and research. Therefore, from a professional perspective it is desirable to continue the tagging studies in the years to come. However, as the studies are also expensive and practically demanding to conduct, it is yet uncertain whether we will continue at the same level in the years to come. In collaboration with Vega municipality and the local moose management, this is something we hope to clarify soon.

Solberg, E.J., Heim, M., Herfindal, I., Rolandsen, C.M., Haanes, H., Lee, A.M., Tremblay, J.-P., Markussen, S.S., Ofstad, E.G., Kvalnes, T. & Sæther, B.-E. 2024. The moose studies on Vega – Experiences from the period 1992-2021. NINA Rapport 2258. The moose studies on Vega consist of a series of research projects on moose ecology, evolution, and management that we conducted on the island of Vega in Nordland in the period 1992-2021. Initially, the studies were started by the Norwegian Institute for Nature Research (NINA) with the purpose of studying the demographic consequences of varying sex and age structure in a harvested moose population and were later expanded with four new projects. These were conducted in collaboration with the Norwegian University of Science and Technology (NTNU) in Trondheim and dealt with 2) moose demography and life history strategies, 3) moose land use and 4) moose genetics and evolution. The latter project was divided into two subsequent subprojects (4a and 4b). In this report, we describe the development of the moose population on Vega in the period 1985-2021 and summarize the results from the different research projects. The projects were based on individual-based data on moose body growth, reproduction, survival, behavior, genetics and kinship. These were obtained by radio-tagging (collars with VHF and GPS transmitters), weighing, measuring, sampling (tissue, feces, blood) and tracking virtually all moose that survived to about eight months of age (the calf's first winter) and all new individuals immigrating to the population. During the study period (1992-2021), we tagged a total of 324 moose on Vega, most of them from helicopters using a dart gun and anesthetic. In addition, we tagged 50 moose from helicopters on the neighboring island of Ylvingen (in Vega municipality) in the period 2009-2013. Moose killed during hunting (age > 4 months) were also weighed and sampled. From tissue samples we extracted DNA that we later used to determine kinship (e.g., parent-offspring), and to estimate population size and structure back to 1985, when the first two moose in modern times colonized Vega from the mainland. Since 1985, the moose population on Vega has varied greatly in size and composition. The population increased until 1994, when there were just over 50 moose on the island before hunting. The population was then regulated to around 30 animals after hunting (winter population) and remained at about the same level for the next 6 years. Over the past 22 years, the population has increased, reaching a preliminary peak in 2019, with 55 individuals after hunting. Moose hunting was first permitted in Vega municipality in 1989 (2 animals) and has since been used by research (1992-1999) and management (2000-2021) to regulate the population to the desired level and composition. A total of 635 moose were harvested in Vega during the period 1989- 2021, most of which were calves (52%). The harvest was highest in 2015, with 34 animals. In addition, we recorded 46 moose (older than 4 months) that died for reasons other than hunting, of which three in connection with tagging (a fourth died in 2022). We also recorded 10 moose that emigrated from Vega during the study period, most of which swam to the mainland or islands in between the mainland and Vega. In the opposite direction, 13 bulls and five cow moose were recorded as immigrants to Vega. The moose population on Vega experience high recruitment rates (e.g., pre-harvest proportion calves) because of high fecundity rates, and at times a very low proportion of bulls. The latter was particularly true in the late half of the 1990s, when the proportion of bulls was greatly reduced as part of the research project on the island. Since then, the proportion of bulls has increased, and the age-specific fecundity rates have declined. The latter is probably related to a corresponding decline in age-specific body masses, potentially due to density dependent food limitation. In addition, increasing temperatures in both summer and winter may have led to a declinie in quality of moose forage and increased energy costs due to thermoregulation. In the first project on Vega, we investigated how biased sex and age structure affect moose body growth and reproduction in two consecutive periods. First, we reduced the average age of bulls (but not the proportion of bulls) and then we reduced the proportion of bulls (but not bull age). Despite significant changes in sex and age structure, we found no changes in the proportion of impreganted moose cows and no changes in bulls' body growth due to early start of breeding. However, we found significant delays in the date of birth, which in turn led to lower body mass of calves in fall and winter. Since older bulls produced more male calves, we also found a close correlation between the average age of bulls in the herd and the proportion of male calves produced. A low proportion of bulls can thus have several unfortunate consequences from a harvesting point of view. In the following years (2000-2021), the population was subject to normal hunting and without major manipulations of the population structure. During this phase, we conducted several studies of moose demography and life history strategies, drawing on increasingly better kinship data based on genetic analyses. Besides an effect of time of birth, we found in these studies, that calf body mass is also affected by maternal age, calf sex, and the number of siblings in the same year (0 or one). This variation persisted throughout life, so that large calves became large individuals even as adults. Body mass is important in the moose's life, as heavy moose cows tend to start reproducing at a young age and can thus produce many calves over a long life. On Vega, the moose are in good condition (high age-specific body mass) and are generally very productive. More than 50% of the moose cows gave birth to calf as two-year-olds (covered as yearlings), of which 7% had twin calves. The most productive moose cows (15-18 calves in their lifetime) on Vega started to reproduce as two-year-olds and stayed alive until well into adulthood (11-15 years). Interestingly, there was also a negative relationship between age of first reproduction and future reproductive success (number of calves), and between age of first reproduction and the probability of surviving the hunting season. The former shows that there are large quality differences between cow moose, while the latter is most likely because hunters rarely shoot the cow from the calf. Moose cows that start reproducing as two-year-olds are therefore more likely to survive the hunt than moose cows that start reproducing as three- or four-year-olds. The variation in reproductive success is greater among bulls than among cows. On Vega, only one third of the bulls succeeded in becoming a father, and even among these there was great variation in the number of offspring. The most productive bulls mated with up to 8 cows during the rutting season, and the most successful bull fathered as many as 44 calves. On average, each bull sired 3.2 calves. The survival was the main reason for the number of offspring produced by bulls. Many bulls were shot at the age of 18 months and thus had little chance of becoming a father. In addition, the weight and antler size of bulls are important for achieving matings. These traits increase with the age of the bull, and consequently, older bulls were more successful than younger bulls during the rut. Even small bulls in their age group can thus experience sirehood if they are allowed to live to a high adult age. This introduces a degree of randomness in terms of which bulls contribute to the next generation, and this randomness becomes more prominent when the proportion of bulls in the population is low. Under such conditions, even genetically weak bulls can sire a large proportion of calves. In the early 2000s and 2010s, we also conducted several studies of moose land use on Vega (spatial dynamics). Using radio collars, we recorded the location of moose at different times of the day and year, and, combined with high resolution resource maps, we studied both habitat use and selection. This type of study became particularly interesting after 2006, when we equipped moose on the island with GPS radiocollars for the first time. These can record many precise positions (± 5 m), which are then stored in the collar and can be sent to NINA via satellite or the cellphone network. Based on these data, we estimated moose home range size in different seasons or the full year, and we determined the most used and preferred habitats by moose on the island. Habitat use shows how the moose distribute themselves across different types of forest, agricultural land, heathland etc. (i.e., habitat types), while the level of habitat selection indicates the use of different habitat types relative to their availability. These studies showed that some habitat types were strongly selected, such as nutrient-rich deciduous forests. Habitat use also varied between males and females, between summer and winter, and to some extent between age groups. In addition, we found a clear effect of human activity on the spatial distribution of moose. This was particularly evident at calving and in the subsequent few weeks. During this period, cow moose stayed as far away from people as possible, or in other 'safe' areas, even though these areas often offered little food. In the last two sub-projects, we conducted a comprehensive mapping of the genetic relationships between moose on Vega and investigated how inbreeding and hunting affect the genetic variation and evolutionary development. The relative isolation of the island, combined with a limited number of immigrants, make the population vulnerable to inbreeding and inbreeding depression. In addition, there are no large predators other than humans on the island, which means that hunting and the choices made by the management and hunters can have a major effect on the evolutionary development of the population. The results show that the moose on Vega was significantly inbred in certain periods because of frequent mating between close relatives. The degree of inbreeding was particularly high in the 2000s, when the average degree of inbreeding in the population corresponds to what we can expect from mating between half-siblings. This corresponds with relatively low immigration of moose from the mainland during this period, leading to an accumulation of inbreeding over generations (i.e., mating between inbred related individuals). Conversely, we found a significant reduction in the degree of inbreeding in the population in years after new individuals immigrated from the mainland. The degree of inbreeding had several consequences for the fitness of moose. This was expressed in delayed calving, reduced calf body mass and reduced twin calf production in the most inbred individuals. However, the reduction was not very large, and calf mass and twin calf production in inbred individuals on Vega were still rather high compared to what is found in most mainland (and outbred) moose populations. The degree of inbreeding on Vega was also lower than expected if moose cows chose their partners by random. Because of the potential adverse effects, animals have often developed strategies to avoid inbreeding (inbreeding avoidance), for example by forcing young males to emigrate, or by recognizing and avoiding relatives. On Vega, we observed that moose often avoided close relatives as mates, but mainly when genetically more distant alternatives were available. In periods of low population density and bull:cow ratio, there were few good alternatives and mating between close relatives occurred more frequently. In these studies, we also investigated how different individuals contributed to future generations and how hunting affected the outcome. Among moose cows, we found selection for early calving because early calving cows gave birth to on average more recruits. In addition, small cows were more successful because their calves were less likely to be shot than calves of larger cows. Among bulls, there was selection for a later birth, as late born male calves had lower weight and lower probability of being shot. This shows that hunting can affect which individuals that pass on their genes (and traits) to subsequent generations. Both the body mass of calves and time of birth had a significant heritable component. Consequently, both traits may evolve, for example as an adaptation to changing environmental conditions or selective harvesting. Although there is considerable uncertainty, our analyses suggest a future reduction in the body weight of female calves on Vega because of higher hunting pressure on calves from larger females, and a similar development for male calves because early born calves are more likely to be shot. A prerequisite is that similar hunting selection will occur on Vega also in the years to come. In addition to characteristics such as size and calving time, hunting can affect the prevalence of various behavioral traits in a moose population. Like humans, moose have different personalities, which to a varying degree can make them susceptible to be shot during hunting. On Vega, we specifically investigated how the use of open (e.g., infields) and closed (e.g., forest) habitat types varied between individuals over time and how these behavioural traits affected their individual fitness. Overall, we found that moose were relatively consistent in their preferences for these habitat types, and there were marked differences in fecundity and mortality between shy individuals that spent more time in the forest and bold individuals that spent more time on infields. As expected, bold individuals were more likely to be killed during the hunt or they were more likely to have a calf killed. However, bold individuals were also more fecund (higher twinning rate) and produced larger calves than shy individuals, on average. This may be because moose can consume more nutritious food on infields than in the forest. As a result, there was no significant difference in the individual fitness of shy and bold moose, with the consequence that these behavioral traits are likely to remain in the population with the same frequency. A prerequisite is that the hunting pressure will be similar and that the nutritional benefits of using the different habitat types remains as now. The moose studies on Vega have provided substantial data and knowledge of great relevance to Norwegian deer management and research. Therefore, from a professional perspective it is desirable to continue the tagging studies in the years to come. However, as the studies are also expensive and practically demanding to conduct, it is yet uncertain whether we will continue at the same level in the years to come. In collaboration with Vega municipality and the local moose management, this is something we hope to clarify soon.